Writing and the city in early China

WANG HAICHENG

Chinese urbanism has a history of more than 5,000 years, and ever since the invention of the Chinese writing system more than 3,000 years ago, the process of urbanization and the uninterrupted transmission of literacy have gone hand in hand.

Precursors of Anyang writing

Let me begin with a quick word about cities and writing in the two centuries before 1350 bce. During the fifteenth and fourteenth centuries a great city flourished 200 kilometers south of Anyang, in today's Zhengzhou. The largest city of its time, it had two walls built of pounded earth. The inner wall, 22 meters thick at the base, has a perimeter of about 7,000 meters and

I would like to thank Norman Yoffee for inviting me to participate in this wonderful project. For penetrating comments on drafts of this chapter I am most grateful to Robert Bagley. I am greatly indebted to Kyle Steinke for redrawing several maps and figures. Cao Dazhi and Yan Shengdong generously sent me their own drawings; Shan Yueying, Song Guoding, Tang Jigen, and Wu Hsiao-yun were instrumental in obtaining photo­graphs.

Table 7.1 Chronological table of early Bronze Age China

Map 7.1 Archaeological sites of the Early Bronze Age mentioned in this chapter (drawn by Kyle Steinke).

encloses an area of more than 400 hectares. The earliest examples of writing known from East Asia were found near Zhengzhou and belong probably to the fourteenth century. They are graphs written in vermilion on clay pots (Figure 7.1). Though few, they clearly belong to the writing system we know a century later from the so-called oracle bone inscriptions at Anyang. The Anyang inscriptions are the first substantial corpus of Chinese writing, but they are display inscriptions; neither at Anyang nor at Zhengzhou does

Figure 7.1 Corpus of Chinese writing from Xiaoshuangqiao, c. fourteenth century bce. The corpus seems to include numerals, titles, kinship terms, and possibly a deity's name as shown on bottom right. The characters in the fourth and the eighth rows are from Xiaoshuangqiao, the second and the sixth rows from the oracle bone inscriptions a century or two later, the third and the seventh rows from bronze inscriptions a little later than the oracle bone inscriptions, the first and fifth rows are their transcriptions in modern Chinese (table after Song Guoding, “Zhengzhou Xiaoshuangqiao yizhi chutu taoqi shang de zhushu” Wenwu 5 [2003], p.

everyday writing survive. Everyday administrative documents were almost certainly written on perishable materials like wood and bamboo strips,^[89] the

standard materials of later times, and these do not begin to survive in the archaeological record until about the fifth century bce. What we have from earlier periods is only inscriptions meant for display of some sort. The graphs in Figure 7.1, done in vermilion for some sort of ritual deposit, were clearly display writing. The polity based at Zhengzhou created an empire, building fortresses to secure new territories as far away as the middle Yangzi region 450 kilometers to the south, and it may well have used writing in the administration of its empire. But the empire was short-lived. Its fortresses were abandoned after a century or so, and eventually its capital too.

The Huanbei period

Material culture, especially elite material culture, strongly suggests that part of the Zhengzhou population, led by its elite, moved north across the Yellow River and settled at the Huanbei site at modern Anyang. The Huanbei site was discovered only recently, so survey and excavation data are very limited, but they do seem to represent a new city created by the organized migration of a population with prior urban experience. Creating the city had to begin with reconnaissance and migration. The next step was to erect temples to house ancestors and palaces to house the king and his family. At the same time, dwellings were built around the royal compound for the king's followers. Sometime later the royal palace and temple com­pound were enclosed by a wall to separate the royal precinct from the lesser dwellings. And finally another wall was built enclosing the royal precinct, the other houses, and some burials within a second rectangle of over 400 hectares, an area comparable to that inside the inner wall at Zhengzhou. Scattered over an area of about 800 square kilometers surrounding the Huanbei site, preliminary survey has found at least twenty small settlements that were probably controlled by the rulers of Huanbei.

Only two bits of writing datable to the Huanbei period have been found so far, both at the Huanbei site. One is a jade inscribed with three characters, perhaps an amulet; the other is a bone object bearing perhaps the name of a person or a lineage (Figure 7.2). No inscribed divination bones have been found. They may turn up in new excavations; but it is also possible that the decision to inscribe the bones used for divination was only made at the next stage, at the beginning of the Yinxu period around 1250 bce. There are hints

Figure 7.2 Two inscribed objects of the Huanbei period. Left: a jade pendant inscribed with three characters, from top to bottom read “great ancestors harm.” From Xiaotun Tomb 331 at Anyang, length 6.7 cm, width 1.6 cm (photograph courtesy of History and Philology, Academia Sinica). Right: a bone fragment inscribed with two characters, perhaps a name. From Huanbei Huayuanzhuang, extant length 5.5 cm, width 1.8-2.8 cm (photograph courtesy of Tang Jigen).

that divinination practice changed at that time as the result of a royal decision taken in the midst of larger changes.

The Yinxu period

Thefounding and peopling of Yinxu

Around the middle of the thirteenth century the temple/palace compounds at Huanbei burned down. Whatever the cause of the fire, the buildings were never rebuilt. Construction work on the city wall, begun a short while before, seems to have stopped at the time of the fire. The site seems to have been abandoned in favor of a new one just across the Huan River, in an area today called Xiaotun, where a Huanbei period bronze foundry and settle­ment were already located. These earlier buildings seem to have been demolished to clear the ground for new royal temples and palaces.

We do not know the reasons for this relocation, but it was surely an act of will on the part of a governing elite, and the responsible agent seems to have been a king named Wu Ding. A number of notable changes may stem from Wu Ding's direct initiative. Here is a list, in descending order of our confidence in connecting them with him: (ι) The construction of a royal cemetery outside the city, the earliest tomb probably being intended for Wu Ding himself (Map 7.2). (2) Certain rapid and dramatic changes of bronze style in Wu Ding's reign, changes that may have occurred simply as by­products of a tremendous expansion of the bronze casting industry (Figure 7.3). (3) The carving of divination records on the divination medium (bovine scapulae and turtle shells, for example Figure 7.4). (4) The addition of emblems and short dedicatory phrases to ritual bronzes (Figure 7.3). (5) The first attestation of horses in China and probably the earliest chariots, both of which were imported from the northwest and used as conspicuous symbols of elite power and status. (6) A distinct increase in the quantity of small carvings done in marble, a scarce material that from this time onward was used almost exclusively by royalty.

The new city that Wu Ding created is called Da Yi Shang (Great Settle­ment Shang) in his divination texts, so we will call its people the Shang people, but we will use the modern name Yinxu to distinguish it from Huanbei (Map 7.2). Unlike Huanbei, Yinxu had no city walls and no clearly demarcated perimeters other than those provided on the north and east by the riverbank. Allowing a riverbank to define city boundaries amounts to abandoning a tradition of orthogonal enclosing walls, a tradition that

Map 7.2 Shang sites at Anyang (based on Niu Shishan, “Zhongguo gudai ducheng de guihua moshi chubu yanjiu,” in Zhongguo shehui kexueyuan kaogu yanjiusuo [ed.], Yinxu yu Shang wenhua: Yinxu kexuefajue 80 zhounian jinian wenji [Beijing: Kexue chubanshe, 2011], p.

Figure 7.3 Two bronze he vessels from the reign of Wu Ding, shown roughly to the same scale. Left: from a foundation deposit shown in Map 7.3 (lower right corner), with an inscription cast under the handle dedicating it to a certain “Father Yi,” probably made by Wu Ding at the beginning of his reign for his father. Height 34 cm (photograph courtesy of Tang Jigen). Right: one of a set of three from Tomb 1001 at the royal cemetery shown in Map 7.2 (horizontal rectangle on upper left), generally believed to belong to Wu Ding; the vessel was therefore made at the end of his reign, and its architectural look makes it almost impossible to recognize its ancestry in the vessel on the left, made at most a few decades before. The set of three were inscribed under their handles “left,” “middle,” and “right,” probably indicating their positions on the altar. Height 71.2 cm (courtesy of the Nezu Bijutsukan Tokyo).

extends back into the Neolithic. A change so significant must have been the king's conscious decision.

The current estimate of Yinxu's size is 30 square kilometers, but how much of that area was actually inhabited during the Yinxu period is unknown. Nevertheless, eighty years of mortuary and settlement archae­ology tell us to visualize the city as an agglomeration of lineage settlements

Figure 7.4 Scapula inlaid with red pigment from the reign of Wu Ding. A series of divinations were recorded on this side and the other side, each beginning with a common question asking about the fortune of the coming week (as in text [7]), followed by the king's correct prognostication and what actually happened, including a death and a chariot accident involving the king. Why the king would display such disasters remains a mystery (after Zhongguo Guojia Bowuguan [ed.], Zhonghua Wenming. [Beijing: Zhongguo Shehui Kexue Chubanshe, 2010], p. 142, by permission of the National Museum of China)

and cemeteries that grew outward from the royal precinct in the northeast corner over a period of two centuries. Each lineage comprised perhaps twenty to fifty families. Each family, or several families together, lived in a courtyard compound. Several of these compounds, varied in size and sometimes just a meter apart, collected around a bigger compound that perhaps belonged to the lineage head or the lineage temple. Figure 7.5 shows a typical lineage settlement containing patio groups and such features of everyday life as wells.

Figure 7.5 A lineage settlement at Xujiaqiao north, Yinxu. The excavated area of the house foundations (numbered 1-16) is approximately 2 hectares. About 400 tombs have been found inside and around the settlement, but to date none is published (from Meng Xianwu and Li Guichang, “Yinxu siheyuan shi jianzhu jizhi kaocha,” Zhongyuan wenwu 5 [2004], p. 28, fig. 2. Redrawn by Kyle Steinke).

This spatial organization was mirrored in the lineage cemetery nearby, in which tombs clustered in three levels can be discerned, perhaps correspond­ing to nuclear family, extended family, and lineage. Each lineage had its own name and ancestors. Its name or emblem was often cast on the ritual vessels buried with lineage members (Figure 7.6). To date more than 15,000 burials have been excavated. Far fewer houses have been found, and houses were subject to continuous rebuilding, so how many lineages inhabited Yinxu at any one time cannot be reconstructed. It was probably fewer than the total number of lineage emblems known to us, about 150. But to get a very rough idea of the Yinxu population, let us suppose all 150 lineages coexisted; if we then assign 100-200 people to each lineage - a number that has some support from the mortuary data - we get a range of 15,000 to 30,000. This is lower than the

Figure 7.6 A selection of lineage emblems or names cast on Yinxu bronzes. Most of the components of these monograms are characters that stand for words, but they are not arranged in linear sequence of an ordinary text (for instance, the one shown in

Figure 7.8), hence it is difficult to ascertain the meaning of each monogram. Many, but not all of these monograms are lineage emblems (from Zheng Ruokui, “Yinxu Dayi Shang zuyi buju chutan,” Zhongyuan wenwu 3 [1995], p. 88, fig. 3).

commonly cited range of 70,000-120,000 but differs by an order of magnitude at most. Both ranges can accommodate references in divination texts to raising armies of 3,000 or 5,000 or even in one case 13,000 men:

[1] Crack-making on dingyou (day 34), Ke divined: “This season, (if) the king raises 5,000 men to campaign against the Tufang, he will receive assist­ance.” (HJ 6409)^[90]

Central planning and control of city layout are not much in evidence at Yinxu, but perhaps the recently exposed intersecting thoroughfares and long waterways along the major axes were designed from the first for wheeled traffic and water supply - the wheel ruts are very clear. The waterways passed through areas with heavy industry such as pottery workshops and bronze foundries (Map 7.2). The thoroughfares seem to have converged on the south side of the royal precinct.

The royal precinct and its divination texts

The royal precinct covers about 70 hectares, with over 100 building founda­tions found so far (Map 7.3). It is in storage pits associated with some of the buildings that most of the inscribed divination bones have been found. At least half of these texts can be dated to the time of relocation from Huanbei to Yinxu, in other words, to the reign of Wu Ding. The sudden appearance of so many divination texts suggests that the decision to inscribe the bones was somehow involved with the other changes taking place at this time. Most of these texts show the king anxiously divining about the fortune of the coming week or about the appropriateness of particular sacrifices to royal ancestors. Here are some examples:

[2] Crack-making on jimao (day 16), Ke divined: “In performing an exorcism for [Lady] Hao to Father Yi, cleave a sheep, offer a pig, pledge ten penned sheep.” (HJ 271)

[3] Crack-making on dingchou (day 14), Xing divined: “The king hosts Father Ding, performs the xie-ritual, no fault.” (HJ 23120)

[4] On the eighth day, indeed decapitated 2,656 persons at Meng. Ninth month... (HJ 7771)

These texts, together with the sacrificial remains found in the precinct and in the royal cemetery, are gruesome testimony to the supreme importance of the royal ancestors in the institution of Shang kingship, and writing figured prominently in the rituals directed toward them.

Archaeological evidence suggests that during the rituals the ancestors were represented not by portraits but by small wooden tablets inscribed with their names. The ancestral cult, as reconstructed from divinations about sacrifices, included a cycle of five rituals performed on a fairly strict schedule. The schedule was complex, prescribing particular sacrifices to particular ancestors on dates determined by their positions in the Shang king list. For example:

Map 7.3 The royal precinct of Yinxu, blocks showing the order in which major building foundations were constructed during the Shang period (Phase I here, perhaps includes a bronze foundry) and the Yinxu period (Phrases II-IV) (based on Du Jinpeng, Yinxu gongdian qu jianzhu jizhi yanjiu [Beijing: Kexue chubanshe, 2010], p. 407, fig. 11-2, redrawn by Kyle Steinke).

[5] jiaxu (day ιι) the yi ritual [for] Shang Jia, yihai (day 12) the yi ritual [for] Bao Yi, bingzi (day 13) the yi ritual [for] Bao Bing, [dingchou] (day 14) the yi ritual for Bao Ding, renwu (day 19) the yi ritual for Shi Ren, guiwei (day 20) the yi ritual for Shi Gui... (HJ 35406)

Besides the continuing ritual cycle, there were many other occasions when sacrifices might be divined about and then offered:

[6] In praying (for rain) to (the ancestors) from Shang Jia (to) Da Yi, Da Ding, Da Jia, Da Geng, Da Wu, Zhong Ding, Grandfather Yi, Grandfather Xin, and Grandfather Ding, the ten ancestors, (we will) lead-in-sacrifice (?) a ram. (HJ 32385)

Week by (ten-day) week, a sacrificial schedule was presented to the ances­tors in writing.^[91] The schedule probably included all the week's sacrifices, not just those of the cycle of five rituals but also others of such types as those seen above in texts [3], [4], and [6]. The ritual specialist had a duty to file a written report to the ancestors, but in substance his report was also a plan for the performance of the sacrifices, a document with an administrative function.

But further than this, the functions of writing in the Anyang ritual system are not well understood. Contrary to what is often stated in general books about Anyang writing or religion, writing was not essential to communi­cation with the spirits. Most of the divination bones found at Anyang are uninscribed. The diviner first carved hollows on one side of the shell or scapula, then applied heat to the hollows in order to produce cracks on the other side. The cracks were then interpreted as the omen's response to a previously announced question. The vast majority of bones cracked for divination are not inscribed, and the bones that do have inscriptions were inscribed after the communication was finished. Some inscriptions were beautifully carved and even inlaid with red or black pigment (Figure 7.4); these clearly had a display function, but we do not know the audience for the display.

What seems likely is that a group of literate diviners interpreted most of the omens and kept records of them on documents made of bamboo or wood slips (a graph that depicts bundled slips actually occurs in the divin­ation texts and in bronze inscriptions). These records, or a digest of them, would regularly be sent to the king to report on divine intentions. Perhaps only when grave matters were concerned would the king himself make the prognostication and record it on the bone, as in the following inscription:

[7] Crack-making on guisi (day 40), Ke divined: “In the next ten days there will be no disasters.” The king read the oracle and said: “There will be calamities; there may be someone bringing alarming news.” When it came to the fifth day, dingyou (day 34), there indeed was someone bringing alarming news from the west. Guo of Zhi reported and said: “The Tufang have attacked in our eastern borders and have seized two settlements. The Gongfang likewise invaded the fields of our western borders.” (HJ 6057)

In rare instances like this we glimpse another channel for gathering information: reports on earthly affairs made by subordinates. This inscription and others like it make it clear that Shang scribes were capable of writing not only reports of enemy attacks but also letters, royal decrees, almost anything the king might require. Guo of Zhi could have made his report in writing on perishable materials. He might even have been required to make his report in writing:

[8]... Junior Servitor Qiang assisted (the king) to attack. Mao [enemy leader] of the Wei [enemy state] was captured, (also captured),.. 20, captives 4, head trophies 1,570, captives of the Meifang 100, horses.., chariots 2, shields 183, quivers 50, arrows... [the remaining text talks about making different kinds of human and animal sacrifice to various royal ancestors. It is broken at the point Junior Servitor Qiang was rewarded by the king]. (HJ 36481)^[92]

[9] On renzi (day 49) the king made cracks and divined: “Hunting at Zhi, going and coming back there will be no harm.” The king read the oracle and said: “Prolonged auspiciousness.” This was used (?). (We) caught foxes 41; mi-deer 8; rhinoceros ι. (HJ 37380)

The Shang king must have spent a great part of his time gathering two kinds of information, divine and human, comparing them, and only then making his final decision and giving his commands. All these activities are likely to have involved writing. The aforementioned practice of presenting the ancestors with written sacrificial schedules might well have been mod­eled on real-world administration. The divination texts make it clear that book-keeping had reached the realm of the ancestors:

[10] Wo brought in 1,000 (shells); LadyJing ritually prepared 40. (Recorded by the diviner) Bin. (HJ 116b)

Considering that our sample of Shang writing, the divination texts, is limited to brief records of the king's questioning of his ancestors, it is remarkable how many traces of book-keeping we find in them, how much careful accounting of the flow of people and materials: 5,000 men, 2,656 human victims, 183 shields, 41 foxes, 1,000 shells... Communication with the spirits is the content of the first inscriptions that survive from ancient China, but communicating with the spirits was not the motive that inspired the inven­tion of Chinese writing. The motive for invention was undoubtedly in the administrative sphere, where the overriding concern was to exert control, and the means of control was to make inventories and create accountability. The immense scale of Shang agriculture, metallurgy, and colonial enter­prise, as revealed by inscriptions and archaeology, argue for heavy adminis­trative use of writing. The inscriptions quoted here are the surviving tip of an administrative iceberg.

Agriculture and book-keeping

[ιι] “If the king orders the Many Yin (officers) to open up fields in the West, [we] will receive crops.” (HJ 33209)

[12] Crack-making on guihai (day 60): “Should the king order the Many Yin (officers) to work on the field at Yu?” Crack-making on yichou: “Should the king order [the Many Yin] to work on the field at Jing? To work on the field at Xun?” (Kaizuka 1959-68, no. 2363)

[13] We will receive the harvest that Fu cultivates at Zi. (HJ 900)

[14]... if [we] greatly order the laborers, saying “Work together in the fields,” [we] will receive harvest. (HJ ι)

Texts like these give us a general idea of Shang agriculture, especially the farming of the royal domain. The king sent officials to allocate fields, presumably to both royal and non-royal lineages, by ridging boundaries and making earth cairns.^[93] This act presupposed a land survey, and surveyors are attested in the divination texts. An accompanying operation was to “build a settlement” in such-and-such a place, sometimes as many as thirty settle­ments in a single campaign. Royal crops were prefixed with qualifiers such as “the king's,” “Shang's,” “the great settlement's,” or “our.” Some of the farms had named locations; others were specified only by cardinal direction. Inscriptions [11]-[13] suggest that the Shang kings knew (at least roughly) where the farms were located and what officers were responsible for them. Where did they obtain this knowledge? It seems likely that the royal house possessed lists of fields and officers. In the corpus of divination texts over a hundred toponyms occur in divinations that inquire about the harvest. Bone and bronze inscriptions mention an official inspection of fields that involved classifying them into four types. The character for field, tian H, may depict a field divided into square parcels by a grid of pathways and/or drainage ditches. If land reclamation was organized by the royal house, as inscription [ιι] suggests, fields might have been laid out in a way that facilitated land survey, but there is no archaeological evidence for the shapes of Shang farms.

Regularly subdivided agricultural landscapes would help the state allocate land to discrete units of farmers.^[94] Archaeological evidence suggests that labor gangs depended on the state for agricultural tools. Large numbers of stone sickles have twice been found in the royal precinct (one pit contained “a thousand” by the excavator's guess; the other had 444). These were most likely made by state workshops for distribution to harvesters. Efficient provision of tools depended on accurate knowledge of the users, their numbers, and their administrative units, information that would also have facilitated the distribution of rations. The king divined about sending offi­cials to inspect granaries, some of which seem to have been located not far from the fields. Could it be that part of the stored grain was used for feeding the local laborers? In some early states the distribution of rations was managed with writing; the Inka state used khipus instead. In either case record-keeping was essential. The bone inscriptions have examples of “the counting of people” (deng ren) related to agriculture and warfare. As we have seen above, sometimes the numbers of persons and goods are specific enough to suggest that careful book-keeping was maintained. But the extant records do not tell us what kind of census information was collected beyond the number of persons, nor do they disclose whether there was any state­wide enumeration of the people. Nevertheless, the records suggest that accounting was a routine feature of state administration. Although the records we have were carved on bone, the sources for the numbers they contain must have been now-perished administrative documents written on wood or bamboo strips.

Book-keeping was not confined to religion, agriculture, war, and hunting. Animal husbandry must have been the main source for the meat consumed by the elite and the spirits. Skeletal remains together with countless divin­ation records show that whole horses, cattle, sheep, dogs, and pigs were offered to the spirits in staggering numbers, and the living must have consumed equal or greater numbers both during rituals and in everyday life. Although it is possible that many domestic animals were acquired through long-distance trade or war, many more would surely have been raised locally, if not at Yinxu itself (where pens have not been found) then in some of the villages in the Huan River Valley. When we read an Early Dynastic Mesopotamian text recording the amount of fodder given to hogs, or an Assyrian shepherd's account of his sheep,^[95] we cannot help wishing that Chinese administrators too had kept their records on durable clay. Yinxu archaeologists have given us many more bones than can be accounted for in the divination texts.

City industries and book-keeping

Map 7.2 shows some of the workshops excavated at Yinxu. Workshops for bone, jade, pottery, and bronze have been found, and there must be a workshop for chariot building somewhere too. At present four bone work­shops are known. A waste pit at one of them contains about half a million fragments of cattle bones dumped there by wheeled vehicles. The fragments from another workshop weigh 32 metric tons. From wasters we can recon­struct the process of manufacturing a bone hairpin step by step, a process that may have involved minute division of labor. The Yinxu bone-working industry involved transferring materials from slaughterhouses to workshops in vast quantities. When we find receipts for quite small numbers of jades and turtle shells carved on a few jades and shells (for example, [ιo]), it is hard to imagine that scribes would not be regularly employed in tracking the vastly larger transactions in bone.

[15] Crack-making on dinghai (day 24), Da [divined]: “... if (we) cast yellow ingots^[96]... making pan basins, the auspicious day will be...” (HJ 29687)

Three bronze foundries of the Yinxu period have been partly excavated. The largest one, in the western part of the city (Map 7.2), is estimated to have covered an area of 5 hectares. There is no doubting the complex division of labor here. One pit contained raw clay for making molds and models. Four pits were floored with charcoal, upon which sat clay cores that had perhaps been covered by mats. These pits were probably used to dry clay models, molds, and cores in the shade. In two semi-subterranean rooms were found unbaked models for big tripod legs, clay molds for casting round vessels more than 1.5 meters in diameter (no bronze vessel of such colossal size has ever been found), and other foundry remains. The debris from this site includes numerous fragments of furnaces, along with molds and models for vessels, weapons, and tools. Many tools for carving the models and polishing the finished products were also recovered, along with tuyeres and charcoal for melting the bronze in the furnaces.

Three groups of semi-subterranean houses were found surrounding the area with foundry debris. So far about ninety have been excavated. The number of rooms varies from one to five, with the one-room house being the most common (forty-five examples) and only one five­room house. The roofed area ranges from 5 to 25 square meters. The variation in size and number of rooms may reflect some sort of social hierarchy of the residents. Except for a small reception area in the multi­room houses, each room has a low earthen bench 0.8 to 1.4 meters wide, presumably a bed just large enough for one or two persons. Hearths and niches were found in both living rooms and bedrooms, with pottery, hairpins, divination bones, and sometimes animal bones from meat consumption.

The houses appear to have been built all at once, early in the Yinxu period, while the foundry debris is dated to a later phase. It is possible that the foundry was in operation from the beginning of the Yinxu period and that the people who lived in these simple houses were foundrymen under state control. Houses for ordinary lineage members in Yinxu were mainly above-ground courtyard houses (Figure 7.5), usu­ally with many pits around them for storage or garbage disposal. Other adjacent features include wells and streets. But at the foundry site, the surroundings of the houses are strangely clean. The lack of storage facilities might signify the workmen's dependence on state rations (though they would have cooked the food themselves: the bread and other dry foods of many early state societies lend themselves to mass-production of the kind found in, say, Egyptian bakeries, but the Chinese prefer to boil cereals). The foundry's specialized and highly skilled workmen were a valuable resource and must have been under close bureaucratic supervision. Scribes presiding over the foundry's operations must have kept rosters and ration lists.

[16] Crack-making on dingsi (day 54), Gen divined: “Should we call upon (someone) to acquire (copper) ingots?” Divined: “Should we not call upon (someone) to acquire (copper) ingots?” The king read the oracle and said: “Auspicious. If...” (broken) (HJ 6567)

Records would also have needed to be kept of the tremendous quantities of copper, tin, lead, and fuel consumed by the foundry, at least the first three of which would have been obtained from long-distance trade. Smelting slags have not been found at Yinxu, so we can be sure that the extraction of metals from ores took place elsewhere (the logical place is at the mines, if fuel was available there, so that only metal, not ore, would have to be transported). But ore sources have not yet been pinpointed with certainty; provenance studies based on analyses of bronze artifacts face many difficulties of interpretation. Without docu­ments like the Old Assyrian letters from Kanesh, little can be said about Shang trade other than that its scale must have been large and the geographic areas involved must have been great. Divination texts that mention “tribute” received by the Shang king might be interpreted as evidence of interregional trade. The tribute is sometimes large numbers of horse, cattle, sheep, jade, and cowry shells.

Colonial enterprise and writing

[17] On renxu (day 59)... order Jiang... to acquire salt... Second month. (HJ 7022)

Recent surveys and excavations in modern Shandong province tell us how the Shang procured one key commodity from a distant source. Ten to thirty kilometers inland from the modern coast of northern Shandong archaeologists have found over 200 seasonal camps for making salt from underground brine. The brine is distributed in a belt about 250 kilometers long that runs between salt water and fresh water further inland (Map 7.4). The camps range in size from 0.4 to 0.7 hectares, each including (i) a brine well; (2) a series of pits for percolation, sedimenta­tion, and evaporation; (3) a closed kiln under a hut for evaporating brine in coarse helmet-shaped pots; (4) a workspace and pits containing brine,

Map 7.4 Shang colonies in Shandong, showing possible traffic routes for the shipment of salt to Anyang, connected by sites yielding bronze ritual vessels. (ι) Niyangtun;

(2) Lanjia; (3) Daguo; (4) Gucheng; (5) Sangjiazhuang; (6) Yujia; (7) Laowa; (8) Subutun; (9) Zhaibian; (10) Huaguan; (ιι) Shijia; (12) Tangshan; (13) Jianxi; (14) Daxinzhuang;

(15) Liujia; (16) Xiaotun; (17) Xiaoli; (18) Hongfan; (19) Xihua; (20) Haozhuang; (21) Gushan (drawn by Yan Shengdong and Kyle Steinke).

also under the hut (Figure 7.7). Ten men would be needed at each camp for one season of work, which would produce about 500 kilograms of salt. Several dozen such camps were clustered together in an area ranging from several square kilometers to several hundred (the little squares in Map 7.4).

During the off season the workmen lived in permanent settlements on the other side of the salt water-fresh water dividing line. These settle­ments, with a three-tier settlement pattern, were clustered at intervals of 2 to 3 kilometers. The settlements of the first tier housed about ten families in an area of 1 hectare; these were closest to the camps. Their inhabitants raised animals and made helmet-shaped pots for the camps, but they did not farm. Their grain, and perhaps timber for constructing huts as well, were imported from the second-tier settlements further removed from the camps. These ranged from 3 to 6 hectares; each was divided into zones for housing, storage, garbage disposal, cemetery, and

Figure 7.7 Reconstruction of a salt-making hut in Shandong, showing how the helmet­shaped pots are placed on a gridding frame (courtesy of Yan Shengdong).

sometimes crafts such as the making of pottery and lithic tools. Sickles, spades for tilling, and knives indicate that farming was a major occupa­tion of the inhabitants. Ninety percent of the animal bones recovered were from domesticated species, suggesting that animal husbandry was another source of food. Minor elite tombs with bronzes were found in a few of these settlements. Two or three settlements located still further inland constituted the third tier. A cemetery near one of them (Subutum in Map 7.1 and Map 7.4, no. 8) had tombs comparable in form, size, and contents with the tombs in the royal cemetery at Yinxu. These are in fact the largest tombs of their time outside the Shang capital.

Few of these camps and settlements were established before the Yinxu period. Their abrupt appearance in northern Shandong was the result of population movement from somewhere else. The material culture of the sites, ranging from burial customs to utilitarian and elite objects, resem­bles that of Yinxu, and some of the lineage names inscribed on bronzes are attested also in divination texts and bronze inscriptions from Yinxu. It seems therefore that the Shang at Yinxu managed to colonize a region ι,ooo kilometers away, at least in part to procure salt. Each year hundreds of metric tons of salt must have been made and shipped inland, while thousands of tons of grain, probably some meat or livestock, and large quantities of timber were transported to the salt-making bases. The archaeological findings that enable us to reconstruct the scale and traffic patterns of salt production here are remarkably clear and exact, and they immediately raise the question: how did the Shang organize this colonial enterprise?

Comparison with other early states suggests that effective control of regions at a distance depends on good communications, either by land, as exemplified by the famous road system of the Inka state,^[97] or by water, as in the Ur III state's river-borne traffic in grains and bitumen,^{[98] [99]} or by a combin­ation of the two, as in the Egyptian exploitation of Nubia, in which naviga­tion on the Nile had to change to overland travel from time to time to bypass cataracts.¹¹ The divination records at Yinxu tell us that the Shang had a similar state-run system of communication.

[18] Crack-making on guihai (day 6o), divined: “Is it today that the king should attack?”... At night the king walked from the third fortress. (HJ 33149)

It appears that there were numbered relay stations and fortresses on major traffic routes terminating at Yinxu. The distribution of sites between Yinxu and Shandong where bronzes have been unearthed suggests two possible overland and two river-borne routes (Map 7.4). A group of divination texts found at Yinxu but made during a year-long military campaign in Shandong by a late Shang king suggests that Shang colonization was backed by military power. Inscriptions in [ι], [8], and [18] speak of campaigns against enemies; others speak of enemy attacks on borders or frontiers (for example, [7]), suggesting that Shang patrolled borders to its domain.

10

II

How did distant outposts communicate with Yinxu? Because the writing system in use at Yinxu had already developed the ability to record continuous speech, its administrative applications could have gone well beyond the making of ledgers for distributing rations. Did the frontier officials use written dispatches like the Semna reports in Egypt?^[100] The content of inscription [7] is certainly comparable to the Semna dispatches.

David Keightley calls the administration of the Shang state “incipient bureaucracy” or “proto-bureaucracy.”^[101] It seems to me that his words “incipient” and “proto-” are too cautious. By the time of the divination texts the thinking behind administrative records had already permeated the realm of the supernatural and bureaucratized the form taken by communication between the living and the spirits. From near the end of the Shang dynasty we have a few lengthy commemorative bronze inscriptions cast by Shang noblemen (Figure 7.8). At this time awards from the king seem to have prompted nobles to make written reports of their achievements to be read by their ancestors. When we read in a bronze vessel that, on such-and-such a day, in reward for such-and-such a service, the king gave a nobleman gifts (carefully specified), whereupon the latter made the inscribed vessel for his ancestors, we may be reminded that in Mesopotamia and Egypt the commonest type of admin­istrative document was the receipt. More importantly, the action of reporting in writing is the quintessence of developed bureaucracy. The two main functions of administrative documents in Mesopotamia and Egypt were to establish accountability and responsibility. Shang bronze inscriptions that list date, event, and participants seem to be performing just those functions in the particular context of the ancestor cult. This new genre of bureaucratized written display was adopted and developed further under the Zhou dynasty, which conquered the Shang shortly after the genre came into being.

Figure 7.8 A late Shang inscribed bronze, the Xiaozi X you in the collection of the Hakutsuru Bijutsukan, Kobe. The rubbing reproduces an inscription cast in the lid. It reads vertically from top right to bottom left:

On the day yisi [day 42], Zi ordered Xiaozi X to go in advance and deliver people [for sacrifice?] to Han. Zi awarded X two strings of cowries. Zi said: “The cowries are in recognition of your merits.” X thereupon used them to make a vessel for Mother Xin. This was in the tenth month; it was when Zi issued the order to observe Mei [enemy leader] of Renfang [Shang enemy in Shandong].

Zi must have been a high-ranking commander serving one of the last two Shang kings. The latter led two (or even three) campaigns in Shangdong in his tenth and fifteenth year and left many divination records made en route, from which scholars have tentatively reconstructed campaign routes that by and large coincide with the suggested routes of shipping salt shown in Map 7.4. It has been convincingly argued that the generous spacing between the columns and tight stacking of characters within each column reflect a fair copy written on four bamboo slips laid side by side. Photograph after Umehara Sueji, Hakutsuru kikkinshu (Hyogo, Japan: Hakutsuru Bijutsukan, 1934), no. 12 (rubbing after Zhongguo shehui kexueyuan kaogu yanjiusuo [ed.], Yin Zhoujinwenjicheng [Beijing: Zhonghua shuju, 1984], no. 5417-1).

Coda

The city ofYinxu was abandoned after the Zhou conquest around 1050 BCE.Modern cities decline and collapse for economic reasons, according to Jane Jacobs,^[102] but in ancient China the death of major cities normally had political rather than economic causes. Dynastic change and the creation of empires propelled the cyclic renewal of urbanization. The renewals preserved some of the innov­ations of earlier cities, writing above all. The writing system and scribes of the Shang were adopted by the Zhou conquerors, who spread literacy to a wider area of China through their own process of urbanization. Writing permeated political, religious, administrative, military, and cultural life within and outside the cities. It was instrumental in constructing urban and rural identities. Yet its most persistent use seems to have been in the sphere of administration. The beliefs that held the moral community together could, it seems, change in an instant - think of the abrupt disappearance of the large-scale human sacrifice routinely practiced at Yinxu - but literate administration survived all political ups and downs.

further readings

Archaeology of early Chinese cities

Bagley, Robert, “Shang Archaeology,” in Michael Loewe and Edward L. Shaughnessy (eds.), The Cambridge History of Ancient China: From the Origins of Civilization to 221 B.C., Cambridge: Cambridge University Press, 1999, pp. 124-231.

Liu, Li, and Xingcan Chen, The Archaeology of China, Cambridge: Cambridge University Press, 2012.

Steinke, Kyle (ed.), Art and Archaeology of the Erligang Civilization, Princeton, NJ: Princeton University Press, 2014.

Early cities in comparative perspective

Houston, Stephen D., Hector Escobedo, Mark Child, Charles Golden, and Rene Mufioz, “The Moral Community: Maya Settlement Transformation at Piedras Negras, Guatemala,” in Monica L. Smith (ed.), The Social Construction of Ancient Cities, Washington, D.C.: Smithsonian Books, 2003, pp. 212-53.

Kemp, Barry, The City of Akhenaten and Nefertiti: Amarna and Its People, London: Thames & Hudson, Ltd., 2012.

Chinese writing

Bagley, Robert, “Anyang Writing and the Origin of the Chinese Writing System,” in Stephen D. Houston (ed.), The First Writing: Script Invention as History and Process, Cambridge: Cambridge University Press, 2004, pp. 190-249.

Keightley, David, “The Shang,” in Michael Loewe and Edward L. Shaughnessy (eds.), The Cambridge History of Ancient China: From the Origins of Civilization to 221 B.C., Cambridge: Cambridge University Press, 1999, pp. 232-91.

Wang, Haicheng, Writing and the Ancient State: Early China in Comparative Perspective, Cambridge: Cambridge University Press, 2014.

Writing in comparative perspective

Baines, John, John Bennet, and Stephen Houston (eds.), The Disappearance of Writing Systems: Perspectives on Literacy and Communication, London: Equinox Publishing Ltd., 2008.

Houston, Stephen D. (ed.), The First Writing: Script Invention as History and Process, Cambridge: Cambridge University Press, 2004.