Horses, Silver, and Cowries: Yunnan in Global Perspective

Present-day Yunnan is a province in Southwest China. Historically, Yunnan maintained close relationships with Southeast Asia, India, and Tibet, as archaeological findings and other studies have confirmed.[1] As an interaction zone among several civilizations, Yunnan was influenced by and had an impact on other cultures. Scholars of China have named a trade route connecting the above regions the “Southwest Silk Road.”[2] This international trade route geographically centered on Yunnan and Upper Burma. Yunnan’s importance, however, was based on far more than simply its location. Like Upper Burma, Yunnan is rich in precious metals such as gold and silver as well as other minerals such as tin, lead, and copper, and other local resources. In addition, Yunnan’s connections with the overland Silk Road and the maritime Silk Road greatly enhanced Yunnan’s role in transregional interactions.

This paper aims to demonstrate the global significance of Yunnan and to redraw the map of early Eurasian communication. While utilizing Chinese scholarship, I supplement Chinese scholars with non-Chinese sources to construct a more comprehensive picture of the Southwest Silk Road that in turn will add a new dimension to the Sino-foreign exchange and Eurasian communication. First, I will present a concise description of the road. Then, focusing on commercial items such as horses, silver, and cowries, I attempt to demonstrate the global importance of Yunnan by illustrating how Yunnan had shaped neighboring societies. Finally, the use of a world-system perspective will contribute to the ongoing world-system debates and add a new dimension to our understanding of Eurasian communications.

Yunnan and Its Trans-Regional Trade: A Critique

Since the early twentieth century, scholarly investigations of the overland Silk Road and the maritime Silk Road have constructed a fundamental basis of the communication within the Eurasian supercontinent. While contributing a great deal to the understanding of ancient East-West exchange, studies of the above two silk routes have more or less overshadowed the third route, the so-called Southwest Silk Road from Southwest China via Burma to India.[3]

The earliest textual source of the Silk Road is Zhang Qian’s exploration in the Western Regions (xiyu) in the late second century BCE, recorded by Sima Qian in his Shi Ji. Nevertheless, Zhang Qian’s report indeed leads to another Silk Road: a road connecting Southwest China with India, where he found Sichuan cloth (Shubu) and bamboo cane (Qiongzhu) in Daxia (Bactria). Emperor Wu of Han (140–87 BCE) then dispatched his envoys and troops to pacify local polities around Yunnan, with the expectation that he could open this road for his sake. His efforts, unfortunately, failed.

Because of Emperor Wu’s attempts, scholars in China for long time have paid a great deal of attention to this road. Many fragmentary and obscure records in Chinese historical writings prior to the Tang dynasty (618–907 CE) referred to the exchange between China and India through jungles, forests, rivers, and mountains from Sichuan, Yunnan, Burma, and Assam to India. We have no firsthand accounts of anyone completing this journey from early periods. Chinese documents after the Tang have detailed records, but they could offer little help for purposes of drawing a map of regions far away from the Chinese empire.

At the turn of twentieth century, sinologists such as Paul Pelliot devoted a great deal of energy to this matter. During World War II Chinese scholars began further exploration. The year 1941 saw two articles on this route by Fang Guoyu and Gu Chunfan, the former in Chinese and the latter in English.[4] Seven years later, the first monograph by Xia Guangnan was published.[5] From then until the1980s, the study of this route stagnated, occasionally occupying only a tiny space in books on ancient Indian or Chinese civilization. Joseph Needham noticed this road, but his conclusion is based on the acceptance of Chinese texts.[6] Although scholars of India or South Asia at times mention the early trade between China and India via Burma, a full exploration was never done. D. P. Singhal in his India and World Civilization, for instance, mentioned overland routes through Nepal and Tibet to China, and through Assam and Upper Burma to Yunnan. His map titled “Ancient Indian Trade Routes” marked only the former.[7] Yu Yingshi’s study on trade in Han China (206 BCE–220 CE) presented a refined argument of the existence of the route. Nevertheless, limited by the nature of his study and the unavailability of the present archaeological discoveries, the whole picture was missing.[8]

Thanks in part to advances in archaeological findings, Chinese scholars since the 1980s have made significant progress on this previously underestimated or ignored subject. Books and articles have been published by the dozen; nonetheless, many questions remain unresolved. These abundant Chinese sources, for instance, cannot locate routes in non-Chinese regions such as Burma and Assam, nor can they tell us the volume of trade. When did the Southwest Silk Road begin to take shape? Many opinions are in contention. Even an agreement of the name for the road has not been reached. Some use Xi’nan Sichouzhilu (Southwest Silk Road), some use Nanfang Lushangzhilu (Southern Overland Route), some use Nanfang Sichouzhilu (South Silk Road), some use Dian-Mian-Yin Gudao (Ancient Roads Connecting Yunnan, Burma, and India), some follow the traditional term Shu-Yandu dao (The Road between Sichuan and India), some use the term Shubuzhilu (The Road of Shu Cloth), and others do not give a specific term. In this paper I use the term “Southwest Silk Road” (SSR) to distinguish it from the overland Silk Road and the maritime Silk Road.[9]

While Chinese scholars have furthered this study, their weaknesses have been as outstanding as their achievements. Most works are case studies that present details but fail to draw a comprehensive picture. There are a few macrostudies, based on Chinese sources—mainly Han Chinese texts—and they can hardly escape the accusation of Sino-centralism. In fact, previous studies have been criticized by Sun Laichen for lacking breadth and depth:[10]”First, they lack breadth as they fail to treat the Sino-Southeast Asian overland interactions as a whole. Second, they lack depth because they are, by and large, descriptive rather than analytical. Third, the source materials utilized in most of these works are largely Chinese, while Southeast Asian sources are ignored. Thus the perspective adopted is usually ‘looking down south’ from China and is inevitably Sinocentric.”[11]

I want to further this critique. Previous studies indeed lack a global view, downplaying the global significance of the Southwest Silk Road. A global perspective of the SSR will give rise to many inspiring questions. For example, what kind of relationship had existed among the three Silk Roads? How did they function together temporally and spatially? Did they come into being a network linked so closely that they brought the Euro-Asian supercontinent into a world-system? If so, when and how? If not, why? Sun Laichen’s recent study has established an excellent case for interactions between the overland route and the maritime Silk Road during the Ming period (1368–1644 CE).[12] His study has influenced my decision to explore earlier times and to think about the tripartite interactions between the three roads.

While a global approach will broaden our view, a scrutiny of local influence of the SSR will highlight its dynamics and significance. How and to what degree did this trade route shape Southeast Asia and other areas involved? Janice Stargardt and Sun Laichen have provided excellent case studies, but how about other areas and periods they have not examined?[13] This paper, by scrutinizing Yunnan, attempts to fill some of these gaps.

General Description of the Southwest Silk Road

The earliest textual record of the SSR came from Chinese sources. In the year 138 BCE, Zhang Qian volunteered to be the Han envoy to the Western Regions (xiyu) where he was supposed to forge an alliance with the Yuezhi people against the Xiongnu, who posed a major threat to the young, ambitious Emperor Wu (140–87 BCE) of the Western Han dynasty (220 BCE–9 CE).

Zhang Qian’s mission to the Western Regions (138–125 BCE) initiated the records of the Silk Road and the Southwest Silk Road. Detained for ten years by the Xiongnu who controlled the roads to the Western Regions, Zhang Qian finally arrived at his destination. It is in Bactria (Daxia, or Ta-hsia in Wade-Giles system) that Zhang Qian found Sichuan items. To cite a passage from Shi Ji,
Southwest of Ta-hsia (Bactria) is the kingdom of Shen-tu [India]. “When I was in Ta-hsia,” Chang Ch’ien reported, “I saw bamboo canes from Chiung and cloth made in the province of Shu.” When I asked the people how they had gotten such articles, they replied, “Our merchants go to buy them in the markets of Shen-tu.” Shen-tu, they told me, lies several thousand li southeast of Ta-hsia. The people cultivate the land and live much like the people of Ta-hsia. The region is said to be hot and damp. The inhabitants ride elephants when they go into battle. The kingdom is situated on a great river.
“We know that Ta-hsia is located twelve-thousand li southwest of China. Now if the kingdom of Shen-tu is situated several thousand li southeast of Ta-hsia and obtains goods which are produced in Shu, it seems to me that it must not be very far away from Shu. At present, if we try to send envoys to Ta-hsia by way of the mountain trails that lead through the territory of the Ch’iang people, they will be molested by the Ch’iang, while if we send them a little farther north, they will be captured by the Hsiung-nu. It would seem that the most direct route, as well as the safest, would be that out of Shu.”[14]

On hearing Zhang Qian’s report, Emperor Wu dispatched four envoys to look for the route, but all failed. However, the Han court had gained a clear understanding of local peoples in Yunnan, as shown in Shi Ji. These earliest records by Sima Qian provided invaluable though obscure sources on the Southwest Silk Road, since it was the only textual record of that period.

The SSR refers to a route network linking Southwest China (Sichuan and Yunnan), with Tibet, Southeast Asia, South Asia, and beyond. The SSR includes four main branches with numerous sub-branches. The first links Sichuan and Yunnan to India, via Burma. It is called the Road of Chuan-Dian-Mian-Yin (Sichuan-Yunnan-Burma-India) or of Shu-Yandu (Sichuan-India) by the Chinese. Because of its enormous significance, scholars sometimes identify it with the SSR. However, other roads also contribute to the formation and function of the SSR. One connects Vietnam with Yunnan; one connects Yunnan with Laos, Thailand, and Cambodia; another is from Yunnan, sometimes via Sichuan, to Tibet and India. Because the main articles traded along this road were tea and horses, it is named Dian-Zang Cha-Ma Gudao (Ancient Road of Tea and Horses between Yunnan and Tibet).

Because of the SSR’s temporal and spatial complexity, a concise introduction to these sections may help readers develop a general idea.

The Shu-Yandu dao (The Road between Sichuan and India) was a major section of the SSR. It started from Chengdu, capital of Sichuan and a symbol of the developed Shu culture that was no less significant than the Shang dynasty. Guanghan, a city about an hour’s drive from Chengdu, was the site of the well-known Sanxingdui relics that exerted a great influence on the bronze culture in Southeast Asia.[15] From Chengdu through Kunming to Dali, the latter two being major commercial and cultural centers of Yunnan, there are two roads. The northern one passed through the following cities: Chengdu, Lingqiong (Qionglai), Lingguan (Lushan), Zuodu (Hanyuan), Qiongdu (Xi-chang), Qingling (today Dayao, entering Yunnan), Dabonong (today Xiangyun), and Yeyu (Dali). Because the Lingguan Pass must be used, so the northern route is called the Road of Lingguan (Lingguan dao). The southern one went through the following cities: Chengdu, Yibin, Zhuti (Zhaotong), Wei County (Qujing), Dian (Kunming), Anning, Chuxiong, and Yeyu (Dali). The two roads joined in Dali, where the so-called Road of Bonan (Bonan dao) commenced, because it had to cross Bonan Mountain. The Road of Bonan passes Yongchang and Tengyue to Burma and India.

The second branch of the SSR was from Yunnan to Vietnam.[16] The Hong River linked Yunnan and northern Vietnam, which accounted for some common features of bronze crafts found in Yunnan and Vietnam.

The route actually was winding with the Yeyu River, today’s Yuan River or Hong River. It started from Miling of Jiaozhi, via Jinsang, Fengu, or Xisui, to Dali where it connects with the Bonan Road.[17] That is why Jia Dan (730–805 CE) of the Tang dynasty named it the “Route between Annam and India” (Annan tong Tianzhu dao).

The third section connected Yunnan with Laos, Thailand, and Cambodia. Indeed southern Yunnan, Upper Burma, and Laos cannot be divided geographically, culturally, or ethnically. Communications among them seem self-evident. Though early Chinese documents are lacking, tribute missions of small states in Laos, Thailand, and Cambodia were by no means rare after the Tang.

The above three sections actually are what Sun Laichen calls the Sino-Southeast Asian overland interactions. Obviously Yunnan was a nucleus of this overland interaction.

Before we turn to the linkage between Yunnan and Tibet, let us first take a look at Chinese sources of the Tang period that illustrate the dynamics of the three southward routes going through Yunnan.

Xuanzang (mid-seventh century) and Yijing (later seventh century), the two famous Chinese Buddhist pilgrims, both mentioned these routes in detail between India and Sichuan, and their records of miles and days are pretty close, which shows that some people of that time were familiar with the routes.[18] Fan Chuo, a military official who served in the Tang dynasty’s Annam Protectorate (Annan Duhufu) had also recorded these roads in his Yunnan Zhi (History of Yunnan), compiled in about 863 CE.[19] But all their contributions were out-detailed by Jia Dan.

A prime minister (zaixiang) in the Tang, Jia Dan in the seventeenth year of the Zhenyuan reign (801 CE) presented the emperor with two books that documented Sino-foreign communications. His books are both missing, but fortunately Xin Tang Shu (New history of the Tang dynasty), edited in the tenth century, retained some of their passages about the seven routes linking China with barbarians of four directions. The sixth route linked Annam with India (Tianzhu). This route started in Tonkin, and continued via Yunnan, through Prome, to Magadha.[20]

There were two ways from Tonkin to Dali: one via the river; the other over land. After arriving at Dali, the routes join and extend to Burma and India. And again it was divided into the southern and northern branches from Yunnan to India. This northern one was about 3,200 li, compared with 5,600 li for the southern route. The southern route seemed too roundabout, but it is important because it not only links Yunnan and Burma, but also connects the maritime Silk Road with the SSR, which explains why merchants bothered to take this longer and winding way.[21]

Because all the above sources are recorded by the Chinese and many of them indeed are cited from official and pro-official histories, it is no wonder that scholars have been able to develop a relatively clear map of the SSR within the Chinese imagined boundaries. In fact, one of the shortcomings of present research concerns the section connecting Burma and Assam with Indian states of which the Chinese did not have detailed descriptions.

Mountains dominate the region from northwestern Burma through Assam to the Ganges Plain. These ranges run parallel north to south and function as natural barriers to communication. Many passes exist, however, and they were presumably used by peoples on both sides. Unfortunately, unlike the Chinese, the indigenous peoples did not leave us with any early documents. Contemporary descriptions, nonetheless, can help us trace these routes in earlier times.

When examining how bullion was shipped from Yunnan and Upper Burma into Bengal during the period 1200–1500, John Deyell has illustrated the intertwined network of three overland routes, in addition to the water route. The migration was via three main land routes from Yongchang, through Upper Burma, westward to Bengal.

The first went from Yung Chang to Momien, crossed the Irrawaddy to Mogaung, went north through the Hukong Valley, across passes in the Patkai Range, to the upper Brahmaputra Valley. This was the eastern frontier of Kamarupa. The second route followed the Shweli River, crossing the Irrawaddy at Tagaung, followed the Chindwin River north, and crossed via the Imole Pass to Manipur. This was the eastern approach to Bangala via Tripura. The third route embarked on the Irrawaddy at Tagaung, Ava, or Pagan, and then passed from Prome over the Arakan Range to Arakan. A variation of this went directly from Pagan to Arakan via the Aeng Pass. This gave access to either a land route northward to Chatisgaon, or embarkation on the coastal trading boats to Bengal.[22]

In his discussion of the Assam-Bengal trade of the medieval era, Nisar Ahmad also details trade routes between the two areas. Three routes lead from Assam to Bengal: one by water and the other two by land. The Brahmaputra River is easily navigable. Of the two land routes, one was from Tezpur (Darang district of Assam) to Lakhnauti (the capital of Bengal Sultans) through the districts of Kamrup and Goalpara, north of the Brahmaputra. The second ran to the south of Brahmaputra, and, after crossing the river, joined the first path. The second path seems favored by merchants who were interested in sea trade, because it connected with the river ports of Bengal. Moreover, Lakhnauti had the line of connection with Tibet via Kamrup. And there was a route from Kashmir to China (Yunnan) via Koh-i-karachal (Kumaon Mountains), Patkai Hills, and the upper districts of Burma, which was joined by a passage from Lakhnauti. Nisar also points out that some portions of the three routes (Lakhnauti-Tezpur, Lakhnauti-Tibet, and Lakhnauti-China) probably were common.[23]

Although both scholars talk about the early medieval ages, earlier routes should have been basically the same, because these routes utilized mountain passes and rivers.

The last section of the SSR is that between Yunnan and Tibet. The Yunnan-Guizhou plateau indeed is the extension of the Tibetan Plateau so that northwestern Yunnan naturally connects with Tibet. In the Ming dynasty Yunnan exported tea to Tibet. Tea started from Puer and traveled via Dali, Lijiang, Zhongdian, Zayu, and Bomi to Lahsa, and from Lahsa to Nepal and India. Because Sichuan also neighbors Tibet, the Yunnan tea could pass through Sichuan to Tibet as the Sichuan tea could be shipped to Tibet through Yunnan. As John Deyell notices, “A separate apparently well-traveled and renowned route led from the region of the Upper Yangtse-Mekong-Shalween Rivers through Tibet. Passes led through Bhutan and Nepal into Kamarupa and Hindustan respectively.”[24]

Known as the Road of Tea and Horses, the Yunnan-Tibet connection started before the Tibetan people began to drink tea around the eighth century. Because Yunnan bordered with Tibet, Burma, and India, goods could take the southward road through Burma and India to Tibet. Fan Chuo had detailed this southern route.[25] Roundabout as it is, the route had been often used before the Tang, because it converged into the Yunnan-Burma-India route. Because this way is much longer, communications between India and Yunnan more often followed the northwestern branch after the eighth century because of the close relationship between Yunnan and Tibet. Indeed, the Tibetan empire once controlled northwestern Yunnan, and many local regimes there accepted Tibetan suzerainty in the seventh and eighth century. During the Ming and Qing period, Tibetan demands for tea were totally dependent on Yunnan and Sichuan.

The above is a very short description of the Southwest Silk Road. While the geographical framework has been introduced, the temporal dimension of the route should not be ignored. The four sections emerged in different times, they functioned variously, and their sub-branches changed courses through history.

We now have a general map of the geographic location and historical development of the Southwest Silk Road, especially before the medieval period. Another key question is what sorts of goods circulated on this road. Again, besides many common goods such as silk, precious metals, and stones valued by ancient peoples, the articles exchanged were full of local features. In addition, articles traded on the SSR changed over time, which makes it more difficult to discuss. If time were omitted, commercial articles in the SSR included shell, jade, precious stones, elephant tusks, horses, lumber, cloth, herbs, spices, salt, tea, gold, silver, copper, tin, lead, cotton, silk, and opium, just to list some that are familiar to us.

The distinct contribution of Yunnan to this traffic includes two items studied here: horses and silver. Both were local products of Yunnan. Through utilizing existing Chinese and other scholarship, I seek to illustrate how trade along the SSR shaped local societies.


Horses were a famous local animal in Yunnan. Along with slaves and oxen, they constituted an important source of wealth for people west of the Dian, according to Sima Qian. Horses entered Yunnan along with prehistorical migrations from Central Asian grasslands. After centuries of adaptation, horses in Yunnan can be classified into two types: one from northwest Yunnan, where the climate is cool; the other from southern Yunnan, where the climate is subtropical.[26] Archaeological evidence shows that horse husbandry was popular as early as the sixth century BCE. Horses were used in wars as well as for transportation.[27]

Horses in Yunnan were smaller and shorter than their northern counterparts. They might not have been good at fighting on the northern frontiers, but they did have some advantages in the Huai River region and the Yangzi River region where the dominant subtropical climate probably favored the Yunnan horses. In addition, some horses in Yunnan were products of the northwestern plateau. That is why Marco Polo concludes, “The best horses here are bred in this province.”[28]

Though horses in Yunnan were well known to the Chinese, there is little evidence that they existed in China before the Han dynasty in the third century bce. Horses were reportedly exported to Sichuan in the early Western Han.[29] Legends recorded in Hua Yang Guo Zhi (HYGZ) and Hou Han Shu (HHS) talk about divine horses (shenma) in Lake Dian, which may imply that high-quality horses had already appeared before the third century CE.[30] One source in HYGZ also suggested the prosperity of husbandry, since 300,000 head of oxen, horses, and sheep were presented to the Han court.[31] Since the third century, Yunnan horses have been visible in Chinese documents. Riding Yunnan/Sichuan horses became a sort of fashion for nobles in the Central Plains, while horses served as war sources for local regimes.[32]

Sources of the Tang/ Nanzhao period document horses in local society. For example, the Nanzhao presented horses as tributary items to the Tang court.[33] Both Yunnan Zhi and Xin Tang Shu (New History of the Tang Dynasty, hereafter XTS) listed the Yuedan horse as the best.[34] Yuedan, in the former Ailao area, was today’s Tengchong. Zhou Qufei of the Song dynasty gave more details about Yunnan horses. He concluded that the farther northwest the horses came from, the better they were. Southern horses, according to his evaluation, were not as durable as northeastern ones. However, he admitted that the best southern horses exceeded northern horses, and were worth several dozen taels of gold.[35] In fact, the Song court highly valued Yunnan horses, seeing them as good as those from the north.[36]

Chinese people from early imperial times desperately desired horses from Central Asia, since most areas of the Chinese realm did not produce horses, an animal of high commercial, but, more important, military, value.[37] South China did not produce horses, which might account for the fact that southern-based kingdoms barely managed to resist northern invasions. Yunnan, south of the Yangzi River, was an important exception, thanks to its unique geography. Hence, if possible, southern Chinese kingdoms were eager to get access to Yunnan horses. Here are the two classic cases: one of the Three-Kingdom period (220–265 CE); the other, as we know, of the Dali/Song period (mid-tenth through mid-thirteenth centuries CE).

Horses in Yunnan gained strategic importance during the third century CE, a time that witnessed the rise of three kingdoms in China, namely, the Wei in north China, the Wu in Jiangnan, and the Shu in Sichuan. Among them, the Wei was relatively powerful. Therefore, there was an alliance between the two southern kingdoms. On the other hand, conflicts between Shu and Wu occasionally took place. One place of contention was Jiaozhou (northern Vietnam). The control of Jiaozhou was extraordinarily crucial to the Wu kingdom, because it would not only give them access to the profitable overseas trade, but also allow them to gain access to horses from Nanzhong (Yunnan) through the Yunnan-Vietnam connection. Furthermore, Jiaozhou was a springboard to Yunnan. The Shu also wanted to take over Jiaozhou, but conditions were not favorable for this undertaking. Yong Kai, the local chieftain who controlled Nanzhong, was in communication with Shi Xie, who headed Jiaozhi and sided with the Wu. Through Shi Xie, Yong Kai established connection with the Wu. Therefore, the immediate task of the Shu was to take over not Jiaozhi but Nanzhong, as it did later.

Wu’s desperate demand for horses can be illustrated with the following case. Sun Quan, the king of Wu (222–252 CE), dispatched envoys to Liaodong (Northeast China) by sea where Gongsun Yuan was in charge. Gongsun, betraying his promise, killed the envoys and presented their heads to the Wei.[38]

Although the Wu kingdom received some horses from the Shu and Wei through official transactions, it obtained horses in large numbers from Nanzhong through Jiaozhi. The Shi family of Jiaozhi paid tribute to the Wu. Horses of course were a major item on the tribute list. It is recorded that several hundred horses each year were presented to the Wu,[39] while trades probably would exceed this figure. When the Wu finally eliminated the power of the Shi family in Jiaozhi in 226 ce and ruled Jiaozhi, we can assume that more horses were moved from Yunnan to the Jiangnan area, where the Wu kingdom was based.

The horse trade between Yunnan and Annam lasted at least to the Tang, according to sources which said that horses were still in trade in the reign of Dezhong (860–873 CE).[40] Sources remain silent in terms of horse in other branches of the SSR at the time.

The Dali/Song period reached an unprecedented degree of horse trade. During the seventh century, a local regime named Nanzhao (lit. “Southern Kingdom”) rose in the Dali region and gradually unified Yunnan. It seems that continuous wars between Nanzhao, Tang China, and Tibet contributed to the collapse of all the three empires. At the end of the ninth century and in the early tenth century, Yunnan was in chaos until the Duan family allied with local chieftains, unified Yunnan, and established the Dali kingdom. The Dali kingdom desired trade with Song China, as trade was a major basis of its economy. That is why the Dali appealed to the Song court for tributary missions on many occasions. The Song court, having seen the troubles between the Nanzhao kingdom and Tang China, no longer intended to impose her suzerainty on the Dali.[41] Several times, the Song court refused the appeal of the Dali for tributes,[42] but its anxious demand for horses compromised the stance of isolation.

The Northern Song (961–1127 CE) confronted powerful kingdoms in the northern frontiers. The Liao dynasty (916–1125 CE), followed by the Jin kingdom (1115–1234 CE), put great pressure on the Northern Song. Finally, the Northern Song was unable to defend the Yellow River region and retreated to the Yangzi River region, starting the Southern Song (1127–1279 CE).

Both the Liao, the Jin, and later the Mongols enjoyed the advantage of cavalry. In order to fight effectively, the Song eagerly demanded good horses. Though the Song made efforts to breed warhorses, it failed to meet the demands of war,[43] so the only alternative was to trade with the nomadic barbarians. The Liao, Jin, and Mongols of course did not like to export horses to the Song, at least not on a large scale. As a result, though the Song did not like contact with the Dali kingdom, it had to, because only the Dali produced horses. The Song in the Xining reign (1068–1077 CE) made it a state policy to exchange silk and tea for horses.[44]

Consequently, the Song officially opened over a dozen markets in Sichuan and Guangxi.[45] Sichuan was a major horse-trading center in the Northern Song of the tenth and eleventh centuries, although the Song court and its frontier officials avoided regular contacts with the Dali, whether commercial or official. Official markets were set up in Lizhou (Hanyuan), Yazhou (Yaan), Jiazhou (Leshan), Rongzhou (Yibin), Luzhou, and Changningjun (Gongxian). Merchants from Dali moved their horses and other local products through the SSR southward to Sichuan. While encouraging southwestern barbarians to drive horses to Sichuan, Chinese people were recruited into Yunnan as well. Yang Zuo, a Jinshi degree holder in Emei (Sichuan), took all his family assets and led several dozen people with silks and substances, found ways through mountains, and finally met local chieftains, announcing Song’s willingness to participate in the horse trade.[46]

However, the Song dynasty’s loss of the Yellow River region made Sichuan a frontier area facing both the northern threat and potential southern intervention from Dali, so that the Song tightened the reins of the horse trade and closed some markets in Sichuan. And the center of the horse trade shifted to Guangxi.

The horse trade in Guangxi began in the Northern Song and entered its golden time in the Southern Song. Several markets were set up, administered by the Offices of Horse Trade (maimasi). Among many horse markets, Hengshanzhai (Tiandong County, Guangxi) was the most famous and lasted the longest, from the Yuanfeng reign (1078–1086 CE) until the Mongols’ occupation of Yunnan in the mid-thirteenth century. The Song state set the quota of horses in Guangxi: 1,500 head each year, however, some years had seen the purchase of 3,000–4,000 head.[47] Whereas the Offices of Horse Trade were established and many officials were assigned exclusively for this state project, local officials were ordered to help. Horse trade had been a major state project both in the Song central court and in local governments.

Unlike the previous trade, horse trading between the Dali and the Song can be utilized to reflect the scale of the trade. Fang Guoyu estimates that the average scale of horse trading in Guangxi was about 70,000 taels of silver, based on 1,500 head per year and 30–70 taels per head.[48] Fang’s calculation was just part of the trade between the Dali and the Song, since it excluded the trade in Sichuan and exchanges of other items. In addition, the number of horses in some years doubled the quota of 1,500 head per year. Therefore, the trade was large in scale.

In 1253 CE the Mongols conquered the Dali kingdom, which probably marked the end of the official horse trade.[49] The lack of horses then might have contributed to the Song’s failure, though it is hard to know exactly what influence it had. But we do know that horses from Yunnan were crucial for Song’s defense and had positive effects on Southern Song’s occasional military successes. Yue Fei, the famous general who had defeated the Jin cavalry several times, had horses from Yunnan in his army.[50] At the same time horse trade was also significant for the Dali kingdom on the grounds that many substantial items as well as luxuries were imported.

The Yunnan horses were exported not only to China, but also to mainland Southeast Asia. Jiaozhi and Zhancheng were two local regimes around today’s Vietnam, and they competed with each other. They both produced horses that were too short to use in wars. At first, they were dependent on the Song horses, many of which indeed migrated from the Dali. The Song, exhausted by the northern conflicts, tried to keep the two southern regimes balanced, avoiding the emergence of a dominant power that could have been another troublemaker. To achieve this goal, the Song intentionally controlled the quality of horse exports.[51] However, Jiaozhi eventually won this contest thanks to its ability to obtain more horses. As a matter of fact, the horses obtained by both Jiaozhi and Zhancheng were mainly from Dali, directly or via the Song. Almost at the same time, the Song cavalry put down a rebellion led by Nong Zhigao in Guangxi. Those circumstances contributed to the spread of cavalry in eastern mainland Southeast Asia, and horses from Yunnan, as the major source, increased their significance in local societies and power struggles.[52]

In a word, it may not be an exaggeration to conclude that horses from Yunnan had to some extent shaped the power struggles east of the Himalayas. A study of horse trading in early Bengal by Ranabir Chakravarti, on the other hand, shows how horses from Yunnan affected Indian states.[53]

Like many other parts of India, Bengal during ancient and early medieval times did not have any indigenous, good quality warhorses. The emergence of Bengal as a regional political entity like the Palas (c. 750–1175 CE) and the Senas (c. 1096–1225 CE) during early medieval times (c. 600–1300 CE) must have increased the demand for war-horses. The northern quarter of Bengal, sometimes called Uttarapatha, had its traditional supply of horses that provided the Palas, the premier regional power in eastern India, with fine warhorses. The Sena kingdom, which succeeded the Palas as the most powerful ruling house in early medieval Bengal and Bihar, also highly valued horses as reflected by the title asvapati (lord of horses) the ruler assumed, along with gajapati (lord of elephants) and marapati (lord of men).

When identifying Tatatri and Bakadasti, horses from the north and northwest and hilly horses from Bhutan or Tibet, branded as the Kohi variety, Ranabir notices that Yunnan was another source of hilly horses, as Marco Polo observed. Marco Polo was aware of the excellent and strong horses in Yunnan. He said Amu “produces many horses and oxen, which are sold to merchants, and conveyed to India” after a forty-five-day journey through Pagan and Bengal.[54]

But the export of horses was not only for Bengal; indeed Bengal served as a transit center of the horse trade. Horses, after arriving in Bengal, were transferred to the eastern Deccan and China. Chinese sources document no fewer than fourteen missions from the Bengal sultanate to China. Horses were the costliest among the twenty-five articles traded from Bengal to China. Bengal envoys were both in their own ships and in the Chinese treasure ships. Therefore, horses played a significant role in the Bengal-China tributary trade.

In fact, Ranabir traces the horse trading in Bengal back to the third century and concludes that both in the third century and during the post-1200 era, horses were brought to Bengal from elsewhere (including Yunnan) and then a number of them were exported as an item of maritime trade. It is certainly not a coincidence that the ages he discusses agree with the flourishing horse trading in Yunnan, especially in the medieval era when large numbers of horses from Yunnan were exported to Song China.

Although we know neither how many horses were traded into Bengal from Yunnan nor how they compared with those from Tibet and the northwestern frontiers, it is valuable to realize that horses from Yunnan played prominent roles in Indian societies.


Like horses, silver from Yunnan (and also from Upper Burma) played a significant role in China and in southern India. Yunnan abounds in metal resources: gold, silver, copper, tin, and lead, to list a few. Some scholars have already pointed out that bronze in Yunnan was shipped to the Central Plains over three thousand years ago, and made into the fine bronze items by the Shang and Zhou people.[55] Silver was another Yunnan item that the Chinese desired.

Ban Gu of the first century already mentioned the Zhuti yin (silver mined in Zhuti County [Zhaotong]) that was valued one and half times as much as other silver.[56] Later sources all list silver as a local product, together with gold and tin. One source reads that Zhuge Liang, the minister of the Shu kingdom, once mentioned the mining of Zhuti silver as an important source of wealth.[57] Another says several dozen silver sites in Nanzhong (Yunnan) paid yearly tribute to the Shu.[58] Although available documents cannot produce a map of silver production in Yunnan, we can assume silver mining was ongoing on the grounds that silver has always been listed as local product. And it is highly possible that the Nanzhao and the Dali kingdoms had been deeply involved in silver production, by either monopolizing or taxing, or both.

The Yuan dynasty did tax silver mining. The Office of Silver Mining (Yinchangguan) was created, and silver taxation was set with yearly quotas. Yunnan and Jiangxi were the two provinces whose silver outputs were the greatest. In the first year of Tianli reign (1328 CE), the silver mining tax (yinke) in Yunnan was 36,784 taels, compared to 23,104 taels in Jiangxi.[59]

Silver mining of Yunnan in the Ming dynasty developed, as reflected by tax on silver mining. Song Yingxing of the sixteenth century pointed out that silver output in Yunnan constituted over half of the national output, more than that mined from other provinces together.[60] Other documents affirm his conclusion. In the second year of Tianshun reign (1458 CE) the quota of silver mining tax in Yunnan was set at 52,380 taels, compared with 21,250 taels in Zhejiang province and 15,120 taels in Fujian province, the latter two provinces second only to Yunnan.[61] Two years later, the silver mining tax in Yunnan reached over 100,000 taels, almost double what it had been.[62] These figures definitely illustrate the significant contribution of silver from Yunnan to the Ming, because the annual average silver mining tax of the nation throughout the Ming period was around 100,000 taels.[63] Quan Hansheng estimates that the total silver mining tax during the years 1390–1520 was 11,395,775 taels.[64] Supposing that silver from Yunnan makes up half of national production, that is 5.7 million taels, and supposing the silver mining tax in the Ming was around 30 percent or more as Quan concludes, Yunnan in the Ming dynasty had produced 19,000,000 taels of silver. James Lee estimates that by the end of Ming, Yunnan had produced 2.5 million kilograms of silver, three quarters of China’s total output, as much as the Portuguese trade brought to Ming China.[65]

As a matter of fact, the silver mining tax in Yunnan had gradually become an enormous burden for local people, and was unable to be completed yearly.[66] Ming China was in the process of silver monetization, which resulted in an insatiable desire for silver. Consequently, silver resources in Yunnan were exhausted, as reflected by numerous official reports.

Scholars have paid a lot attention to “the transition from the coin economy of the early imperial era to the silver economy of the later imperial period [that] marked a crucial watershed in the evolution of Chinese society, economy and culture.”[67] Scholarship on seventeenth-century China likewise has bifurcated into the “early modern” approach, which emphasizes the economic stimulus of silver imports, and the “crisis thesis,” which underscores the dire consequences of China’s dependency on the world economy.[68] While the massive inflow of silver engendered a great transformation in China’s economy, the shift from bronze coin to a silver standard already was under way long before the influx of foreign silver in the late Ming.[69] Hence arises the question, what role did Yunnan silver play in this shift?

As revealed, Yunnan silver in the Yuan and early Ming constituted a major amount of the national production, and can be compared with the New World silver import in terms of scale as James Lee states. Certainly we can conclude that Yunnan silver contributed to this shift, although the degree remains unexplored. If so, then scholars may like to reconsider the impact of silver imports from the New World. Or at least, Yunnan silver should be taken into account for China’s monetary transition and economic development from the Yuan onward. Furthermore, the case of Yunnan silver obviously sheds light on the interaction and impact of the Chinese incorporation of local societies. Local resources were utilized to serve the national interest. To put it another way, it is fair to conclude that the inland prosperity was built at the price of exhausting the local resources of frontier/peripheral areas, though this article will not examine this issue.

Again, let us turn to Bengal. John Deyell’s study on medieval Bengal (1200–1500 CE) overlaps with the Yuan-Ming time.[70] He points out the fact that during this period, Bengal states had gradually achieved silver monetization. “Due to the total absence of indigenous sources of silver in Bengal,” he concludes, minting, export, industrial uses, savings, and attribution losses of silver in circulation during 1200–1500 “were directly all dependent on the rate of importation of silver.”[71]

Hence, the crucial question is where the silver was imported into Bengal. Deyell has examined the sites of gold and silver mines in East and Southeast Asia and concludes that contemporarily available silver was from Yunnan and the northern Shan states, because silver sources in Siberia, Manchuria, Japan, and Hu’nan in the mid-Yangzi region were too far away to serve as sources for medieval Bengal. In addition, accessible gold sources basically concentrated on the Tibet-Sichuan-Yunnan borderlands as well.

Naturally, Deyell turns to study the road through which silver had been shipped from Yunnan and Burma to Bengal and westward. He identifies land routes and water routes connecting southwest China and frontier areas of Bengal. Although sea routes or the maritime Silk Road played a great role in West-East exchange even before the arrival of Europeans, Deyell has actually demonstrated that land routes contributed to silver shipment more than sea routes. Several reasons he lists accounted for the advantages of land routes. First, because of the limited stock, silver was more valuable than in modern times. And small quantities with relatively high value were good for animal transport as well as for ships, while land routes were safer than those over the sea. Furthermore, land routes were considerably shorter than sea routes, regardless of the fact that sea routes also required silver to be shipped to the ports over land roads. There was another problem for maritime transport: the Yuan and Ming dynasties prohibited the export of precious metals.

In addition, Deyell notices that the indigenous peoples in Yunnan and Burma had mastered the technology of silver mining and smelting and played agents of trade. For example, many Chinese sources mentioned the indigenous group named Jinchi, or “Gold Teeth,” which Marco Polo had met. And Bawdwinggyi, the name of the major silver site, means “Great Silver Mines.” Citing from other studies, Deyell emphasizes the close relationship between the Nanzhao and its following Dali regimes in Yunnan with states in Burma.

In a word, Deyell believes that he has established “sufficient and necessary conditions for the existence of bullion movements during the period of the late thirteenth to the early sixteenth centuries from the Shan states or Yunnan to the Bengal sultanate.”[72] And his study joins the demonstration that trade along the SSR had shaped the Bengal states, at least with regard to economy.

The Cowry Money System in Yunnan: A Global Analysis

Trade along the SSR circulated diverse goods, but what was the medium of trade, or what sorts of things functioned as money? Contrasted with the prosperity and dynamics of trade along the SSR, scholars seem disappointed to “find that the records (mostly epigraphic) of these countries do not furnish much information on different economic professions, nor is there any reference to a Khmer or any other coinage, though one can hardly doubt the circulation of coinage on the progressive economic life in those countries.”[73] In fact, barter had been the major form of trade until the very recent period of European colonialism.[74]

There were not many currencies in this trade network. Gold, silver, cloth, and salt certainly sometimes functioned as money. Nonetheless, their roles indeed were dual, namely, commodity money. Many reasons account for the barter trade along the SSR. For example, there was no single empire strong enough to control the road and to carry out her currency policy. Indigenous long-distance trade as transit trade was controlled by local elites and completed by many merchants. While Marxist scholars of China criticize that the lack of common currency reflected the low degree of commercial economy, it instead revealed the interdependence and dynamics of local economies.

However, there was a common currency circulated on the SSR: cowry money. Cowries originally came from the Maldives and had for long time been exported to Bengal, and from there migrated throughout the Old World. Cowry money was used in India; Southeast Asia including Arakan, Martaban, Pegu, Siam, Laos, and Burma; and Yunnan.[75] Because of their features, cowries became currency in transactions of small value and penetrated into daily life.[76] James Heimann concludes that in India, while metal currencies dominated the trans-market economies, cowries were counterparts in the local markets.[77]

Chinese travelers, such as those in Zheng He’s Treasure Fleet, noticed cowry money in Southeast Asia when cowry money had disappeared in China much earlier, except in Yunnan.[78] Tombs and relics dated to the Xia (twenty-first through sixteenth centuries BCE), Shang (sixteenth through eleventh centuries BCE), and Zhou (eleventh through third centuries BCE) periods have presented a lot of shells. Cowries were not indigenous around the Yellow River, but were imported from coastal areas or might come through Central Asia. After the Spring-Autumn Period (770–476 BCE), cowries gradually began to retreat from Chinese markets. When Qin Shihuang, the first emperor of China, unified China in 221 BCE, he standardized diverse measures, moneys, and characters prevalent in previous states. As a result, cowries were replaced with iron and copper coins. Since then cowries have been used only as ornaments.

The case of cowries in Yunnan was different from that of China proper in terms of their long-term existence and their diverse usage. From 1955 to 1972, archaeologists unearthed a large number of cowries in tombs in Yunnan, amounting to more than 260,000 pieces and over 700 kilograms.[79] All these tombs are dated before the Qin unification in 221 BCE and demonstrate that cowries were present in Yunnan before the late third century BCE. Were these cowries moved into Yunnan as a sort of money or a sort of valuable item with high esteem? Scholars hold different opinions.[80] Michele Pirazzoli-t’Serstevens “has demonstrated that the most frequent cowry species found in Tien tombs are Cypraea annulus Lin. This seems to have been a special highly prized form of money, a status marker, and a certain form of prestige goods, accumulated as stores of value and used in intersocietal exchanges between elites, exclusively.”[81] Hence, these cowries serve as strong evidence for the SSR, or trade of luxuries between Yunnan and the Indian Ocean zone.

Probably it is worthwhile to mention one more time the cowries unearthed in Sanxingdui (Sichuan), dated over three thousand years ago—the same age as cowries found in the Shang tombs of the Central Plains. It is said that what was found in the Sanxingdui matches those in Yunnan, which, together with the geographical implication, strongly suggest that the source of cowries in Sichuan was the Indian Ocean and that they came into China through Yunnan.

While the paucity of textual sources allows us no further conclusion, it is confusing that very few cowries have been found in Yunnan by archaeologists that date from after the Western Han. It seems that the cowry trade suddenly was disrupted. The only reasonable explanation is the Han’s military control over Yunnan. Both the Western Han and the Eastern Han made efforts to subjugate Yunnan. When the Ailao people submitted themselves in 69 CE, Han China controlled most of present-day Yunnan, with Yongchang (Baoshan) as a frontier station. The Han’s presence in Yunnan certainly affected the local economy. Though it is difficult to reckon the degree, the disappearance of cowries and the presence of Han’s Wuzhuqian, a popular coin, in archaeological sites served as solid evidence. Wuzhuqian are unearthed even in Yongchang. The case of cowries shows that political changes as a matter of fact had a great influence on economic orientation in Yunnan during the Han time.

It was during the period when China lost authority over Yunnan that cowries again entered Yunnan in large numbers and a cowry system took shape. But the issue of the exact date remains unknown, since the first Chinese text recording cowry money in Yunnan is Xin Xang Shu (New history of the Tang dynasty). Fan Chuo’s Yunnan Zhi, written around 864 and assumably a firsthand source, did mention cowries, but as ornaments rather than money. Hence Pelliot concludes it is difficult to know details before the tenth century.[82] However, Fang Guoyu, incomparably familiar with Yunnan sources, points out that Fan Chuo, though he was in Annam at the time, used many records from previous works, sometimes two centuries earlier than Fan’s era.[83]

Based on Fang Guoyu’s research, Vogel cautiously concludes that cowry money was used in Yunnan in the early ninth century, which is implied by the popular acceptance of Buddhism in the Nanzhao while leaving space for an earlier speculation.

The Song sources again remain silent, which Vogel wrongly accredits to the political independence of the Dali kingdom.[84] As discussed previously, the Song intentionally reduced its connection with Yunnan and refused Dali’s appeals for tributary missions. But the Yuan and Ming have left abundant sources of cowry money in Yunnan. Diverse usages of cowry money are revealed in those texts. Cowries were used in land purchases as well as in small transactions. Cowries also were used to pay taxes, to make donations to Buddhist churches, and to pay government salaries. In addition, cowries were stored as treasures. Contrary to the common image, cowries in Yunnan served not only in small transactions, but also in large activities, such as house sales. In short, cowry money fulfilled the functions of modern abstract money.[85]

Vogel’s study has illustrated that the basic denominational system of Yunnan cowry money had remained unchanged between the ninth and seventeenth centuries. Moreover, Vogel has uncovered a close relationship of denominational and reckoning systems, compared with those in Bengal and Siam, which serves as another basis of the cowry system in the Indian Ocean trade network.[86]

Two questions intrigue scholars in terms of cowry money in Yunnan: Why did cowry money last so long in Yunnan, over one thousand years longer than in the Central Plains? And why did the cowry system in Yunnan collapse suddenly in the mid-seventeenth century?

A lot of studies have been done, and, in general, two schools are debating with each other. The first school emphasizes that the penetration and control of central states (that is, Chinese governments) used power to forbid and abolish cowry currency, which, in turn, helped the consolidation of the central presence on the periphery. Therefore, state policies are the key to understanding the disappearance of the cowry monetary system.

As early as 1948, Jiang Yingliang, one of the pioneers in Yunnan studies, pointed out that since the Yuan-Ming era, Yunnan had been under the direct rule of China. As a result, the relationship between Yunnan and central states overcame that between Yunnan and Siam. Consequently, silver and copper coins, the Chinese official currency, took the place of cowry as the “naturally economic choice.”[87] Jiang also stated that the lack of copper coins in Yunnan partially constituted another reason for the long-standing use of the cowry money, and that copper coins produced in large amounts in the end of Ming and early Qing facilitated the replacement.

Yang Shouchuan instead focuses on the growing markets that he regards as the result of the Han immigration encouraged by the state. The small value of cowries could not meet the demands of the growing markets when many relatively valuable items were introduced. Silver and copper became the main media of trade. The emergence of the private mining industry, another impact of state policy, is listed as a factor also. Since the mid-Ming, the state relaxed its mining monopoly in Yunnan, which to a great deal contributed to the rise of private copper mining and the growth of copper output.[88]

Furthermore, Yang stresses the policy of the Daxi regime at the end of the Ming dynasty, since the disappearance of the cowry coincided with the rule of the Daxi regime in Yunnan. The Daxi regime was established by Zhang Xianzhong, leader of a peasant rebellion at the end of the Ming dynasty. After Zhang’s death, Sun Kewang, a general of Zhang’s, occupied Yunnan. Sun ordered the production of copper cash for his regime, a legitimate symbol of new power, and the use of cowries was forbidden. It is recorded that eighteen large minting furnaces were set up, and all kinds of taxes, levies, and trades were completed using silver and copper.[89]

Focusing on internal changes caused by the state policies, the first school argues that cowry currency was popular before the mid-seventeenth century because of low social productivity, and that it disappeared in the mid-seventeenth century because of the prosperity of the commodity economy.[90] In short, this school believes that the replacement of cowries with copper cash was the result of “natural economic law” represented by state policies. The retreat of cowry money, on the one hand, was the result of central penetration; on the other hand, it helped and symbolized the central penetration.

Contrasted with the first school, the second school extends its eyes into the globalizing modern capitalist world. With a global perspective, its conclusions are diametrically contrasted.

Fang Guoyu points out that cowries in Yunnan were shipped from coastal areas. Therefore, even when the state forbade using cowries as money as the medium of trade, cowries continued to enter Yunnan. So, unlike the first school, Fang argues that the state policy violated the natural economic law.[91]

Instead of turning to state policy, Fang examines the source of cowries, namely, the coastal areas in Southeast Asia. Fang points out that the long-term existence of cowries in Yunnan was because of Yunnan’s close commercial relationship with those areas where what happened consequently explained the change of Yunnan cowry system. The advent of European capitalism in South Asia and Southeast Asia, Fang Guoyu argues, broke the previous trade network, which left a negative influence on the commercial relationship between Yunnan and coastal areas. Previous trade, as a result, decreased and declined. Cowry money, the symbol of trade, could not be maintained. Therefore, Fang Guoyu concludes that the collapse of trade system around South Asia and Southeast Asia resulting from European colonialism that functioned as the decisive factor.[92]

Like Fang Guoyu, Zhang Bincai puts the cowry system in the global context within which he pursues the question as to why the cowry system collapsed in Yunnan over a relatively short period of time (1660s through 1680s). Zhang questions the influence of state policies or the growing internal markets of Yunnan that he thinks were insufficient to account for the quick disappearance of cowry money. Following Fang Guoyu, Zhang has examined the expansion of European capitalism in Southeast and South Asia. However, while Fang argues that European capitalism destroyed the existing trade system so that Yunnan’s maritime connection was cut, Zhang argues that European capitalism neither broke the local trade network nor intentionally blocked the supply of cowries. Indeed, Zhang finds that the collapse of cowry money in Yunnan is an unexpected consequence of European commercial capitalism.

Sharply contrasted with Fang Guoyu, Zhang points out that the trade had been intensified after the coming of the Europeans until boundaries were demarcated after 1949. However, the dramatically increasing slave trade led to the growing demand of cowries. As a result, the cowry prices increased so much that Yunnan could not afford them. Hence, cowries in the Maldives were no longer available for Yunnan.[93]

It seems that both schools are right, because internal and global changes both contributed to the collapse of cowry money in Yunnan in the mid-seventeenth century. Vogel points out that while the value of the cowry increased in Bengal, in Yunnan it was devalued. As a result, it was no longer profitable to ship cowries into Yunnan.[94] And I speculate that the huge price gap may have resulted in the outflow of cowries from Yunnan back to the Indian Ocean.

Unfortunately, Vogel does not pursue the reason for the devaluation of the cowry. But what else could have caused the devaluation but Ming China’s control over Yunnan? Throughout the Ming dynasty, over two million Han migrants moved to Yunnan, which dramatically changed Yunnan’s demography. By the end of the Ming, Han Chinese constituted the ethnic majority in Yunnan, with a population of about three million in the reign of Tianqi (1621–1627 CE).[95] The introduction of an agrarian economy and social traditions obviously clashed with the local economic system, including the monetary system. The Ming attempted three times to establish mints in Yunnan, aiming to replace cowries with copper cash. Though not as successful as expected, these attempts, to greater or lesser degrees, disintegrated cowry money and its credibility. Hence the devaluation of cowry was a logical result.

Generally speaking, the cowry’s long-term existence was the result of the SSR that made Yunnan part of the Indian Ocean trade structure.[96] And its dramatic disappearance from Yunnan, to a great extent, was the result of an expanding European modern world-system into the Indian Ocean. The collapse of the Yunnan cowry system indeed yields some significant insights for world-system theories.

Yunnan and the World-System Debate

I regard China as a precapitalist world-system—or world-economy, as Wallerstein calls it—that incorporated Yunnan.[97] But was Yunnan a frontier area or an independent world-system, a section of another world-system or an external area over which the two world-economies were contending? The cowry system in Yunnan and other strong evidence of Yunnan’s connection with Southeast Asia and South Asia may suggest that Yunnan belonged to the Indian Ocean economy before the Ming period.

Before the Mongol conquest, Yunnan generally was independent, although it had been conquered occasionally by China. Since the end of the Han dynasty, Yunnan at least was autonomous, and when the Nanzhao and Dali regimes were established, they were independent from China. But as the Mongol military campaign had successfully brought Yunnan into China, could the Yuan dynasty be regarded as Chinese?

However, cowry money lasted another four centuries in Yunnan after the Mongol conquest in the mid-thirteenth century. What does this suggest? It suggests that although Yunnan was subject to China politically, economically Yunnan was still more closely associated with the Indian Ocean. Even political or military control could not be exaggerated. During the Ming and Qing dynasties, China had a lot of troubles with local frontier tribes and regimes, and China did not always win. The boundaries between China and Southeast Asia were obscure or never existed until demarcated by the modern nation-states in the mid-twentieth century.

The cowry system, representing the trade network along the SSR, indeed has raised intriguing issues for contemporary discussions of world-system theories. Immanuel Wallerstein contends that the modern world-system has dominated the world for five hundred years as Europe, followed by North America, had exercised hegemony.[98] Janet L. Abu-Lughod, in search for the roots of European world-system, has found another world-system between 1250–1350.[99] Andre Gunder Frank and Barry Gills insist rather that a five-thousand-year world-system had existed in the world.[100] One hot issue of their confrontations is whether or not a world-system existed before 1250.[101]

The year 1250 having been chosen by Abu-Lughod as the beginning year of her world-system is interesting, since three years later the Mongols conquered Yunnan and began to penetrate mainland Southeast Asia. The Mongols’ military conquests in Burma and Vietnam, although not as successful as in Yunnan, to some extent facilitated local communications and cross-regional interactions. In this sense, the case of Yunnan seems to affirm the 1250–1350 world-system.

Nonetheless, readers may have already seen the closeness between Yunnan, Southeast Asia, and South India before the Mongols. Silver migration to Bengal on a large scale, according to John Deyell, should have occurred at least half a century before 1250. Horses, again, seem to have moved from Yunnan to Bengal much earlier.[102] And the early Middle Ages saw the rising local powers around Southeast Asia and South India, a result from interactions of overseas and overland trades.[103] James Heimann has illustrated that “an ordered system of ratios” between cowries and specific metal currencies had lasted since the post-Gupta period until the nineteenth century,[104] and demonstrated how “trade in the Indian Ocean integrated local production/consumption patterns and currency development, resulting in a specific Indian Ocean ‘world-economy.'”[105] Hence, local geopolitical units in the Indian Ocean zone, just like those in the Mediterranean, could be understood only “in terms of their interdependency within that wider network.”[106] While Heimman’s argument centers on the Maldives-Bengal cowry trade, Yunnan definitely belonged to his Indian Ocean world-economy, since the similar denominational and reckoning systems extended from Bengal to Siam and Yunnan.[107] If so, does the case of Yunnan support Wallerstein’s several world-economies before 1250 or Frank’s one world-system? Or could the 1250–1350 world-system be pushed back earlier since the Yunnan trade is ignored in Abu-Lughod’s research? I certainly do not have a conclusion, but try to emphasize that the study of Yunnan, a seemly peripheral area but indeed a bridge connecting several civilizations, would help understand volumes and degrees of civilizational interactions. And it is these interactions that eventually formulated the “ecumenical world system.”[108]

Furthermore, Heimann has shown how the expansion of the European world-system into the Indian Ocean adopted and finally bankrupted the existing cowry system. The incorporation of the cowry system into the slave trade had contrasted influences on the Maldives-Bengal, the center of the cowry system, and on Yunnan, the very marginal area of the Indian Ocean world-economy. While the Maldive-Bengal trade further developed because of the increasing demand of cowries, Yunnan, a receiver of cowries, was the first victim. Yunnan was no longer able to afford the booming price of cowries. That is why within only two or three decades cowries in Yunnan, like a balloon pierced by a needle, immediately evaporated from local markets. The collapse of the cowry system in Yunnan shows that the global modern world-system, while incorporating the Indian Ocean precapitalist world-system, Yunnan, part of the Indian Ocean precapitalist world-system, or a peripheral region overlapped by two world-economies, was pushed to be incorporated into the Chinese world-system. Therefore, the Chinese incorporation of Yunnan should first and foremost be credited to the Mongol conquest, a result from the long-term conflicts among China and Central Asia frontiers, and then to some extent to the globalizing European world-system. While the first event was military and political, and the second was economic, neither was Chinese.

But this is just one possible conclusion on the ground that the Southwest Silk Road had an influential role in the Indian Ocean zone. If we extend our view to the three Silk Roads, some facts probably lead to the five-thousand-year world-system.

Three Passages: A World Network?

Like the other two Silk Roads, the Southwest Silk Road had lasted over two thousand years, contributing much to local peoples and societies. Furthermore, the three Silk Roads supplementarily developed interactions between the West and the East, creating a web that had penetrated Eurasia.

Scholars of Southwest China have not only noticed and demonstrated that the SSR was one of the three main Sino-foreign communications, but also have begun to realize the close connections of the three Silk Roads. Indeed, a few argue that the three roads have constituted a world network. Shen Xu, for example, points out that “the Southwestern Overland Silk Route was one of the three main passages by which ancient China made contacts with the outside world, [which] served as a link that linked up the maritime Silk Road and the Northern Overland Silk Route, thus forming an integrated network system of China’s trade, communications, and cultural flow with the outside world.”[109] Sun Laichen, in his research on Southeast Asian overland interaction, has discussed the cooperation of the maritime Silk Road and the Southwest Silk Road in the Ming dynasty, as Janice Stargardt has done for the medieval age. If we examine coastal mainland Southeast Asia, we can argue that the overland routes are the extensions of the overseas trade, or vice versa. And it is the same case when we discuss the interaction between the overland Silk Road and the SSR.

Starting from Yunnan or from Chengdu (Sichuan), the SSR could go either northward by internal highways to Chang’an, the beginning city of the Silk Road,[110] or northwestward, entering the Western Regions; or it could enter Tibet. Tibet indeed was an interaction zone connecting the Western Regions, the Central Plains, Sichuan, Yunnan, India, and Nepal.[111] Winding south, the SSR could enter India, either directly or through the maritime Silk Road, and then go northward, joining the Silk Road. Therefore, the two roads were connected in many cases. However, in the eyes of local peoples, for example, people in Northern India, it would be nonsense to talk about the conjunction of the two Silk Roads—for them there was only one road/network. So is the case for peoples in coastal mainland Southeast Asia, or people in Sichuan, or in Tibet—to them there was nothing but one trade route system. Therefore, when we talk about the “east” and the “west,” we should bear in mind that these words meant something entirely different to the ancient peoples than they do to us, because they did not see any boundaries, either politically or economically. Demarcated international boundaries emerged only in modern times.

In the following I will argue that the three Silk Roads had constituted a network and had worked together for the communications within Eurasia or eastern Eurasia, the area of my study.

First, geographically the three roads constitute paralleled passages within Eurasia. From north to south, the three roads went through and penetrated the continent, and with their sub-branches criss-crossed the continent.

Furthermore, temporally the three roads worked supplementarily. It seems that overland roads had emerged earlier than the overseas route. Though we do not know which overland road came into being earlier, obviously the overland Silk Road had contributed much more in the beginning. However, the wars around Central Asia sometimes blocked the migration of silks, and merchants then turned to the SSR and the maritime Silk Road. One example is in the first century CE when the Eastern Han lost control over the Western Regions: many tributary missions have been recorded through either the maritime Silk Road or the SSR, a solid evidence of this shift. Since the Tang dynasty, the maritime Silk Road had been the major communication while the SSR was also developed. It is highly probable the two roads were mutually stimulating, as reflected by the rise of local states, cities, and towns along mainland Southeast Asia simultaneously. When the overland Silk Road declined after the collapse of the Mongol empire, the maritime Silk Road had already functioned as the major passage. On the other hand, the SSR should not be underestimated. During the Ming-Qing, Chinese overland expansion into Southeast Asia was a major state move, and the Chinese tributary system was considerably revealed here. Finally, when Europeans colonized Southeast Asia, the volume of trade in the SSR began a new era, a fact present-day Yunnan eagerly desires.

In conclusion, many influential cultural migrations were the result of the three roads together, silk, paper and papermaking, gunpowder, and Buddhism, to list a few.

The spread of Buddhism is the symbol of Sino-Indian communication. The diversity of Buddhist sutras in China sheds light on their various sources. It is commonly accepted that Buddhism came first through the overland Silk Road and then through the maritime Silk Road, while some scholars suggest either the maritime Silk Road or the SSR. Within all the arguments, the fact that all the three passages contributed to this cultural migration seems to be ignored. The existing Buddhist sutras in China are useful in terms of this issue. There are five kinds of Buddhist sutras that are waiting for exploration: sutras in Chinese; sutras in Tibetan or Mongolian; sutras found in Xinjiang (the West Regions) in ancient local languages; pattra sutras (beiyijing) by the Thai people in Yunnan; and sutras found in Tibet or Xinjiang in ancient Indian languages.[112]

Second, the diversity of religions in Yunnan is very illustrative, too. Theravada and Mahayana both exist, as well as Tibetan Buddhism, or Mijiao in Chinese. Theravada certainly was from Upper Burma, Mahayana was from China, and Mijiao in Yunnan was from Tibet. Furthermore, Nestorianism and Islam arrived in Yunnan from the north, as did as Confucianism and Daoism. It is amazing that so many religions are found in such a small area, a fact again demonstrating Yunnan’s multiple connections and geographical significance.

Finally, let us look at how Chinese and Indian monks came and went. Chinese monks took the Western Regions, or Tibet, or Yunnan, or the South China Sea to India. Therefore all the three passages were used. When Yi Jing visited East India, he found that a monastery had been built to honor over twenty Chinese monks in Yunnan. Indian monks also took a route through Upper Burma to visit China. During the Nanzhao reign, Buddhism was accepted and took the place of Daoism as the state religion. Many Indian monks reportedly contributed to this transformation. Eight kings of the Dali regime gave up the throne and spent the rest of their lives as monks. Around the transition of the thirteenth century, Zhikong, an Indian monk, arrived in Yunnan, visited China, and finally stayed in Korea.[113]

Paper and papermaking, as opposed to the eastward spread of Buddhism, will shed light on how the three passages jointly functioned.

Before the introduction of paper, Indians used Bhuja (Baetula Bhojpattr), or cotton clothing, or parchments, or tada-tala (Borassus flabelliformis) to write, which was observed by both Chinese and Arabic travelers. Because of the inconvenience of using these materials, monks more often taught their disciples by oral transmission. Most Buddhist scriptures taken to China by Chinese monks such as Xuanzang were on pattra, called beiyejing in Chinese, and these scriptures when translated into Chinese were written on paper.[114] In the mid-second century, paper had been used in the Western Regions. Written languages on those papers found include Chinese and non-Chinese, for example, Sanskrit. Based on Chinese textual documents, Ji Xianlin concludes that at least before the end of seventh century, paper had been used in India. But paper manufacturing had arrived there a little late.[115]

Ji’s discussion focuses on archaeological evidence found in the Western Regions and implies that paper and paper manufactoring spread to India by the Silk Road. However, Ji did not give an earlier limit of the spread of paper in India. Could it be earlier than the second half of the seventh century? And could paper spread to India through sea routes?

Huang Shengzhang points out that Xuanzang, who had extensive knowledge of India, never mentioned paper in his nineteen-year stay in India (626–645 CE). Therefore, paper must have spread into India in the mid-seventh century. During this period a new road connecting China and India was invented, the Tibet-Nepal (Tubo-Niboluo Dao), thanks to the official marriage between Nepal, Tibet, and Tang. The marriage of Bri-btsum, the Nepalese princess, with Sron btsan sgampo in 639 CE was followed by the arrival of the Wencheng princess two years later from the Tang court. Political marriages embody and enhance the increasing communications via Tibet. Two years later (643 CE), Li Yibiao and Wang Ce, Tang’s envoys, took this route, arriving in Magadha. Xuan Zao, a Chinese pilgrim, took this route in 665 CE, but was unable to return because the route was blocked by local tribes in Qinghai. In 674 CE Xuan Zao met Yijing in India. All the documents lead to the speculation that around 650 CE paper manufacturing spread into Tibet, which Huang Shengzhang believes, because in 649 CE Tang China approved Tibet’s quest for all kinds of smiths, which must have included papersmiths.[116]

But the paper issue has not been fully solved. Li Xiaocen noticed that several different styles of papermaking existed on the Indian subcontinent. One is popular in India-Nepal that is similar to that used in Tibet. Another is popular in Bengal that is similar to that used in inland China, that is, mulberry bark paper (sangpizhi). Li examines the paper along the SSR, finding that mulberry bark paper had dominated there. He also finds the similarity of paper manufacturing in Yunnan and in Bengal. Hence he argues that paper manufacturing in Bengal was spread along the SSR.

While Huang Shengzhang rules out the contribution of the maritime Silk Road to the spread of paper manufacturing, some Chinese documents refer to paper in Java, where paper must have passed if it took the maritime Silk Road. We cannot rule out the possibility of paper migration by sea. As a matter of fact, all the three roads probably played a role in the spread of paper.[117]

Gunpowder, in addition to paper and printing, was a major contribution of ancient Chinese civilization. The westward spread of these technologies to some extent was crucial for industrialization.[118] The Mongol conquest was credited for the presence of gunpowder in the Middle East, Europe, and northwestern India.[119] But what about Southeast Asia and South India? Scholars have speculated the overseas route and overland contacts through Burma, which have been partially illustrated by Sun Laichen’s work on the spread of gunpowder technology in the Ming period, were responsible.[120]

Ming China had experienced a “military revolution” that helped its expansion into Southeast Asia, at least in the beginning, thanks to its mastery of gunpowder. However, military campaigns and trades immediately spread the knowledge of firearms into “the Maw Shans, Lan Na, and Dai Viet from the end of the 14th century, over one hundred and twenty years before the arrival of European firearms in 1511 CE.”[121] Here again we have seen the cooperative contribution of the Silk Roads.

Finally, three cases enhance the impression of the networks constituted by the three passages. One is the bubonic plague, or Black Death, that decimated the population in Eurasia; another is Marco Polo; and the last is Zheng He. All are related to Yunnan. The disastrous influence of the Mongols was recorded throughout Eurasia, but the number of victims of the Mongols’ slaughters could not compare with that of the bubonic plague, which was probably brought by the Mongol cavalry from Upper Burma and Yunnan.[122] Millions of people died, and it tremendously changed the trajectory of world history.

Marco Polo was probably the only man who traveled all three passages. Arriving in the Mongol court by the Silk Road, he was dispatched to Burma, experienced the Southwest Silk Road, and finally took the maritime Silk Road back home.[123]

In the case of Zheng He, his ancestors were from Central Asia, and lived in Yunnan as a result of the Mongol conquest.[124] The Ming conquest then brought him to Beijing, where he became a favorite eunuch of Prince Yan, who later usurped the throne and dispatched the Treasure Fleet headed by Zheng He. Zheng’s family once more symbolizes the triangle connection among the three passages.

The above discussion to greater or lesser degrees may affirm the conclusion that there was at least a system of trade networks before the European domination, if not Andre Gunder Frank’s five-thousand-year world-system. And the contribution of the SSR to Eurasian communication can by no means be underestimated.


This paper has introduced the Southwest Silk Road by utilizing many Chinese sources and some non-Chinese studies.[125] I argue that the Southwest Silk Road had played an active and dynamic role for long-distance trade between China, Southeast Asia, and South Asia for over two thousand years. The trade along the road left significant marks on the regional societies and cross-regional interactions. Moreover, I argue that three Silk Roads had constituted a network (or, if not a network, at least they connected to each other), supplementing each other temporally and spatially, which would shed light on the understanding of the Sino-foreign /East-West communications and thus the whole Eurasian continent. Finally, in terms of Yunnan, I argue that Yunnan should be seen as part of Southeast Asia and the Indian Ocean until the “modern” period. The cowry money system in Yunnan has demonstrated that the successful incorporation of the Chinese could not be understood until Yunnan had been examined in a global perspective.

The SSR, illustrating Yunnan’s global contribution, has raised many interesting questions about Yunnan. The evidence that Yunnan had been culturally and economically inclined to Southeast Asia and the Indian Ocean seems convincing. But why was Yunnan unable to retain her independence, as Burma, Vietnam, or Thailand managed to do? How did Yunnan, part of the Indian Ocean trade world, become a part of China, a giant world-economy? Such questions indicate that only a world-history approach can fully explicate issues that have previously been examined totally within the confines of national histories.

This paper shows the significance of using a regional and global approaches to understanding the SSR and Yunnan history. The well-imagined and/or established national and regional boundaries within the academy to some degree have been barriers to a full, deep, and dynamic picture not only of our world, but also of frontier areas such as Yunnan. At this point, it is time to take seriously Frank’s appeal for world-system history.[126]


* This paper is a revision of my presentations in the World History Center Reunion (June 2002, Boston) and the Association for Asian Studies Annual Conference (March 2003, New York). I would like to express my gratitude to Professor Christina Gilmartin, Adam McKeown, Lu Ren, Liu Xinru, Sun Laichen, Patrick Manning, John Wills, Cheng Yinghong, Andre Gunder Frank, Parker James, and the anonymous reviewer for JWH.

1. Yunnan became a province of China only after the conquest of the Mongols in the mid-thirteenth century. William Skinner put Yunnan and Guizhou in his Yunnan-Guizhou macroregion of late imperial China. Historical Yunnan was much different from present Yunnan province. William Skinner, ed., The City in Late Imperial China (Stanford University Press, 1977); for the discussion of historical changes of Yunnan, see my forthcoming dissertation.

2. Chinese studies of this road are extensive, but few Western scholars have noticed this road. For a bibliography see Sun Laichen, “Ming-Southeast Asian Overland Interactions, 1368–1644” (Ph.D. diss., University of Michigan, 2000); and my forthcoming dissertation. I use the term “Southwest Silk Road” not because of silk export through the route, but because I want to emphasize that this ignored international trade networks and parallel it with the overland Silk Road and the maritime Silk Road.

3. For example, in his recent dissertation, Sun Laichen has accused the “maritime mentality” of previous scholars of Southeast Asia who “have overwhelmingly studied elements coming from China, India, and Europe via the sea.” See Sun Laichen 2000, pp. 7–8.

4. Fang Guoyu, “Yunnan yu Yindu Miandian zhi Gudai Jiaotong” (Ancient communication between Yunnan, Burma and India),” in Works of Fang Guoyu, vol. 4, ed. Lin Chaomin (Kunming: Yunnan Jiaoyu Chubanshe), pp. 338–369; Gu Chunfan (Kuo Tsung-fei), “A Brief History of the Trade Routes between Burma, Indochina and Yunnan,” T’ien Hsia Monthly 13, no. 1 (1941): 9–32.

5. Xia Guangnan, Zhongyinmiandao Jiaotongshi (History of the route among China, Burma, and India), (Beijing: Zhonghuashuju, 1948).

6. Joseph Needham, Science and Civilization in China, vol. 1 (Cambridge, UK: Cambridge University, 1954), p. 174.

7. D. P. Singhal, India and World Civilization, vol. 1(Michigan State University Press, 1969), pp. 294, 422.

8. Yu Yingshi, Trade and Expansion in Han China (University of California Press, 1967); Joseph Needham’s Science and Civilization in China (Cambridge University Press, 1974; see vol. 5, pp. 237–240) may be another example.

9. Probably we should call the Silk Road the North Silk Road, which would show the geographical locations of the three roads from the north to the south and from land to the sea.

10. Shen Xu, “Xi’nan Sichouzhilu Gailun” (On the Southwest Silk Road), in Southwest China Cultural Studies, ed. Yunnan Social Science Institute, vol. 1 (Kunming: Yunnan Minzu Press, 1996), p. 5; Sun Laichen 2000, pp. 3–4.

11. Sun Laichen 2000, p. 3.

12. Sun Laichen 2000.

13. Janice Stargardt, “Burma’s Economic and Diplomatic Relations with India and China from Medieval Sources,” Journal of Economic and Social History of the Orient 14 (1971): 38–62; Sun

14. The first character of the word “shen-du” should be pronounced in Chinese pinyin system as “yan” or “juan,” though in most cases it is represented as “shen.” It is supposedly a phonetic translation of “Indus” or “India.” Previously many Western scholars as well as Chinese missed it. Shiji (Records of the grand historian of China), juan 123 (translated from the Shih Chi of Ssu-ma Ch’ien by Burton Watson, Volume II: The Age of Emperor Wu 140–circa 100 B.C. [New York and London: Columbia University Press, 1961]).Laichen 2000.

15. Charles Higham, The Bronze Age of Southwest Asia (Cambridge University Press, 1999). The discovery of Sanxingdui has been a major contribution to and demonstration of the plural origin theory of the Chinese civilization raised in the1980s.

16. For historical communications between China and Vietnam, see, for instance, Chen Yulong, Lidai Zhongyue Jiaotongdaoli Kao (Studies of Sino-Vietnamese routes) (Kaifeng: He’nan Renming Chubanshe, 1987).

17. For this road, see Paul Pelliot, Jiaoguangyindu Yindu Liangdao kao (Deux itineraires de Chine en Inde à lapu du VIII e Siecle), Trans. Feng Chengjun (Beijing: Zhonghuashuju, 1955); Yan Gengwang, “Hanjin Shidai Dianyue Tongdao Kao” (A study of the communication between Yunnan and Vietnam during the Han and Chin dynasties), Journal of the Institute of Chinese Studies of the University of Hong Kong 8, no. 1 (1976): 24–38;”Tangdai Dianyue Tongdao Bian” (A study of the communication between Yunnan and Vietnam during the Tang dynasty), Journal of the Institute of Chinese Studies of the University of Hong Kong 8, no. 1 (1976): 39–50;and Fang Guoyu 2001, 4: 370–383.

18. For records of Yijing and Xuanzang, see Fang Guoyu 2001, 4: 338–369.

19. Though Yunnan zhi was compiled by Fan Chuo at that time, he relied heavily on earlier sources. For a full discussion of this work, see Fang Guoyu 2001, 2: 367–394. Yunnan-zhi is also called Manshu (Records of southern barbarians). Fan Chuo survived when Nanzhao conquered Annam, and recorded what he heard and read, though he did not travel to Yunnan.

20. For the discussion of this route, see Pelliot 1955; Yan Gengwang 1976 a–b; and Fang Guoyu 2001, 2: 657–684; 4: 338–369.

21. Walter Liebenthal questions why people took the southern winding way to Kalinga instead of traveling a couple of days by boat. The reason lies in the trade. Travelers, indeed merchants, were driven to seek profit, not the shortcut. Walter Liebenthal Walter, “The Ancient Burma Road—A Legend?” Journal of the Greater India Society 15, no. 1 (1956): 10.

22. John Deyell, “The China Connection: Problems of Silver Supply in Medieval Bengal,” in Money and the Market in India 1100–1700, ed. Sanjay Subrahmanyam (Delhi: Oxford University Press,1994), p. 128.

23. Nisar Ahmad, “Assam–Bengal Trade in the Medieval Period, a Numismatic Perspective,” Journal of the Economic and Social History of the Orient 23 (1996): 176–177.

24. Deyell, p. 128.

25. Zhao Lufu, Yunnanzhi Jiaoshi (Annotations to Yunnan Zhi) (Beijing: Zhongguo She-huikexue Chubanshe, 1985), pp. 86–87; For discussions, see Feng Hanyong, “Chuanzangxian shi Xi’nan Zuizao Guoji Tongdao Kao” (The Sichuan-Tibet route: The earliest international road in southwest China), Zhongguo Zangxue (China Tibetology), no. 5 (1988): 147–156;Lu Ren 1997, 102–104;and Shen Xu, “Lishi shang de Dianzang Jiaotong” (Communications between Yunnan and Tibet in history), in Southwest China Cultural Studies, ed. Yunnan Social Science Institute, vol. 3 (Kunming: Yunnan Minzu Press, 1998), pp. 100–128.

26. Pony and mule caravans constituted the major form of long-distance trade, though when the caravan emerged is unclear. Elephant caravans were popular, too.

27. Wang Ningsheng, “Gudai Yunnan de Yangmaye” (Horse husbandry in ancient Yunnan), Sixiang Zhanxian, no. 3 (1980): 34.

28. Marco Polo, The Travels of Marco Polo, ed. Manuel Komroff, ill. Withold Gordon (New York: Liveright Publishing Corp, 2002), p. 194.

29.SJ, juan 129; Han Shu (HS), juan 28.

30. HHS, juan 86; HYGZ, juan 4.

31. HYGZ, juan 4.

32. Wang Ningsheng 1980, p. 35.

33. Fang Guoyu, ed., Yunnan Shiliao Congkan (Series of Chinese Documents on Yunnan) (hereafter YNSLCK), vol. 2 (Kunming: Yunnandaxue Chubanshe, 1998), p. 85.

34.YNSLCK, 2: 67;XTS, p. 222.

35.YNSLCK, 2: 252–253.

36. Cf. Fang Guoyu 2001, 2: 431.

37. For the role of horses in Chinese history, see H. G. Creel, “The Role of Horse in Chinese History,” The American Historical Review 70, no. 3 (1965): 647–672.

38.San Guo Zhi (SGZ), juan 47.

39. Ibid., juan 49.

40. Cf. Lu Ren 1997, 50; Zi Zhi Tong Jian (ZZTJ), juan 249.

41. Many imperial scholars believe that the Yunnan-based Nanzhao kingdom played a crucial role in Tang China’s collapse.

42. Fang Guoyu 2001, 2: 395–405.

43. Creel 1965, 670; Song Shi (SS), 198: 4944–4946.

44.SS, 198.

45. Lin Wenxun, “Songdai Xi’nan Diqu de Shima yu Minzu Guanxi” (The horse trade and ethnic relations in southwest China of the Song dynasty), Sixiang Zhanxian, no. 2, (1989) 66–72.

46. Yang’s story is recorded in Xu zi zhi tong jian chang bian, juan 267. Yang once was housed in Dayunnanyi, a postal inn Fang Guoyu said was located in Yaozhou (Yaoan) where Yang read a trade route network written by someone describing the trade route east to Rongzhou (Yibin, Sichuan), west to Yandu (India), southeast to Jiaozhi (Annam), Northeast to Chengdu, north to the Daxueshan (Big Snow Mountains), and south to the sea. Furthermore, the miles were detailed. If true, it illustrates the trade along the SSR was very familiar to local people in Yunnan. Also see YNSLCK, 2: 244–247.

47.YNSLCK, 2: 250.

48. Fang Guoyu 2001, 2: 450.

49. For Yunnan horses in the Yuan-Ming-Qing period, see Wang Ningsheng 1980. Horses were still the major item in the tributes that local chieftains had paid to the Yuan, Ming, and Qing courts. Central states, on the other hand, continued to purchase horses from them too.

50. Cf. Wang Ningsheng 1980, p. 37.

51. Jiang Tianjian, Beisong Shima zhi Yanjiu (A Study on the Horse Trade in the Northern Song), (Taiwan: Guoli Bianyiguan, 1994), p. 309.

52. Jiang Tianjian 1994, p. 309.

53. Ranabir Chakravarti, “Early Medieval Bengal and the Trade in Horses: A Note,” Journal of Economic and Social History of the Orient, 42, no. 2 (1999): 194–211. The following is a summary of his study.

54. Marco Polo 2002, p. 210. Scholars hold different opinions of where Amu (Aniu) was. Fang Guoyu and Lin Chaomin argue that Amu was today’s Tionghai, southern Yunnan. For discussions, see Fang Guoyu and Lin Chaomin, Makeboluo Xingji Yunnan Shidi Congkao (Historical and geographical studies of Marco Polo’s travel in Yunnan), (Kunming: Yunnan Minzu Chubanshe, 1994). Marco Polo mentioned horses and oxen that had been famous local exports to Sichuan since the Qin-Han period.

55. Jin Zhengyao, “Wanshang Zhongyuan Qingtong de Kuangliao Laiyuan Yanjiu” (A study on the sources of bronzes of late Shang period in the Central Plains), in Kexueshi Lunji (Works on the history of sciences), ed. Fang Lizhi (Hefei: Zhongguo Keji Daxue Chubanshe, 1987), pp. 365–386; Li Xiaocen, “Shangzhou Zhongyuan Qingtongqi Kuangliao Laiyuan de zai Yanjiu” (A further study on the sources of ores for the Central Plains bronze vessels of the Shang-Zhou period), Ziran Kexueshi Yanjiu (Studies in the history of natural sciences)12, no. 3 (1993): 264–267.

56.HS, 24: 1178.

57.HYGZ, p. 4.

58.YNSLCK, 1: 205.

59.Yuan Shi (History of the Yuan dynasty), juan 94: 1596.

60. Song Yingxing, Tiangong Kaiwu, ann. Zhong Guangyan (Guangzhou: Guangdong Renmin Chubanshe, 1976), pp. 343–344.

61. Li Chunnong, ed., Yunnan Shiliao Xuanbian (Selected sources on Yunnan), (Kunming: Yunnan Minzu Chubanshe, 1998), p. 535.

62. Ibid.

63. Quan Hansheng, “Mingdai de Yinke yu Yinchanye” (China’s silver mining tax and silver output during the Ming dynasty), Xinya Shuyuan Xueshu Niankan, no. 9 (1967): 245–267.

64. Ibid.

65. James Lee, “State-Regulated Industry in Qing China, The Yunnan Mining Industry: A Regional Economic Cycle, 1700–1850,” paper presented at the 1984 Conference on Spatial and Temporal Trends and Cycles in Chinese Economic History 980–1980, sponsored by the ACLS and SSRC at Bellagio, Italy, 17–23 August 1984, p. 3; cf. Jacqueline Armijo-Hussein, “Sayyid’Ajall Shams Al-Din: A Muslim from Central Asia, Serving the Mongols in China and Bringing ‘Civilization’ to Yunnan” (unpublished diss., Harvard University, 1996), p. 179.

66. Li Chunnong 1998, 536–537.

67. Richard Von Glahn, Fountain of Fortune: Money and Monetary Policy in China, 1000–1700 (Berkeley: University of California Press, 1996), p. 1.

68. Ibid., p. 7.

69. Ibid., p. 8.

70. John Deyell 1994.

71. Ibid., p. 119.

72. Ibid., p. 133.

73. U. Thakur, “A Study in Barter and Exchange in Ancient India,” Journal of Economic and Social History of the Orient 15 (1972): 297–315.

74. U. Thakur (1972) concludes that barter constituted the main system of trade in India until the eleventh/twelfth centuries. In the case of Yunnan, barter lasted even longer as we can see in the trade between Song China and Dali. Copper cash did not dominate markets in Yunnan until the mid-seventeenth century when cowry money stopped. During the nineteenth and twentieth centuries, barter was still popular in Yunnan foreign trade.

75. For the spread of cowries in East Asia, see Namio Egami, “Migration of Cowrie-Shell Culture in East Asia,” Acta Asiatica 26 (1974): 1–52.

76. For the advantages of cowry money and cowries in slave trade, see Jan Hogendorn and Marion Johnson, The Shell Money of the Slave Trade (Cambridge: Cambridge University Press, 1986).

77. James Heimann, “Small Changes and Ballast: Cowrie Trade and Usage as an Example of Indian Ocean Economic History,” South Asia 3, no. 1 (1980): 56.

78. Zheng He’s Treasure Fleet had visited the Maldives several times. See Roderich Ptak, “The Maldive and Laccadive Islands (Liu-shan) in Ming Records,” Journal of the American Oriental Society 107, no. 4 (1987): 675–694.

79. Yang Shouchuan, “Beibi Yanjiu: Zhongyuan yu Yunnan yong Haibei zuo Huobi de Lishi (Studies on Cowry: Historical Survey on Cowry as Currency in the Central Plains and in Yunnan),” in Beibi Yanjiu (Studies on cowry currency), ed. Yang Shouchuan (Kunming: Yunnan University Press, 1997).

80. The late Spring-Autumn Period (770–476 BCE) or the Warring States period (475–222 BCE), the Qin-Han period, and the ninth century are the main opinions. See Yang Shouchuan, ed., Beibi Yanjiu (Studies on cowry currency), (Kunming: Yunnan University Press, 1997).

81. Michele Pirazzoli-t’Serstevens, “Cowry and Chinese Copper Cash as Prestige Goods in Dian,” in Southeast Asian Archaeology 1990: Proceedings of the Third Conference of the European Association of Southeast Asian Archaeologists, ed. Ian Glover (Center for SouthEast Asian Studies, University of Hull, 1992), p. 49.

82. Paul Pelliot, Notes on Marco Polo, 3 vols. (Paris, 1959), pp. 531–563.

83. Fang Guoyu, Yunnan Shiliao Mulu Gaishuo (A general introduction to sources on Yunnan), vols. 1–3 (Beijing: Zhonghuashuju, 1984), p. 2; Fang Guoyu 2001, 2: 373–375.

84. Hans Ulrich Vogel, “Cowry Trade and Its Role in the Economy of Yunnan: From the Ninth to the Mid-Seventh Century (Part I),” Journal of the Economic and Social History of the Orient 36, no. 3 (1993): 200.

85. Hans Ulrich Vogel, “Cowry Trade and Its Role in the Economy of Yunnan: From the Ninth to the Mid-Seventh Century (Part II),” Journal of the Economic and Social History of the Orient 36, no. 3 (1993): 319.

86. Ibid., pp. 246–247.

87. Jiang Yingliang, “Yunnan Yong Bei Kao” (On cowries in Yunnan), in Beibi Yanjiu (Studies on cowry currency), ed. Yang Shouchuan (Kunming: Yunnan University Press, 1997), pp. 81–93.

88. Yang Shouchuan, “Yunnan yong Haibei zuo Huobi de Lishi Kaocha” (Historical survey of cowry in Yunnan), in Beibi Yanjiu (Studies on cowry currency), ed. Yang Shouchuan (Kunming: Yunnan University Press, 1997), pp. 122–123.

89. Ibid., pp. 122–123.

90. Yang Shouchuan, “Lun Mingqingzhiji Yunnan ‘Feibeishiqian’de Yuanyin” (On the reason for the replacement of cowry by copper as currency in Yunnan in the Ming and Qing dynasties), in Beibi Yanjiu (Studies on cowry currency), ed. Yang Shouchuan (Kunming: Yunnan University Press, 1997), pp. 162–164.

91. Fang Guoyu, “Yunnan yong Bei zuo Huobi de Shidai ji Bei de Laiyuan” (The date and source of cowry as currency in Yunnan), in Beibi Yanjiu (Studies on cowry currency), ed. Yang Shouchuan (Kunming: Yunnan University Press, 1997), p. 54.

92. Ibid., p. 56.

93. Zhang Bincai, “Shiqi Shiji Yunnan Beibi Bengkui de Yuanyin” (Reasons for the collapse of Yunnan cowry currency in the seventeeth century), in Beibi Yanjiu (Studies on cowry currency), ed. Yang Shouchuan, (Kunming: Yunnan University Press, 1997), pp. 172–208.

94. Vogel 1993.

95. For Han Chinese immigration to Yunnan in the Ming, see James Lee, “The Legacy of Immigration in Southwest China, 1250–1850,” Annales de demographie historique (1982): 279–304;and Lu Ren, Bianqian yu Jiaorong—Mingdai Yunnan Hanzu Yimin Yanjiu (Transformations and incorporation: A study on the Han migration in Yunnan in the Ming dynasty), (Kunming: Yunnan Jiaoyu Chubanshe, 2001).

96. For the Indian Ocean trade structure, see James Heimann 1980, pp. 48–69.

97. For the definition of “world-economy” and how it differs from “world-system,” see Immanuel Wallerstein, The Modern World System, vols. 1–3 (New York: Academic Press, 1974, 1980, and 1988).

98. Ibid.

99. Janet L. Abu-Lughod, Before European Hegemony (New York and Oxford: Oxford University Press, 1989).

100. Andre Gunder Frank, “A Theoretical Introduction to Five Thousand Years of World System History,” Review 13, no. 2 (1990): 155–248; Andre Gunder Frank and Barry K. Gills, “The 5,000-Year World System,” in The World System: Five Hundred Years or Five Thousand? ed. Andre Gunder Frank and Barry K. Gills (London and New York: Routledge, 1996), pp. 3–58.

101. For world-system theories and debates, see Andre Gunder Frank and Barry K. Gills 1996; and articles in Review.

102. Ranabir Chakravarti 1999, pp. 194–211.

103. Abu-Lughod has noticed this fact, but she accredits it only to the overseas trade, a maritime mentality labeled by Sun Laichen. Janice Stargardt has presented an excellent illustration how overseas and overland trades shaped local powers in coastal and Upper Burma. See Stargardt 1971.

104. Heimann 1980, pp. 56–58.

105. Ibid., p. 48.

106. Ibid.

107. Abu-Lughod indeed puts Yunnan into her East Asian subsystem. I believe that during 1250–1350, Yunnan at least should be regarded as part of the overlapping area between Circuit VII (the Bengal Bay Region) and Circuit VIII (East Asia). See Abu-Lughod 1989, p. 34.

108. For this phrase, see William McNeill, The Rise of the West (Chicago & London: University of Chicago Press, 1991), p. xxii.

109. Shen Xu 1994, synopsis.

110. When regarding Chang’an as the beginning city, we probably ignore the “internal” trade network in China. But if examining the Silk Road, should we not forget the political map? Do we need the word “internal” or “international”?

111. For the Tibetan foreign routes and trades in the early twentieth century, for instance, Hatsuo Yamagata, ed., Xizang Tonglan (A comprehensive view of Tibet), (Taibei: Huawen Shuju, 1969); for the early medieval, see Christopher I. Beckwith, The Tibetan Empire in Central Asia (Princeton: Princeton University Press, 1987); and Zhang Yun, Silu Wenhua: Tubo Juan (The Silk Road cultures: Tibet), (Hangzhou: Zhejiang Renmin Chubanshe, 1995). For the route between Chang’an, Tibet, and Nepal, see Lu YaoGuang, ed., Tangbo Gudao Kaochaji (Records of the fieldwork on the Tang-Tibet ancient road), (Xi’an: Shanxi Luyou Chubanshe, 1989).

112. Jin Kemu, “Santan Bijiao Wehua” (The third talk on comparative cultures), in Wang Shuying, ed., Zhongyin Wenhua Jiaoliu yu Bijiao (Communication and comparison of the Indian and Chinese civilizations), (Beijing: Huaqiao Chubanshe, 1994), pp. 114–120.

113. For Zhikong, see YNSLCK, 3: 254–255; Qi Qingfu, “Zhikong Youdian Jian Zhengxushi Kao” (A study on Zhikong in Yunnan and the building of Zhengxu monastery), Yunnan Shehuikexue, no. 2 (1995): 88–94; Xiao Yaohui, “Zhonghan, Hanzhong Zhikong Yanjiu Xueshu Taolunhui Zhongshu” (A summary of Sino-Korea studies on Zhikong), Yunnan Shehuikexue, no. 4 (1998): 92–94.

114. Ji Xianlin, “Zhonguo Cansi Shuru Yindu Wenti de Chubu Yanjiu” (An elementary study of the spread of Chinese silk to India), Lishi Yanjiu (History study), no. 4 (1955): 51–94.

115. Ji Xianlin 1955.

116. Huang Shengzhang, “Guanyu Zhonguo Zhi he Zaozhifa Chuanru Yinbacidalu de Luxian” (The spread route of paper and papermaking into the subcontinent of India), Lishi Yanjiu (History study), no. 1 (1980): 113–132.

117. Li Xiaocen, “Zhongguo Zhi he Zaozhifa Chuanru Yinbacidalu de Luxian” (The spread route of paper and papermaking into the subcontinent of India), Lishi Yanjiu (History study), no. 2 (1992): 130–133.

118. For the invention of gunpowder and its spread to the West, see, for instance, Feng Hansheng, “Huoyao de Faxian jiqi Chuanbo” (The discovery and spread of gunpowder), in Feng Hansheng Lunzhu Jicui (Selected works of Feng Hansheng), (Beijing: Zhonhuashuju, 1984), pp. 225–74; “Yisilanjiaoguo wei Huoyao you Zhonguo Chuanru Ouzhou de Qiaoliang” (The Islamic world was the bridge of the spread of gunpowder from China to Europe), in Feng Hansheng Lunzhu Jicui (Selected works of Feng Hansheng), (Beijing: Zhonhuashuju, 1984), pp. 275–326; Needham 1986; and Iqtidar Alam Khan, “Coming of Gunpowder to the Islamic World and North India: Spot on the Role of the Mongols,” Journal of Asian History 30, no. 1 (1996): 27–45.

119. Sun Laichen 2000, p. 28.

120. Ibid., pp. 28–76.

121. Ibid., p. 75.

122. For this speculation, see William McNeill, Plagues and People (New York: Anchor Press, 1976), pp. 160–164.

123. For Marco Polo’s travel along the SSR, see Fang Guoyu and Lin Chaomin 1994.

124. Many Central Asians served in Yunnan, for example, Sayyid’Ajall Shams Al-Din, the first civilian governor of Yunnan in the Yuan dynasty. As a Muslim, he succeeded in consolidating the Mongol control over Yunnan, and won the support and respect of local peoples. As such, he has been regarded as a model both by the Muslims and by the Confucians. For his experience, see Armijo-Hussein 1996.

125. A strength of my work is that it reduces the Sinocentric approach in this field, although the exploration of non-Chinese sources in this paper is insufficient. Hence I am looking forward to active responses from Indian and Southeast Asian experts.

126. Andre Gunder Frank, “A Plea for World System History,” Journal of World History 2, no. 1 (1991): 1–28.



By: Bin Yang



Are you passionate about history?

Share that passion with the world. Join us: Contribute to History Cooperative