C H A P T E R T H I R T E E N THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA Nicola Di Cosmo The northern frontier of China has long been recognized as something more than a simple line separating natural zones, political entities, or ethnic groups.* This frontier has been represented as the birthplace of independent cultures and the habitat of peoples whose lifestyle, economic activities, social customs, and religious beliefs became, from the Bronze Age onward, gradu- ally but increasingly distant from the civilization of the Central Plain. This distinct cultural region, often called the "Northern Zone" of China, com- prises the interlocking desert, steppe, and forest regions from Heilongjiang and Jilin in the east to Xinjiang in the west.1 The frontier between China and the north has also been envisaged as a bundle of routes and avenues of communications through which peoples, ideas, goods, and faiths flowed incessantly between West and East. In economic terms, it provided the Chinese with a source of foreign goods as well as a market for domestic production. The process by which the northern frontier acquired these qualifications was a long one. While its complexities cannot be captured in a single image, the Great Wall — this symbolic and material line that came into existence as * I wish to acknowledge funding received from the Milton Fund, which has made possible the collection of research material and a fact-finding trip to Xinjiang. Among the colleagues that have made suggestions and helped me with gathering sources, I wish to thank Robert Bagley, Emma Bunker, Louisa G. Fitzger- ald Huber, Fred Hiebert, Jessica Rawson, and the archaeologists from Xinjiang: Wang Binghua, Wang Ping, Idris from the Institute of Archaeology, Dr. Zhao from the Society for Altaic Studies (Altai City), and the Hi Cultural Bureau. A special thanks to my friend Cui Yanhu for his help in organizing my research in Xinjiang, and to Yangjin Pak for his knowledgeable assistance. All mistakes are, of course, my own. 1 For uses and definitions of the term "Northern Zone," see Lin Yiln, "A Reexamination of the Relation- ship Between Bronzes of the Shang Culture and of the Northern Zone," in Studies ofShang Archae- ology, ed. K. C. Chang (New Haven, Conn.: Yale University Press, 1986), pp. 257-73. See too William Watson, Cultural Frontiers in Ancient East Asia (Edinburgh: Edinburgh University Press, 1971), p. 63; idem, Inner Asia and China in the Prc-Han Period (London: School of Oriental and African Studies, 1969), p. 16. 885 Cambridge Histories Online © Cambridge University Press, 2008 886 NICOLA DI COSMO a unified system of fortifications with the establishment of the Qin empire in 221 B.C. - can be seen as the culmination of a long process of cultural dif- ferentiation that embraces several aspects. The first concerns the frontier's material culture. As Neolithic communi- ties learned to use and transform their territory, different applications of economic potential and intellectual abilities created cultural, social, and economic differences that are visible through the relics of their civilizations. It was during the Bronze Age that a distinct northern culture emerged. During the latter part of the Shang dynasty (ca. 1200—1050 B.C.), northern China already featured a clearly discernible cultural complex undeniably dis- tinct from that of the Central Plain (Zhongyuan ^JM). This Northern Complex cannot be regarded as a single culture; rather different communi- ties shared a similar inventory of bronze objects across a wide area. This inventory allows us to establish broad connections with bronze civilizations of North Asia, West Asia, and China. Second, human adaptation to an environment more arid than that of the lower Yellow River and Wei tl River valleys, such as that of Inner Mon- golia and Gansu, historically emphasized animal breeding over farming technology. Across the northern frontier a steady transition from an agriculture-based to a pastoral-based economy took place beginning in the late second millennium B.C. While hunting, fishing, and farming remained actively pursued, animal remains indicate the gradual expansion of domesti- cated cattle, sheep, and horses. Domesticated animals were not new, but larger herds demanded new ways of management. Beginning in about the eighth century B.C., throughout Inner Asia horse-riding pastoral communi- ties appear, giving origin to warrior societies. Known by the Greeks as Scythi- ans in the western end of Asia, their cultural expansion was by no means limited to the Pontic steppe to the north of the Black Sea, but extended across the Eurasian steppe belt. Equestrian pastoral peoples, who may be broadly defined as "early nomads," were present in northern China and can be regarded as cultural forerunners of the Xiongnu &}$., the Huns, and the later Turco-Mongol nomads. The use of the horse in war became widespread in northern China well before it was adopted by the Chinese. Finally, in liter- ary records frontier peoples came to represent a sort of alter ego that con- tributed, by providing a contrasting image, to the formation of China's cultural identity. The relationship between China and the nomads appears to have been of secondary importance to Chinese history until it exploded into one of its most critical issues, during the Qin—Han period. The emergence of the Xiongnu empire, in 209 B.C., struck the newly born Chinese empire with unprecedented strength, forcing upon it the realization that the north had Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 887 become a major antagonist, politically, militarily, and culturally. Because of this, scholars have had great difficulties explaining the origin of the Xiongnu and related nomadic cultures in northern China. Countless efforts to identify some of die alien peoples (Xianyun ##£, Rong $c, Qiang ^£, and Di $C in particular) that figure prominently in pre- Han written records with pastoral nomadic cultures have so far failed to yield firm results. Pastoral nomads become historically identifiable only during the late Warring States period. This has created the impression that their sudden appearance on the stage of Chinese history was a product of the creation of a unified Chinese state. However, archaeological evidence has now shown that the presence of nomadic pastoralists on the northern frontier of China was by no means a sudden phenomenon and that the genesis of distinct northern cultures can be traced back to the Shang period, if not earlier. The interface between China and the Northern Zone was a dynamic factor that featured prominently in their respective histories. From the Central Plain the geographical configuration and natural topography of the north could be modified by building roads and fortifications, by subjugating native peoples, and by establishing settlements. On the other side, northern peoples formed independent political entities that entered Chinese history in their own right as military opponents, political allies, and trading partners, stimulating inno- vations and cultural exchanges. However, written sources for this historical phenomenon are hazy at best. As northern peoples emerge from prehistori- cal obscurity, their history can be tentatively reconstructed only through suc- cinct mentions of battles and alliances that supply more questions than answers. Recently, archaeological work has provided a referential framework that allows us to trace the cultural progress of the Northern Zone from the Shang through the Warring States periods. Though the temptation to cloak archaeological finds in historical garb should be avoided, and though the asymmetry between textual and material records leaves room for wide dif- ferences of interpretations, the contours of this broad phenomenon have become less blurred. This chapter, however, cannot be a history of contacts between China and Central Asia over two millennia; neither can it say anything definitive about the evolution of ancient nomadic cultures in today's northern China. Rather, it aims to provide a broad narrative of the northern frontier with a particu- lar question in mind: what was the genesis of the Xiongnu steppe empire as a historical phenomenon? This question, so fundamental for the history of China, is the leading issue that this chapter attempts to address. Cambridge Histories Online © Cambridge University Press, 2008 888 NICOLA DI COSMO PERIODIZATION OF THE NORTHERN ZONE CULTURES The appearance of a distinct Xiongnu culture in northern China is com- monly dated to the end of the fourth century B.C. Western scholars have often attributed it to patterns of continental migrations of Central Asian pastoral nomads, at times even identified linguistically as Altaic or Indo- European. By contrast, the traditional Chinese ethnogenealogy of the Xiongnu, traceable back to the Shi ji, refers to them as the descendants of ancient northern peoples.2 Archaeological studies in China have by and large emphasized continuity and gradual evolution over sudden ruptures and influences from Central Asia, and Chinese scholars insist that the origin and evolution of the Xiongnu people be placed within the context of the autochthonous formation of pastoral cultures in northern China. Neverthe- less, though detailed evidence is still wanting, connections between the Northern Zone and South Siberia, in particular the Karasuk and Tagar cul- tures, are beyond doubt, and it is safe to say that the evolution of pastoral cultures in northern China, while displaying an original and distinctive history, was not an isolated phenomenon. The Northern Zone cultural complex was an active participant in the continent-wide evolution of a North Eurasian pastoral nomadic culture whose fundamental character and premises it shared with other complexes across the Eurasian steppe and forest regions. Its position within this process is, however, not clear. Given the uncertainty of the connections, caution requires us to limit the scope of this study to the Sino-Mongolian frontier zone (Shanxi, Shaanxi, Hebei, northern Henan, and Inner Mongolia), flanked by northeastern (Liaoning and Heilongjiang) and northwestern extensions (Gansu, Ningxia, parts of Qinghai and Xinjiang). Within this broad Northern Zone, archaeological remains reveal the existence of regional and local features according to period and territorial distribution. Various efforts have been directed at showing the internal coherence of the evolution of the Northern Zone, or "Ordos," culture. Excavations carried out in Inner Mongolia and adjacent areas have shown that the presence of pas- toralists in the Northern Zone can be dated to the Shang period and can be divided into three stages of development: Bronze Age (Shang, Western Zhou, and Spring and Autumn periods); Early Iron Age (Warring States period); Iron Age (Former and Latter Han). Stages prior to the late Warring States period are referred to, by Chinese scholars, as early or proto-Xiongnu Shiji, no (Beijing: Zhonghua shuju, 1985 [1959]), pp. 2879-82. This "genealogy" was rejected by both Bernhard Karlgren, "Some Weapons and Tools of the Yin Dynasty," BMFEA, 17 (1945): 141; and Otto Maenchen-Helfen, "Archaistic Names of the Hiung-nu," Central Asiatic Journal 6 (1961): 248-61. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA cultures.3 An alternative four-phase periodization, based on bronze typology, includes a first phase, the late Shang and early Western Zhou (thirteenth to ninth centuries B.C.), with bronzes characterized by dagger and knife termi- nals in the shape of animal and bird heads; a second phase, from the late Western Zhou to the mid-Spring and Autumn period (ninth to sixth centuries B.C.), represented by the bronzes of the Upper Xiajiadian MMfB culture; a third phase, from the late Spring and Autumn to the end of the Warring States (sixth to third century B.C.), characterized by rectangular bronze plaques with animal designs in relief or openwork, and by animalfiguresin the round; and a fourth phase, during the Former and Later Han dynasties, also featuring rec- tangular belt ornaments with animal motifs in openwork.4 This first periodization compresses approximately ten centuries (1500—500 B.C.) within the same category of "Bronze Age," thus obscuring the variety and complexity of socioeconomic, technological, and cultural changes that took place in the Northern Zone during that period. Moreover, both peri- odizations tend to convey the impression that such a development was wholly endogenous and to discard the extensive contacts between the Northern Zone and Central Asia as a powerful stimulus to change and development. Finally, these periodizations emphasize the material culture almost exclu- sively, with socioeconomic and historical considerations playing only ancil- lary roles. An approach that combines social and economic considerations and information drawn from historical records may provide a fuller appreci- ation of cultural progress in the Northern Zone. This progress can be divided into four periods. The Second Millenium B.C. A frontier, understood as a geographical area penetrated by cultures with dif- ferent characteristics and cutting across various political and social units, can be discerned prior to and during the Shang dynasty, as some cultures of the Northern Zone are believed to have already been pastoral, though probably not yet engaged in pastoral nomadism, that is, a socioeconomic system based on a fixed migratory cycle and overwhelmingly dependent on animal prod- ucts. More likely, the earliest representatives of the Northern Zone complex were mixed communities of shepherds and farmers that also practiced exten- sive hunting. At this early stage the frontier appears to have been a broad ' Tian Guangjin, "Jinnianlai Nei Menggu diqu de Xiongnu kaogu," Kaogu xuebao 1983.1: 7-24; idem, "Taohongbala de Xiongnu mu," Kaogu xuebao 1976.1: IJI. 4 Wu En, "Woguo beifang gudai dongwu wenshi," Kaogu xuebao 1990.4: 409—37. See also Emma Bunker, "Ancient Ordos Bronzes," in Ancient Chinese and Ordos Bronzes, ed. Jassica Rawson and Emma Bunker (Hong Kong: Museum of Art, 1990), pp. 291—307. Cambridge Histories Online © Cambridge University Press, 2008 890 NICOLA DI COSMO belt of cultural transition between the Shang civilization and the bronze cul- tures of Central Asia and South Siberia. Though northern peoples displayed different cultural traits, such as distinct ceramic objects, burial customs, habitations, and economic structures, certain characteristic elements, such as bronze weapons, became widely adopted throughout the Northern Zone, and this common metallurgical culture was different and independent from that of the Central Plain. The date of 1000 B.C. seems a suitable chronolog- ical marker to end this phase. The Zhou conquest caused a political realign- ment of the Central Plain, which affected foreign relations. The Zhou house had a very different relationship with some of the earlier enemies of the Shang; for instance, they intermarried with the Jiang 31 clan, probably related to the Qiang.5 Moreover, the early Zhou kings achieved a period of peace on the borders. Western Zhou to Early Spring and Autumn (ca. 1000-650 B.C.) During the Western Zhou, the appearance of a northeastern cultural complex characterized by advanced metallurgical techniques marked an expansion of the Northern Zone, in particular toward the east (Heilongjiang, Jilin, Liao- ning). The Western Zhou court carried out intense political and military activities in the north, east, and west against peoples whose archaeological identification is still debated. Among the multitude of alien peoples, at times friendly and more often hostile, the funerary inventories are consistent with the emergence of aristocratic warrior elites and a broad productive basis, whose main force was represented by increasingly more specialized pastoralists. Mid-Spring and Autumn to Early Warring States (ca. 650-350 B.C.) A third phase of the history of the northern frontier can be dated from the middle or late seventh century to the mid-fourth century. This period wit- nessed the appearance of fully developed steppe pastoral nomads who may be regarded as the cultural ancestors of the Xiongnu and can be termed "early nomadic." Early nomadic cultures spread in the Eurasian steppe beginning from the eighth century B.C. and lasted through the third century. It is inter- esting that approximately the same chronology applies to the Scythians, known from Herodotus's Histories as the powerful nomadic warriors of the North Pontic steppe, and to the eastern steppe cultures variously termed On the relationship between Jiang and Zhou see Wang Zhonghan, ed., Zhongguo minzu shi (Beijing: Zhongguo shehui kexue, 1994), p. 125. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 89I Altaic Scythian,6 Scytho-Siberian, or Saka. Even though there were funda- mental ways in which nomadic groups over such a vast territory differed, the terms "Scythian" and "Scythic" have been widely adopted to describe a special phase that followed the widespread diffusion of mounted nomadism, characterized by the presence of special weapons, horse gear, and animal art in the form of metal plaques. Archaeologists have used the term "Scythic continuum" in a broad cultural sense to indicate the early nomadic cultures of the Eurasian steppe.7 The term "Scydiic" draws attention to the fact that there are elements - shapes of weapons, vessels, and ornaments, as well as lifestyle - common to both the eastern and the western ends of the Eurasian steppe region. However, the extension and variety of sites across Asia makes Scythian and Scythic terms too broad to be viable, and the more neutral "early nomadic" is preferable, since the cultures of the Northern Zone cannot be directly asso- ciated with either the historical Scythians or any specific archaeological culture defined as Saka or Scytho-Siberian.8 Mentions of "proto-Xiongnu" cultures should also be avoided to eschew inferences of a close genetic rela- tionship with any single group of nomads that would appear centuries later, since no such relationship can be confirmed to date, and since the early nomads of the Northern Zone could just as well be claimed to be the pre- decessors of other nomadic groups, such as the Dong Hu ^L#!, Lin Hu Pf~ #3, and Wusun ,%%&. These early nomadic communities achieved a dominant position through- out the Eurasian steppe region, and their metal culture displayed a singular uniformity of features. In the Northern Zone, the transition to pastoral nomadism occurred gradually. The Saka culture in Xinjiang, the Shajing 0? # culture in Gansu, the Ordos complex in Inner Mongolia, and the Upper Xiajiadian culture of Liaoning, all point to a transition from mixed agropas- toral to predominantly or exclusively pastoral nomadic cultures. From the seventh century onwards, objects related to improved horse management and horse riding, such as the bit, cheekpieces, horse masks, and bell ornaments, became ever more widespread and sophisticated. Iron metallurgy is likely to 6 See, e.g., S. I. Rudenko and N. M. Rudenko, hkusstvo skifov Altaia (Moscow: Gos. muzeia izobrazi- tel'nikh iskussrv im. A. M. Pushkina, 1949); M. I. Artamanov, "Skifb-sibirskoie iskusstvo tsverinogo stili (osnovniie etapy i napravleniia)," in Problemy skifikoi arkheologii, ed. P. D. Liberov and V. I. Guliaev (Moscow: Nauka, 1971), pp. 24-54; and idem, Sokrovishcha Sakov: Amu-Dar'inskii klad, altaiskie kurgany minusinskit bronzy, sibirskoe zoloto (Moscow: Iskusstvo, 1973). 7 William Watson, Cultural Frontiers in Ancient East Asia, p. 101; Roman Kenk, Crabfunde der Skythen- zeit aus Tuva, Siid-Sibirien (Munich: Beck, 1986); and E. A. Novgorodova et al., Ulangom: Ein skythen- zeitliches GrdberfeUl in der Mongolei (Wiesbaden: Harrassowitz, 1982). ' For uses of the term "early nomads" for the eastern" steppe people of the Scythian period, see Mikhail P. Gryaznov, The Ancient Civilization of Southern Siberia (New York: Cowles, 1969), chapter 3, "The Age of the Early Nomads"; and idem, Der Grofikurgan von Arzan in Tuva, Sudsibirien (Munich: Beck, 1984). pp. 76-7- Cambridge Histories Online © Cambridge University Press, 2008 892 NICOLA DI COSMO have spread in the Northern Zone before its general appearance in the Central Plain, with iron objects found in Xinjiang in the ninth century B.C. and in Inner Mongolia and the northeast by no later than the mid-seventh century B.C. The changes that were taking place in the north, however, were not imme- diately noticeable to historians and chroniclers in China. This was due in part to the mode of historical writing itself, not yet mature enough to be concerned with anthropological and ethnographic detail. But it is also pos- sible that the nomads inhabited a "deep" frontier not yet discovered by the people of the Central Plain. The presence of mixed archaeological remains on the borders of China, and the lack of reference to mounted pastoralists in the Chinese sources, probably indicate that alien communities of mixed shepherds and agriculturalists acted as a buffer between the Central Plain and the nomadic lands. Late Warring States to Qin (ca. 350—209 B.C.) The fourth period starts in the late Warring States and leads to the founda- tion of the Xiongnu state in 209 B.C. In 307 B.C. King Wuling of Zhao SS S£1U, influenced by the nomads, introduced cavalry into his army. The famous debate that accompanied this event provides the first historical information on pastoral nomads - known as Hu #3 - along the northern borders of China. The reduction of the number of the contending states in China to a few strong ones, and the northward expansion of Qin, Zhao, and Yan ?!&, resulted in the assimilation of several peoples, such as the Rong and the Di, within the orbit of the Zhongyuan. This eliminated the protective screen that those people had supplied between China and the unambigu- ously nomadic, warlike groups further north, who suddenly appear in the Chinese sources beginning at this time. It is in this phase that China started to develop a knowledge of Inner Asia and a deeper understanding of the surrounding cultures. Even though, during the last phase of the Warring States, Hu nomads and Xiongnu played a relatively minor role in the history of the Central States, nevertheless their presence and activity on the Chinese northern border cannot be denied. The frontier reached its historical definition with the creation of the Great Walls, which separated the northern states of Qin, Zhao, and Yan from the nomads. Wall building was not limited to the northern regions, but the northern walls and fortresses remained and were strengthened after Qin's unification of the realm, whereas fortifications between states disappeared. Qin continued to expand northward after 221 B.C. The colonization of the Ordos region led by General Meng Tian HMS in 215 B.C. caused a widespread Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 893 dislocation of Xiongnu tribesmen from their ancestral land, the ensuing political and economic crisis creating among the Xiongnu the conditions for the rise of a new leadership. This led to the formation of a powerful north- ern empire, founded by Maodun H$I in 209, just as China was about to plunge into a civil war. When the newly born Han dynasty came to confront the Xiongnu, it found that the balance of power had been reversed. The Xiongnu were able to impose their rule in the north, the Great Wall being declared the frontier between the people "with bows and arrows" and those "with hats and girdles." THE NORTHERN FRONTIER IN THE SECOND MILLENNIUM B.C. The Northern Zone Complex: Defining Characteristics A cultural "complex" with characteristics unquestionably different from those of the Central Plain emerged during the Shang period in the Northern Zone. This term should not suggest a homogeneous culture, but rather a broad area in which different people shared certain common traits, in particular their bronze inventory. This metallurgical tradition typifies the north and marks the cultural boundary with the civilization of the Central Plain. Most char- acteristic of the Northern Zone complex are bronze weapons, probably indi- cating that the development of metallurgy was linked to the rise of military elites and to increased warfare — possibly resulting from competition for eco- nomic resources. Among the weapons, the most representative are daggers, knives, axes, mirrors, and a "bow-shaped" object. DAGGERS. Daggers, or "short swords," are characterized for the most part by the integral casting of hilt and double-edged blade and by a relatively narrow and straight hand-guard (Fig. 13.1a). The early types, dated to the middle and late period of the Shang dynasty, display a characteristic curved hilt, often decorated with geometric designs, featuring a terminal in the shape of an animal head, usually a horse, ram, eagle, or ibex. Others have perfo- rated hilts, or straight hilts with grooves ending in a rattle. KNIVES. Whereas Shang bronze knives normally have short stems inserted into a handle of a different material, all northern-type bronze knives of this period have an integrally cast hilt (Fig. 13.1b). The spine of the knife has an arched shape and is wider than the grip. Between the blade and the grip there is often a small tongue. Pommels come in many shapes; the most character- istic are the mushroom, animal head, and various shapes of rings, or loops. Handles are decorated with geometric motifs similar to those of the daggers. Cambridge Histories Online © Cambridge University Press, 2008 894 NICOLA DI COSMO Figure 13.1. Northern Zone bronzes of the second and early first millennium B.C. A: Bronzes of the Karasuk culture, Southern Siberia. B, C, D: Respectively, daggers, knives, and axes of the Northern Zone complex. After: Wu En Jl,®. "Yin zhi Zhou chu de beifang qingtong qi" Kaogu xuebao 1985.2, figs. 1, 2, 3, 5. Drawings by Li Xiating. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 895 AXES W I T H TUBULAR SOCKETS. The axe's blade is typically long and thick, with a relatively narrow cutting edge clearly different from the fan- shaped axe of the Shang. Their main characteristic, however, is a tubular socket set perpendicularly to the blade (Fig. 13.1c and d). In early axes, the socket can be longer than the width of the body. This hafting system is very different from the predominant Shang method of attaching the handle to a protruding flat tang. Tubular axes have been found in Hebei (Chaodaogou SHttJtf, Qinglong Wfl county), in Shanxi, (Gaohong IRJ«I, Liulin WbW county, and Chujiayu tfaMMi, Shilou 5 f t county), and at Shang sites, such as Dasikong ^CiUS:.9 In Gaohong two axes were found together with a dagger with a rattle pommel, a spearhead, a helmet, three knives each with a double-ring head, and other small objects.10 M I R R O R S . Round bronze disks, usually defined as mirrors in Chinese archaeology, are also part of a northern heritage (Fig. 13.id). Typically they have a smooth surface on one side; on the other, which may carry surface decoration, they have a central knob handle. A Qinghai mirror decorated on the back with a star-shaped design suggests a solar cult, possibly of Central Asian origin. Mirrors found in Anyang tombs, such as those in Fu Hao's #§ £F tomb, have decorative motifs that are not consonant with the artistic vocabulary of the Shang.11 Other mirrors found in Shang burials in central Shaanxi together with a ding $\A vessel, curved knives, and gold earrings, suggest associations with non-Chinese cultures.12 Finally, a mirror has been found in another burial together with two bronze jue Wt vessels with the char- acter Qiang H inscribed on them.13 This evidence connects the mirror to a distinctive northern culture, and possibly to the Qiang people. Only in the mid-seventh century did the bronze mirror become part of the Chinese native tradition. " B O W - S H A P E D " OBJECT. This curious object is comprised of a slightly bent decorated central bar and curved lateral sections. Various hypotheses have been adduced as to its use (Fig. 13.id).14 Found in the Yinxu Ixffi culture ' Kaogu 1962.12: 644-65; Wu Zhenlu, "Baode xian xin faxian de Yindai qingtong qi," Wenwu 1972.4: 62-6; Yang Shaoshun, "Shanxi Liulin xian Gaohong faxian Shangdai tongqi," Kaogu 1981.3: 211-12; Wenwu 1981.8: 49—53; Ma Dezhi et al., "Yijiuwusan nian Anyang Dasikong cun fejue baogao," Kaogu xuebao 1955.9: 25-90. 10 Yang Shaoshun, "Shanxi Liulin xian Gaohong faxian Shangdai tongqi." " See pp. 194-202 above, "The Tomb of Fu Hao"; Diane M. O'Donoghue, "Reflection and Reception: The Origins of the Mirror in Bronze Age China," BMFEA 62 (1990): 5—25. " Yao Shengmin, "Shaanxi Chunhua xian chutu de Shang Zhou qingtongqi," Kaogu yu wenwu 1986.5:12—22. " Wenwu 1986.11: 7,fig.11.5. M Karlgren, "Weapons and Took," p. 112, is probably wrong in considering it a kind of yoke, but right in rejecting the old hypothesis that it was a "banner bell." Max Loehr, "Weapons and Tools from Cambridge Histories Online © Cambridge University Press, 2008 896 NICOLA DI COSMO with end rattles and horse heads, and in the Minusinsk region with a much simpler decoration of knobs, it was probably invented in the Northern Zone and thereafter transmitted to both China and South Siberia. Other objects also regarded as characteristic of this particular culture are a distinctive type of spoon (Fig. 13.id) and helmet. The spoons have rings on the handle with attached pendants. The helmets are undecorated, the sides coming down to cover the ears; they have a ring on top and holes on the bottom to the right and left.1' ANIMAL STYLE. The term "Animal Style" indicates a decorative style and an artistic tradition shared across the Northern Zone from the thirteenth century onwards. Common both to the Karasuk culture of South Siberia (1200-800 B.C.) and the early nomadic cultures of Central Asia, the style consists of various representations of animals on bronze vessels, weapons, and tools. At this early stage the Animal Style is expressed in the Northern Zone mainly in the form of ornamental animal heads in the round attached to the end of knife handles and dagger hilts.16 Geographic Distribution During the Shang, and in particular the late Shang period, the Northern Zone included the territory of northern Shaanxi, Shanxi, and northern Henan; in the east it reached as far as the Liaodong coast; in the north it reached western Liaoning and Inner Mongolia; and in the west it extended to Gansu, Ningxia, and Xinjiang (Map 13.1). That elements of the bronze culture of the Northern Zone are present as far as Transbaikalia, Mongolia, the Altai region, South Siberia (Minusinsk River basin), and Tuva is evidence of the extraordinary reach of this cultural complex. In recent years the exca- vation of archaeological sites in North China allows us to identify a number of cultures or cultural features located within the Northern Zone complex. Anyang, and Siberian Analogies," American Journal ofArchaeology 53 (1949): 138, relates it to bows and quivers. However, the "bow-shaped" object appears commonly on so-called deer stones - anthropo- morphic steles with carvings representing stylized deer — attached to the belt. On this basis, Lin Yiin, "A Reexamination," p. 263, suggests that it was used as a "reins-holder" by drivers of horse-drawn char- iots and horse drivers. Yet the iconography of the deer stones does not show it in combination with chariots or horses. See also Qin jianming, "Shang Zhou 'gongxingqi' wei 'qiling' shuo," Kaogu 19953: 256-8. " A. Kovalev, " 'Karasuk-dolche', Hirschsteine und die Nomaden der chinesischen Annalen im Alterum," in Maoqinggou: Ein eisenzeitlkhes Grdberfeld in der Ordos-Rcgion (Innere Mongolei), ed. Thomas Hollman and Georg W. Kossack (Mainz: Philipp von Zabern, 1992), pp. 60—1. " On the animal style the following works may be consulted: Karl Jettmar, Art of the Steppes (New York: Crown, 1967); and Anatoly Martinov, The Ancient Art of Northern Asia, trans, and ed. Demitri B. Shimkin and Edith M. Shimkin (Urbana: University of Illinois Press, 1991). On the origin and earli- est occurrences of the Animal Style, see Yakov A. Sher, "On the Sources of the Scythic Animal Style," Arctic Anthropology 25, 2 (1988): 47-60. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 897 7T Northern Zone Sites of the Second Millennium B.C. Map 13.1. Northern Zone: Archaeological sites of the second millennium B.C. These are the Lower Xiajiadian culture, the northern-style bronzes, and Bai- jinbao fijfeU (Heilongjiang) culture in the northeast; the Zhukaigou 5fclH| $f and Chaodaogou cultures, together with mixed Shang and Northern Zone sites, in the north-central sector; and the Qijia WM, Xindian ¥$5, and Siwa ^F'& cultures in the northwestern portion, including present-day Gansu and Ningxia provinces. LOWER XIAJIADIAN. Approximately contemporary with the last phases of the Zhukaigou culture, but located further east, is the Lower Xiajiadian culture (ca. 2000-1300 B.C.), essentially a pre-Shang culture, though partially overlapping with the early Shang. Lower Xiajiadian sites extend across south- eastern Inner Mongolia, Liaoning, and northern Hebei, and reveal a Bronze Age culture at the initial phase of transition toward metalworking. People lived in settlements, the economy of which was firmly agricultural, based on millet. This was supplemented with hunting of deer and stock raising, as shown by remains of sheep, cattle, and, in particular, pigs. Metalwork was limited to small objects, such as rings, knives, and handles. In contrast, stone and bone production attained a high level. The southern limit of the Lower Cambridge Histories Online © Cambridge University Press, 2008 898 NICOLA DI COSMO Xiajiadian culture is located in Hebei (Yixian JaU and Laishui ISftzR coun- ties). The whole Beijing region, however, forms a large belt where the Lower Xiajiadian and the Shang cultures met.17 NORTHEASTERN BRONZES. Bronzes found in the northeast that are typo- logically similar to the Northern Zone bronzes attributed to the Shang and early Western Zhou periods have been found mostly in cache and storage pits; they consist primarily of weapons: battle-axes, socketed axes, knives, and daggers. These findings are concentrated in western Liaoning; the eastern part of Liaoning seems to have been marginal to the distribution of northern-type bronzes. In eastern Liaoning, together with knives, socketed axes, battle-axes, and a distinctive type of socketed dagger, only one type of northern-style dagger has been found, with hollow hilt and blade in a single casting. Socketed daggers are not found in the rest of the Northern Zone. The battle-axes with long and narrow sockets and straight blades are similar to those found in Chaodaogou. Unusual items, which nevertheless belong to the same Northern Complex, are a dagger hilt that terminates in the shape of a human head, as well as vessel lids and chariot ornaments. In the north- east, as in the rest of the Northern Zone, bronzes are not associated with pottery.18 ZHUKAIGOU. The importance of the Zhukaigou culture lies in its role as the reputed progenitor of the "Ordos bronze culture" and, by extension, as the first Northern Zone culture. It extended to northern and central Inner Mongolia, northern Shaanxi, and northern Shanxi, with the Ordos region at its center.'9 Bronze objects dated to the last period of its existence (ca. 1500 B.C.) have been excavated. These point to the indigenous production of typical Northern Zone items, such as bronze daggers, together with typical Shang dagger-axes (ge %) and integrally cast knives with terminal ring and upward-turned point that reveal both Shang and northern features. The people of Zhukaigou were agriculturalists, with a main staple of millet. They also raised sheep, pigs, and cattle. The transition to metalworking occurred around the end of the third millennium, which also coincides with a higher level attained in the ceramic industry. It is in this period that certain motifs appear, such as the snake pattern and the flower-shaped edge of the // & 17 Li Jinghan, "Shilun Xiajiadian xiaceng wenhua de fenqi he leixing," in Zhongguo Kaogu xuihui diyi ci 0979) nianhui lumuenji (Beijing: Wenwu, 1980), pp. 163—70. 18 Zhai Defang, "Zhongguo beifang diqu qingtong duanjian fenqun yanjiu," Kaogu xuebao 1988.3: 277—99. O " t n e distribution of bronze cultures in Liaoning, see Xu Yulin, "Liaoning Shang Zhou shiqi de qingtong wenhua," in Kaoguxue wenhua lunwenji, ed. Su Bingqi (Beijing: Wenwu, 1993), vol. 3, pp. 311-34. '' Kaogu 1988.3: 301—32; Wu En, "Zhukaigou wenhua de faxian ji qi yiyi," in Zhongguo kaoguxue luncong (Beijing: Kexue, 1995), pp. 256-66. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 899 vessel, the latter of which archaeologists regard as characteristic of later nomadic peoples of this area.20 Interestingly in this area in die first half of the second millennium, people already used oracle-bone divination, a prac- tice that came to be closely associated with Shang culture and statecraft. Shang ritual vessels, such as ding and jue, and weapons appear here during the Erligang and Erlitou periods." This may suggest that around the mid- second millennium B.C., there was a northward movement of Shang culture or that contacts between the local people and the Shang increased. CHAODAOGOU CULTURE. The Chaodaogou culture," located between the bend of the Yellow River and the Liao IS River drainage basin, extends across northern Hebei, Shanxi, Shaanxi, and Henan. It is characterized mostly by funerary sites that have yielded identical or closely related objects, among which the most characteristic are bronze weapons, in particular daggers, knives, and axes. The type site is Chaodaogou (Qinglong WII county, Hebei), excavated in May 1961.23 The bronzes found here have con- tributed greatly to defining the Northern Zone as a distinct cultural complex. These are a dagger with decorated handle and ram-head pommel, an axe with tubular socket, and four knives each with arched back and a pommel in the form of a rattle, while a fifth knife has a ram-head knob. Another typical site of this culture is Linzheyu #JS|ll® (Baode i&W* county, Shaanxi), a burial ground that yielded a dagger with grooved hilt and rattle pommel, bronze plaques with spiral designs, small rattles in bronze, small bells, horse harness, two axes with tubular sockets, as well as bronze ritual vessels.24 BAIJINBAO CULTURE. Roughly contemporary with the Chaodaogou culture is the Baijinbao culture in Heilongjiang, which spans the late second and the early first millennium.2' The type site of this culture is located in Zhaoyuan $£M county, Heilongjiang, and was excavated in 1974. Centered in the plain of the Sungari-Nonni River system, its distribution extends west 20 Tian Guangjin and Guo Suxin, "E'erduosi shi qingtong qi de yuanyuan," Kaogu xuebao 1988.3: 260. On the snake motif, see Li Shuicheng, "Zhongguo beifang didai de shewenqi yanjiu," Wenwu 1992.1: 50-7. 11 See Katheryn Linduff, "Here Today and Gone Tomorrow: The Emergence and Demise of Bronze Pro- ducing Cultures Outside the Central Plain." Paper presented at the Conference on Chinese History and Archaeology, Academia Sinica, Taipei, Taiwan, January 4—8, 1994. 11 For the definition of Chaodaogou as an archaeological culture, see Kovalev, " 'Karasuk-dolche', Hirschsteine und die Nomaden der chinesischen Annalen im Alterum," pp. 48—61. *' Kaogu 1962.12: 644—5. M Wu Zhenlu, "Baode xian xin faxiande Yindai qingtong qi," Wenwu 1972.4: 62-6. *' On the Baijinbao type site, see Kaogu 1980.4: 311—24. On the relationship between Baijinbao and the Neolithic Hanshu 9 I # culture, see Du Xingzhi, "Shi lun Hanshu wenhua he Baijinbao wenhua," Beifang wenwu 1986.4: 21—5. For a general survey of the Baijinbao culture, see Jia Weiming, "Guanyu Baijinbao leixing fenqi de tansuo," Beifang wenwu 1986.1: 11-15, 52. Cambridge Histories Online © Cambridge University Press, 2008 9OO NICOLA DI COSMO to the border between Inner Mongolia and Heilongjiang, south to Baicheng E=J$t county in Liaoning, north to Ang'angxi ^pt^'iH, Fuyu H-fS, and Nen- jiang $&.tL counties in Heilongjiang, and east to Harbin, Bin county %ff>> and Bayan Bji: county.26 Bronze objects are typically small and include knives, buttons, rings, and earrings. Casting molds have also been found. Two semisubterranean houses excavated at Baijinbao indicate settled life. Some tools are made of stone (polished axes, adzes, and scrapers), but most are made of bone or shell. The abundance of fishing- and hunting-related tools, such as harpoons, spears, projectile points, knives, and scrapers, indi- cate that fishing and hunting were the main activities of this culture, as we would expect in a region rich in rivers and forests. The chronology of the culture is based on the pottery, such as li I?I tripods, hu si jars, and guan ^ pots. The Baijinbao ceramics are characteristically embellished with a decor executed in dotted lines that includes geometric designs or animals, such as frogs, sheep, and deer. The finding of a typical Baijinbao guan pot in an Upper Xiajiadian (eleventh to fourth centuries B.C.) site in southeastern Inner Mongolia points to close relations between these two areas in the early Western Zhou period.27 The latest date of the Baijinbao culture is believed to be the early Spring and Autumn period (eighth to seventh centuries B.C.). NORTHERN CULTURES IN THE GANSU~NINGXIA REGION. T h e Gansu—Ningxia region has had a long and rich history of archaeological investigation. European surveys in the 1920s and 1930s by Sven Hedin, Teil- hard de Chardin, and, especially, J. Gunnar Andersson recognized a number of early cultures and established an initial chronology for the cultures of Qijia, Yangshao WaS (Machang -^i^), Xindian, Siwa, and Shajing. Later Chinese excavations, in the 1930s and 1940s, however, led to the recognition of the Qijia as an Early Bronze Age culture.28 Throughout the 1950s and 1960s extensive surveys and excavations were conducted in the Weihe fSM, Jinghe MM, Taohe MM, and Huangshui flzK River valleys. Further studies in the 1970s and 1980s synthesized the material collected thus far and tried not only to establish a better chronology, but also to assess the extent of contacts with neighboring regions.29 The Qijia culture is regarded as one of the earliest bronze cultures and is dated as early as 2000 B.C. Though its main sites are located in present-day 16 Tan Yingjie and Zhao Shandong, "Song Nen pingyuan qingtong wenhua chuyi," Zhongguo kaogu xuehui di si ci nianhui lunwenji ipSj (Beijing: Wenwu, 1985), p. 199. " Jia Hong'en, "Wengniute qi Dapaozi qingtong duanjian mu," Wenwu 1984.2: 50-4. 28 On these cultures see K. C. Chang, The Archaeology of Ancient China, 4th ed. (New Haven, Conn.: Yale University Press, 1985), pp. 138-50. " Shui Tao, "Xibei diqu shiqian kaogu yanjiu de huigu yu qianzhan," in Kaoguxue wenhua lunji, ed. Su Bingqi (Beijing: Wenwu, 1993), vol. 3, pp. 1-11. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 9OI Gansu province, such as in Huangniangniangtai §L$Ll)$.ci, Qinweijia ijiiJt W-, and Dahezhuang ^cf°T^±, its distribution is very broad, reaching north and east up to Inner Mongolia, the upper Yellow River valley, and the upper Weihe and Huangshui River valleys.30 It was a sedentary culture, based on agriculture. Pig raising seems to have been important, given the relevance of the pig in sacrifices. Oracle divination was also practiced. An important feature of the Qijia culture is the presence of numerous domesticated horses. According to recent scholarship some of the metal artifacts recovered from Qijia sites, in particular knives and axes, might point to a connection with Siberian and Central Asian cultures, in particular the Seima—Turbino complex.31 Though these connections are still hypothetical at present, it is plausible that future research may further corroborate the existence of early contacts between the Qijia culture and Central Asia. During the second part of the Shang dynasty, three almost contemporary and intersecting cultures appeared in Gansu, Ningxia, and northern Qinghai, known as Siwa, Xindian, and Kayue -£$). These cultures, succeeding the Early Bronze Age Qijia, show the presence of more advanced bronzes, in par- ticular weapons. The Xindian culture dates between the first half of the second millennium and the late Western Zhou period, and is regarded as later than Machang but earlier than Siwa. Its area of distribution is the Hehuang yBJ?M River valley and the eastern part of Gansu. It originated in the late Qijia, with which the pottery bears evidence of cultural continuity. The early Siwa sites are dis- tributed in the same area. They were contemporary neighboring cultures, located in close proximity, though each followed its own discrete develop- ment. The Xindian culture later expanded toward the west and came closer to the Kayue culture, by which it was possibly absorbed.32 The Siwa culture is dated to the middle and late second millennium; its eastern expansion may have resulted in a close interaction with the Zhou culture. It takes its name from the site of Siwashan TF'^LLJ (in Lintao E6$l: county, Gansu) and was discovered in 1924.33 An important site is Xujianian %£MM (in Zhuanglang $±'M county, eastern Gansu), a large cemetery where 104 graves were excavated, 7 of which contained human sacrifices. Two *" On the Qijia sites and their distribution, see Hu Qianying, "Shilun Qijia wenhua de butong leixing ji qi yuanliu," Kaoguyu wenwu 1980.3: 77—82, 33; Xie Duanju, "Shilun Qijia wenhua," Kaoguyu wenwu 1981.3: 79-80. 11 Louisa G. Fitzgerald Huber, "Qijia and Erlitou: The Question of Contacts with Distant Cultures," EC 20 (1995): 17-67; see also An Zhimin, "Shilun Zhongguo de zaoqi tongqi," Kaogu 1993.12: nio-19. 11 Zhang Xuezheng et al., "Xindian wenhua yanjiu," in Kaoguxue wenhua lunji, ed. Su Bingqi (Beijing: Wenwu, 1993), vol. 3, pp. 122—52. " K. C. Chang, The Archaeology of Ancient China, pp. 384—5; An Zhimin, "The Bronze Age in the Eastern Parts of Central Asia," in History of Civilizations of Central Asia. Vol 1: The Dawn of Civilization: Ear- liest Times to 700 B.C., ed. A. H. Dani and V. M. Masson (Paris: UNESCO, 1992), pp. 327-8. Cambridge Histories Online © Cambridge University Press, 2008 9O2 NICOLA DI COSMO chariot and horse pits have also been found. The weapons recovered include dagger-axes, spearheads, and knives, all of whose shapes are similar to Western Zhou types.34 Relationship with the Shang Civilization Much evidence of contacts between the Shang and the Northern Zone comes from the discovery of Northern Zone bronzes in Shang tombs excavated in the Anyang area. These include a bronze knife with animal-head pommel found at Houjiazhuang •fUs^IE, another knife with a ring head together with a Shang pickaxe found at Xiaotun /JNlti, and a pickaxe with shorter tubular socket unearthed together with a clay tripod and a piece of jade in 1953 at Dasikong." Some of the most important evidence of contact between the Shang and northern cultures comes from the tomb of Fu Hao, the consort of King Wu Ding (ca. 1200 B.C.), excavated in 1976. It contained several items, such as a northern-style knife with an ibex head, that certainly did not originate in the Shang culture. There are also four bronze mirrors and a bronze hairpin with no equivalent in the Central Plain. Through laboratory "fingerprinting," a connection has been established between a large number of the jades found in the tomb of Fu Hao and nephrite sources in Xinjiang. Since opinions differed as to the provenance of the 755 jade objects exca- vated, many exquisitely crafted, over 300 pieces were sent to various labora- tories in Beijing and Anyang, including that of the Chinese Academy of Social Sciences. Examination is reported to indicate that all but 3 pieces came from quarries in Xinjiang.36 Outside Anyang there are several other sites with mixed findings. For instance, in Yantou $SH (Suide tSiW* county, Shaanxi) a Northern Zone knife has been found together with Shang bronze vessels and a ge dagger-axe, and a Western Zhou burial in Baifu S'/? (Changping i H ^ county, Beijing) yielded a northern-style arched-back knife with geometric patterns on the handle and a ring pommel with three small knobs.37 During the Yinxu period there was increased warfare between the Shang and the Northern Zone. By this period, the Northern Zone complex had evolved into a number of dis- crete cultural centers, which had established a network of contacts both between them and with the Central Plain. M Kaogu 1982.6: 584-90. " Li Ji, "Ji Xiaotun chutude qingtong qi," Zhongguo kaogu xuebao 4 (1949): 1—70. These were found in tomb M164. Kovalev, " 'Karasuk-dolche', Hirschsteine und die Nomaden der chinesischen Annalen im Alterum," p. 54. 16 Wang Binghua. "Xi Han yiqian Xinjiang he Zhongyuan diqu lishi guanxi kaosu," in Wang Binghua, Sichou zhi lu kaogu yanjiu (Urumqi: Xinjiang renmin, 1993), p. 167. " Wenwu 1975.2: 82—4. For a description of finds from Shilou, a related site in Shanxi, see p. 225, this volume. Kaogu 1976.4: 246-58, 228. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 903 THE CHARIOT. Numerous studies have suggested that the chariot was imported into China from the west, possibly around the thirteenth century B.C.38 Wagons and carts were first made in the Near East in the third mil- lennium B.C., as bronze tools and the domestication of the horse made pos- sible the conception and technical realization of horse-drawn wheeled vehicles. Chariots should be distinguished from four-wheeled wagons and two-wheeled carts. Like wagons, carts were used to transport men and goods: they had solid or spoked wheels and a central axle on which the passenger box rested. Chariots had spoked wheels and a rear axle on which a box nor- mally holding no more than two people rested.39 Recent discoveries related to the Andronovo sites have revealed fully formed chariots with spoked wheels of the Sintashta-Petrovka culture, which may date, according to recent studies, as early as 2026 B.C.4° These are technically and conceptually very similar to chariots found both in West Asia (such as the Lchashen site, Armenia, Russian Federation, in the Caucasus), and East Asia, such as the chariots unearthed at Anyang. Though based on preexisting models of wheeled vehicles, the war chariot seems to have been developed by the agropastoralists of the Andronovo culture. This successful culture was advanced in animal domestication and breeding and mastered the art of bronze metallurgy to the point that craftsmen were able to manipulate alloys so that the quality of the bronze would be harder or tougher according to the specific function of weapons and tools. Indeed, economic success and the development of the war chariot may have been the basic factors accounting for the rapid spread of this culture across the Eurasian steppe from the Urals to South Siberia. The first Chinese chariots have been found in burials of the Shang dynasty at Anyang, together with horses and drivers, who served as sacrificial victims. This type of vehicle was used by the aristocracy for display, hunting, and war. It was made of a central pole with one horse harnessed on each side, and a box with two spoked wheels attached to the end of the pole. The box was typically rectangular or oval. The chariot appears in China already fully formed, unpreceded by stages of development.41 There appear to have been 18 Franz Hancar, Das Pferd in prdhistorischer und friibcr historischer Zeit (Wien: Herold, 1956); Magda- lene von Dewall, Pferd und Wagen im fruhen China (Bonn: Habelt, 1964); Edward L. Shaughnessy, "Historical Perspectives on the Introduction of the Chariot into China," HJAS 48, 1 (1988): 189-237; Stuart Piggott, "Chinese Chariotry: An Outsider's View." In Arts ofthe Eurasian Steppelands, ed. Philip Denwood. Colloquies on Art and Archaeology in Asia no. 7 (London: Percival David Foundation, 1978), pp. 32—51; M. A. Littauer and J. H. Crouwel, Wheeled Vehicles and Ridden Animals in the Ancient Near East (Leiden: Brill, 1979). " Stuart Piggott, The Earliest Wheeled Transport: From the Atlantic Coast to the Caspian Sea (Ithaca, N.Y.: Cornell University Press, 1983), p. 95. 40 David Anthony and Nikolai B. Vinogradov, "Birth of the Chariot," Archaeology 48, 2 (1995): p. 38. " Yang Baocheng, "Yin dai chezi de faxian yu fuyuan," Kaogu 1984.6: 546—55; Kaogu 1984.6: 505—9. Cambridge Histories Online © Cambridge University Press, 2008 904 NICOLA DI COSMO no other wheeled vehicles, such as wagons or carts, pulled by cattle or equids. During the Zhou dynasty, chariots became a common feature of the funer- ary inventory of the richest tombs, and chariotry was the central core of both Zhou and foreign armies. Few chariot remains have been found in the Northern Zone.42 However, the presence of chariots in this region is documented by petroglyphs from South Siberia, the Altai region, the Tianshan ^clll Mountains, and the Yinshan Pit ill Mountains in Inner Mongolia. For instance, a rock carving from the Yinshan Mountains illustrates a hunting scene where a hunter was shooting game after having dismounted from a chariot with eight-spoked wheels, pulled by two horses.43 This is identical with another drawing recov- ered from an incised bone fragment depicting the same scene, with the addi- tional presence of two dogs, found at Nanshan'gen jfjlllfll, dating to the eighth century B.C. or earlier. Their depiction, where chariot and horses are represented flat, is remarkably similar. If we can assume that petroglyphs in a similar style found in regions culturally and geographically akin can be attributed to the same period, then it is possible to hypothesize that the pet- roglyph from Yinshan is earlier than the eighth century B.C., thus providing a later chronological limit. The earlier limit should be the twenty-first century B.C., as this is the date of the earliest chariots found in the necro- polises of the Andronovo culture.44 The presence of similar drawings on deer stones, together with drawings of daggers with bent hilt common in the Northern Zone during the Shang period, suggest a dating of the petroglyphs to the mid-second millennium B.C.4S The petroglyphs, as well as the actual chariots found in the Sintashta burials, indicate essentially the same design and technical characteristics as the Chinese chariot, which is also very similar to a model found in the Cau- casus at Lchashen and dated to the late second millennium B.C.46 It is there- fore probable that the war chariot with lightweight box and spoked wheels, 42 For instance, the remains of wooden wheels found in Nuomuhong 58;fc)ft (Dulan SPS8 county, Qinghai) and tentatively dated to the mid-second millennium B.C. (Kaogu xuebao 1963.1: pi. 3); see also Jenny F. So and Emma C. Bunker, Traders and Raiders on China's Northern Frontier (Washington, D.C.: Arthur Sackler Gallery, Smithsonian Institution; and Seattle: University of Washington Press, 1996), p. 26. 4) Gai Shanlin, "Cong Nei Meng Yinshan yanhua kan gudai beifang youmu minzu de lishi gongxian," in Sichou zhi lu yanhua yishu, ed. Zhou Jinbao (Urumqi: Xinjiang Renmin, 1993). For other examples of chariot petroglyphs, see Shaughnessy, "Historical Perspectives on the Introduction of the Chariot into China," pp. 202—203. 44 Elena E. Kuzmina, "Les steppes de l'Asie centrale a l'Epoque du bronze: La culture d'Andronovo," Dossiers d'archiologie 185 (September 1993): 82-9. 45 Wu En, "A Preliminary Study of the Art of the Upper Xiajiadian Culture." The International Acade- mic Conference of Archaeological Cultures of the Northern Chinese Ancient Nations (Collected Papers), ed. Zhongguo kaogu wenqu yanjiusuo (Huhhot, August 11-18, 1992). 4 ' Two chariots were found in barrows 9 and 11. See Piggott, The Earliest Wheeled Transport, p. 95. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 905 pulled by two horses, originated in Central Asia and was later adopted by the peripheral civilizations of China, Mesopotamia, and the Caucasus.47 Relationship with Northern and Central Asia The relationship between the Northern Zone and the bronze cultures of South Siberia remains problematic.48 From the twelfth to the eighth centuries B.C., a new culture known as Karasuk dominates the Altai Mountains and Minusinsk Basin. The Karasuk people, like their northern Chinese neigh- bors, had a mixed economy that, although mainly based on livestock, also relied on agriculture and other supporting activities.49 Bones of antelope and deer point to extensive hunting, whereas cattle and horse remains show that the Karasuk were devoted to animal husbandry, their main productive activ- ity. Their metal inventory presents many points in common with the North- ern Zone bronzes of North China. Among the knives we find the type with hunched back. The daggers are also similar to the Ordos style, with a short guard. The pickaxes display tubular sockets for halting such as those of the Northern Zone, though the blade presents a pointed cutting edge that may have been derived from a Shang prototype. Arrowheads are also similar to those found in Anyang. The similarities triggered a long-standing dispute as to the influence, and primacy, of one culture over the other. Seminal work by Russian archaeolo- gists in South Siberia50 and subsequent analysis of the Karasuk and Shang affinities showed that the Anyang finds represented the earliest occurrences of certain Karasuk types. These studies also pointed to the presence of a new racial type in Karasuk, akin to the population of North China. As a result, the hypothesis that the appearance of a distinct bronze inventory pointed to 17 This is the opinion of the majority of scholars. See in particular Shaughnessy, "Historical Perspectives on the Introduction of the Chariot into China"; Piggott, "Chinese Chariotry: An Outsider's View." Most Chinese scholars, however, either disagree or remain noncommittal as to the hypothesis of the exogenous origin of the chariot; e.g., see Lu Liancheng, "Chariot and Horse Burials in Ancient China," Antiquity 67, 257 (1993): 814-38. ** From the viewpoint of the development of metallurgy, the eastern zone of the steppe, of which one of the most dynamic centers was located in the Sayano-Altai region, embraced also Mongolia and Trans- baikalia. This is what Chernyk defines as the Central Asian Metallurgical Province. The development of metallurgy and its diffusion throughout the region should probably be seen in connection with the development of nomadism, since more technologically advanced communities may have advanced more quickly along the road of pastoral specialization, and in turn their increased mobility may have facilitated die diffusion of dieir technology. For the extension and metallurgical development of this region in the Bronze Age, see E. N. Chernykh, Ancient Metallurgy in the USSR, trans. Sarah Wright (Cambridge University Press, 1992), pp. 270-1. •" A. P. Oldadnikov, "Inner Asia at the Dawn of History" in The Cambridge History of Inner Asia, ed. Denis Sinor (Cambridge University Press, 1990), pp. 85-8. 50 S. V. Kiselev, Drevniaia istoriia luzhnoi Sibiri (Moscow: Nauka, 1951). Cambridge Histories Online © Cambridge University Press, 2008 906 NICOLA DI COSMO a migration from China to Central Asia in the eleventh or tenth century B.C. was formulated.'1 Recent studies, however, show that the Shang metallurgical tradition is separate from that of the Northern Zone complex.51 That the Northern Complex was an independent cultural unit seems now to be generally accepted." However, the Northern Zone acted as a filter as well as a link between China on the one hand and Central and North Asia on the other; many typical features, especially bronze decorations, originated here and were later transmitted elsewhere.54 Others see a closer relationship between the Northern Zone and South Siberia, regarding the Northern Zone together with the Sayano—Altai region, Mongolia, Transbaikalia, and northwestern China (Xinjiang), as part of the Central Asian Metallurgical Province. In view of the scarcity of archaeological information from Mongolia and Xinjiang, the full extent of the relationship among these areas, and between these areas and China, remains vague. Yet the similarity between some Central Asian forms and the Seima—Turbino knives with animal terminals and socketed celts may suggest an initial Western stimulus to the metallurgy that developed in the Sayano—Altai region and other zones of the Central Asian Metallurgical Province. There was also a possible symbiosis between Central Asian metallurgy and "true Chinese examples of high-quality casting," especially with respect to weapons and ritual objects. In a later period, typical artifacts of this broad Central Asian zone gradually penetrated the west." At the present stage of research it is only possible to say that the Northern Zone was home to a distinctive metallurgical culture with close ties not only with China, but also with areas to the West, in particular with the Sayano—Altai region. The economy of the Northern Zone peoples also resembles that of early pastoralists documented for the Karasuk people. It therefore seems that a process of gradual economic differentiation developed in which more mobile cultures specialized in raising livestock, whereas settled cultures retained mixed economic forms with emphasis on farming. sl Karl Jettrnar, "The Karasuk Culture and Its South-Eastern Affinities," BMFEA 22 (1950): 116-23. " Lin Yun, "A Reexamination," p. 272. On the relationship between Xinjiang and neighboring cultures, an excellent overview is provided by Chen Kwang-tzuu and Frederick T. Hiebert, "The Late Prehis- tory of Xinjiang in Relation to Its Neighbors," Journal of World Prehistory, 9, 2 (1995): 243-300. " For a general survey of the Northern Zone bronzes and their archaeological importnace, see also Wu En, "Yin zhi Zhou chu de beifang qingtong qi," Kaogu xuebao 1985.2: 135—56. 54 Tian Guangjin, "Jinnianlai Nei Menggu diqu de Xiongnu kaogu," Kaogu xuebao 1983.1: 23. " Chernykh, Ancient Metallurgy in the USSR, pp. 269-73. The similarity between the Ordos-Karasuk and the Seima-Turbino artifacts is also recognized by Jettmar, "The Karasuk Culture and Its South- Eastern Affinities," p. 119. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 9O7 Northern and Western Peoples in the Historical Records According to the "Yu Gong" MM (Tribute of Yu) section of the Shang shu jo]* (Venerated documents),'6 ever since the beginning of the Xia JE dynasty the Central Plain was surrounded by alien peoples, living in marshes or on mountains, who brought tribute to the Central Plain. Among the various tribute-bearing peoples are some whose description sounds familiarly nomadic, such as the felt-wearing people of Xiqing HIP,, Kunlun S^r, and Qusou liffi. The Xiqing are said to have lived by the Huan $1 River, not far from the Shang. Among the tribute items are furs and hides, bronze and other metals, including iron, steel, and, possibly, gold, special wood to make bows, stone and whistling (possibly bone) arrowheads. Yet there is no reason to believe that the "Yu Gong" preserves genuine information about real pas- toral nomads living next to the Central Plain in a remote past. At most, we might say that the "Yu gong" furnishes a mythical past with features con- temporary with the period of its composition (possibly the third century B.C.), when mounted nomads had become a common sight on Chinas north- ern borders. Shang oracle bones and written documents of the Zhou dynasty contain a considerable amount of data concerning the names of peoples against whom the Shang fought or with whom they entered into relations.57 These people are usually referred to as fang ~JT (country), preceded by a character probably indicating their ethnic name.'8 Given the paucity of information and the subsequent uncertainty in the identification of these people, we cannot be sure that they formed political unions. Certainly some of them were quite powerful and militarily threatening, but the repeated appearance of a certain name for a relatively long period does not mean that all the people recognized by the Shang under that name were at any given time organized, like the Shang, in a single political and social structure. It is more likely that these definitions indicated certain broadly similar ethnic groups that may or may not have had internal political cohesion. The Gongfang T=T^ were probably located to the northwest of the Shang in northern Shaanxi, northern Shanxi, and possibly even the Ordos 14 On the Shang shu, see Michael Loewe, ed., Early Chinese Texts: A Bibliographical Guide (Berkeley: Society for the Study of Early China and the Institute of East Asian Studies, University of California., 1993). PP- 376-89. Although it contains some of the very earliest of transmitted texts, much of it was written at a later date, the "Yu gong" probably dating as late as the Qin dynasty. 57 On the relations between the Shang and surrounding states, and Shang historical geography, see Chen Mengjia, Yinxu bucizongshu (Beijing: Kexue, 1956), pp. 249—312; Li Xueqin, Yindaidilijianlun (Beijing: Kexue, 1959); Shima Kunio, Inkyo bokuji kenkyu (Hirosaki: Hirosaki daigaku Chugokugaku kenkyukai, 1958). PP- 349-4*3- * Li Xueqin, Yindai dilijianlun, p. 80. Cambridge Histories Online © Cambridge University Press, 2008 908 NICOLA DI COSMO area.59 They maintained frequent interaction with the Shang, especially during the reign of King Wu Ding, when several campaigns were undertaken against them. More resilient enemies may have been the Guifang %LJ], who are said to have resisted Wu Ding's attacks for three years.6" Another group mentioned often is that of the Tufang i.'Jj, probably located in northern Shanxi and extinguished by Wu Ding's conquests. Another group becomes more important toward the end of the Shang, namely, the Renfang J\.~fj, probably located in the middle and lower valley of the Huai 'M River.6' King Di Yi led several expeditions against them, and a chief of the Renfang is said to have been eventually captured and sacrificed to the gods. The most constant enemy of the Shang was the Qiangfang zfcti, or simply Qiang.62 Because of the presence of elements for "sheep" and "man" in the graph qiang, they are described in later sources as "shepherds."63 Their terri- tory extended through southern Shanxi, northern Henan, and northern Shaanxi. The Shang often mounted expeditions against them, capturing slaves and sacrificial victims. Qiang prisoners were also skilled in the prepa- ration of oracle bones.64 The Qiang also seem to have been horse breeders, as some groups among them are called Ma Qiang H§^£ or Duo Ma Qiang ^J§;^i (Qiang with many horses).6s Finally, the Qiang had a close relation- ship with the Zhou, who were intermittently at war with the Shang. The many wars fought by the Shang in the north and west make it clear that the frontier was anything but static and may explain why contacts with the Northern Zone seem to have been particularly active during the Yinxu period. It is in the twelfth century that we have records of increased military activity. The Shang state, surrounded by hostile people, fought and possibly subjugated some of the northern groups, incorporating alien cultural tradi- tions. Whether these people were ethnically or linguistically different cannot be said, but they certainly represented cultures that were neither purely local nor self-enclosed. Indeed, their hinterland extended far to the north and west. By attaining access to this northern complex, the Shang established indirect links with Central Asia. Even at this early stage of Chinese history, the pres- " On the Gongfang in particular, and on Shang historical geography in general, see Edward L. Shaugh- nessy, "Historical Geography and the Extent of Early Chinese Kingdoms," AM, 3d series, 2 (1989): pt. 2, 1-13. 60 K. C . C h a n g , Shang Civilization ( N e w H a v e n , C o n n . : Yale University Press, 1980), p . 12. 61 C h e n Mengjia, Yinxu buci zongshu, p . 304. '* Ibid., pp. 276-82. '' Duan Yucai, Duan shi shuo wenjie zi zhu (Shanghai: Sao Yeshan fang, 1928), 7, pp. 14b—15a; cf. Duan Yucai, Kundoku Setsumon Kaiji chu, ed. and trans. Ozaki Yujiro WJ%tk—%$, vol. 2, pp. 728-30 (Tokyo: Tokai Daigaku, 1986). ** On the Qiang at Anyang, see Hu Houxuan, "Jiaguwen suojian Yindai nuli de fan yapo douzheng," Kaogu xuebao 1976.1: 8—10. On the Qiang in general, see Shirakawa Shizuka, "Kyozoku ko," Kokottu kinbungaku ronsb<) (1958): 1—16. <! Chen Mengjia, Yinxu buci zongshu, pp. 276—7. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 909 ence of a different metallurgical tradition, the many layers of interaction between the north and the Central Plain (possibly responsible for the trans- mission of the chariot and other cultural features), and the presence of mixed communities bespeak the importance of the north in Shang history. WESTERN ZHOU TO EARLY SPRING AND AUTUMN (CA. IOOO-65O B.C.) Transition to Pastoral Nomadism At the end of the nineteenth century, the hypothesis was advanced that "the domestication of animals was possible only under the conditions of a seden- tary way of life."66 This presupposes both a long process of experimentation and accumulation of technical knowledge, and the existence of other sources of production to generate the surplus in fodder and grains needed to feed the animals. Human communities raised animals — sheep, cattle, onagers, pigs, and dogs - as a complement to farming. They then learned to use them either for direct productive activity or for other purposes, such as transportation. Steppe oases provided a natural environment equally favorable to agricul- ture and animal husbandry. Surrounded by steppe pastureland, these areas imposed fewer restrictions on stock raising than valley agriculture, where an imbalance between the human and animal element could lead to disastrous consequences. The oasis dwellers specializing in stock breeding eventually separated themselves from their original environment and became nomadic pastoralists. In areas contiguous to farming communities, nomads remained to a certain extent dependent on them for agricultural and handicraft prod- ucts. To a varying degree nomads also continued to practice some forms of agriculture in their winter quarters.67 The role of horses in the transition from sedentary herding to nomadic pastoralism is crucial. Possibly the first equid to be tamed was not the horse but the more docile onager.68 Paleozoological evidence based on tooth bitwear suggests that the horse was first domesticated and possibly ridden in the 66 S. I. Vejnshtein, "The Problem of the Origin and Formation of the Economic-Cultural Type of Pas- toral Nomads in the Moderate Belt of Eurasia," in The Nomadic Alternative, ed. W. Weissleder (The Hague: Mouton, 1978), pp. 127-33. See also Anatoli Khazanov, Nomads of the Outside World (Cam- bridge University Press, 1984), p. 89. 67 Iwamura Shinobu, "Nomad and Farmer in Central Asia," Acta Asiatica 3 (1962): 45—50; Nicola Di Cosmo, "Ancient Inner Asian Nomads: Their Economic Basis and Its Significance in Chinese History," JAS 53, 4 (1994): 1092-126. 68 J. F. Downs, "Origin and Spread of Riding in the Near East and Central Asia," American Anthropolo- gist, 63 (1961): 1193-6. Cambridge Histories Online © Cambridge University Press, 2008 9IO NICOLA DI COSMO fourth millennium B.C. in the Ukraine.6' Like the onager, it was first used as a draught animal, being mounted and ridden only later. The first commu- nities to breed horses were mainly agricultural, also raising pigs, cattle, and sheep. The horse was, then, just another member of the animal stock of early farmer-pastoralists. Although there are differences of opinion, it seems that horseback riding in the strictest sense of the word, that is, riding astride, spread in West and Central Asia around the early part of the second millennium B.C. or possi- bly at the end of the third millennium B.C.7° The early pastoral communi- ties were by no means nomadic in the sense that the later steppe nomads were, characterized by the military use of the horse. Although some were more or less mobile, following their herds on wheeled carts, or perhaps even mounted on horses, their pastoralism cannot be defined as nomadic — that is, a regular seasonal migratory cycle based on grass-producing environments at varying altitudes - but rather as "herder husbandry" or at most semi- nomadism. They still had settlements and depended on agricultural pro- duction. The vigorous expansion of the oasis civilizations of Central Asia in the third and second millennia B.C., based on coexisting agricultural and pas- toral production and increasingly closer contacts with neighboring people, led to the spread of enhanced metalwork and transportation technology to the steppe, where the environment was more favorable to the development of a pastoral economy. By the third millennium B.C., farmer-herders had already begun to use various alloys and traveled in wheeled vehicles.7' During this period "the struggle for forcible redistribution of pasture and accumu- lated wealth gives rise, at a certain stage, to a type of militarization of society that found expression and progress in the production of weapons."72 Archaeological and linguistic evidence shows that preconditions for the emergence of pastoral nomadism were present in the middle of the second millennium B.C. in both the European and Kazakh steppe. Nevertheless, the transition to pastoral nomadism practiced by horseback riders was not com- pleted until the beginning of the first millennium B.C.73 69 Marsha Levine, "Dereivka and the Problem of Horse Domestication," Antiquity 64 (1990): 727-40; Dmitriy Telehin, Dereivka: A Settlement and Cemetery of Copper Age Horse Keepers on the Middle Dnieper (Oxford: British Archaeological Reports International Series no. 287, 1986); David W. Anthony and Dorcas R. Brown, " T h e Origins o f Horseback Riding," Antiquity 65 (1991): 22-38; Khazanov, Nomads of the Outside World, p . 91. 70 Littauer a n d Crouwel, Wheeled Vehicles and Ridden Animals in the Ancient Near East, pp. 45-7, 65-8. Anthony and Vinogradov, "Birth of the Chariot," p. 40. 71 Anthony and Vinogradov, "Birth of the Chariot," pp. 36-41. 71 V. M. Masson and T. F. Taylor, "Soviet Archaeology in the steppe Zone: Introduction," Antiquity 63 (1989): 780. 7J Aryan terminology appears in a Hittite work on horsemanship of the fourteenth century, concerned with the training of horses hitched to chariots, which may indicate that this type of training was a technique developed by steppe Indo-European peoples. See Pentti Aalto, "The Horse in Central Asian Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 9II The eventual emergence of pastoral nomadism has been attributed to several factors: the increasing number of livestock and the accumulated expe- rience of a more progressive, mobile pastoral husbandry; changes in the climate leading to progressive desiccation as a result of which formerly seden- tary cultivators and catde breeders became full nomads; and finally transi- tion to a nomadic life as a result of overpopulation.74 It is also possible that nomadism emerged among forest hunters, who borrowed animals from their sedentary neighbors and then, after they began to use the horse, moved into the steppe.7S In South Siberia the transitional period that preceded die emergence of a more homogeneous nomadic culture is represented by the Karasuk cultural complex of the twelfth to the eighth centuries B.C. This period witnessed a rapid development of metallurgic technology based on a wide range of alloys and the use of stock breeding as the main economic activity. While no def- inite explanation can be provided for the appearance of the mounted steppe culture around the end of the second and beginning of the first millennium B.C., this is generally seen as the result of an internal process, rather than one caused by external factors, such as invasions or mass migrations. It was the combination of many factors over a long period of time that eventually pro- duced an economic and cultural complex singularly well adapted to the arid conditions of the steppe. However, the speed with which people turned to fully developed pastoral nomadism is still subject to debate. For some, once the first pastoral communities emerged, the spread of nomadism ensued rapidly. By the eighdi century B.C., many people in different parts of the steppe had taken to nomadism. At this time a military aristocracy was formed that concentrated in its hands a higher percentage of the common wealth (mostly weapons, ornaments, and, especially, animals) removed from the rest of the community. The rise of a warrior class is related to a general increase in aggressive warfare among pastoralists, aimed at securing pastures. Agri- cultural production was considerably reduced and was preserved only at the tribes' winter pastures. Others have proposed a far more gradual transition that required several centuries.76 Nomadic Cultures," Studia Orientalia 46 (1975): 4-7. See also Hancar, Das Pferd in prdhistorischer und Friiher historischer Zeit, pp. 551-63. 74 G. E. Markov, "Problems of Social Change Among the Asiatic Nomads," in The Nomadic Alternative, ed. W. Weissleder (The Hague: Mouton, 1978), p. 306. 71 Vejnshtein, "The Problem of the Origin and Formation of the Economic-Cultural Type of pastoral Nomads in the Moderate Belt of Eurasia," pp. 130-1. 76 These various viewpoints are expressed in the following works: V. Masson, "Central Asia in the Early Iron Age: Dynamics of Cultural, Social and Economic Development," Information Bulletin. UNESCO International Association for the Study of the Cultures of Central Asia 9 (1985): 59-68; Gryaznov, The Ancient Civilization of Southern Siberia, pp. 131—2; Sergei Rudenko, Kul'tura naseleniia tsentral'nogo Altaia v skifikoe vremia (Moscow: Nauka, 1960), p. 197. Cambridge Histories Online © Cambridge University Press, 2008 912 NICOLA DI COSMO Evidence of Pastoralism in the Northern Zone From the Western Zhou to the late Warring States period, an increasing number of horse fittings and ornaments is found throughout the Northern Zone. The amount and variety of horse gear, together with the presence of horse sacrifice within the funerary assemblage, betoken the growing impor- tance of the horse in both the economic and symbolic spheres. There are several indications that the early use of metal horse fittings and riding in the northeast may be related to the development of the Upper Xiajiadian culture. Since saddles are not found in burials and stirrups had not yet been invented, it is difficult to say whether horse-related findings are to be associated with horseback riding or with other uses of the horse, such as pulling a cart or a chariot. There is however an indication in the Upper Xiajiadian site at Nanshan'gen that horses were in fact ridden, as figures of horseback hunters are represented on a bronze ring in the act of pursuing a hare. This find shows the existence of horse-riding people in the northeast as early as the eighth to seventh century B.C. Other horse-related findings consist of bronze bits, cheekpieces, rein rings, ornaments, and chamfrons. With the exceptions of the cheekpieces, which were sometimes made of bone or wood, they were all made of bronze. The ornaments include disc-shaped ornaments, bells, luan IB bells, and some head ornaments and masks. Iron ornaments appear regularly only in the Warring States and Former Han periods. Throughout the geographic distribution of the Upper Xiajiadian culture — eastern Inner Mongolia, western Liaoning, Hebei, and the Beijing area — horse fittings appear prominently in Western Zhou and Spring and Autumn funerary assemblages. However, specific evidence of horse riding is not doc- umented in the southern part of this area until the sixth century, which is the date assigned to special riding bits found in Yanqing 3i£Ji, Beijing.77 In the Central Plain we find no real evidence that horses were ridden before the fourth century B.C.78 This may point either to a progressive southward movement of horse-riding communities or to a slow diffusion of horse riding toward the south. 77 Emma C. Bunker, "Unprovenanced Artifacts Belonging to the Pastoral Tribes of Inner Mongolia and North China During the 8th—ist Century B.C.," in The International Academic Conference ofArchaeo- logical Cultures ofthe Northern Chinese Ancient Nations (Collected Papers), ed. Zhongguo kaogu wenwu yanjiusuo (Huhhot, August n—18,1992); see also So and Bunker, Traders and Raiders on China's North- ern Frontier, p. 19. 78 H. G. Creel, "The Role of the Horse in Chinese History," American Historical Review 70, 3 (1965): 647—72, 649. See also C. S. Goodrich, "Riding Astride and the Saddle in Ancient China," HJAS 44, 2 (1984): 279-306. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 913 EARLY IRON TECHNOLOGY. In the area of present-day Xinjiang we find the earliest appearance of iron of the Northern Zone. The cemetery site of Chawuhu M^r¥ Pass reveals some of the earliest iron remains.79 On the basis of multiple calibrated radiocarbon datings, this site has been attributed to a period from the tenth to the seventh century B.C. It consists of stone mounds with multiple burials encircled within a ring of stones. The funer- ary assemblage includes gold, bronze, and iron objects. Among the bronze objects are knives with ringheads, a spearhead, and horse bits. A bone cheek- piece in the form of a ramhead is representative of early Animal Style. Iron objects are few and small, such as an awl and a ring. The extensive evidence of animal sacrifices buried in sacrificial pits, either separate from or together with human remains, points to a culture that is no longer agricultural and has a surplus of animals to spare. Among the funerary objects no agricultural tools have been found. The existence of iron in this region at a time that precedes its first appear- ance in the Central Plain (possibly in the sixth century B.C.), is confirmed by analogous findings in Qunbake fP^Eini (Luntai ^a county), the Pamirs, and the area near Urumqi.80 Comparable iron and bronze knives found in the Chust culture in Ferghana, and skeletal remains of Europoid stock, point to a connection with the Pamirs and Ferghana regions. The far northwest was not the only area close to the Northern Zone that had iron at an early period. Iron dated to the end of the second millennium B.C. has been found along the Amur River, in the Maritime Territory of Russia, while by the ninth century B.C. there was rather elaborate iron met- allurgy, producing knives, daggers, and armor.8' Direct connections between this area's ferrous metallurgy and that of the Northern Zone have not yet been established for the earlier period, but there are indications that relations existed between Transbaikalia and the Chinese northeast possibly following ancient routes of communication through the forests of Manchuria and the large waterways that run north to south: the Sungari, Nonni, and Liao Rivers.82 The question of the spread of iron technology in South Siberia is quite complex, and there is no unified opinion. The date usually assigned to the 79 Kaogu xuebao 1988.1: 75-99. *° Kaogu xuebao 1981.2: 199-216. 11 A. P. Derevianko, Rannyi zheUznyi vek Priamur'ia (Novosibirsk: Nauka, 1973); Pavel Markovich Dolukhanov et al., "Radiocarbon Dates of the Institute of Archaeology, 2," Radiocarbon 12 (1970): '30-55- 81 At a later period, usually defined as Hunno-Sarmatian in Russian scholarship (second century B.C. to second century A.D.), there is evidence of close contacts between Heilongjiang and Transbaikalia; see D. L. Brodianskii, "Krovnovsko-Khunnskie paralleli," in Drevnee Zabaikal'e i ego kul'turnye sviazi, ed. P. B. Konovalov (Novosibirsk: Academiia Nauk USSR, 1985), pp. 46—50. Cambridge Histories Online © Cambridge University Press, 2008 914 NICOLA DI COSMO Early Iron Age in Central Asia (Transoxiana) is the beginning of the first mil- lennium B.C. The same date usually applies to the Early Iron Age in the steppe regions of Kazakhstan, Tuva, South Siberia, and Mongolia, even though sites of this period in the Kazakh steppe do not contain iron artifacts and iron metallurgy developed in Mongolia only from the middle of the first millennium B.C. Nevertheless, iron undoubtedly existed in Tuva at least from the eighth century B.C., as documented by the finds of the Arzan royal burial and other kurgans of the early nomadic period.8' In the part of the Northern Zone that is closer to China, and in particu- lar in the north-central sector, iron was present at a time roughly compara- ble with the general period of diffusion throughout the steppe region. An iron knife was found at the Taohongbala $li,tEBJ& burial site (Ordos, Inner Mongolia), chronologically close to the "Scythian" period of the western and central Asian steppe regions.84 Distribution of Northern Cultures NORTHEAST: U P P E R XIAJIADIAN CULTURE. The Upper Xiajiadian culture's geographical extension reaches in the north to the Sira Moren River basin, up to the eastern side of the Great Khingan Mountains. The south- ern boundary is formed by the Luan HI River, Yan M Mountains, and the Qilaotu -b^-lfil Mountains. The eastern boundary is the basin of the Liao River; the western boundary is the area of Zhaowudameng Bg^JHK! in Inner Mongolia. It therefore extends over the three provinces of Inner Mongolia, Liaoning, and Hebei (see Map 13.2). Its chronological limits have been the object of considerable discussion among Chinese archaeologists.8' The general consensus today is that it lasted approximately eight centuries, from the eleventh to the fourth century B.C. The first excavation that identified the Upper Xiajiadian as a separate culture took place at Chifeng ff^., in Inner Mongolia, where distinctive fea- '' On the Iron Age in Central Asia see: A Askarov, "The Beginning of Iron Age in Transoxiana," in History of Civilizations of Central Asia. Vol i: The Dawn of Civilization: Earliest Times to joo B.C., ed. A. H. Dani and V. M. Masson (Paris: Unesco, 1992), p. 457; A. Askarov, V. Volkov, and N. Ser-Odjav, "Pastoral and Nomadic Tribes at the Beginning of the First Millennium B.C.," in ibid., pp. 459—75; Grjaznov, Der Grojlkurgan von Arzan in Tuva, Siidsibirien, Kenk, Grabfundt der Skythenzeit aus Tuva, Siid-Sibirien. We should also note, however, that according to Martynova the presence of iron artifacts in the Shestakocvo cemetery (Minusinsk region) is a completely new phenomenon associated with a new cultural complex, attributed to the Hunnic (Xiongnu) culture and dated to the last two centuries of the first millennium B.C. See Galina S. Martynova, "The Beginning of the Hunnic Epoch in South Siberia," Arctic Anthropology 25, 2 (1988): 74. *• Tian Guangjin, "Taohongbala de Xiongnu mu," pp. 131—42. 8 ' Jin Fengyi, "Xiajiadian shangceng wenhua ji qi zushu wenti," Kaogu xuebao 1987.2: 177-208. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 915 Northern Zone Sites of the Western & Early Eastern Zhou Periods (1,000-650 B.C.) Map 13.2 Northern Zone: Archaeological sites of the Western and early Eastern Zhou periods (1000-650 B.C.) tures were documented that differentiated it from the contemporary bronze culture of the Ordos region. Three chronological phases can be outlined based on characteristic objects such as the "wobbly," or curved-blade daggers.86 Sites of the older period lasting throughout the Western Zhou period are located mostly in Linxi # ! f county in northern Liaoning. The type site for this period is the ancient mine of Dajing 'XR, which we find in connection with a small settlement consisting of a few semisubterranean houses. The mine includes at least forty mining shafts, near one of which the remains of twelve smelting furnaces have been found. Most of the mining tools are made of stone, with very few small bronze objects (a drill and an arrowhead).87 The extensive remains of wild animals suggest an economy still heavily dependent on hunting. The cemetery at Dapaozi ifcffll^P in the Onggut Banner (Wengniuteqi w ^^M, Inner Mongolia) also belongs to this earlier phase. The pottery found here has a close relation to that of Baijinbao, in Heilongjiang. The northern location of these early sites, and the connection between the Dapaozi and K On the classification of northeastern swords, see Lin Yun, "Zhongguo Dongbeixi tong jian chu lun," Kaogu xuebao 1980.1: 139-61; Jin Fengyi, "Lun Zhongguo dongbei diqu han quren qingtong duanjian de wenhua yicun," Kaogu xuebao 1981.4: 387—426 (part 1), and 1 (1983): 39—54 (part 2). 87 Wenwu ziliao congkan 7 (1983): 138—46. Cambridge Histories Online © Cambridge University Press, 2008 916 NICOLA DI COSMO the Baijinbao pottery, may suggest a southward movement of northern peoples associated with a hoe-agriculture- and hunting-based economy. The second phase of the Upper Xiajiadian culture (850-750 B.C.) is iden- tified with the Nanshan'gen type site (in Ningcheng 3pt$t county, Inner Mon- golia), excavated in 1958 and 1961 (Fig. 13.2). Chronologically it is placed between the late Western Zhou and the early Spring and Autumn periods, with a lower limit of approximately 750 B.C. The graves are rectangular earthen pits with stone slabs lining the walls. The funerary assemblage reveals a considerable quantity of bronze objects and some golden ornaments. The characteristic Upper Xiajiadian assemblage of this phase includes bronze weapons, horse fittings, and mirrors. Animal bones of pigs, dogs, cattle, sheep, and deer have been found in considerable quantity. The contempo- rary settlement and the finding of a hoe indicates the presence of agriculture. But what is most remarkable about Nanshan'gen is the presence of horse- back riding, suggested by the already mentioned bronze ring with two figures of horseback riders chasing a hare, and by a varied and sophisticated inven- tory related to horse and chariot technology: harness with cheekpieces, two different types of bit, and tinkling bells. Moreover, an incised bone plaque also recovered at Nanshan'gen shows a hunting scene with the use of chari- ots pulled by horses. The bronze weapons unearthed at sites of the second phase of Upper Xiajiadian include bronze daggers, knives, axes, spearheads, arrowheads, shields, and helmets. Human figures represented on the hilt of a dagger can be associated with the plates bearing designs of human faces that were unearthed at Shiertaiyingzi "t~— taH^T". This period also boasts two additional important sites: Dongnangou 3fE]^! M (in Pingchuan TJl| county, Hebei), and the Zhoujiadi JWJ^itfe cemetery (in Aohan M'M Banner, Inner Mongolia).88 The first was excavated in 1964-65. Of the eleven tombs excavated, only four had burial goods, which suggests the presence of different social groups. Their funerary assemblage, made primarily of bronze weapons, small ornaments, and bone beads, is far less rich than that at Nanshan'gen, but very similar in bronze typology. The second was excavated in 1981 and has been assigned to the Upper Xiajiadian culture of the Spring and Autumn period. Here are fewer weapons and more ornaments, such as bronze plaques and pendants. A special feature of the burial custom is that the face of the dead was covered with sackcloth deco- rated with turquoise beads and bronze buttons. This characteristic trait may have significance with respect to the ethnic distribution of northern peoples and is sometimes attributed to the Shan Rong liljJJ. In the Xiajiadian culture a special place was reserved for dogs, whose 88 Kaogu 1977.1: 51-5; Kaogu 1984.5: 417-26. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 917 Figure 13.2. Bronze objects of the Upper Xiajiadian culture. After: Zhu Yonggang "Xiajiadian shangceng wenhua de chubu yanjiu" H^Ji5i:^}3tiLW:#Ji£'5'f3?. In Kaoguxue wenhua lunji ^~£ ^JtitUlM ed. by Su Bingqi j ? 3 * ^ , (Beijing: Wenwu, 1987), p. 109. Cambridge Histories Online © Cambridge University Press, 2008 918 NICOLA DI COSMO remains have been found in both dwelling sites and sacrificial pits. They also appear as a decorative motif on bronze weapons (a ge dagger-axe and a knife) and are depicted on the aforementioned bone plaque representing a hunting scene. NORTH-CENTRAL REGION. The most representative Western Zhou sites of the so-called Ordos bronze culture of the north-central zone are Lingtai I P in northern Gansu, and Changping il^f and Yanqing MJS, both near Beijing, though Ordos-type finds are also present in Upper Xiajiadian sites, such as Dongnangou and Nanshan'gen tomb 101. During the Western Zhou and early Spring and Autumn periods the Ordos bronze culture went through considerable changes. A general increase in the volume of animal bones, and the presence of a larger number of domesticated animals in the funerary burials, suggest a general tendency toward the expansion of pastoral economies in the Northern Zone.89 Bronze weapons remained prominent in funerary assemblages, but underwent stylistic modifications. For instance, the earlier animal-head and rattle pommels became rare and were gradually replaced by mushroom pommels. GANSU-QINGHAI REGION. The Kayue culture of Gansu and Qinghai has been regarded in the past as the successor to the Xindian culture, but recent research has showed that the two were separate and to a certain degree con- temporary cultural complexes. The Kayue culture, the first appearance of which is roughly the beginning of the Shang period, and which continues to the Han dynasty, shows a gradual evolution from a mixed farming and pas- toral culture with settled life to a predominantly nomadic economy. This transition is reflected not only in an increased number of animal bones and sacrifices, but also in the composition of the animal stock. In the early period (Shangsun _hi^ type), the pig was the usual sacrificial offering, but it disap- peared in the middle period (Ahatela P5JPH#£A type), giving way to cattle and horse sacrifices.90 In the third phase (Dahuazhongzhuang ^ ^ 4 " ^ type), bronze spears, knives, and ge dagger-axes replaced the earlier axes, showing a higher degree of ornamentation. Remains from a late Qijia or early Kayue site in Huangjiazhai H:M-M (Datong ;fci! county, Qinghai), such as a decor of deer figures on a hollow bone tube, a bronze bird figure, and hooves of sacrificed horses, suggest early contacts with Siberian or Central Asian cultures.9' 19 Cui Xuan, "Nei Menggu xian Qin shiqi xumu yicun shulun," Nei Menggu shehui kexue i (1988): 69-74. 90 Gao Donglu, "Liielun Kayue Wenhua," in Kaoguxue wenhua lunji, ed. Su Bingqi (Beijing: Wenwu, 1993), vol. 3, pp. I53-65- " Kaogu 1994.3: 193-205. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 919 Historical Survey For the period that stretches from the early Western Zhou to the middle of the Spring and Autumn period, historical documents reveal a complex situ- ation on the northern borders. The Zhou fought various battles against foes who appear to have been their military equals. Some of these eventually forced the rulers of China to evacuate their capital. The period opens with a series of hostilities between the Zhou and the Guifang $LJT, one of the northern peoples whom the later historical tradi- tion identifies as ancestors of the Xiongnu.92 It ends, in 650 B.C., with the last mention of the Northern Rong JLiS, and with the rise of the Di %k, who thereafter became the most important northern power. The middle of the seventh century B.C. is a turning point in the political balance of the north- ern frontier, possibly more significant in the history of the northern frontier than the end of the Western Zhou (771 B.C.). GUIFANG. Records in the Zhushu jinian f t f t ^ S ^ (Bamboo annals) state that a people by the name of Gui Rong M.3&, certainly the same as the Guifang, were already at war with the Zhou in Shang times, being attacked in the thirty-fifth year of King Wu Yi KZ1 (1119 B.C., according to the chronology given in that text) by the Zhou leader Jili SjsM.93 The Guifang "proper" who appear during the reign of King Kang M (1005/03-978 B.C.) were probably located to the northeast of the Zhou territory. According to the inscription of the Xiao Yu ding WNJ£JPI, cast in the twenty-fifth year of King Kang (979 B.C.), after two successful battles against the Guifang, captives were brought to the Zhou temple and offered to the king. The pris- oners numbered over 13,000 and included four chiefs, the chiefs being exe- cuted. A large amount of booty was also collected. In general the Guifang do not seem to have posed any serious threat to the Zhou frontier and must have been either conquered by the Zhou at an early stage or dissolved polit- ically, since their name soon disappears from the patchy sources at our disposal. XIANYUN $S#t. The Shijingt^M. (Classic of poetry) contains four songs that mention military engagements between the Zhou and the Xianyun. As examples of epic poetry, they relate feats of ancient heroes against alien foes. The song "Cai qi" 7RIS (Gathering sow thistle) extols the deeds of Fang Shu Wang Guowei, "Guifang Kunyi Xianyun kao." in Haining Wangjing'an xianshengyishu, vol. 5, Guan- tangjilin, chapter 13, "Shilin" 5, pp. 1-20. Zhushu jinian (Sibu beiyao ed.) 1, 17b (James Legge, The Chinese Classics. Vol. y. The Shoo King [London: Henry Frowde, 1865), Prolegomena, p. 138). Cambridge Histories Online © Cambridge University Press, 2008 92O NICOLA DI COSMO , who apparently led as many as 3,000 chariots in battle against the Xianyun.94 The impression created in the songs is one of trepidation, urgency, and great relief and jubilation at the final victory of the Zhou troops. The song "Liu yue" f\M (Sixth month) provides geographical information that allows us to place the battlefield very close to the center of the Zhou state,95 between the lower reaches of the Jing M and Luo ?& Rivers and the Wei River valley. Although scholars dispute the exact date of the attacks, most place it during the reign of King Xuan m. (827/25-782 B.C.). Written records place thefirstXianyun incursions against the Zhou in 840 B.C., the fourteenth year of reign of King Li M,96 when they reached the capital itself. Three years earlier the Western Rong (Xi Rong MJSC) had also launched an attack; it is possible that Xianyun and Western Rong may have been the same people, indicated in the first case by a generic term meaning "warlike tribes of the west" and in the second case by their actual ethnonym. They attacked again in 823 B.c, the fifth year of reign of King Xuan.97 On the basis of their military tactics which were characterized by sudden attacks and could only have been carried out by highly mobile troops, most likely on horseback, some scholars have related the sudden appearance of the Xianyun to the general rise of mounted nomadism in the steppe region and to the specific appearance of Scythians and Cimmerians migrating from the west.98 No definite evidence, however, supports the hypothesis that they were nomadic warriors. In fact, new evidence indicates that the Xianyun fought, like the Zhou, on horse-drawn chariots. This is based on the inscription on the Duo You ding^isLfk, unearthed in 1980 nearXi'an M:£, which tells of a Zhou mil- itary campaign, probably in 816 B.C., against Xianyun forces that had attacked a Jing M garrison in the lower Ordos region.99 The Xianyun's pres- sure against the northern frontier was undoubtedly serious; however, there is no evidence that the large size of their armies or their increased mobility can be related to the emergence of a new type of warrior, namely, the mounted nomad armed with bow and arrow. 9< Shi jing, 10/2, 8b (Legge, The Chinese Classics. Vol. 4: The She King, pp. 284-7). " Shi Jing, 10/2, 2a (Legge, The She King, pp. 281-4). 96 Zhushu jinian, 2, p. 8a (Legge, The Shoo King, Prolegomena, p. 154). 97 This date is based on the Xi Jiapan ^ ¥ $ inscription; Ma Chengyuan ed., ShangZhou qingtongqi mingwen xuan (Beijing: Wenwu, 1988), no. 437 (Shaughnessy, Sources of Western Zhou History, p. 141); Guo Moruo, Liang Zhou jinwenci daxi lulu kaoshi (Beijing: Kexue, 1958), vol. 7, p. 143b. 98 Yaroslav Prujek, Chinese Statelets and the Northern Barbarians in the Period 1400—300 B.C. (Dordrecht: Reidel, 1971), p p . 119—35. 99 E. Shaughnessy, "The Date of the 'Duo You Ding and Its Significance." £C9-io (1983-5): 55-69; Li Xueqin, "Lun Duo you ding de shidai ji yiyi," Renwen zazhi 1981.6: 87-92; Tian Xingnong and Luo Zhongru, "Duo you ding de faxian ji qi mingwen shishi," Renwen zazhi 1981.4: 115-18. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 921 RONG jft. The term rong^L is often applied in Chinese sources to warlike foreigners. Its general meaning relates to "martial" and "military," "war" and "weapons."100 A widely accepted view is that the term "Rong" was used in Chinese sources as a blanket word that included many alien peoples around and within the territory occupied by the Zhou, without a specific ethnic con- notation, thus including people not only of the north and west but also of the south - such as the Rong Man J&Wt."" Because of the ambiguity of the term, it is not clear whether it was also used to indicate peoples known by other names, such as the Xianyun. It certainly appears to indicate more than a single people during the reign of King Mu H, who defeated the Quan Rong jtl& in the twelfth year of his reign and the following year attacked the Western Rong and the Xu Rong #j£. 102 These events bespeak a phase of expansion under King Mu, whose journey to the west was romanticized in the fourth century B.C. work Mu Tianzi zhuan Wt3i^~W (Biography of the Son of Heaven Mu).103 Based on the available records, relations between the Zhou and the Rong do not seem to have been hostile until the seventh year of King Yi M (859 B.C.), when the Rong of Taiyuan ;fcJ® attacked the area of the Zhou capital. At this time the Zhou royal family gradually came to depend on other noble families to defend the realm. In 854 Guo Gong W.'fc attacked the Rong, cap- turing a thousand horses. Under the reign of King Li M, the power of the dynasty was in decline, and both Western Rong and Xianyun launched attacks deep into Zhou territory. A more energetic policy was followed by King Xuan (827/25—782), during whose period of reign the Zhou gradually incorporated some of the Rong within their own borders. In the fourth year of his reign, King Xuan ordered Qin to attack the Western Rong, who retreated. More expeditions against them finally led to their submission and to territorial gains.104 Throughout the end of King Xuan's reign there were repeated engagements against the Rong. Particularly significant seem to have been the expedition in either 790 or 788 (the thirty-eighth year of King Xuan's reign) by Jin W against the lo ° Everard D. H. Fraser and James H. S. Lockhart, Index to the Tso Chuan (London: Oxford University Press, 1930), p. 165. In the Zuo zhuan, rong is also used in the sense of "war chariot" in the phrase yu rongfflW(to drive a war chariot) and in the compound rong che #M (war chariot). 101 This opinion has been challenged by Edwin G. Pulleyblank, who regards the Rong as a Tibeto-Burman group related to the Qiang people, on the one hand, and to the founders of the Zhou dynasty, on the other. See Pulleyblank, "The Chinese and Their Neighbors in Prehistoric and Early Historic Times," in The Origins of Chinese Civilization, ed. David N. Keightley (Berkeley: University of California Press, 1983), pp. 416-23. 101 Zhushu jinian, i, pp. 4b—5a (Legge, The Shoo King, Prolegomena, p. 150). '"' Loewe, Early Chinese Texts, pp. 342—6. 104 Hou Han shu, 87, pp. 2871—2. Cambridge Histories Online © Cambridge University Press, 2008 922 NICOLA DI COSMO Northern Rong and the king's expedition the following year against the Rong of the Jiang # clan, who were utterly destroyed.105 The final period of the Western Zhou, under the reign of King You ffl (781-771 B.C.), was marked by increasing instability on the northern frontier and by a series of attacks by the Quan Rong. The Zhou defenses were overrun, the capital invaded, the king killed, and the court forced to move to the city of Luo $£, marking the beginning of the Spring and Autumn period, in 770 B.C. During the Western Zhou the various Rong were scattered over a broad area that encompassed the northern and western areas of the Wei River valley, the Fen :&f River valley, and the Taiyuan j&JM region. They were therefore distributed in present-day northern Shaanxi, northern Shanxi, and Hebei, up to the Taihang jsi'fi Mountains. With few exceptions, their attacks against the Zhou do not seem to have been particularly effective. Like the Xianyun, they most probably used chariots, but a record for 714 B.C. shows that they also fought on foot.106 There is no evidence to identify them with the later nomads. Rather, their military pressure on the Zhou borders might be attrib- uted to unrecorded events taking place in the north that set a large number of people in motion and that may be related with the appearance of pastoral nomads. The later arrival of the Di populations along the northern borders of the Central Plain may have also been connected with large-scale migra- tory movements. Archaeological data concerning the development of horse riding in the northeast, north, and northwest may suggest that the pressure of the Rong, and later Di, on the Zhou northern frontier may have been the indirect effect of nomadic expansion in areas such as western Liaoning, Inner Mongolia, and Ningxia. During the late Western Zhou, groups of Rong had settled in territory nominally controlled by the Zhou king and were interspersed among the political centers of the Central Plain. During the Eastern Zhou, some Rong tribes were located in the neighborhood of the state of Lu # . The North- ern (Bei ^t), or Mountain (Shan ill) Rong attacked the states of Zheng HP in 714 B.C. and Qi 3^ in 706 B.C.107 These names indicate relatively broad groups or confederations whose ethnic or political cohesiveness is difficult to determine. For instance, the Shan Rong included a people, known as the Wuzhong M$&, whose location has been traditionally placed by some near 105 Guoyu (Sibu beiyao ed.), 1 ("Zhou yu shang"), 9a. "* Zuo zhuan, 4 (Yin 9), 14b (James Legge, The Chinese Classics. Vol$: The Ch'un Ts'ew with the Tso Chuen [London: Triibner, 1872]) p. 18; Yang Bojun, ed., Chunqiu Zuo zhuan zhu [Beijing: Zhonghua, 1990], p. 65). 107 Zuo zhuan, 4 (Yin 9), 146 (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 28; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 65); Zuo zhuan, 6 (Huan 6), 21a (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 49; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 113). Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 923 present-day Beijing, but who were probably located in the Taiyuan region, bordering on the state of Jin W-ias However, other Wuzhong groups appear to have inhabited the region north of the state of Yan.109 The record of a Wuzhong leader presenting the lord of Jin with tiger and leopard skins indi- cates that they may have been hunters who lived in or close to a forest envi- ronment, perhaps northern Hebei and Liaoning."° The Zuo zhuan mentions repeated military clashes between Zhou and Rong forces during the reign of Lu Zhuang Gong Hf-St^ (693—662 B.C.)- In 674, Zhuang Gong chased Rong raiders. Two years later Qi attacked them, and in 670 B.C. the Rong attacked the smaller state of Cao W. In 668 B.C., Zhuang Gong again attacked the Rong. In 666 B.C., Jin Xian Gong ^affifc^ took Rong women as wives.1" Qi continued to battle Rong peoples, and in 664 B.C. Qi again obtained a victory against the Shan Rong."2 These events show that Rong communities were dispersed over a broad territory across the northern frontier of the Eastern Zhou and that a large number of them had even settled within or to the south of the state of Jin. It is possible that these Rong tribes retained a degree of autonomy or inde- pendence, but this was eroded and eventually eliminated by the relentless expansion of the Central States. By the mid-seventh century, the Rong, repeatedly defeated by Jin and hard pressed in the north by the rapidly growing power of the Di, were for the most part incorporated by Jin and Qi. By 662 the Di conquered the Taiyuan plains and replaced the Rong in this important economic zone to the north of Jin."3 The Quan Rong, located north of the Wei River valley were defeated by Guo It in 660 and 658 B.C. At the same time, the Di intensified their attacks, first against Xing ffl and then against Wey $i. The people of Wey escaped from their city but were pursued by the Di and massacred near the Yellow ' Prusek, Chinese Statelets, p. 21. ' See map in ibid., p. 120. Zuo zhuan, 29 (Xiang 4), (22a; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 936; Legge, The Ch'un Ts'ew with the Tso Chuen, p. 424). In chronological order: Zuo zhuan, 9 (Zhuang 18), 14b, (Legge, The Ch'un Ts'ew with The Tso Chuen, p. 97; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 208); Chunqiu, 9 (Zhuang 20), 19a (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 100; Yang Bojun ed., Chunqiu Zuo zhuan zhu, p. 214); Chunqiu, 10 (Zhuang 24), 3b (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 107; Yang Bojun, ed., Chunqiu Zuo Zhuan zhu, p. 228); Chunqiu, 10 (Zhuang 26), 9a (Legge, The Ch'un Ts'ew with the Tso Chuen, p. no; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 233); Zuo zhuan, 10 (Zhuang 28), 13a (Legge The Ch'un Ts'ew with the Tso Chuen, p. 114; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 239). Zuo zhuan, 10 (Zhuang 30), 18b (Legge, The Ch'un Ts'ew with the Tso Chuen, pp. 117—18; Yang Bojun, ed., Chunqiu Zuo Zhuan zhu, p. 246). The attacks of Qin against die Rong are related in a later entry in the Zhu zhuan, where it is said that Qin, greedy for land, had persecuted the Rong and forced them to migrate and seek the protection of Jin during the period of reign of Duke Hui S (650-636 B.C.). See Zuo zhuan, 32 (Xiang 14), 8b (Legge, The Ch'un Ts'ew with the Tso Chuen, pp. 463—4; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, pp. 1005-7). Cambridge Histories Online © Cambridge University Press, 2008 924 NICOLA DI COSMO River, only a few hundred people managing to survive."4 The Di threat was so great that it prompted a renewed unity of the states of the Central Plain, which formed a league in 659 B.C. to save Xing. The last mention we have of the Bei Rong is in the year 650 B.C., when they were attacked by Qi and Xu Wr; the "Bei Rong" mentioned here possi- bly refers to the same people as the Shan Rong who attacked Yan ?RI in 664. After this event, the term "Rong" appears vestigially, often in compound with Di to indicate generic non-Chinese peoples to the north and west. Culturally, the Rong seem to have had a keen sense of their difference from the Hua-Xia people. This is plainly expressed by a Rong leader, who is reported to have said that "in what we drink and eat, and in the way we dress, all of us Rong are different from the Chinese [Hua]. We do not exchange gifts with them, and do not understand each others language."1" However, it is not possible to identify them according to distinct ethnic or linguistic affiliations. Possibly they were hunting-farming communities that retained customs and forms of political organization markedly different from the Chinese. They were also associated with pastoral activities, but it is highly unlikely that they had a mature nomadic steppe economy. They also lacked unity, thus eventually becoming part of the Zhou political "galaxy," playing a gradually less important role in the political and military history of the Eastern Zhou. MID-SPRING AND AUTUMN PERIOD TO THE MID-WARRING STATES (CA. 65O-35O B.C.) Development of Early Nomadic Cultures in Northern China The first historical steppe nomads, the Scythians, inhabited the steppe north of the Black Sea from about the eighth century B.C. Pastoral nomadism was their main economic activity, and their society was ruled by a class of mounted warriors, who in Herodotus's Histories are called Royal Scythians. Below this aristocracy were other groups, such as the "agricultural," "nomadic," and "free" Scythians. Archaeologically, early nomadic cultures close to the Scythian, characterized by "specific equestrian armaments, horse trappings and the Animal Style,""6 became dominant throughout the steppe region of Central Asia in the first half of the first millennium B.C. In the 114 Zuo zhuan, n (Min 2), 9a (Legge, The Ch'un Ts'ew with the Tso Chuen, pp. 129-30; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, pp. 265—6). '" Zuo zhuan, 32 (Xiang 14), 10b (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 464; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 1007). '" Karl Jettmar, "The Origins of Chinese Civilization: Soviet Views," in The Origins of Chinese Civi- lization, ed. D. N. Keightley (Berkeley: University of California Press, 1983), p. 224. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 925 northern area of the eastern steppe, which includes South Siberia, the Altai region, Mongolia, and Transbaikalia, evidence suggests that the early nomadic phase may have started as early as the eighth century B.C."7 Two separate centers existed in this region. To the east, in Transbaikalia and Mongolia's southern Gobi region, is a complex characterized by cist- stone tombs, bronze knives with characteristic human and animal decora- tions on the handle, and a Northern Asiatic anthropological type, similar to that of the Xiongnu burials of Noin Ula."8 In Western Mongolia, the Altai region, and Tuva, there are timber-chamber burials similar to those of Pazyryk, as well as petroglyphs, bronze objects, and decorations in the Animal Style typical of the steppe region of Kazakhstan, Tuva, and South Siberia; the totemic sculptures known as "deer stones"; and a racial type with Europoid characteristics.11' This cultural separation can also be traced along the Chinese frontier. Anthropological studies have ascertained the Europoid character of nomadic cultures of Xinjiang, such as the Wusun Mffi and Saka.uo Discoveries at burial sites in Ningxia show that around 500 B.C. the North Asian Mon- goloid component of the population increased considerably and was associ- ated with nomadic cultural forms.121 This suggests the existence of two different anthropological and cultural complexes that came to share a similar way of life and cultural features. Unambiguous elements of horse-riding nomadic culture appear in the eastern steppe before the eighth century B.C., and specific early nomadic cultural features found in northern China — char- acteristic shapes and "animal" decorations of knives, daggers, and belt buckles — belong to the same cultural universe as the contemporary cultures of the South Siberian region. In the eighth and seventh centuries B.C., the needle of the Northern Zone's cultural orientation pointed to the north and west. Questions concerning the historical context in which archaeological data from the Northern Zone should be read have suggested the movement of various Central Asian nomads to the Inner Asian borders of China, some- times thought to have provided the impulse for the transition to nomadism 117 Jettmar. Art of the Steppes, p. 143; Kiselev, Dreviniaia istoriia luzhnoi Sibiri, pp. 302-3. 118 S. I. Rudenko, Die Kultur der Hsiung-nu unddie Hiigelgraber von Noin Ula (Bonn: Habelt, 1969), pp. 99-106. "' E. Nowgorodowa, "Mongolie de l'e'poque du 'style animale.'" in Ethnologic und Ceschichte. Festschrift fur Karl Jettmar, ed. Peter Snoy (Wiesbaden: Harrassowitz, 1983), pp. 440—4. On the Pazyryk culture, see S. I. Rudenko, Frozen Tombs ofSiberia: The Pazyryk Burials of Iron Age Horsemen (London: Dent, 1970). On Siberian and East Asian animal styles, see Shu Takahama, "Early Scytho-Siberian Animal Style in East Asia," Bulletin of the Ancient Orient Museum 5 (1983): 45—51. "° Han Kangxin, "Xinjiang Kongjiaohe Gumugou mudi rengu yanjiu," Kaogu xuebao 1986.1: 361—84; Han Kangxin and Pan Qifeng, "Xinjiang Zhaosu Wusun mu gu renleixue cailiao de yanjiu," Kaogu xuebao 1987.4: 503—23. 111 Han Kangxin, "Ningxia Pengpu Yujiazhuang mudi rengu zhongxi tedian zhi yanjiu," Kaogu xuebao 1995.1: 107-25. Cambridge Histories Online © Cambridge University Press, 2008 926 NICOLA DI COSMO in northern China.'12 However, it is more plausible that though the initial impulse may have come from large-scale migratory movements, the evolu- tion toward fully developed pastoral nomadism soon acquired, in the various areas comprising the Northern Zone, a distinct localflavor,as it blended with pre-existing cultures. The chronological discrepancy between the archaeolog- icalexistence of a more or less homogeneous "early nomadic" bronze culture in the steppe in the eighth century and the later historical appearance of mounted nomads in Chinese sources, in the fourth century B.C., may be due to the slow expansion of contacts between these cultures and China, ham- pered by the presence of intermediate sedentary peoples. Between the seventh and the fourth century B.C., contacts between the Central Plain and the far northwest were established.12' The presence of Chinese silk and lacquer in the burials of the Altai culture of Pazyryk of the fifth century B.C. demon- strates the existence of at least indirect contacts. Moreover, the continued preference for, and even growth in the use of, bronze to produce funerary and ritual objects, even where iron metallurgy was available, separates the northern region, and the eastern steppe in general, from developments in western Central Asia, where ferrous metallurgy gradually replaced bronze.124 The discovery of early Chinese knife-coins in Inner Mongolia dated to the sixth century has been interpreted as evidence of a certain degree of trade between China and the north.12' The archaeological cultures of northern China display a variety of centers that suggest two levels of development. At one level, represented by the metal production and use of ritual and functional tools and weapons similar to the early nomadic complex, they can be related to the continental phenomenon of the spread of nomadic pastoral societies. At a different level, represented by pottery and burial customs, we find a variety of local traditions. Possibly nomadic groups expanded and became dominant over communities that eventually adopted their technology. Alternatively, we may suppose a paral- lel development of the same technology within contiguous but different local traditions, as they progressed on the road of pastoral specialization. "' Gustav Haloun, Seit warm kannten die Chinesen die Tocharer oder Indogermanen iiberhaupt (Leipzig: Verlag der Asia Major, 1926); R. Heine-Geldern, "Das Tocharenproblem und die Pontische Wan- derung," Saeculum 2 (1951): 225. According to Jettmar's more cautious opinion, the transition to nomadism was influenced by a general "contact" with a zone of unrest to the south and west of the steppes; movement in the area of the Volga River seems to have led this process, which was only real- ized over a fairly extended period of time. See Jettmar, Art of the Steppes, p. 215. llJ On the issue of the first appearance of nomads in the Chinese sources, see Owen Lattimore, Inner Asian Frontiers of China (Boston: Beacon, 1962), pp. 60, 341; Pulleyblank, "The Chinese and Their Neighbors in Prehistoric and Early Historic Times," pp. 449—450; Yii Ying-shih, "The Hsiung-nu," in The Cambridge History of Early Inner Asia, ed. Denis Sinor (Cambridge University Press, 1990), p. 118. IU Chernykh, Ancient Metallurgy in the USSR, p. 271. lli William Watson, Cultural Frontiers in Ancient East Asia, pp. 101-2. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 927 Major Archaeological Cultures NORTHEASTERN ZONE. The heart of the late Bronze Age and Iron Age cultures, in Heilongjiang, is to be found at the confluence of the Sungari (Songhua fcTts) and Nonni (Nenjiang) Rivers, known as Song-Nen Plain (see Map 13.3)."6 In this area rich in forests, arable land, and waterways, we find iron metallurgy that can be dated to die eighth century B.C. or even earlier, appearing together with the full blossoming of bronze production. The two most important cultures for this period are Pingyang ^F"'^ and Hanshu 'Still II, both of which present rich metal assemblages of bronze and iron. The Pingyang culture has been identified in burials in southwestern Heilongjiang and eastern Inner Mongolia,127 but no settlements associated with this culture have been found. The type site is Pingyang (Tailai ?H5|5. county, Heilongjiang), which consists of two cemeteries, Zhuanchang SJMic and Zhandou WL-4-, excavated respectively in 1984 and 1985 and dating from the late Spring and Autumn to the middle Warring States period. The burials present very similar characteristics. Most of them are simple rectangular earthen pit graves, often used for multiple burials. A few tombs are T-shaped pit graves. Animal sacrifice was common, in particular dog and horse, with possible preference for the dog. The funerary inventory includes objects in bronze, iron, and gold, a large variety of pottery, and other tools made of bone or stone. Among the bronzes decorative elements prevail and include Animal Style plaques, buttons, and circular discs. The iron items are mostly objects of daily use, such as arrowheads, scrapers, and spearheads, though "tube ornaments" have also been found. The gold objects include three ear- rings and two plates. Most of the finds come from the first cemetery. Other burial goods include stone, agate, turquoise, bone, antler, ivory, and shell. The importance of archery is demonstrated by the discovery of more than 50 bow ends and 240 arrowheads. On the basis of stratigraphic analysis and typological comparison with Bai- jinbao and Sanjiazi H l ^ i 1 (Qiqihaer WWtfiW city, Heilongjiang), archae- ologists have proposed four phases of development, spanning from the late Spring and Autumn to the late Warring States. Some of the mortuary prac- tices are similar to those found at Sanjiazi, such as the coexistence of primary '** Tan Ying-jie et al., "The Bronze Age in the Song Nen Plain," in The Archaeology of Northeast China: Beyond the Great Wall, ed. Sarah M. Nelson (London: Routledge, 1995), pp. 225—50. Tan Yingjie and Zhao Shandong, "Song Nen pingyuan qingcong wenhua chuyi," in Zhongguo kaogu xuthui disici nianhui lunwenji 198} (Beijing: Wenwu, 1985), pp. 196-202. Yang Hu, Tan Yingjie, Zhang Taixiang, "Heilongjiang gudai wenhua chulun," in Zhongguo kaogu xuthui diyici nianhui lunwenji 1979 (Beijing: Wenwu, 1980), pp. 80-96. Li Chenqi et al., "Song Nen pingyuan qingtong yu chuxing zaoqi tieqi shidai wenhua leixing de yanjiu," Beifang wenwu 1994.1: 2-9. 117 Pingyang muzang (Beijing: Wenwu, 1990). Cambridge Histories Online © Cambridge University Press, 2008 Northern Zone Sites of the "Early Nomadic" Period (650-350 B.C.) Map 13.3. Northern Zone: Archaeological sites of the "early nomadic period" (ca. 650-350 B.C.). Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 929 and secondary burials, animal sacrifice, and the custom of covering the deceased's face with bronze buttons. The latter custom is characteristic of sites that as we have seen above, are attributed to the Shan Rong. The type of assemblage suggests a mixed hunting-pastoral economy, with considerable use of metal. It is possible that this area was a center of metal production, but it is doubtful whether its people had achieved a full transition to pas- toral nomadism. The artistic vocabulary, nonetheless, includes the classic motifs of the Animal Style plaques and small ornamental objects. This area, rich in rivers and forests, may have also been important as a route of com- munication between northeastern China and Transbaikalia and Mongolia. The cemetery of Sanjiazi is dated to the Warring States period."8 It is similar to Pingyang, and includes multiple burials in rectangular earthen pit graves, with evidence of sacrifices of horses and dogs. The metal remains include objects in bronze, iron, gold, and silver. Here, too, bronze is used almost exclusively for ornamental objects, whereas iron is also used for weapons and tools. Asforprecious metals, only one golden pendant and one silver earring have been found. A number of bow ends and bone arrowheads points to a developed hunting economy. Silk from China has also been found, suggesting that by the Warring States period, Heilongjiang had some relations with the Central Plain, possibly the state of Yan. The presence of several semisubterranean dwellings and carbonized grains indicates that the Hanshu II culture was a settled one. Located in the same area as the Hanshu culture, it forms an independent complex, extending into the Song—Nen Plain.129 It displays a bronze production that is varied and large, but is typically limited to small tools and ornaments, such as knives, arrowheads, awls, earrings, and buttons. Large bronze objects, such as daggers, are absent. Over fifty clay molds for bronze objects have been recov- ered, revealing local manufacture of spears, buckles, arrowheads, and horse- shaped ornamental plaques. Sandstone molds have also been found for axes and fishing hooks. But even more interesting is the recovery of iron-socketed axes and knives, which are similar to those found in the Central Plain during the Warring States period. The main economic activity was agriculture, but the presence offishinghooks, boat-shaped pottery bowls, pots decorated with painted net patterns, and the large number of fish bones also suggests that fishing was widely practiced. The burial site at Erkeqian —~&W£ (Nehe Wl'ffl county), excavated in 1985,130 has also been associated with this culture. The metal funerary assemblage fea- tures mainly small bronze objects, such as knives, bells, plates, buttons, and 118 Kaogu 1988.12: 1090-8. "* Dongbei kaogu yu lishi igSz.u 136-40. "° An Lu and Jia Weiming, "Heilongjiang Nehe Erkeqian mudi ji qi wenti tantao," Beifang wenwu 1986.2: 2-8. Cambridge Histories Online © Cambridge University Press, 2008 93O NICOLA DI COSMO earrings, as well as some rusted iron objects, among which daggers and knives can be identified. The burials are rectangular earthen pits, in several cases with remains of a male and a female in the same burial and sacrifices of dogs and horses. At this site we find again two types of relics, one belonging to an earlier stratum, probably later than the Baojinbao culture, the other being roughly contemporary with the Hanshu II culture. The difference is marked not only in the pottery assemblage, which in the later burials includes more vessels and decorative patterns, but also in the metal assemblage. In the second, or upper, layer, an iron dagger has been found that is very similar in shape to the bronze daggers of the Northern Zone.'3' The earlier phase has been dated from the Spring and Autumn period, whereas the second phase seems to belong to the Warring States. What the finds suggest is a pattern of cultural evolution, from the Spring and Autumn to the late Warring States, not dissimilar from the one observed in the Ordos region. This seems to emphasize two elements. The first is the creation of a wider and more stable network of contacts through which both artistic motifs and technical innovations could travel rapidly. This network included, in its southern fringes, adjacent parts of China, such as the state of Yan. The second is a more liberal use of iron; bronze was still widespread but increasingly limited to decorative objects. The molds indicate an advanced level of production and possibly show that these sites were centers of cultural diffusion within the Northern Zone. In Liaoning and eastern Inner Mongolia the period ca. 650-350 B.C. cor- responds to the last stage of the Upper Xiajiadian culture. It is best repre- sented by the Shiertaiyingzi "t"—cJ^f site (Chaoyang 4PJI, Liaoning), a cemetery in use since the Western Zhou, whose upper layer is dated to the early and middle part of the Warring States period, that is, ca. 450—350 B.C. The metal inventory is entirely of bronze, and includes curved-blade daggers, mirrors with multiple knobs, knives, arrowheads, ornaments in various shapes, belt hooks, and buckles.'32 Because of the site's location, archaeolo- gists often associate this culture with the people known in historical docu- ments as Dong Hu Jfl$j, but without direct evidence such identification is purely speculative. The cemetery found at Tiejianggou WLTLM (Aohan Banner, Inner Mon- golia), which is located in the eastern part of Inner Mongolia and within the area of distribution of the Upper Xiajiadian culture, is representative.'33 Five tombs of this cemetery have been excavated; they can be divided into two "' Zhao Shantong, "Heilongjiang Guandi yizhi faxian de muzang," Kaogu 1965.1: 45-6. 1)1 Zhu Gui, "Liaoning Chaoyang Shiertaiyingzi qingtong duanjianmu," Kaogu xuebao 1960.1: 63-71. '" This was excavated in May 1991. See Shao Guotian, "Aohanqi Tiejiangguo Zhanguo mudi diaocha jianbao," Nei Menggu wenwu kaogu 1991.1-2: 84-90. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 931 groups. The first group (three tombs) yielded funerary objects such as bronze artifacts, pottery, and stone ornaments, whereas no burial goods were recov- ered from the second. The bronze objects include three knives and various Animal Style ornaments, belt hooks, buttons, earrings, and arrowheads. These burials are comparatively poor. Since a number of decorative and styl- istic elements link this site with Xiaobaiyang / h f i i l (Xuanhua rn.it county, Hebei), its occupants may been have southern immigrants, pushed north — and eventually subjugated — as a consequence of Yan's expansion. The area was conquered by Yan in 299 B.C., the terminal date of the site. NORTH-CENTRAL ZONE. The north-central frontier in this period pre- sents a fluid picture. Some sites display traits that already foreshadow the appearance of a Xiongnu culture. Others show a lesser degree of change with respect to the previous period. Though ethnic or historical attributions are highly speculative, these sites are frequently associated with the Shan Rong and the Di. Archaeological sites attributed to the Shan Rong are scarcely consistent with any specific culture. Generally speaking, these are burial grounds whose system of interment differs from that of the Central Plain, and whose funer- ary inventory shows mixed derivation and cultural affiliation. Many of these sites are located to the northeast of the Central Plain. The cemetery sites at Hushiha Paotaishan !%itUn]&alU (Luanping iiFF- county, Hebei) and at Jundushan ¥ # l l l (Yanqing county, Beijing),'34 have been dated to the late Spring and Autumn and early Warring States periods and attributed to the Shan Rong. The chief common characteristic of these sites is that while they show a clear association with the culture of the Central Plain, several dis- coveries link them with Northern Zone cultures. The Hebei site, for instance, participates in the last phase of the Upper Xiajiadian culture, whereas in the Yanqing burials a great number of weapons and horse fittings are found that show a clear association with the Ordos—Xiongnu bronze culture. Similar objects are also found in northern Hebei. Moreover, there are several types of burials, such as earthen pits, stone chambers, and wooden coffins. In some cases there are combinations of both stone chambers and wooden coffins. One particular aspect of the burial custom consists of covering the face of the deceased with sackcloth decorated with bronze buttons, similar to the custom found in Heilongjiang. Other cemeteries, to the north of the Beijing area, are also characterized by a combination of more than one burial practice. At a site in Baotou 'S.H county, Inner Mongolia, excavated in 1988 and attributed to the Lin Hu, 1)4 Wenuiu ziliao congkan 1983.7: 67—74; Wenwu 1989.8: 17—35. 43- Cambridge Histories Online © Cambridge University Press, 2008 932. NICOLA DI COSMO both rectangular earthen pits and catacomb-style burials were found.'35 Animals were sacrificed and buried with the deceased on secondary plat- forms. The funerary assemblage is dominated by bronze ornaments, such as buckles, plaques, rings, buttons, earrings, and other small ornaments. In addition to these, there are tools and weapons, such as a spoon, an arrow- head, belt hooks, buckles, and knives. The bronze buckles are in typical northern Animal Style, similar to finds from Maoqinggou %StM (Inner Mongolia). The three-winged arrowhead found here has a very large distri- bution, including Inner Mongolia (Liangcheng W.$L county), Hebei (Bei- xinbao ^G^%., in Huailai |§5(£ county), and Liaoning (Zhengjiawazi JtP^ %S.-f, Shenyang 'MM).')6 Similar specimens have also been found in Iron Age burials (sixth century B.C.) in Transbaikalia. Several elements, such as a bronze semiannular pendant, similar to a silver one found in Guyuan EIIK county, Ningxia, and the shape of the catacomb burials, common to the Xindian and Kayue cultures, also suggest contacts with the northwest. While the assemblage in Baotou exhibits similarities with later Xiongnu sites, it is remarkable for the absence of some of the most typical elements of later nomadic cultures: iron, gold, and horse fittings, as well as the daggers, pickaxes, and plaques typical of the Ordos bronzes. Another Inner Mongolian site attributed to a northern people, the North- ern Di, is that of Guoxianyaozi ^ H ^ ^ (Liangcheng county, Inner Mon- golia), excavated in May-July 1983.'37 Archaeologists believe three phases of development can be identified, from the late Spring and Autumn to the early Warring States. The burial practices are similar to those at Jundushan. They include rectangular vertical earthen pits — sometimes provided with head niches and secondary platforms, wooden coffins, stone chambers, or a com- bination of both. Interments are single in extended supine position with the head to the east. Animal sacrifices were practiced, and males were typically buried with horses, deer, or sheep, whereas females were buried with sacri- fices of cattle and sheep. However, we do not find in Guoxianyaozi the custom of covering the face of the deceased with sackcloth decorated with bronze buttons. The assemblage consists mostly of bronze ornaments, such as buckles, plaques, buttons, bells, rings, and earrings. Among the tools, we find two knives and a pickaxe. The plaques are particularly abundant - forty-four - and are in both Geometric Style and Animal Style. The buckles and the button ornaments establish a context for this site that is typical of the Ordos region. Similar buckles were found at contemporary or later Ordos sites, such as Taohongbala flfe&EBJi, Fanjiayaozi fS^gg-? (Helin'geer W#t&lf), and '" Net Menggu wenwu kaogu 1991.1.: 13—14. 1)6 O n Zhengjiawazi, see Kaogu xutbac 1975.1: 141—56. '" Kaogu xuebao 1989.1: 57-81. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 933 Xigoupan WMffi. Similar bronze bells have been found at Beixinbao.1'8 Among the other materials, stone beads are particularly numerous, with some of turquoise and one of agate. The bone arrowheads are similar to the bronze examples found in Taohongbala and are probably a poorer prototype. Despite all the similarities with die early Xiongnu sites, specimens of the most advanced technology available in the area at this time, such as the iron daggers and horse fittings found at Taohongbala and Maoqinggou, do not appear here. Although horses were bred and used in sacrifices, the people of Guoxianyaozi do not seem to have had a highly developed horse- riding culture. Their metal inventory indicates a people rather different from the typical early nomads and points to a pastoral-hunting community that had established contacts with other more powerful mounted nomads who were gradually penetrating the area and establishing themselves throughout the steppe belt of Inner Mongolia and especially in the Ordos region. In Yanqing the metal inventory is dominated by bronze weapons and horse fit- tings. Whereas ornaments such as plaques, belt hooks, buckles, and bells are reminiscent of the Guoxianyaozi site, the discovery of about a hundred daggers with straight blade, ge dagger-axes, and axes indicates a southern extension of a martial horse-riding community of the Ordos type. They are also likely to have had trade relations with the Central Plain, as knife-coins have been found from die late Spring and Autumn and early Warring States. The earliest Xiongnu-type bronze and early iron sites in the Ordos region are the cemeteries of Taohongbala and Maoqinggou (Inner Mongolia). The excavation of Taohongbala has brought to light a large number of small ornamental objects, which include plaques, buckles, rings, button-like and pea-shaped bronze decorative elements, and ornaments with a double-bird motif.'39 Stylistic affinities connect Taohongbala not only with Warring States sites, but also with earlier Upper Xiajiadian sites. Similar bronze plaques have been found, for instance, at Nanshan'gen. Bronze daggers in the so-called Antennae Style {chujiao shifflfiljK) are widespread and found, among others, at Beixinbao (Hebei) and Fanjiayaozi (Inner Mongolia). The ring ornaments are similar to those seen in Fanjiayaozi. The metal inventory also includes a pair of gold earrings like those seen in Nanshan'gen and Beixinbao. Taohongbala was originally regarded as a site of the Bai Di &ik, and has only later come to be recognized as a Xiongnu site. This is undoubtedly due l3< Kaogu 1966.5: 231—42. '" Tian Guangjin, "Taohongbala de Xiongnu mu." The site consists of seven tombs excavated in 1973. In the original report published in 1976, Taohongbala was dated to the Warring States and regarded as a Xiongnu site on the basis of typological similarities widi Xiongnu sites in Inner Mongolia, such as Fanjiayaozi, and the presence of iron objects. In the reprint of 1986 die site is attributed to die late Spring and Autumn period on the basis of carbon dating to approximately the sixth to fifth century Cambridge Histories Online © Cambridge University Press, 2008 934 NICOLA DI COSMO to the difficulty of attributing a site to a people — the Xiongnu — who appear in the historical records only two centuries later. However, the pottery found here, and in particular the brown single-ear guan Wt pots, handmade and fired at low temperature, shows continuity with the Xiongnu sites of the Warring States period, such as Xigoupan and Aluchaideng WH^^i:, which . present a type of gray pottery that is a more refined development of the earlier brown variety. This suggests that radical changes in the population of the area, if they occurred at all, must be dated to the late Spring and Autumn rather than to the mid-Warring States period, when the Xiongnu make their appearance in the written sources. The other important Xiongnu site is that of Maoqinggou,140 which pre- sents a four-phase chronological evolution. The first phase, dated to the late Spring and Autumn period, contains pottery and bronze items. The treat- ment of the body, shape of the burial, and animal sacrifice have clear paral- lels at the previously discussed Guoxianyaozi site. Other similarities are seen in the absence of iron and the presence of a large number of bronze orna- mental plates. But there are several discordant elements, such as the presence, in Maoqinggou, of a bronze dagger with double-bird-head pommel, of a bronze bit, and of belthooks, all of which are missing in Guoxianyaozi. It is possible that the early people of Maoqinggou belonged to the same cultural milieu as those of Guoxianyaozi but were starting to develop in the direc- tion of a more specialized pastoral nomadism. The later period of Maoqinggou, inclusive of early, middle, and late Warring States — phases II, III, and IV, respectively — shows a great differ- ence with respect to phase I, such that archaeologists have attributed the first period to the Di and the subsequent ones to the Loufan $£'M, a people who were certainly horse-riding steppe nomads and culturally related to the Xiongnu. In these later burials a large number of iron objects, including daggers, pickaxes, one knife, ornamental plates, and belthooks, have been found (Fig. 13.3). The typical jar with a small mouth, round belly, and round bottom of the early period is replaced with a similar example with a flat bottom. In the upper layer a knife-coin has been found, which again indi- cates the existence of trade with China. The appearance of iron is incre- mental, used not only for weapons and tools but also for ornamental plates. This site is thought to have been abandoned at the beginning of the third century as a result of occupation by Zhao. The site's economy shows both pastoral and agricultural elements. Remains of a settlement, kilns, and pottery are next to the cemetery, whose funerary goods show the same military and ornamental inventory that is w See Tian Guanjin and Guo Suxin, eds., O'erduosi qingtong qi, pp. 227-315. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 935 Figure 13.3. Bronze daggers and Animal Style plaque from Maoqinggou. From: Tian Guanjin H S & and Guo Suxin f i ^ f r , eds., O'erduosi qingtong qi "$>W%%\TtWlWi, plates XXVI, XXVIII, LXVII (Beijing: Wenwu, 1986). Cambridge Histories Online © Cambridge University Press, 2008 NICOLA DI COSMO characteristic of the Northern Zone. The weapons are mostly bronze daggers and arrowheads in the earlier part of the site, with more iron objects in the later graves. Ornaments are in the animal style, with the bird as a favorite subject. Particularly common are the bronze plates with a double-bird design and a plate shaped in the form of a bird. Besides these, there are objects related to horse management, such as rein rings and bits. In sum, we find here an assemblage typical of an early nomadic culture and closely related to that of Taohongbala. The split between the earliest occupancy and the later tombs seems to support the hypothesis of a gradual affirmation of pastoral nomadism in this area. In the case of Maoqinggou, the shift must have taken place in the course of the sixth century. Given the similarities in mortuary practices and pottery types between the two phases, an internal evolution is likely to have occurred, though the new technology may also point to intru- sive elements. However, the hypothesis that a new aristocracy of "Scythian- type" nomads might have extended its rule to this area is, at the present stage, insufficiently supported, requiring the existence of an original homeland that cannot presently be established. Another site that can be attributed to the early nomads is Hulusitai Hf # W\X (Wulate Zhonghou Lianhe Qi H f e ^ f t W ^ S l , Inner Mongolia).14" This is dated to the early Warring States (fifth to fourth century B.C.) and belongs to a group of transitional sites between the late Spring and Autumn and the middle to late Warring States Xiongnu sites, which also include Fanjiayaozi and Shuijiangoumen T K ^ ' ^ P ^ (Tumote Youqi i.H^.'WfcM, Inner Mongolia).'42 The assemblage is very similar to those at Taohongbala and Maoqinggou, but presents also more advanced elements, which are found in later Warring States Ordos sites (Zhunge'er ^PI&M Banner) such as Yulongtai 3Lf$kJSL, Xigoupan, and Sujigou M18?#. The bronze tools and weapons such as bronze daggers, arrowheads, knives, axes, and pickaxes are very close to late Spring and Autumn types. Horse fittings are also similar to the earlier types. Innovations appear mostly in the area of ornamental objects, as in the case of decorative waist belts, which foreshadow the golden hu ^ belt of the later period. Modifications in die production of traditional objects were also carried out, and certain features were standardized, as in the case of the wing-shaped dagger guard.143 This site has been attributed to the Xiongnu. This analysis of the Ordos and contiguous areas seems to indicate the exis- tence of non-horse-riding communities living together, or in close proxim- ity, with more advanced Scythic, early nomadic people already adept in the "' Ta la and Liang Jinming, "Hulusitai Xiongnu mu," Wenwu 1980.7: 11—12. *** On the latter see Tian Guanjin and Guo Suxin, eds., O'erduoti qingtong qi, pp. 220-1. M) Tian Guangjin, "Jinnianlai Nei Menggu diqu de Xiongnu kaogu." Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 937 use of the horse for riding and war. In Hebei, Liaoning, and Inner Mongo- lia, a radical transformation was taking place, which must surely be associ- ated with the rise of new types of societies, both more mobile and militarily superior. It is not excluded that at least part of this transformation may have been caused by the arrival of northern people from Mongolia and Trans- baikalia. They brought new types of burial systems,144 such as the wooden and stone-cist coffins, and a new burial inventory, in which symbols of a warlike, horse-riding culture predominated. It is this movement that may have been responsible for the sudden acceleration of pressure on the north- ern frontier from people such as the Chi Di 7?F$k, Bai Di, and Shan Rong in the mid-seventh century B.C. The so-called pre-Xiongnu culture, therefore, should be seen as a syn- chronic evolution of different core areas where a true nomadic aristocracy established itself either by migration or internal evolution. Throughout the steppe and mountain areas of the northern region, increasingly homogeneous material culture, religious beliefs, and rituals were adopted, some of which coexisted and blended with the mortuary practices of preexisting and neigh- boring people. Due probably to increased contacts with China, the charac- ter of this aristocracy gradually began to shift from a notion of power and status symbolized by weapons and tools, to one in which wealth, accumu- lated in precious metals and stones, horses, and ornamental art, became its predominant pursuit. NORTHWESTERN ZONE GANSU. The Hexi MM Corridor, an arid region in northern Gansu between the Yellow River and the steppe region of the eastern Tianshan region, is home to the Shajing ijflt culture, distributed over Minqin SStl, Yongchang TKH, Gulang "£"'?&, and Yongdeng TRS counties.145 Its chronol- ogy with respect to other regional cultures such as Siwa, Xindian, and Kayue, is not clear, but it is probably later than Xindian, partly overlapping with the Kayue culture, and probably dating from the Spring and Autumn to the Warring States. Its type site is Shajing cun ^J^^f (Minqin county) exca- vated in 1923—4 by Andersson.'46 It consists of a fortified dwelling site and a cemetery of forty graves. The settlement is surrounded by an earthen wall, and among the metal finds both in the settlement and in the funerary assem- blage we find small bronze items: spearheads, arrowheads, knives, and orna- ui On Xiongnu burials, see S. Minyaev, "Niche Grave Burials of the Xiong-nu Period in Central Asia," Information Bulletin. International Association for the Cultures of Central Asia, 17 (1990): 91-99; S. Minyaev. "On the Origin of the Hiung-nu." Information Bulletin. International Association for the Cul- tures of Central Asia 9 (1985): 69-78. '•" K. C. Chang, The Archaeology ofAncient China, pp. 407-8. See also Kaogu yu wenwu 1981.4: 34-6. 144 J. C. Andersson, "Researches into the Prehistory of the Chinese," BMFEA 15 (1943): 197-215. Cambridge Histories Online © Cambridge University Press, 2008 NICOLA DI COSMO ments. Among the other remains, cowrie shells, turquoise beads, and marble rings were used as ornaments. One of these, a three-lobed object with spiral design, is similar to one found in a Warring States tomb at Luanping ifl^F- county, Hebei.147 A similar fortified settlement has also been found at San- jiaocheng Hft$c (Yongchang county),148 where a late Warring States date is suggested by the presence of an iron hoe. More iron objects have been unearthed at Yushugou Ho®^ (Yongdeng county), which marks the southern extension of the Shajing culture. At this site, one tomb has been excavated; it presents a number of features roughly comparable to those of the Ordos sites of the Spring and Autumn period. Animal sacrifices of horse, sheep, and cattle are evident. The bronze objects include mainly ornaments in the animal style (eagle, deer, and dog), but also a chariot axle end. The iron production is limited to tools, such as an object in the shape of a spade, a spearhead, and a drill. The ornaments can be com- pared to those of Northern Zone sites of the Spring and Autumn and Warring States periods such as Taohongbala (round ornaments), the Zhongshan "^lll state in Pingshan T'lil, Hebei (openwork round ornaments with a whirlwind design), and Xigoupan (eagle head). The appearance of a reddish, coarse- grained jar with a double loop on the shoulders typical of Shajing shows the affiliation of this site with the Shajing culture. The presence of settlements and agricultural tools reveals a farming culture. These people also bred animals, as indicated by the animal sacrifices. The presence of fortifications suggests conflicts with neighboring peoples, who were probably nomads. Moreover, ornaments closely related to those of the Ordos and Hebei Northern Zone sites, such as cowrie shells and turquoise beads, are evidence of trade with those areas. NINGXIA. The ancient cemetery at Yanglang H§§P (Guyuan county, Ningxia), dated to the Eastern Zhou period, shows strong similarities with the early Ordos Xiongnu sites of Taohongbala and Maoqinggou.149 The pre- dominant burial style is the catacomb grave, so called because of its L-shaped configuration, with the body placed in a lateral locule, or niche, typically lower than the main shaft of the grave, and sealed with stones. Over 800 objects of bronze, iron, gold, silver, and bone, in addition to 2,000 beads, were found. These were not specifically made for funerary purposes, but were typically objects of daily use that had belonged to the dead. They reveal strong local characteristics, particularly in the bone items and clothing orna- ments. The graves are divided stratigraphically into two periods, the first from the late Spring and Autumn and early Warring States, and the second 147 K. C , Chang, The Archaeology of Ancient China, p. 407. 148 Kaogu 1984.7:598—601. 149 Kaogu xuebao 1993.1: 13-56. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 939 Figure 13.4. Bronze sword handle and Animal Style ornamental objects from Yanglang. After: Xu Cheng 'SffS., Li Jinzeng $if|Jp, "Dong Zhou shiqi de Rong Di qingtong wenhua" Jfcffl Kaogu xuebao 1993.1: 1-11, figs. 3 and 4. Drawings by Li Xiating. from the late Warring States. Similarities across both periods point to con- tinuous presence by the same people: grave shape and construction, body orientation, and burial custom. The tombs of the first period contain a greater number of bronze artifacts. Among the weapons and tools we find ge dagger-axes, spearheads, daggers, knives, arrowheads, pickaxes, drills, and chisels, and a large number of orna- ments, such as buckles, belt ornaments, earrings, and belt hooks (Fig. 13.4). The daggers, in the classic Antennae Style, are found only in graves of this Cambridge Histories Online © Cambridge University Press, 2008 94° NICOLA DI COSMO period. At this time, iron was not widely used, though it was certainly known, as shown by the finding of fragments of an iron sword (tomb IM3). Horse and chariot fittings are present in the early period, but not in large quantity. Among the precious metals, only silver earrings are found in the earlier graves (tomb IIIM3). Noteworthy in the funerary assemblage is the presence of bones of sacri- ficed animals, in particular skulls of horses, bovines, and sheep. Agriculture may also have been practiced, but evidence is limited. Pottery is also scarce. All data indicate that the people of Yanglang were predominantly pastoral- ists, whose considerable wealth is evident from the resources allocated to burial rituals in terms of labor, animals, and objects. Almost every grave con- tained funerary goods — usually more than ten objects and several with over fifty. Similar sites have been found in Guyuan county, at Pengbao 1&%. and Shilacun 5?l!ltt.ISO The majority of the burials in Pengbao are T-shaped cat- acomb graves. The Pengbao cemetery is particularly important since, of the thirty-one graves excavated, twenty-seven were undisturbed. Animal sacrifice was practiced and documented by the presence of heads and hooves of horses, sheep, and cattle at both sites. The type of burials, divided between T-shaped catacomb types and vertical-pit graves may again point, as in the case of some Ordos sites, to a cultural admixture that possibly reflected the cohabitation, and even fusion, of different groups. Objects recovered from both types are very similar, including weapons, ornaments, and horse fittings, all in bronze. The absence of iron objects may indicate that these artifacts were deemed unsuitable as funerary objects. The ge dagger-axe is often present in assem- blages of this period, as is the short straight-edge dagger. Horse technology was fairly advanced in Pengbao and included bronze bits, masks, and a bridle frontal piece. Though bronze weapons and tools are predominant, the greater role played by the mounted horse in this culture, the large selection of Animal Style ornaments, and some gold finds foreshadow the type of changes in the funerary assemblage that were to take place in the middle and late Warring States period.1'1 Another transitional site is a burial ground found in Zhongning 4 " ^ county.1'2 The two graves excavated here in 1983 are attributed to the early Warring States period. The burials are single rectangular earthen pit graves, 1S ° Zhong Kan, "Guyuan xian Pengbao Chunqiu Zhanguo muzang," Zhongguo kaoguxue nianjian 19SS (1989): 255—6. Luo Feng, "Ningxia Guyuan Shilacun faxian yizuo Zhanguo mu," Kaoguxue jikan 3 (1983): 130—1, 142. Luo Feng and Han Kongle, "Ningxia Guyuan jinnian faxian de beifang xi qingtong qi," Kaogu 1990.5: 403—18. "' On the Pengbao Yujiazhuang burial site, see Kaogu xuebao 1995.1: 79-107. On the physical character- istics of its inhabitants, see Han Kangxin, "Ningxia Pengbao Yujiazhuang mudi rengu zhongxi tedian zhi yanjiu," Kaogu xuebao 1995.1: 107-25. '" Kaogu 19879: 773-7- Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 941 with a body orientation from east to west, a type of burial close to that of the contemporary sites of Yanqing county and Maoqinggou. There is evi- dence of horse sacrifice, and the metal artifacts include bronze and gold. The bronzes consist of weapons and tools (daggers, knives, pickaxes, axes, and arrowheads), horse fittings, and ornaments. The weapons display traditional or even archaic features that would place them typologically in the late Spring and Autumn period or even earlier. The presence of a round golden plate and of many ornaments and horsefittings— bits, chamfrons, ornamental bells — however, suggests a later date, closer to the early Warring States period. In general, these early nomadic sites in Ningxia share important cultural traits with the Ordos and suggest the presence of a similar military aristoc- racy. From the seventh to the sixth century the herds probably increased, as shown by the ample animal remains. Chamfrons and bits, though still limited in number, indicate a progressively more important role of the horse, used not only for transportation and herding, but also war. In all these aspects the Northern Zone resembles the general evolutional pattern of the Eurasian early nomadic universe. XINJIANG. The silk and laquer of obvious Chinese provenance unearthed at Pazyryk and at the Alagou II cemetery show that the region to the far northwest of the Central Plain had a degree of interaction with China in this period. Archaeological studies have allowed a partial mapping of the pres- ence of nomadic people referred to as Saka in Xinjiang from the eighth to third century B.C. (Map 13.4).1" The term Saka is the Iranian form of the Greek "Scythian," which entered Chinese sources as Sai H, read sBk in ancient Chinese.'54 Historical infor- mation on the Saka is contained in the Han shu (History of the [Former] Han) biographies, "Zhang Qian Li Guangli zhuan" Wc^^M.^'Jffi and "Xiyu zhuan" M^$M-. According to these the Saka was the original inhabitants of the land (to the west of the Xiongnu) that were later invaded and conquered by the Wusun (the Great Yuezhi M R) and finally overtaken by the Xiongnu, in the course of their war against the Yuezhi.1" The archaeological "Saka '" On the Saka culture in Xinjiang, see Wang Binghua, "Gudai Xinjiang Sairen lishi gouchen," Xinjiang shehui kexue yanjiu 1985.16: 8-19. This has been translated by Corinne Debaine-Francfort and pub- lished as "Recherches historiques pre'liminaires sur les Saka du Xinjiang ancien," Arts Asiatiques 42 (1987): 31-44. IM See Bernhard Karlgren, Analytic Dictionary of Chinese and Sino-Japanese (Paris: Librarie orientaliste Paul Geudiner, 1923), p. 233, idem, Grammata Serica Rtcensa, BMFEA 29 (1957): no. 908a, p. 240. A. F. P. Hulsewe' and Michael Loewe, China in Central Asia: The Early Stage 12$ B.C.-A.D. 2}. An Anno- tated Translation of the Chapters 61 and 96 of the History of the Former Han Dynasty, Sinica Leidensia 14 (Leiden: Brill, 1979), p. 104. "' Han shu, 61, 2692; 96A, pp. 3897, 3884; 96B, p. 3901. See also Hulsewe' and Loewe, China in Central Asia. Although in recent Chinese publications the characters fl K are given the reading "Yuezhi" (e.g., Wang Liqi, ed., Ship zhuyi [Xi'an: San Qin, 1988], vol. 4, p. 2591), scholars in China often read the Cambridge Histories Online © Cambridge University Press, 2008 942 NICOLA DI COSMO ' \ . Northern Zone Sites ^ Z . . Balkhash \ in Xinjiang 0 -•• J \ .Nileke Zhaosu County \ '•T Al ago u V ^^" Qunbake . m S / t~ • • ^ s .-- ; 7 Tarim River N ( \ —. - - . # Xiangbaobao r 0 200 mi 0 200 km / \ (V y * -- f' / " • ) ) Map 13.4. Northern Zone: Archaeological sites in Xinjiang. culture" has been based primarily on the discovery of a cache of bronzes in Xinyuan W\W- county (Hi ffiM, Xinjiang), in 1983. However, no Saka sites have been excavated. One of the most interesting bronzes is a small statue (42 cm) of a genuflecting warrior, holding something (now lost) and wearing a high hat with a flat circular rim ending in a point turned downward in the front (Fig. 13.5a). The physical features of the man, who is naked with the exception of a kilt-type skirt, are unquestionably Europoid. Among the other finds are a square bronze basin with zoomorphic motifs, a large fu ^ caul- dron, and two heavy rings with facing animal heads (Fig. 13.5b). Most bronzes show clear connections with South Siberia, the Altai region, and Central Asia, but the cauldron, cast in sectional molds, points to a Chinese technique that character H (yue) as rou, a variant of ^ ; hence, the people's name is read Rouzhi, not Yuezhi. For instance, in the Zhongguo Sichou zhi lu cidian, ed. Xue Li (Urumqi: Xinjiang Renmin, 1994), the entry "Yuezhi Dudu fu" fl RfPfFfl? is listed under the reading rou (Index, p. 57); on the other hand, the same entry is found under yue in Xinjiang lishi cidian, ed. Ji Dachun (Urumqi: Xinjiang Renmin, 1993), p. 750. I could not find specific research to help with this issue, but the variant B for ^ is attested to in the Kangxi zidian, section "Wei" S=, part 3 (xia T). Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 943 Figure 13.5. Statuette of warrior (a), and bronze cauldron (b), Saka culture, Xinjiang. From Mu Shunying 8$l:5& et al., Zhongguo Xinjianggudai yishu/The Ancient Art in Xinjiang, China c :l ( SilfiIlS'fti£i$(, pp. 44-6. (Urumqi: Xinjiang Art and Photography Press, 1994). must presumably have been imported through the Northern Complex in the Western Zhou or Spring and Autumn periods. North of the Tianshan Mountains of western Xinjiang, a poorly known culture represented by large earthen kurgans, all visible on the surface, has been attributed to the Wu Sun, a people who, according to historical sources, moved to this area from Gansu only during the Former Han under pressure from Xiongnu westward expansion. The anthropological type has been rec- ognized as Europoid, and the dating has been thought to be around 550—250 B.C. One of the few excavated Wu Sun cemeteries (Xiata Xt^, in Zhaosu Bg M county, Xinjiang) has revealed different types of burial customs, dated to three different phases."6 The earliest phase, probably pre-Han, exhibits a funerary chamber with earthen walls and an entrance reinforced with wooden poles, whereas the later tombs make greater use of timber. The assemblage includes small bronze and iron objects. Iron and gold objects appear in greater quantity in burials of the later period. Findings of silk, undated, may be contemporary with the Pazyryk finds and point to possible contacts with the Central Plain. Another example of Saka culture is the Xiangbaobao W R 3 cemetery in 116 Wenwu 1962.7-8: 98-102. Cambridge Histories Online © Cambridge University Press, 2008 944 NICOLA DI COSMO the Pamir region (Tashkurgan, Xinjiang),'57 where forty tombs have been investigated. Two different types of burial customs were evenly distributed: interment in stone kurgans, and cremation. The latter custom is attributed to the Qiang. The funerary assemblage includes only small bronzes, almost all ornamental, some of which, such as belt plaques, resemble ones from the Ordos region. In general, this site shows signs of partly settled, partly nomadic habitation and a relatively poor grave inventory. The people are Europoid of the Indo-Afghan type common to regions of Central Asia. The later Saka phase, represented by the Alagou MikM II culture of the Alagou necropolis (Toksun county, Xinjiang) to the south of the Tianshan range, is dated to the Warring States and Former Han periods.1'8 The Alagou I phase, attributed to the Gushi MlSF people, already displays elements char- acteristic of a pastoral culture. Alagou II, however, has a far richer funerary inventory, including large bronzes such as a square basin similar to the one found in Hi, decorative plaques in gold and silver, small iron knives, lacquer, and silk. The decoration on the ornamental plaques, with facing tigers, recumbent felines, and wolf heads in gold and silver, belongs fully to the Ordos artistic idiom. This decorative art and the presence of luxury goods imported from China hint at the presence, in Xinjiang, of a possible evolution in the funer- ary inventory from bronzes used for practical or ritual purposes, such as weapons and vessels, to ornamental objects and the use of precious metals. A similar pattern can also be discerned in the Ordos region. Archaeologists date the use of copper mines found in Xinjiang in Nileke RLW]J& county (Hi) to the period 700-490 B.C. This dating places the mines in the context of the bronzes found in the Hi region and attributed to a Saka cultural sphere. The prolonged use of the mines indicates that these sites were centers of metallurgical production, and their extensive exploitation may have been an important, perhaps decisive, factor in the expansion of Saka culture in the region. Metal Artifacts Associated with Early Nomadic Sites BRONZE P R O D U C T I O N . Bronze objects from Scythian sites in northern China comprise fairly typical nomadic objects, which have a wide distribu- tion throughout the Eurasian steppe belt. Among these the most character- istic are the straight blade double-edged dagger, horse gear, large ritual cauldrons, ornamental plaques, and belt buckles. The daggers from Ordos sites such as Taohongbala and Maoqinggou are '" Kaogu xuebao 1981.2:199-116. "* Wenwu 1981.1: 18-22. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 945 different from those of the Shang and early Zhou periods. A new and rapidly spread ornamental feature is the dagger pommel with facing bird heads, or Antennae Style. This motif is very common to the north and west, and is found widely in South Siberia. Horse fittings increase in quantity, and new horse bits appear, showing signs of experimentation and technological progress. The considerable number of horse masks, or chamfrons - triangular and round bronze plates worn on the head of die horse for protection - indicate diat the horse was not only used for transportation or herding, but also was ridden in battle. Bronze cauldrons, most probably used to cook the meat of sacrificed animals, have a very wide distribution along the northern frontier, extend- ing even into Central Asia and eastern Europe.1" In nordhern China they are found in Heilongjiang, Hebei, Shanxi, Shaanxi, Gansu, and Xinjiang. They have a circular section, straight sides, a round bottom over a conic foot, and round handles often decorated with mushroom-shaped knobs. Other caul- drons have three feet with zoomorphic motifs. Pottery prototypes of these cauldrons can be seen at die Kokel' cemetery in Tuva (South Siberia).'60 But the most characteristic elements of this culture are the Animal Style plaques and buckles. The animal combat motif, which shows West Asian influences,"5' appeared suddenly throughout the Ordos region. Most fre- quently this takes the form of wolves, tigers, and leopards attacking large herbivores, such as bucks and bulls, the whole scene being inscribed in rectangular or circular frames and represented in openwork over a flat surface. The boar is prominent in the animal "pantheon" of die steppe artistic vocab- ulary, as are birds of prey. Odier popular features are small statuettes of animals in the round represented in various positions, such as the kneeling deer, coiled or crouched leopard, or recumbent horse, as well as bird-shaped plaques and double-bird designs. Plates decorated with abstract motifs were linked together to form a metal belt, a distinct component of Xiongnu attire commonly found in "pre-Xiongnu" and Xiongnu sites. Occasionally, orna- mental objects were cast in gold and inlaid with turquoise, thus bearing a striking resemblance to Scytho-Siberian gold artifacts from South Siberia and Central Asia. "' On the bronze cauldron, see Liu Li, "Tong fii kao," Kaoguyu wenwu 1987.3: 60-5. For a comprehen- sive study on bronze cauldrons across Eurasia, see Miklos Erdy, "Hun and Xiong-nu Type Caldrons Finds Throughout Eurasia," Eurasian Studies Yearbook 67 (1995): 5—94- l&> Roman Kenk, Das Crdberfeld der hunno-sarmatischen Zeit von Kokel', Tuva, Sud-Sibirien (Munich: Beck, 1984), pp. 60, 109-42. 141 The Near Eastern origin of the animal style is rejected by several modern scholars, but western Asian motifs, in particular from Assysian and Achaemenid an, are assumed to have played a role in the evo- lution of the animal style throughout Eurasia, including the Northern Zone. See Sher "On the Sources of the Scythic Animal Style." p. 55; and William Watson, Cultural Frontiers in Ancient East Asia, p. 112. Cambridge Histories Online © Cambridge University Press, 2008 NICOLA DI COSMO IRON METALLURGY. Iron appears fairly early in the Northern Zone, point- ing to the introduction of iron metallurgy to China from the north. The ear- liest sites with iron are associated with the Scytho-Siberian sites in the Altai Mountains (Xinjiang) and can be dated around the ninth century B.C. Typ- ically these are small objects, suggesting that the use of iron was still rare. In the central and eastern parts of the Northern Zone, iron objects came into wider usage in the seventh century. One of the earliest weapon-related uses of iron can be seen in the bronze-hilt iron swords found in Ningxia.'62 Iron items have also been found in late Spring and Autumn burials in Inner Mongolia. One iron sword has been found in burials 1 and 2 at Taohong- bala, together with bronze weapons and horse ornaments. The earliest com- plete iron objects are daggers and pickaxes, confirming an early use of iron for weapons and tools, whereas bronze was preferred for ornaments until a much later date. Both in the Ordos and in the Gansu—Ningxia regions, iron is far more common in the Warring States period, with iron ornamental plates, belthooks, and tools. Tomb 2 at Xigoupan yielded a sword, a ladle, a drill, and several horse-related items, such as a horse bit and two cheekpieces. A similar inventory has been recovered from a burial of the same period in the Ordos region, at Yulongtai. Larger iron implements, such as a ding at tomb 2 at Budonggou MfPIflf, started to be made only in the Former Han period. In the Ningxia area, the Yanglang site is particularly rich in iron tools and weapons, dating partly to the late Spring and Autumn, but mostly to the late Warring States period. The earliest finds, in grave I3, include an iron sword, two rings, and two belt ornaments. From later burials other items have been unearthed, such as a bronze-hilt iron sword, a complete iron sword, knives, rings, horse bits, belt ornaments, cheekpieces, a spear, and an ornamental plate. The bronze-hilted iron sword can be associated with the earlier site of Langwozikeng ?lli§i~F"irL, and only appears in the Ningxia—Gansu region. Most iron objects in Gansu are also associated with Warring States or possibly earlier sites. An interesting site is that of Yuanjia MM (Pingzi W~~F, Ningxian 3^S^ county), where an iron spearhead has been recovered together with bronze objects that include both weapons - a ge dagger-axe, a dagger, arrowheads - and ornaments.'6' This site includes a horse sacrificial pit that contained bronze bells, chariot finials, horse-head pole tops, and horse-head ornamental plates. The presence of weapons, and Animal Style decorative motifs, as well as the importance attributed to the horse, qualify this as a typical early nomadic site. In Houzhuang %M. (Zhengning I E ^ county, " ! Kaogu xucbao 1993.1:13—56; Zhou Xinghua, "Ningxia Zhongwei xian Langwozikeng de qingtong duan- jian muqun," Kaogu 1989.11: 971—80. "' On this and the following sites, see Liu Dezhen and Xu Junju, "Gansu Qingyang Chunqiu Zhanguo muzang de qingli," Kaogu 1988.5: 413-24. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 947 Gansu), a bronze-hilt iron sword has been found in a similar type of burial together with tools and horse-related ornaments, along with several bronze weapons. A clear relationship between the two sites can be established by the presence of almost identical ge dagger-axes and distinctive small ornamental statues representing a kneeling deer. It is difficult to estimate the extent to which the use of iron was wide- spread in the Northern Zone. It does not seem, however, to have been limited to mounted pastoralists or associated exclusively with cultures in which the horse had attained a central role. Iron knives have been found in different burials belonging to the Shajing culture in Yongchang county, Gansu, where cattle and sheep sacrifices seem to predominate over the horses.'64 In a later Shajing burial site at Yushugou (Yongdeng county, Gansu), there are two iron spade-shaped objects, a spear, and a drill.16' The rest of the bronze assem- blage does not include weapons, but mostly ornaments and one chariot axle end. The ornaments show a connection between this site and the Ordos region (Xigoupan), the Ningxia area (Yanglang), and even the northern Chinese state of Zhongshan (Pingshan). Bronze and, in the steppe, gold were considered to be more suitable for ornamental or ritual purposes. The use of bronze in the manufacture of hilts of iron swords suggests that it was appreciated for its hardness but not for its beauty;166 this is possibly one reason why few iron artifacts are found in graves. Historical Survey DI $C. The Di appear regularly in traditional sources from the mid-seventh century B.C. They were divided into two large groups: the Bai Di, located in the west, and the Chi Di, located in the east. Arguments as to a possible identification of Di people with "Scythian"-type early nomads have been based on very little textual evidence, such as the often quoted statement that "the Rong and the Di are continually changing their residence, they treasure material objects as valuable but give little importance to land; their land can be purchased.'"67 But this is hardly a conclusive proof, and no horse riding is reported among the Di. On the contrary, they are described as foot soldiers.'68 164 Kaogu xuebao 1990.2: 205-37. l6f Kaogu yu wenwu 1981.4: 34-6. M Jessica Rawson, "Jade and Gold: Some Sources of Ancient Chinese Jade Design," Oritentations 26, 6 (1995): 26-37. <6? Zuo zhuan, 29 (Xiang 4), 5b (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 424; Yang Bojun, ed., Chunaiu Zuo zhuan zhu, p. 939). 'a Zuo zhuan. 41 (Zhao 1), 19a (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 579; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 1215). Cambridge Histories Online © Cambridge University Press, 2008 NICOLA DI COSMO The appearance of the Di on the northern borders of China suggests a large southbound migration of people displacing the sparse Rong commu- nities, which were pushed further south and either fell under the control of the Chinese states, or survived semi-independently in the interstices between them. The relationship between the Di and the Chinese states was not just that of mere border conflicts. Throughout the Spring and Autumn period the Di played a complex, multifaceted role in interstate politics. They pro- vided a safe haven for Chinese runaways, often victims of factional struggles or defeated pretenders to local lordships. From the frequent mentions of pas- sionate appeals to all central states to unite against the Di, we can see that, militarily, they posed a threat that was perceived by the Zhou political community as being potentially more dangerous than the internal conflicts among the Chinese states. However, considerations of realpolitik often pre- vailed over feelings of cultural and ethnic brotherhood, and Chinese states attacked by the Di were often left to fend for themselves. The most vicious wars against the Di were those waged by the state of Jin, bent since 660 B.C. on a campaign of annihilation that eventually paid off in 594 and 593 B.C., with the destruction of several Chi Di groups.'69 This attack probably took place in conjunction with an internal crisis of the Di, as there is evidence of famine and political dissent among them. In the fol- lowing years, Jin engaged in all sorts of ploys and stratagems in its fight against them and completed its "subjugation" of the Di in 541 B.C.,'70 thus establishing Chinese political supremacy in the north. However, fighting con- tinued against statelets set up by the Xianyu $$Jlt - a tribe or state within the larger entity of the Di people — who were repeatedly attacked by Jin in the latter part of the sixth century. In this war, victories were by no means one-sided; in 507 B.C. Xianyu troops convincingly defeated the Jin army. Foreign relations between Chinese states and the Di involved the estab- lishment of marriage ties, and the Chinese custom of exchanging kin members from princely households as virtual hostages was also observed by the Di.17' Regular treaties were concluded between Di and Chinese states, such as the one in 640 B.C. between Qi and the Di, both agreeing to join forces against Wey. Peace agreements were also ratified through treaties, as in 628, when the Di requested peace from Wey, or in 601, when peace was arranged between the Bai Di and Jin, who then proceeded to attack Qin.'7* 149 Zuo zhuan, 24 (Xuan 16), 13a (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 330; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, pp. 767—8). 170 PruSek, Chinese Statelets, p. 172. 171 Zuo zhuan, 15 (Xi 24), 16b, (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 191; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 416). 171 On treaties, see Zuo zhuan, 14 (Xi 20), 25a (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 178; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 387); Zuo zhuan, 17 (Xi 32), 10b (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 220; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 489); Zuo zhuan, 22 (Xuan 8), 7a Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 949 In the political arena the Di did not behave differently from Chinese states, and generally do not seem to have been any more vicious or untrustworthy than any other political protagonist at this time. Indeed, the tactics adopted by Jin in its anti-Di campaigns, though highly effective, can hardly be com- mended as paragons of fair play. Although it is not clear exactly in what way the Di differed from the Chinese, a difference was repeatedly noted by Chinese chroniclers. When in 661 B.C. the Di invaded the state of Xing ffl, Guan Jingzhong l i ^ f t said to Qi Hou *$&, "The Di and the Rong are like wolves and can never be satisfied; all the Xia states are closely related [to Qi], and none should be abandoned; to rest in idleness is a poison that should not be cherished."173 According to an even more scathing judgment, the Di all conformed to the following "four evils": Those whose ears cannot hear the harmony of thefivesounds are deaf; those whose eyes cannot distinguish among thefivecolors are blind; those whose minds do not conform to the standards of virtue and righteousness are perverse; those whose mouths do not speak words of loyalty and faith are foolish chatterers.'74 This clearly placed them beyond the pale of Chinese civilization. In later sources we also find analogous remarks on the "diversity" of the Di. For instance, the state of Shu is described in the Zhanguo ce^MW- (Strat- agems of the Warring States) as "a remote country of the west that still observes the old usages of the Rong and Di."'75 The state of Qin was accused by its enemies of sharing the same customs and moral qualitites as the Rong and Di: it had the heart of a tiger or a wolf, was greedy and cruel, untrust- worthy when it came to making a profit, and did not behave according to protocol and virtuous conduct.'76 By the time of the Warring States, the various Di peoples who had settled along the northern Chinese territories during the Zhou dynasty had devel- oped into relatively small, independent frontier centers. The most important was the state of Zhongshan 't1 lil, which in the written sources is referred to as a state of the Bai Di. Created by the Xianyu, it was attacked by Wei Wen Hou i t ^ t ^ in 408 B.C., conquered in 406 B.C., and ruled by Wei for about forty years. In 377 it regained its independence and continued to exist until (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 302; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 695). •" Zuo zhuan, 11 (Min i)> ib (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 124; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 256). 174 Zuo zhuan, 15 (Xi 24), 21a (Legge, The Ch'un Ts'ew with the Tso Chuen, p. 192; Yang Bojun, ed., Chunqiu Zuo zhuan zhu, p. 425). 171 Zhanguo ce, annotated by Liu Xiang Wl°] (Shanghai: Shanghai guji, 1978), vol. 1, p. 117; see also J. I. Crump, Chan-Kuo Ts'e (Oxford: Clarendon Press, 1970), p. 67. On the composition of the Zhanguo ce, see Loewe, Early Chinese Texts, pp. 1-11. 174 Zhanguo ce, vol. 2, p. 869 (Crump, Chan-Kuo Ts'e, p. 436). Cambridge Histories Online © Cambridge University Press, 2008 95° NICOLA DI COSMO 295 B.C. This state had fortified cities and an army with a thousand war char- iots and very capable troops.'77 Archaeological research has shown that at least from the end of the fifth century B.C., Zhongshan was fully within the sphere of Chinese civilization. Its bronze production, especially at Pingshan ^Plil, reveals its complete absorption within the culture of the Central Plain.'78 Still, no matter how "Chinese" the rulers of Zhongshan were, references to their diversity indicate that for a long time they were not accepted as one of the states of the Hua-Xia cultural sphere.'79 The fall of Zhongshan in 295 B.C. did not put an end to the history of the Di. Some of these groups were attacked by General Tian Dan IB W- of Qi during the reign of King Xiang H of Qi (r. 283-265 B.C.).'80 Since by this time Zhao had already conquered Zhongshan and Qi was cut off from the northern territories, we must assume that either some Di people lived between the states of Yan, Zhao, and Qi, or that either Zhao or Yan allowed Qi to go through their territory to attack the Di people living in the north. Either hypothesis would suggest that Di kingdoms continued to exist until a later date and were gradually absorbed by the Chinese northern states. The narrative of the war between Qi and the Di shows clearly that it was a long siege war, which indicates that the Di were politically organized into city- states. Though horses were imported from Dai, and this was often referred to as Di territory, its inhabitants were probably not nomads. Taking advan- tage of the abundance of grassland, they may have bred horses for export and military purposes. Horses had to be used for chariots by all armies, and by the end of the fourth century B.C. Chinese states were already adopting cavalry warfare, which meant a rising demand for horses. Because of their closer relationship with the steppe areas, the people of Dai may have adopted cavalry and bred horses even earlier. Though production of good horses should not imply that the people of Dai were nomads, it points to closer contacts between the northern Chinese frontier and peoples of the steppe. In general, we can see that during the eighth, seventh, and sixth centuries B.C., the Di settled in the great plains running from the loop of the Yellow River and the Ordos territory to the Taiyuan plain, with some also living in northern Hebei as far east as the state of Yan. They therefore may have created an effective buffer, for several centuries, between the Zhou states and the nomads of the northern territories. The gradual encroachment of central states on the northern region, and 177 Zhanguo ce, vol. i, p. 436 (Crump, Chan-Kuo Ts'e, p. 200). '7* Li Xueqin, "Pingshan muzangqun yu Zhongshan guo de wenhua," Wenwu 1979.1: 37—41; trans, in Chinese Archaeological Abstracts 3, ed. Albert Dien, Jeffrey Riegel, and Nancy Price (Los Angeles: Uni- versity of California, Institute of Archaeology, 1985), pp. 804-8. 179 Zhanguo ce, vol. 3, pp. 1170-4 (Crump, Chan-Kuo Ts'e, pp. 574-6). "° Zhanguo ce, vol. 1, pp. 467—70 (Crump, Chan-Kuo Ts'e, pp. 113—14). Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 951 their subjugation and incorporation of Di and other frontier peoples, even- tually brought China into direct contact with the nomads, primarily in the Ordos region. Relations between the Hu nomads and the state of Zhao prompted the adoption of cavalry. During the period between the end of the fourth century and the mid-third century, non-Chinese city-states continued to struggle for independence vis-a-vis China. Within the Chinese states a consciousness of the deep cultural divide between them and the Di did not soften even by the end of the Warring States. On the other hand, the dif- ferences between the Hua-Xia and Rong— Di peoples should not be confused with the conflict between China and the nomads. The written sources present the Rong and Di as communities politically organized on a tribal and territorial basis, centered around fortified settlements, often in the guise of city-states. LATE WARRING STATES TO QIN (CA. 3 5 0 - 2 0 9 B.C.) The final stage in the pre-imperial history of the northern frontier is a period of direct contacts between nomadic peoples and China. The acceleration of the northern expansion of the states of Zhao, Yan, and Qin caused the rapid absorption of the pastoral and semipastoral peoples who in the past had acted as buffers between the Central Plain and the northern steppe. As we have seen with the case of Zhongshan, from around the fifth century B.C. the dif- ferences between peoples such as the Di and the central states became less and less relevant, both in terms of political structures and cultural founda- tions. The people who inhabited these border regions were mostly settled; they lived in fortified cities, and continued to export pastoral products, animal husbandry having long become their main economic pursuit. It was probably due to pressure from the northern nomads that in the mid-fourth century B.C., the state of Zhongshan started the construction of frontier for- tifications that was the prelude to the building of the Great Wall. Thus, we can take this approximate date as the beginning of our fourth phase in the history of the northern frontier. The evidence available today, from both the written sources and archae- ological investigation, suggests that it was the shrinkage of the intermediate area inhabited by semipastoral people, gradually converted to or absorbed within the Chinese sphere, that eventually brought the northern states into direct contact with the nomads. Contacts may have occurred long before, but firm evidence of a strong impact of northern nomads upon historical developments in the south must be dated to the end of the fourth century B.C. This is epitomized by the appearance of a new type of foreigner, the Hu #3. This term, whatever its origin, soon came to indicate an "anthropologi- Cambridge Histories Online © Cambridge University Press, 2008 9J2 NICOLA DI COSMO cal type" rather than a specific group or tribe, which the records allow us to identify as early steppe nomads. The Hu were the source of the introduction of cavalry in China. Aside from military developments, trade relations and diplomatic contacts that are scarcely hinted at in the written sources can now be documented archaeologically. The large amount of gold found in third century B.C. Xiongnu tombs in the Ordos region reveals a possible shift from a purely military aristocracy to a leadership that engaged in trade as well as war and profited greatly from commerce with China. Politically, the nomads appear on the historical scene in 318 B.C., and the Chinese recording of the native term used by the Xiongnu for their leader, shanyu H^f, betrays the existence of diplomatic exchanges.181 The combined direct and circumstantial evidence of military fortifica- tions, trade, and diplomatic exchanges points to the middle to late fourth century B.C. as the period when closer relations were established. These remained essentially stable during this period and clearly bore advantages for both sides, though increased friction occurred between Zhao and the Xiongnu toward the middle of the third century. During the third century the expansionist power of stronger and larger Chinese states in the south kept pressing the nomads, pushing the frontier progressively northward. Finally, as soon as the First Emperor of Qin had concluded his unification of the central states, he dispatched a powerful army to the Ordos with orders to occupy and colonize it. The fierce and proud Xiongnu reaction to Qin's encroachment led to the foundation of a new leadership and the creation of an immense nomadic power that would soon become a formidable oppo- nent of the young Chinese empire. Archaeological Cultures of the Northern Zone During the Late Warring States The "closing in" between the northern cultures and the Chinese zone accel- erated rapidly during the last part of the Warring States. From the fourth to the third centuries B.C., contacts with China became more significant. In art, the distinctive elements of early nomadic cultures, though still predominant and retaining their northern flavor, blended with different symbols — trees, mountains — which affected substantially earlier stylistic models.'82 '" Zhanguo ce, vol. 3, p. 1129. The correct reading of the character W-, normally read dan, has been the object of some debate. Though the reading shanyu WT has been accepted for some time. Chinese scholars usually prefer chanyu. See Wang Liqi, ed., Shiji zhuyi, vol. 4, p. 2317. "* Esther Jacobson, "Beyond the Frontier: A Reconsideration of Cultural Interchange Between China and the Early Nomads," EC\] (1988): 201-40. The presence of distinctive Chinese motifs in north- ern art has led some to believe that there was a Chinese production of artistic metalwork specifically Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 953 The sites thus far attributed to the Xiongnu are mostly concentrated in the Ordos region. They are also the richest sites in the Northern Zone. Other sites, similar in style, funerary inventory, and material culture, have also been excavated in the northeast and northwest, particularly in Ningxia. A common characteristic is the variety of burial styles on the same site, which may indicate increasing social differentiation or, alterna- tively, the cohabitation of different groups merging into a more composite society, as a result of forced displacement, migrations, or simply an enlarge- ment of the range of action of human communities because of widespread adoption of a horse-based nomadic economy. Precious metals predominate in the aristocratic burials of this period, thus revealing the presence of a social elite no longer purely military in nature. Fewer weapons were buried, and the use of iron became more common. But these changes, as profound as they are, do not appear to be intrusive. The consensus today is that the nomadic cultures of the Northern Zone in the late Warring States were directly linked to earlier inhabitants such as those of the lower strata of Taohongbala and Maoqinggou. There is no doubt, however, that the culture represented by these earlier sites expanded tremendously and, in certain areas such as the northeast, replaced earlier cultures such as Upper Xiajiadian. The Ordos region was the central area of the flourishing of nomadic culture in the fourth and third centuries B.C. The sites are of two types: those used before the Warring States, which show a degree of continuity with the earlier periods, and those that can be dated only to this later period. The sites of Taohongbala and Maoqinggou are representative of the first group, whereas Yulongtai, Xigoupan, Aluchaideng, and others belong to the second (Map 13.5). Iron technology became far more widespread, with a much broader inventory of iron objects. However, iron did not completely replace bronze artifacts. In particular, iron was used for certain types of weapons and horse fittings. Antennae Style iron daggers, similar to earlier bronze daggers, and iron swords similar to those of the Central Plain are found both over a broader area and in larger numbers with respect to the preceding period. Horse bits and chamfrons came more frequently to be made out of iron. Furthermore, the bronze pickaxe was generally replaced with one of iron. Another characteristic of the period is a tendency toward standardization. Traditional weapons and implements started to acquire standard features, such as the hole in the handle of bronze knives and a wavy line decoration designed for the nordiern markets or that there were Chinese artisans among the nomads. For a full illustration of this viewpoint, see So and Bunker, Traders and Raiders on China's Northern Frontier. Cambridge Histories Online © Cambridge University Press, 2008 954 NICOLA DI COSMO ) (. aanjiaxi Northern Zone Sites / of the 4th Phase —i (350-209 B.C.) ^^—-^ • — * ' x S(- / j .—^^ Maoqinggo 11 (late phase > . - r-, J. s Aluchaideng i - Taohongbala* (lafe phase) ^-^^ ^\ • °* Xigoupan) / , - % Z? ' Sujigou (Yulongtai ^f S~ ' -" r~ \ r-'' \f k F /r r j / *'~^ I? \ / / f r -r"~r ^-^ . •r YELLOW 1 I v~ •• i .—.y j •— • ^ .(' c. o N 150 km - ISOn ^"\ "^S ? •' 1 Map 13.5. Northern Zone: Archaeological sites of the fourth phase (350—309 B.C.)- on the shoulders of pots. Among the decorative features of this period, we see a decisive increase in Animal Style belt buckles and plates. Often these plates, round or rectangular, depict human activities.18' Scenes of animal combat both realistic and stylized became more common, as well as artisti- cally sophisticated. By far the most stunning feature of this period is the presence of extra- ordinarily rich burial sites, with hundreds of precious objects, mostly gold and silver ornaments. In Aluchaideng (Hangjin !TL!$ Banner, Inner Mongo- lia),'84 located to the north of Taohongbala, two ancient tombs of the late Warring States period were unearthed. They are remarkable for the extraordinary number of precious objects and the artistic value of the orna- ments (Fig. 13.6). Altogether 218 gold and 5 silver objects were found. Among these the most important are a gold headdress set, or crown, composed of 4 '*' Emma Bunker, "The Anecdotal Plaques of the Eastern Steppe Regions," in Arts of the Eurasian Step- pelands. ed. Philip Denwood (London: Percival David Foundation, 1978), pp. 121-42. '** Tian Guangjin and Guo Suxin, "Nei Menggu Aluchaideng raxian de Xiongnu yiwu," Kaogu 1980.4: 333-8, 364, 368; idem, "Aluchaideng faxian de jin yin qi," in O'erduosi qingtong qi, pp. 342-50. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 955 Figure 13.6. Gold ornaments, Xiongnu culture, Aluchaideng. From Tian Guanjin EBRjit and Guo Suxin SBSSr, eds., O'erduosi qingtong qi fPSf£»rWfflSS, plates I, IV, XVI (Beijing: Wenwu, 1986). pieces (a skullcap and three headbands); 4 rectangular ornamental gold plates illustrating a tiger assaulting a bull or cow; 12 ornamental plaques with designs of tigers and birds inlaid with precious stones; 55 Animal Style orna- mental plates with representations of tigers, birds, sheep, hedgehogs, and 2 tiger heads; and 45 rectangular gold buckles. In addition, ornaments in the shape of buttons, small tubular objects, and necklaces were also found. Because of its richness, this site is regarded as a royal burial of a chieftain of Cambridge Histories Online © Cambridge University Press, 2008 956 NICOLA DI COSMO the Lin Hu people, who presumably inhabited this region in the late Warring States. A closely related site excavated in 1979 and in 1980 is located at Xigoupan (Zhunge'er Banner, Inner Mongolia).'85 The first investigation revealed three tombs of the late Warring States period, while the second brought to light eight tombs and a nearby settlement dated to the Former Han period. That the first three tombs present a widely varied funerary assemblage, even though they belong to the same period, is a possible indication of growing social differentiation. Tomb 2 is the richest, with funerary goods dominated by gold and silver ornaments. Although fewer in number compared with Aluchaideng, they are equally impressive and include several Animal Style ornamental plates and decorative objects, at times depicting either realistic or fantastic animals. Decorations in silver, lead, and bronze are also present. Weapons and tools, such as a sword, a ladle, horse bits, and cheekpieces, are made of iron. Tomb 1 contains a few bronze objects and iron remains. Tomb 3 contains bronze ornaments typical of an earlier stage, including weapons, buckles, and other ornamental objects, comparable with the Taohongbala assemblage. The sharp differentiation in the funerary assemblage seems to reflect distinction in social status rather than ethnic differences and may refer to the establishment of a rich aristocracy. The iron objects appear to be mostly weapons and tools of daily use, with no evident ritual or economic signifi- cance. From an early period, agricultural tools such as hoes, adzes, and pick- axes were made of iron, becoming progressively more common.'86 The existence of a settlement in the vicinity of the cemetery shows the presence of an agricultural or seminomadic people. It is possible that commoners or poor members of the tribe were buried with just a few iron tools. On the other hand, richer people were buried with finer and more prestigious bronze weapons and ornaments, which had a long tradition as mortuary objects. Only the highest elite accumulated enough wealth to be accompanied in death by an ostentatious array of glittering gold and silver jewelry, some of which was imported from China, as is clear from the presence of a Chinese inscription on the back of a golden plaque, and of Chinese characters on silver rein rings. The latter may refer to a workshop located in the state of Zhao. The inscription on the plaque, which indicates its weight, has been l!i Wcnwu 1980.7: 1-10. NeiMenggu wenwu kaogu 1981.1: 15-27. M From Wuhuan S f i burials of the Han period in Xichagou B8)fc) (Xifeng H B , Liaoning), an iron axe, iron pickaxe, iron adze, iron hoe, and other agricultural implements have been recovered. See Lin Gan, "Guanyu Yanjiu Zhongguo gudai beifang minzu wenhua shi de wo jian," Nti Menggu daxut xucbao 1988.1: 3. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 957 attributed to the state of Qin. It is possible that such objects were used as currency in commercial exchanges with the nomads, from whom China imported horses, cattle, and other typical pastoral products. Stylistically, on both Aluchaideng and Xigoupan ornaments we find a vast gamut of animals depicted in relief or in the round, including horses, cattle, sheep, tigers, eagles, deer, and fantastic beasts. Reclining horses and kneeling deer are particularly representative of this period, as are scenes of animal combat. Other Xiongnu graves of the late Warring States period were less richly adorned, such as those at Yulongtai (Zhunge'er Banner, Inner Mongolia), excavated in 1975.l87 The burial custom was identical with that seen at Tao- hongbala and included sacrifices of horse and sheep. The funerary assem- blage consisted of bronze, iron, and silver artifacts, but no pottery. Horse fittings included a bit and bridle stopper, both in iron, in addition to cheek- pieces made of bone. The silver necklace is typical of later objects, and the iron artifacts, such as the pickaxe and horse bit, are more developed than those found in Taohongbala. The number of chariot fittings, which include seven animal-shaped finials in bronze representing lambs, antelope, deer, and horses, and two axle ends, indicate that the chariot was still in use. Other weapons and tools — two knives, one adze, two axes, and one arrowhead — are all in bronze. A similar but richer tomb was investigated in 1984 in the Ordos region at Shihuigou ^GMM (Yijinhuoluo ^ifeS?& Banner, Inner Mongolia),'88 in which there was an abundance of silver objects: ornamental plaques, buttons, and various other animal-shaped decorations. A new type of Animal Style motif is represented by the combat between two tigers. The inlay technique was advanced, as we find an iron-set gilded bronze ornament in the shape of a turtle, and gold- and silver-inlaid iron artifacts. Generally the style and motifs belong to the repertory of the mature Warring State Ordos art. However, iron-set gilded bronze and gold- and silver-inlaid iron objects are rare and betray a different origin. It is possible that the technique came from China, while the artifacts were made locally, but the very rarity of these finds strongly suggests that they were imported, which would confirm the devel- opment of steady commercial relations between the Central Plain and the adjacent northern steppe areas. Finally, a site of some interest is Sujigou (Zhunge'er Banner).'8' This site 187 Kaogu 1977.2: in—14. ** Net Menggu wrnwu kaogu 1992.1—2: 91-6. '* Gai Shanlin, "Nei Menggu Yikezhaomeng Zhunge'er qi Sujigou chutu yibi," Wenwu 1965.2: 44—5, rpt. in Tian Guangjin and Guo Suxin, eds., O'erduosi qingtong qi, pp. 372-4. In the original report, this site was dated to the Han; that dating was omitted in the reprint. Cambridge Histories Online © Cambridge University Press, 2008 NICOLA DI COSMO Figure 13.7. Animal Style bronze ornaments, Xiongnu culture, from Sujigou. From Tian Guanjin rBSfjfe and Guo Suxin H$5KSf, eds., O'erduosi qingtong qi ?PBfi£$WfSI#ir, plates XI, XII, XIV (Beijing: Wenwu, 1986). had been disturbed in the past, and only objects found by local residents were recovered, in the early 1960s. The singular fact is that the objects are mostly bronze pole tops in the shape of animals, including a crane head, a sheep head, a feline cub, two kneeling horses, and a wolf head (Fig. 13.7). Their style shows yet another instance of the variety of applications of Animal Style ornaments in the Ordos culture. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 959 The progressive increase in the use of iron can be seen by looking at Former Han sites such as Budonggou (Yikezhaomeng IP^BSSI, Inner Mon- golia), where there is a vast inventory of iron tools and weapons. Iron was mostly reserved for vessels such as tripods and cauldrons, for weapons such as swords, knives, and arrowheads, for horse fittings such as horse bits, rings, and chamfrons, and,finally,for ornamental objects such as belt plates. Bronze was still the principal material to be used for decorative and ornamental pur- poses. Related sites, from the viewpoint of their material culture, have been found in northern Shaanxi, as at Nalin'gaotu ^ J ^ i l i ^ (Shenmu -tt^fc county, Shaanxi). Dated to the late Warring States, a Xiongnu grave exca- vated here in 1956 yielded a large number of gold, silver, and bronze orna- mental objects.'90 Here too animal sacrifices were practiced, and skulls of horses, cattle, and sheep accompanied the deceased. The subjects of the orna- mental plaques are mostly tigers and deer. One gold object represents a fan- tastic animal in the shape of a deer. A gilded silver dagger handle that is particularly rare, and possibly imported, is the only military object recovered from the tomb. Other sites in Shenmu county, such as Lijiapan ^=^ffl¥ and Laolongchi ^-ifjlftil, lack silver and gold ornaments; their metal inventory is limited to a few buckles, ornaments, and one dagger. In Yanglang we observe a situation similar to that of Maoqinggou and Taohongbala. Namely, the later tombs, dated to the late Warring States period, show a predominance in the use of iron, then widely available for weapons, tools, and ornaments. Though not on the scale of the Ordos, some gold objects appear in the funerary assemblage. Finally, examples of both horse gear (bits, chamfrons, bronze and bone cheekpieces, and harness orna- ments) and chariot fittings (shaft ornaments, axle cuffs, and hubs) increase dramatically in number. Ornamental pole tops and plaques representing animal combat are also typical of this later assemblage. In the Northern Zone, and particularly in the area closer to the Great Wall (from Ningxia and Gansu to Inner Mongolia and the northeast), then, the composition of metal assemblages seems to indicate a common pattern of development: a phase in which bronze weapons predominated — a sign of the formation of a warrior aristocracy - gave way to a stage marked by the extensive presence of horse fittings and ornaments, which point to techno- logical advances in transportation and warfare, as well as to changes in the taste and possibly social and political functions of the elite. During this time not only do we find the widespread use of iron and more elaborate func- l9 ° Dai Yingxin and Sun Jiaxiang, "Shaanxi Shenmu xian chum Xiongnu wenwu," Wenwu 1983.n: 23-30. Cambridge Histories Online © Cambridge University Press, 2008 960 NICOLA DI COSMO tional goods, but social status is more often expressed through the presence of precious objects. Artifacts related to chariots, usually regarded as typical status symbols of ancient China, and to horses, which of course held primary importance in nomadic societies, came together to represent power and wealth enjoyed in life. This change in funerary assemblage may very well point to the emergence of a new class of aristocrats, whose position in society was proportional to their success in managing relations with China and other neighbors. Such relations were not only political and diplomatic; they also carried strong commercial connotations. A tendency toward a commercialization of relations with China can also be found in earlier sites in Yanqing county (Beijing) dated to the Spring and Autumn period and attributed to the Shan Rong.191 Here the presence of gold is consistent and regular. Even more significantly, coins have been found that indicate a degree of monetary exchange. This area was subsequently incorporated into the state of Yan, though it continued to have a dual cultural composition for a long time. These sites can be seen as the first instance of a trend toward commercial- ization of the frontier and the possible transformation of the upper echelons of nomadic society from a purely warrior aristocracy into diplomatic and commercial agents that monopolized or to some degree controlled border exchanges with China to their own profit. This trend reached its highest point during the late Warring States period, thanks to the establishment of more direct contacts between nomads and the northern Chinese states. Relations Between Northern Nomads and Central States The earliest textual evidence of direct contact between the Xiongnu and China is found in the year 318 B.C.,'92 when the Xiongnu are said to have been part of a joint force with Hann ?$, Zhao, Wei lit, Yan ?£&, and Qi W that attacked Qin. Further and more detailed evidence of a direct connec- tion between the Hu nomads and Chinese states can be found in the famous debate held in 307 B.C. at the court of King Willing of Zhao i S l ! (r. 325-299 B.C.). In the course of this debate the monarch supported the adoption of mounted cavalry and archery against the myopic conservatism of his advi- sors.1" This change in military thinking was not due exclusively to the need to repel nomadic assailants - though this may have been a consideration - but mostly to the king's eagerness to gain an advantage against other Chinese states by employing new military tactics and technology. The king's main "'' Wenwu 1989.8: 17-35, 43- ' " Shi ji, 5, p. 207. '" Zhanguo ce, vol. 1, p. 653-67. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 961 goal was to turn part of his Chinese troops into mounted warriors, to be deployed on the borders with both Chinese (Hann, Qin, and Yan) and north- ern nomadic states (Hu and Loufan).194 Since the greatest threat to the exis- tence of Zhao came from other Chinese states, primarily Qin, it is questionable whether many of his new military units were used to contain the various nomads. During the late Warring States period, Yan, Zhao, and Qin expanded their territory mostly at the expense of the northern peoples, who were in a posi- tion of military inferiority. Under King Zhao BS (r. 306-251 B.C.), Qin expanded into the territory of the Yiqu Rong H I S S , possibly the remnant of semipastoral tribes, acquiring the later commanderies of Longxi SiH, Beidi Jtife, and Shang Jb, and building "long walls" as a protection against the Hu. During the reign of King Wuling, Zhao defeated the Lin Hu and Loufan to the north and also built a wall from Dai ft, at the foot of the Yinshan Mountains, to Gaoque MIS, thereby establishing the comman- deries of Yunzhong 9rf3,Yanmen iff PI, and Dai ft. To the east, the state of Yan entertained diplomatic relations with the Hu through General Qin Kai fUm, then attacked them by surprise, defeating the Dong Hu and pushing them back "a thousand //." Yan also built a wall that went from Zaoyang tit® to Xiangping M5?- to protect itself against the Hu and created the commanderies of Shanggu Jt@, Yuyang i&P§, Youbeiping ^Jb^rS Liaoxi 2I1S, and Liaodong J S ^ . At this point Qin, Zhao, and Yan, three of the seven states of "the people who wore caps and girdles," bordered on the Xiongnu.1" The militarization of the frontier was due to the robust territorial expan- sion of the three northern Chinese states, all determined to protect their newly acquired lands. The state of Yan expanded mainly in the northeast and occupied both the maritime region north of the Liaodong Gulf and the Liaodong Peninsula, including, to the west, a large portion of what is today Hebei province. After conquering Zhongshan in 295 B.C., Zhao con- tinued its drive to the north and built a series of fortifications along the northern bank of the great bend of the Yellow River, where it encircles the Ordos steppe in a great loop, thus creating a Chinese enclave deep into nomad territory. The state of Qin also expanded into the Ordos, in the Hetao "MS region. Its line of fortifications ran from the Shang commandery in eastern Hetao to Longxi commandery in southern Gansu, along a northeast to southwest line. Longxi was the westernmost point of China's northern frontier. 194 Zhanguo ce, vol. z, p. 657. ' " Shi ji, n o , pp. 2885—6. Cambridge Histories Online © Cambridge University Press, 2008 962 NICOLA DI COSMO In the course of the third century, Zhao continued to engage in a war of attrition with the Xiongnu. Entrusted by the state of Zhao with the defense of the northern border in the commanderies of Dai and Yanmen, General Li Mu ^ % assumed a defensive posture; he was criticized for his passivity, even though the frontier was not penetrated by Xiongnu attacks. Under pres- sure of criticism and intimations of cowardice, Li Mu led an army of 1,300 war chariots, 13,000 cavalry, 50,000 select infantry, and 100,000 expert archers against the Xiongnu. He succeeded in drawing them into a trap, crushing them. This feat vindicated his honor and was followed by victories over the Dong Hu and the Lin Hu.'?6 Several factors possibly contributed to this shift toward aggressive military policies. Soldiers stationed on the frontier pressed their commanders into active engagements in order to profit from the spoils of war. Court politics could also have an effect on frontier defense. The pattern of military rela- tions between nomads and Chinese in this period should therefore not be seen as a unilateral series of nomadic raids against Chinese soldiers and set- tlers, but rather as a war of attrition carried out between displaced nomads and a body of occupation troops who would often take the initiative and launch raids into nomadic territory. Chinese Knowledge of the Northern Peoples In the sources that can be dated to the Warring States period, there is little indication that the Central Plain statesmen and intellectuals were interested in the life and history of their northern neighbors. In most of the best-known pseudogeographical treatises of this period, such as the Shan hai jing li|"/§M (Classic of the mountains and seas),'97 the space surround- ing the Central Plain was the abode of surreal beings, inhabitants of a fan- tastic world. The rationalist attitude of the Han historians tended to reject information and accounts of this sort as being purely fictitious and untrust- worthy. Other works were less inclined to supernatural description; never- theless, the information they provide is far from reliable. This is true of the Mu Tianzi zhuan ©^"pflf, a biographical work of the fourth century B.C. that describes the travels of King Mu of Zhou (r. 956-918 B.C.) to visit the Queen Mother of the West, Xi Wang Mu B i t . 1 " In philosophical and historical works, such as the Lie shi chunqiu § K # $ ( (Mr. Lii's Spring and 196 Shiji, 81, pp. 2449-50 (William H. Nienhauser, Jr., ed., The Grand Scribe's Records. Vol. 7: The Memoirs ofPre-Han China, by Ssu-ma Ch'ien [Bloomington: Indiana University Press, 1994], p. 271). 197 Loewe, Early Chinese Texts, 357-67. •* Ibid., 342-6. Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 963 Autumn [annals]), the northern peoples lack distinctive ethnographic fea- tures. Only the collection of fictionalized historical accounts known as the Zhanguo ce provides us with a description of their riding gear and their ability as archers. However, from information that enters the texts anecdotally and almost unconsciously, we can grasp a few glimpses of Chinese acquaintance with the nomads, and of a knowledge of them that must have been widespread. A pre-Qin notion of the political organization of the nomads can be seen in a sentence of the Zhanguo ce, where it is said that Hu and Yue i§ (a people of the south) were divided into many groups that did not understand each other's languages; yet when threatened by a common enemy, they would all unite and fight together.1" This statement seems to imply that by the end of the fourth century, the terms Hu (for the north) and Yue (for the south) were used as broad "anthropological" categories applied to various political entities — clans, tribes, or even states — that claimed different origins and spoke different languages; in case of need, however, these barriers could be overcome, and a political unity found. Applied to the nomads, this seems to refer to the formation of tribal con- federations, whose first historical example was the creation of the Xiongnu empire. Commercial and diplomatic interaction between the Central Plain and the Hu is documented both in the Zhanguo ce and in the Mu Tianzi zhuan. The former mentions the importation from the north of horses and furs.100 The second, while a fictional account, mentions information that must have originated in actual practices and customs. In his encounters with several foreign chiefs and dignitaries, King Mu conducted gift exchanges that must have been characteristic of fourth century B.C. relations between China and northern pastoralists. While the information in this work points unques- tionably to dealings with peoples who were predominantly pastoral, it also shows the existence of mixed products from both herding and farming activ- • ities. The largest "gift" received by King Mu was a herd of cattle and a flock of sheep numbered by the thousand. Even more valuable, and always men- tioned at the head of any list of gifts, are horses, numbering in the several hundreds. Both cultures valued horses, which had long been used in China for military and ceremonial purposes, but it was the north that provided a surplus for sale to China. A third item, which appears often but not always, is that of cereals, such as millet. Other items include wine and other animals, such as dogs and goats. The gifts presented by King Mu in exchange include Zhanguo ce, vol. 3, p. mo. ' Zhanguo ce, vol. i, p. 178, and vol. 2, p. 608. Cambridge Histories Online © Cambridge University Press, 2008 964 NICOLA DI COSMO mostly precious artifacts. The mention of a silver deer, a silver bird, and golden deer presumably refers to the small sculptures or plaques with animals shown in relief that appear so commonly in Ordos art. Other items include necklaces of gold or precious stone beads, peals, gold bullion, belts adorned with precious shells, and sometimes fine horses in a team (four of the same color), probably meant to be hitched to a royal carriage and used for display. Finally, women were exchanged as wives, a detail that points to the role of bride giving as an instrument of diplomacy.101 The information contained in the Mu Tianzi zhuan substantiates the pattern of exchanges mentioned earlier — in particular the importation of horses. On the other hand, references to precious gifts of Animal Style objects are supported by archaeological finds in Xiongnu tombs of golden plaques and other objects manufactured in China. It is possible to discern a model of diplomatic and economic exchange beneath the fiction. The Rise of the Xiongnu In 215 B.C., the First Emperor of Qin ordered a campaign against the Hu (Xiongnu) in the Ordos region. This was not prompted by any real threat or any preexisting situation of belligerence. The alleged reason for the expe- dition was a preemptive strike to prevent the fulfillment of a prophecy according to which "Hu" was going to bring about the downfall of the dynasty.202 In fact, the First Emperor was pursuing a policy of territorial expansion, as land was needed both to demobilize and resettle the great number of men who had been serving in the Qin army, as well as defeated enemy troops, and to resettle the dispossessed families of war refugees. The operation was undertaken on a massive scale. General Meng Tian was sent north with an army numbering as many as 300,000 troops, according to some records, or 100,000 according to others. His task was to invade and seize the whole Hetao region, thus making the northern bend of the Yellow River into the new northern frontier of the Qin empire. Forty-four fortified counties were built along the river and settled with convicts sentenced to guard the borders. Moreover, a road was built between Jiuyuan fll& (near present-day Baotou, Inner Mongolia) and Yunyang SSI in order to link the "" See Re"mi Mathieu, Le Mu Tianzi zhuan: Traduction annotie — Etude critique (Paris: College de France, Institut des hautes Etudes chinoises, 1978). 201 Shi ji, 6, p. 252. In fact, the son of the First Emperor, the short-lived last ruler of Qin, was called Huhai SSS (William H, Nienhauser, Jr., ed., with Tsia-fa Cheng, Zongli Lu, and Robert Reynolds, trans., The Grand Scribe's Records. Vol. r. The Basic Annals ofPre-Han China by Ssu-ma Ch'ien [Bloom- ington: Indiana University Press, 1994], p. 271). Cambridge Histories Online © Cambridge University Press, 2008 THE NORTHERN FRONTIER IN PRE-IMPERIAL CHINA 965 border with the metropolitan area. Border defenses were erected throughout the north from Lintao, in the western part of Gansu, to Liaodong.20' Beyond the Yellow River, Meng Tian extended Qin control over the Yang PH Moun- tains in Inner Mongolia.204 The leader of the Xiongnu, Touman U S , fled north. In 211 B.C., Qin transferred 30,000 families of colonists to Beihe JbM and Yuzhong ^34", north of the Yellow River in what is today Inner Mongolia.105 It is within this context of political and military emergency, compounded with the economic crisis derived from the loss of pasturelands, that the uni- fication of the Xiongnu took place. The formation of a tribal confederation under a single charismatic leader was accomplished in various phases. The first phase saw the emergence of a new leader. This was Maodun, son of Touman, who claimed the title of shanyu in 209 B.C. His rise took place ini- tially as the result of an act of patricide and seizure of power, accomplished after having created an efficient, blindly loyal, militarily disciplined corps of personal bodyguards. The account of the killing of Touman by his son sug- gests a struggle between an old aristocracy, evidently unable to meet the chal- lenge presented by the Chinese invasion, and the junior leaders, who joined together irrespective of established hierarchies. Values consisting primarily of military prowess, obedience to the leader, and personal ambition were embraced by the new leadership. The first task the young leader faced was the challenge posed by other nomadic confederations, the Dong Hu and the Yuezhi. As a loose confederation of tribes, the Xiongnu had existed at least a century before the unification of China. The rise of Maodun, therefore, does not bring about the creation of a new historical subject, but rather points to its political reorganization in order to meet a military challenge. Meng Tian's seizure of the Ordos and displacement of the Xiongnu must have caused widespread relocation and migration, upsetting the established territorial makeup and balance of power among the peoples of the Northern Zone. This crisis produced a violent change of leadership, which allowed the Xiongnu not only to overcome the crisis, but also to defeat their enemies and consolidate their position over the steppe region in Inner Mongolia and eastern Manchuria. In 210 B.C., the death of the First Emperor and the forced suicide of General Meng Tian paralyzed the Qin politically and militarily "* Shiji, no, p. 2886. 234 Shi ji, 88. pp. 2565-6. On archaeological investigations of the Great Wall, see Tang Xiaofeng, "Nei Menggu xibei bu Qin Han chang cheng diaocha ji," Wenwu 1977.5: 16-22 (trans, as "A Report of the Investigation on the Great Wall of the Qin-Han Period in the Northwest Sector of Inner Mongolia," in Chinese Archaeological Abstracts 3, pp. 959—65). "* Shiji, 6, p. 259 (Nienhauser, ed., The Grand Scribe's Records, vol. 1, p. 151). Cambridge Histories Online © Cambridge University Press, 2008 966 NICOLA DI COSMO and threw China into renewed civil war, preventing Chinese armies from taking effective action against the Xiongnu. Soon the defense line established by Qin was overtaken by the nomads, determined to reconquer their lost ter- ritories. The subsequent growth of the Xiongnu state into an empire marked the beginning of a new stage in the history of the northern frontier, which had now been denned for many centuries to come. Cambridge Histories Online © Cambridge University Press, 2008