15 Multilingualism and Lingua Franca in the Ancient Chinese World

There is no linguistic evidence for the presence of any kind of multilingualism in China for any time prior to the political unification of the empire in 221 BCE. By the same token there is nothing in China that could be called a lingua franca at this early date.¹ To be sure, a complete absence of multilingualism would be a very unlikely circumstance in any society, past or present. The problem in the pre-imperial Chinese case is that, apart from a few brief and hopelessly uninformative anecdotes, there is virtually no documentary evidence that could give any clue to what kind of multilingual environment must have existed.² Even later, during the first three centuries of the empire, there is little more in the transmitted textual legacy than scattered passing reference to presumably non-Chinese(-speaking) groups of people.³ It is only with the advent of Buddhism in the early second century CE that we find concrete evidence for the use of known languages other than Chinese in China, but even here the linguistic context is highly circumscribed and does not allow any significant inferences about the transmission of knowledge, much less about any kind of (“old world”) globalization that might follow from a genuinely multilingual environment. In the absence of real evidence, in particular the absence of written texts in any language other than Chinese, there is little that the sinologist can do, except look with envy on the diverse textual and linguistic riches of India, the Ancient Near East and the classical Mediterranean that are readily available to his philological brethren working in those areas.

The historical period begins very late in China in comparison with Egypt and the Ancient Near East, by a measure of close to two millennia. The earliest known texts written in Chinese date from about 1200 BCE, and for the next thousand or more years, there is nothing known from the Chinese world written in any language other than Chinese. We might admire the pre-eminence of the sinitic cultural hegemony that this homogeneous Chinese literary and linguistic monolith bespeaks, but at the same time we lament the absence of any record of anyone speaking anything other than Chinese, and of how such people might have interacted with the dominant Chinese political, social and linguistic order. In the transmitted literature of the centuries just before the unification of the empire in 221 BCE we find only occasional, anecdotal hints of the existence of what may have been non-Chinese people speaking non-Chinese languages. The following well-known passage from the Mencius (Mengzi 孟子, late fourth century BCE), in which Mencius is reproaching a disciple who has expressed an interest in a heterodox southern school, is representative.

Mencius 3A.4

吾聞用夏變夷者, 未聞變於夷者也. [...] 南蠻鴃舌之人非先王之道. [...] 吾聞出於幽谷遷于喬木者, 未聞下喬木而入于幽谷者.⁴

I have heard of taking advantage of Chinese culture to transform the “backward,” but I have never yet heard of being transformed into the “backward.” [...] The shrike-tongued people of the southern backwaters repudiate the proper ways of the former kings. [...] I have heard of emerging out of a dark valley to perch in a stately tree; but I have never yet heard of descending from a stately tree to enter into a dark valley.

From the evidence of this passage we can see that for Mencius the distinction between “Chinese” and “non-Chinese” is entirely cultural, and the “shrike-tongued” people of the south that he disparagingly refers to could be speaking a dialect or form of Chinese that was considered rustic and inelegant, and therefore uncultured, rather than a non-Chinese language. We have no way of knowing which was the case, because the text is silent on ethnic and linguistic details. It seems to have been only the cultural distinction that was important to the Chinese political and social elite of the time, not the ethnic or linguistic distinction.

The nature of the evidence changes markedly about a hundred and fifty years later, during the Western Han dynasty (206 BCE–CE 25), but we still do not find any useful linguistic data whereby we can know anything in depth about the non-Chinese language environment of early China. To be sure, the Chinese histories from the first century BCE on include numerous accounts of groups of people with whom the Chinese came, directly or indirectly, into contact and whom we presume to be non-Chinese. These accounts often occur collected together in sections of the histories under the heading “Western Regions” (xi yu 西域). Some of the Chinese names of these people can be confidently identified with ethnonyms known from Classical western sources, chiefly Herodotus or Ptolemy, or from slightly later Sogdian or Khotanese sources, and on that basis we can infer some measure of a multilingual environment.⁵ The accounts are sometimes rich in what we would now call cultural and ethnographic information, but none of them includes enough linguistic data to allow any certain knowledge about the ethnic or linguistic identities of those groups whose Chinese names resist identification.⁶ These accounts reflect a period when the Chinese state was eager to extend its political, economic and martial sway into the far northwest, in the direction of Central Asia and of the lucrative trade routes around and through the Tarim Basin. As a consequence, the official written historical records of encounters with non-Chinese people on the periphery of the expanding Chinese domain tend to describe chiefly the regions of modern Gansu and Xinjiang provinces and areas stretching further west into Central Asia itself. One of the earliest and most important of such records is the richly detailed account of an envoy named Zhang Qian 張騫, who in about 130 BCE was sent by the Han emperor Wudi 武帝 on a mission to the northwest frontier to make contact with the so-called Greater Yuezhi 大月氏 people.⁷

The presumed purpose of Zhang Qian’s mission was to forge an alliance with the Yuezhi against the people known as the Xiongnu 匈奴. According to the Chinese sources, a few surrendered Xiongnu captives had informed the Han court that the Xiongnu had some years earlier defeated the Yuezhi and had made a drinking vessel out of the Yuezhi king’s skull. The Yuezhi then, having fled west out of the area of the Gansu corridor, sans their hapless and headless king, are reputed, not surprisingly, to have harbored great enmity toward the Xiongnu. The Chinese historical accounts thus portray the Xiongnu as the common adversary of the Chinese and the Yuezhi alike and suggest that when the Han court heard these reports, they saw an opportunity to take advantage of the social and political discord in the far northwest by allying with the Yuezhi against the more formidable Xiongnu. This is the ostensible background to the decision to send Zhang Qian to seek an alliance with the Yuezhi in order to mount a joint offensive against the Xiongnu.

Exactly why the Han court was so eager to involve itself in the contentious relations of non-Chinese people on the northwest frontier, hoping to join with the Yuezhi in an offensive against the Xiongnu, is never made very clear by the Chinese sources. The usual explanation starts from a presumption that the Xiongnu were horse-riding, marauding barbarian nomads, invading China from the murky depths of inner Asia in great aggressive hordes to plunder and pillage the northwestern fringes of the settled Chinese state. This is for the most part probably very much a dramatic fiction. There is no real evidence to tell us who the Xiongnu were or where they came from apart from these transmitted Chinese historical records of the late second century BCE. By this time it is already an established historical convention that the Xiongnu were uncultured, unwashed and unwelcome “barbarians” who showed up on the Chinese frontier from some remote, unknown origin in Siberia or Central Asia, but there is in fact no objective basis for this picture.

The real reason for the Han court’s distress at the unification of Xiongnu power on the frontier and the defeat of the Yuezhi was in all likelihood the consequent interruption of the smooth-running trade routes that had operated under cooperative Yuezhi control from Xinjiang, through the Gansu corridor, reaching to the center of Han civilization in the Wei River valley, modern Shaanxi province. With the defeat of the Yuezhi those trade routes had fallen into hostile Xiongnu hands, a hostility that was probably marked at least in part by a Xiongnu attempt to take advantage of their position as trade-route middlemen and commodities brokers, to put a modern label on it, to profit themselves at the expense of the elite Chinese consumers, who were chiefly affiliated with the Han ruling house and court. Zhang Qian was likely not sent to seek an alliance with the Yuezhi against the Xiongnu simply for military purposes; the more important motive was probably to try to outflank the Xiongnu in Gansu and to restore the lost trade routes through territory not under Xiongnu control.⁸

As it happened, the Han court was unsuccessful in its efforts to establish an alliance with the Yuezhi. The historical record says that on his way to make contact with the defeated Yuezhi people, Zhang Qian was captured by the Xiongnu, detained for a decade, given a Xiongnu wife and produced a number of Sino-Xiongnu offspring before he managed to escape and proceed in his quest to meet up with the Yuezhi. The Chinese histories are careful to mention that through all of this Zhang Qian never lost possession of his Han court credentials, inter alia reminding us of the importance of the Chinese / “barbarian” distinction. When Zhang Qian finally made contact with the Yuezhi, it was some two or three thousand li further west than he expected, in the area between the lower reaches of the Jaxartes and Oxus rivers. He discovered that, after their defeat at the hands of the Xiongnu some decades earlier and their migration to the west, the Yuezhi had conquered the Graeco-Indian state of Bactria in northern India and established their own court there. The Chinese historical sources refer to Bactria as Da-xia 大夏, the name by which it would have been known to Zhang Qian. The Shiji 123 account records this as follows:

大月氏在大宛西可二三千里. 居媯水北. 其南則大夏, 西則安息, 北則康居. [...] 故時彊. 輕匈奴. 及冒頓立攻破月氏. 至匈奴老上單于殺月氏王以其頭為飲器. 始月氏居敦煌祁連間. 及為匈奴所敗乃遠去. 過 宛西擊大夏而臣之. 遂都媯水北為王庭. 其餘小眾不能去者保南山羌. 號小月支.

The Great Yuezhi were located to the west of Da-yuan (*ddah-ʔwanʔ = Tocharia) perhaps two or three thousand li. They dwelt north of the Gui (Oxus) river. To their south was Da-xia (Bactria), to their west was An-xi (*ʔʔan-sək = Arsacids, i.e., the Parthians) and to their north was Kang-ju (*kkhaŋ-ka) [...] In former times they were strong and treated the Xiongnu dismissively. When Mao-dun was established [as the Xiongnu chieftain] he attacked and smashed the Yuezhi. Later the Xiongnu senior elder Chan-yu killed the Yuezhi king and made his head into a drinking vessel. Before that the Yuezhi had lived in the area between Dun-huang and Qi-lian. It was only when they came to suffer defeat at the hands of the Xiongnu that they moved far away. They passed to the west of [Da-]yuan and attacked Da-xia (Bactria), which state they then subjugated. They then located their royal court to the north of the Gui (Oxus) river. A small number of them remained behind and took refuge in the areas of Nan-shan and Qiang-zhong. These were called the Lesser Yue-zhi.

This part of the Chinese historical record is notable because the defeat of Bactria by the Yuezhi is the first event in human history known to be recorded in both Chinese and western sources. The Chinese account is found in the Shiji, chap. 123, as given here and the western in the extant geographical books of Strabo (11.8.2, first c. CE) and in Trogus (late first c. BCE, preserved in the third-century work of Justinus).⁹ To be sure, the actual historical facts are likely to be considerably more complicated and less clear-cut than the simple picture that the Shiji account implies, but there can be little doubt all the same that the Chinese text testifies to a historiographical awareness, however imperfect, in Han China of the vicissitudes of Greek rule in Bactria.¹⁰ As interesting as this may be for the historian, it still does not provide anything in the way of linguistic data to open a window on multilingualism or on the use of a lingua franca in China proper. Clearly it shows that the Chinese of the Han period were in extended contact with people who were surely speaking what were undoubtedly non-Chinese languages, but we have no way of knowing from the Chinese sources what those languages might have been or what linguistic circumstances prevailed among the people of the north and northwest frontiers.

The official Chinese histories, chiefly the Shiji and the Hanshu,¹¹ tell us a good deal about the customs and habits of the Xiongnu people, as well as of people from many of the other places mentioned in the various accounts of the far northwest region. The Xiongnu are described as having no fixed settlements, but rather moving about with their livestock in search of water and grasslands. In confrontations they are said to have no shame in running away when it is to their advantage to do so. They take whatever benefits they can from wherever they happen to be, and they know nothing of (Chinese) rituals or proprieties. They eat the flesh of their domestic animals, the young and strong among them getting the richest food and the elderly getting the leftovers. They have no surnames and therefore cannot heed the paramount (Chinese) proscription against marrying someone from the same clan. On the death of a father, a son will marry his step-mother, and on the death of a brother, the remaining brothers will take the deceased brother’s wives as their own. And, crucially for our purposes, we are told that they have no written language and that they therefore conclude contracts verbally.

Expressed thusly, the Chinese historiographical record bespeaks, from the traditional Chinese perspective, a considerable cultural disdain for the Xiongnu. The picture is drawn in a way that highlights the great differences in cultural conventions and standards between the Xiongnu and the Chinese, and it is presented exclusively from a Sino-centric point of view. As with the Mencius passage mentioned earlier, the contrast between Chinese and non-Chinese is seen entirely in cultural terms, not in any way that even hints at recognizing the linguistic differences as pertinent. Multilingualism was of no concern to the Chinese court or to the Chinese historians and apparently had no impact on the Chinese attitude toward their non-Chinese neighbors. A kind of multiculturalism instead was the focus, evident not as a genuinely objective interest in the cultural diversity of their known world, but as a strongly judgmental distinction between the cultured Chinese of the Wei and Yellow river homeland and the uncultured, presumably non-Chinese, people of the periphery.¹²

Apart from religious texts that accompanied the introduction of Buddhism from India in the second half of the Eastern Han period (roughly 100–200), it is another three centuries before we find anything in the Chinese record that purports to register real linguistic data of a non-Chinese people, and even here the actual non-Chinese language data are in a seventh century commentary, not in the primary text. Chapter 86 of the Hou Han shu 後漢書 (History of the Later [Eastern] Han Dynasty), compiled by Fan Ye 范曄 (395–451) in the first half of the fifth century, records the Chinese version of three “songs” from a place called Bailang 白狼, the area where the modern provinces of Sichuan and Gansu share a border, north of modern Chengdu. The passage says that because the “king” (wáng 王) of Bailang “ardently yearned for the civilizing influences of Han culture and was eager to show allegiance to the Han state” (mù huà guī yì 慕化歸義), he presented these three songs to the local Chinese prefectural official. The Li Hsien 李賢 (651–684) commentary to this passage says that in compiling his text Fan Ye relied on the Chinese language version of these songs found in the Dong guan (Han) ji 東觀 (漢) 記, a work compiled ca. 175, now extant only fragmentarily (the pertinent parts here are not extant), which included both the Chinese and the Bailang language versions of the songs. The commentary then proceeds to give the Bailang language version of the songs. The people and by implication the language are identified in the Hou Han shu text as Yi 夷, a non-specific ethnonym that in earlier periods referred to a presumably non-Chinese people located in the northeast, but here is used to refer to the Bailang people in the northwest. Modern linguistic analyses of the Bailang songs suggest a Trans-Himalayan / Tibeto-Burman language, though to be sure the exclusively lexical nature of the data, written in Chinese characters, with all of the phonetic imprecision that that entails, renders any linguistic identification more than a little speculative.¹³ While the commentary does, to be sure, purport to give real foreign language data and the Hou Han shu text itself makes passing reference to the difficulty of understanding Bailang words and of the unfamiliarity of the flora and fauna in the Bailang region, all the same the contextual focus is entirely on the reputed yearning of the Bailang king and his people to submit themselves to the glory of the Han. Any real linguistic interest is decidedly secondary.

About a century after the compilation of the Hou Han shu (but two centuries earlier than the actual Bailang lingusitic data) we find a record of a non-Chinese language in a Buddhist historical context that reflects the Central Asian linguistic world rather than the somewhat more isolated world of the Bailang. In the Gaoseng zhuan 高僧傳 “Biographies of Eminent Buddhist Monks” (compiled in the early sixth century by Hui Jiao 慧皎 [497–544]) account of the renowned Buddhist monk Fo-tu Deng 佛圖澄, we have a passage that is said to represent the Jie 羯 (Early Middle Chinese [hereafter EMC] kät) language.¹⁴ Fo-tu Deng is speaking to Shi Le 石勒, the non-Chinese military figure who in the early fourth century founded the Later Zhao state in north China:

Gaoseng zhuan chap. 9:

[佛圖] 澄曰

相輪鈴音云「秀支替戾岡, 僕谷劬禿當」. 此羯語也. 秀支軍也. 替戾岡出也. 僕谷劉曜胡位也. 劬禿當捉也. 此言軍出捉得曜也.

Fo-tu Deng 佛圖澄 said [to Shi Le]:

The sounds of the transmigration wheel hand-bells say to me “/suw-tɕi thεj-lεj-kang bəwk-kəwk guə-thəwk-tang/.” This is the Jie (EMC kät = Ket?) language.

秀支 /suw-tɕi/ means ‘army’

替戾岡 /thεj-lεj-kang/ means ‘to come forth’

僕谷 /bəwk-kəwk/ means ‘the “barbarian” seat of Liu Yao’

劬禿當 /guə-thəwk-tang/ means ‘to capture’

Thus, this line means “The army will come forth and capture Liu Yao.”

Fo-tu Deng’s comment is clearly intended to suggest that these sounds, emanating from a kind of Buddhist hand-bell and said to resemble the Jie language, predict a victory for Shi Le over his adversary, Liu Yao, and indeed Shi Le won just such a victory in 328 and executed Liu Yao in 329. Here again the passage is more interesting historically than it is useful linguistically. If the passage does indeed represent in some approximate way one or more sentences in a real (non-Chinese) language, and if the individual word-for-word glosses that the text gives are accurate, we can say at least that the language seems to have a “subject - object - verb” (S-O-V) word order, typical of the various Altaic languages that are known to have been spoken in these areas several centuries later. We might also notice, for what it’s worth, that both verb forms, as identified by the glosses, seem to end in /-ang/. Curious as these observations might be, they tell us next to nothing about what language this might really have been.

The ten-word foreign-language passage is usually referred to as the “Xiongnu couplet,” presuming, or at least implying that the language represented is Xiongnu.¹⁵ But the text says explicitly that the language is 羯 Jie. The usual explanation for this is to presume that the Jie were a linguistically, and by implication ethnically, distinct sub-group of the Xiongnu. The salient further implication is that the Xiongnu was a confederation of ethnically and linguistically diverse groups of non-Chinese people.¹⁶ More than fifty years ago E. G. Pulleyblank proposed that the name Jie < EMC kät, may represent the ethnonym Ket, the name of a people who speak a now nearly extinct Yeniseian language in the area of the Yenisei River in Siberia.¹⁷ Alexander Vovin has recently allowed for this possibility, though the linguistic data are sparse and therefore any conclusion must remain speculative.¹⁸ Early in the twentieth century, Shiratori Kurakichi tried to show that the language was an early form of Turkic, but this is no longer believed to be the case.¹⁹ Pulleyblank suggested that the ethnonym ‘Ket’ and the EMC word kät itself both seem to mean “stone,” and that this is not unrelated to the fact that the Chinese surname Shi 石 of Shi Le, the person to whom Fo-tu Deng confides this ostensible Ket language prediction and who in the course of events defeats Liu Yao, also means, in Chinese, “stone.”²⁰ Moreover, Shi Le’s biography in Jinshu 晉書 “History of the Jin State,” (chap. 104) says that his ancestors came from a place called Qiang-qu 羌渠, EMC khang-ga, very likely a variant form of the name Kang-ju 康居, OC *kk^haŋ-ka, that occurs frequently in earlier, Han period, texts. Kang-ju is the state that included the largest part of Sogdiana in antiquity.²¹ Somewhat to the northeast was the area called in Sogdian Čač, later called by the Turkic name Tashkent “Petrapolis” or “Stone City” and known in later Chinese texts as Shi Guo 石國 “stone country.” People who came to China from Kang-ju often took the surname Shi 石 “stone,” if not Kang 康.²²

How these data fit together to reveal unambiguously the linguistic identity of Shi Le is difficult to say. The historical record says that he was a Xiongnu, the language of the Gaoseng zhuan and Jin shu passages suggest, tenuously at best, that he was Ket, that is, perhaps Yeniseian, or at least that he was familiar with that language, and the name of his ancestral home seems to be associated closely with Sogdians. Historically, Shi Le was a subordinate to the Xiongnu confederation that sacked the Chinese capital of Luoyang in 311 under the leadership of Liu Cong 劉聰. This is the sacking of Luoyang 洛陽 to which the second of the Sogdian ancient letters is usually thought to refer.²³ It is said to have been carried out by people whom the writer of the letter called xwn, a term that is now fairly universally identified with the name ‘Hun.’²⁴ The city of Luoyang is called in the letter srγ, that is, saraγ, a Sogdian transcription of 洛 luo < OC *s-rak-, the first syllable of the name Luoyang 洛陽.²⁵ Shi Le himself was the leader of the attack a year later on the city of 鄴 Yeh, the so-called second capital of the Chinese Jin state.²⁶ The Chinese historical record says that Shi Le was a Xiongnu; the Sogdian letter calls him by implication a Hun. But the linguistic data of the couplet as well as the evidence of Shi Le’s own name suggest either a Yeniseian or perhaps even a Sogdian identity (see footnote 22). Whatever his linguistic and ethnic identity, he was in the eyes of the Chinese historiographers a “barbarian.” Apart from this exceptional short account in the Gaoseng zhuan, the texts are not concerned in the least with what language he spoke. But they take considerable care to describe him as a descendant of a Xiongnu clan, in particular a non-noble clan, and to record the history of his rise to power and his complete rejection of Chinese political and cultural values.

The moral of this story, as far as multilingualism in early China is concerned, is that names such as Xiongnu in the Chinese written historical record, and perhaps xwn in the Sogdian letters as well, seem to have been as much culturally grounded, impressionistic epithets as they were real ethnic or linguistic designations, if not more.²⁷ In an early provisional program for the meeting out of which the present paper arose, the first session was called “General reflections on the psychology of multilingualism.” In the final version of the program that became “General reflections on multilingualism.” But thinking about multilingualism in the Chinese case from a “psychology of” perspective can be suggestive. “Multilingualism” is of no consequence; “multiculturalism” is what counts, and the multi- part reduces itself to a pairing of unequal parts, a distinction of the “Greek / barbarian” kind that the late Malcolm Hyman mentioned in a 2006 paper, reminding us of Plato’s famous observation that the dichotomy “Greek / barbarian” leads to the erroneous belief that there is a natural class of barbarians.²⁸ Almost the same thing can be said, I think, mutatis mutandis for the dichotomy “Chinese / Xiongnu,” at least in a sense psychologically, that is, in respect of the cultural connotation that the term “Xiongnu” has when it is used in early Chinese historical accounts. For the Chinese, cultural identity is what mattered; linguistic identity was of little consequence.

The same near silence that we find in early written Chinese records about non-Chinese languages extends also to non-Chinese orthographies. To be sure, there is no clear evidence of writing of any kind other than Chinese in China prior to the advent of Buddhism in the early second century CE, so there may well have been nothing for the Chinese to have noticed in this regard at such an early period. In fact the historical records describing those various frontier groups with whom the Han court had direct or indirect contact often mention explicitly the absence of any kind of writing among those people. In view of the generally high regard in which the Chinese traditionally hold the Chinese script as a cultural achievement of the first order, such explicit observations about the absence of writing among the non-Chinese “others” are likely a further reflection of the perceived cultural gap between China and the peoples of the periphery.²⁹

Accompanying the arrival of Buddhist missionaries and the establishment of Buddhist communities in the major cities of Han China by the second century CE, there were, of course, Buddhist texts. These were predominantly, perhaps exclusively, written in the vernacular Northwest Indic dialect known as Gāndhārī, using the Kharosthī script, an orthography entirely different from and unrelated to Chinese.³⁰ In the first few centuries CE, owing to the formidable and enduring impact of Buddhism in China, in particular the demand for Chinese translations of Buddhist texts, which inevitably incorporated Chinese transcriptions of a myriad of specialized Buddhist names and technical terms, more than a few representatives of the Chinese learned world must have become intimately familiar with Indic abugida-type orthography, even if motivated by missionary zeal rather than scholarly interest (the two do not always go together).³¹ The Indic script in its Buddhist context is therefore the exception to the enduring Chinese disinterest in foreign writing systems, which seems generally to have prevailed on a par with the comparable disinterest, apparent already several centuries earlier, in foreign languages.

Thanks to the predominance of Sogdian traders and merchants across Central Asia and in China, use of the Sogdian language is likely to have been widespread already for several centuries by this time, but nothing in the Chinese historical record attests explicitly to the presence of the Sogdian script having come before Chinese eyes prior to the sixth century.³² Extant epigraphic evidence shows that by the later half of the sixth century the Chinese were in direct contact with the Sogdian language and script. The tomb of a Sogdian merchant (Figure 15.1) who died in CE 579, found within the suburbs of the modern city of Xi’an 西安 (= pre-modern Chang’an 長安), preserves a large portion of a stone lintel with a bilingual Sogdian-Chinese text (Figures 15.2 and 15.3). This is the earliest Sogdian-Chinese bilingual document known. The Chinese portion of the lintel inscription refers to the entombed person as shi jun 史君 ‘Master Shi’ and identifies him as the sabao 薩保 of Liang 涼 Prefecture of the Da Zhou 大周 “Great Zhou” state.³³ The Sogdian portion gives his name as Wirkak.³⁴

Remarkably, the Sogdian script here is clearly written vertically, though Sun Fuxi points out that, all the same, “it must be read horizontally.”³⁵ Sun presumably means that in effect one must rotate the text ninety degrees for “normal” reading purposes. Nevertheless, the epigraphic appearance of the Sodgian script here would seem to represent the nascent point of an orthographic development that is usually not thought to have occurred prior to the mid-Tang, some two centuries later, at the time of the An Lushan rebellion, and that prefigures by several centuries evidence of the script’s modification and eventual use for the vertical writing of Uighur, Mongolian and Manchu. The Chinese part of this bilingual inscription, though legible enough (where not damaged or missing), is said to have likely been written by a Sogdian “who was not very familiar with [...] Chinese.”³⁶ This would suggest that the Sogdians had some reason to attend to the Chinese language and script, learning it well enough and holding it in high enough regard to include a Chinese component as a part of the inscription in a Sogdian tomb. There is no comparable evidence that the Chinese reciprocated that interest or regard toward the Sogdian language or script.

The apparent disinterest on the part of the Chinese toward the languages, and especially the scripts, of the foreign people with whom they interacted in the mediaeval and early modern periods seems generally to have followed unchanged from that same attitude of the earlier periods. The major difference between these later periods and earlier ones is that by the end of the first millennium the non-Chinese states of the north had developed their own writing systems, whereas those from earlier periods did not for the most part have writing of any kind. By the time of the Tangut (Xi Xia 西夏), Khitan (Liao 遼) and Jurchen (Jin 金) states of the far north in the tenth, eleventh and twelfth centuries, and especially with the Mongols in the thirteenth century, the major non-Chinese states had acquired writing systems. The Tangut (linguistically Tibeto-Burman), the Khitans (linguistically Mongol) and the Jurchen (linguistically Tungusic) all were states dominated by non-Chinese speaking (as the indigenous language) ruling houses.³⁷ By the time that they gained political control over a part of north China all three had devised “siniform” writing systems, based impressionistically on Chinese (Figures 15.4, 15.5 and 15.6). None of these writing systems actually operated or was structured like Chinese writing, and none is even now completely deciphered, though both Tangut and Jurchen are reasonably well understood. The Liao shi 遼史 and the Jin shi 金史, the official Chinese histories of the Liao and Jin states respectively, both conclude with short treatises (in Chinese) on their respective languages, mostly consisting of lists of administrative and cultural terms or material goods in the Khitan, resp. Jurchen, languages transcribed in Chinese characters. Both of these histories were compiled in the middle of the fourteenth century, well after the fall of the states in question, by the history bureau of the succeeding Yuan / Mongol dynasty.³⁸ In other words, only when the Mongols had achieved dynastic supremacy over the whole of China was any interest shown in these non-Chinese languages, and even here there is next to nothing said about their scripts.

With the success of the Yuan / Mongol dynasty itself the picture changes once again. The Yuan is one of the best-documented multilingual states in Chinese history. Apart from the culturally dominant Chinese language, Mongol was the principal language of the ruling house and people. Beyond those two, there were significant populations of Tibetan and Turkic speakers, as well as descendants of the already polyglot Tangut, Khitan and Jurchen states mentioned above. Recognizing the multilingual nature of his empire and the difficulty of using any of the already established writing systems (chiefly Chinese, Uighur and Tibetan; the siniform scripts were marginal at best) for administrative purposes that served all of these different linguistic groups, Qubilai Qagan decreed that a new writing system be created that would accommodate all of the languages of his empire equally effectively. We read in the Yuan shi 元史, the Chinese offical dynastic history of the Yuan / Mongol dynasty, the following mandate:

至元六年詔頒行於天下. 詔曰朕惟字以書言. 言以紀事. 此古今之通制. 我國家肇基朔方. 俗尚簡古. 未遑制作. 凡施用文字因用漢楷及畏吾字以達本朝之言. 考諸遼金以及遐方諸國. 例各有字. 今文治寖興而字書有闕. 於一代制度寔為未備. 故特命國師八思巴創為 蒙古新字. 譯寫一切文字. 期於順言達事而已. 自今以往凡有璽書頒降者並用蒙古新字. 仍各以其國字副之.

In the sixth year of the Ultimate Origin reign period (of Qubilai Qagan, = CE 1269) an imperial edict was distributed throughout the Subcelestial Realm. The edict said:

We aver that —

By means of characters we write words and by means of words we record matters. This has been the ubiquitous practice from antiquity to the present. Our State finds its first foundation in the Far North. Our customs remain still simple and old-fashioned. We have never had the leisure to devise or create [anything new.] In general when we made use of writing we relied on using the “square-shaped” characters of the Han or the script of the Uighurs in order to express the statements of the Present Court. If we look at this in connection with the Liao (Khitan) state, or the Jin (Jurchen) state, or any of the still remoter states, the fact is that each has its own script. For us, literacy and administrative order both have gradually come to be fully developed, yet as far as having documents written in our own script, here we are still wanting. For the practices and institutions of our particular era, this is the thing that has still not been made available.

Therefore–

We especially command the State Preceptor, [the] ‘Phags-pa [Lama], to devise a new Mongol Script, to accommodate the transcription of all other writing systems [of the empire]. We look simply to reflect our language fluently and register our affairs effectively. From this time on in general all documents bearing the imperial seal that are handed down for distribution will use this new Mongol Script, at the same time maintaining the script of the state in question alongside as a supplement to it.³⁹

Specifically, the assignment was given to a Tibetan Lama, usually referred to as the ‘Phags-pa Lama, a title that means ‘the glorious Lama,’ who was named by Qubilai Qagan to serve as the guo shi 國史 “State Preceptor” at the Mongol court. The script that he created, modeled generally on Tibetan, is known as the ‘Phags-pa script.⁴⁰ The script was to serve the dual proposes of (i) providing the Mongol Empire with its own writing system and at the same time (ii) providing a way to write with a single script all of the languages of the empire. In order to satisfy both of these demands, the script actually was devised with language-specific sets of secondary orthographic distinctions to accommodate the different languages that it was to be used to write. Each language had, in other words, “its own transcriptional conventions.”⁴¹

Whatever ancillary influences may have been in play, the ‘Phags-pa script is clearly based chiefly on Classical Tibetan, a horizontally written abugida-type of writing with secondary superscript and subscript letters arranged vertically relative to the primary letter. The remarkable feature of the ‘Phags-pa script from the perspective of the development of writing systems is that it is a vertically written script in which the super- and sub-script secondary graphs of the Tibetan abugida have become nearly full-fledged “letters” of an alphabet-type script. In other words, based on a script that was graphically two-dimensional, incorporating both a primary horizontal and a secondary vertical alignment of its graphs, the ‘Phags-pa Lama devised a one-dimensional script; i.e., a script with only a vertical alignment of graphs, in effect largely eradicating the relative status difference between the primary and secondary letters. Even so, the transformation into an alphabet was not quite complete; the ‘Phags-pa script retains the abugida feature whereby a single consonant written without any vowel sign at all is inherently to be read as a syllabic consonant + /a/ vowel, and the secondary letters are attached by graphic ligatures to their associated primary graphs, even though written vertically in alignment with them.

The ‘Phags-pa script was intended to be widely used throughout the Mongol empire for all kinds of administrative, literary, ceremonial and practical purposes, but that intention seems not to have been entirely realized. The script is known in large part from extant monumental and epigraphic uses from the Mongol period (Figure 15.7).

Most recently the ‘Phags-pa script was included in a ceremonial silk banner presented to the Max Planck Institute for the History of Science (Berlin) by the research group of the Mongolian Academy of Sciences (Ulanbator) on the occasion of their visit to Berlin in March, 2011 (Figure 15.8). This is a nice example of the continuing use of a very old script in a formal, classical context.

With the fall of the Mongol dynasty political control over China came into the hands of the indigenous Chinese ruling house of Zhu 朱, founders of the Ming dynasty, and once again interest in foreign languages and foreign writing systems waned. To be sure, the Ming court established a “translation bureau,” known as the “Si yi guan” 四夷館 “Bureau for Matters of the Fourfold Foreign Peoples” (later the nearly homophonous “Si yi guan” 四譯館 “Bureau for Translation of the Fourfold Peoples”) that was responsible for translation and interpretation services in diplomatic dealings with the non-Chinese people of the realm. But this was an entirely practical matter and reveals no evidence of any intrinsic interest in either the languages or writing systems of any of these foreign people.⁴²

Only with the demise of the Ming and the founding of the non-Chinese Manchu Qing dynasty did a genuine concern with the multilingual nature of the Chinese state again emerge. And once more it was spurred by the establishment of a non-Chinese political state. The Manchu court saw early on the need for multilingual resources. In 1673, less than thirty years after the founding of the Qing dynasty, the Kangxi Emperor Shengzu ordered the compilation of a Manchu lexicon that would take fully into account the needs of Chinese translators and interpreters. This lexicon, which was not completed until 1708, was called the Yu zhi Qing wen jian 御製清文鑑 (Imperially Commissioned Manchu-Chinese Dictionary), but was in fact a monolingual Manchu language lexicon arranged by topic. Two years later work was begun again under court aegis on a bilingual Manchu-Mongolian dictionary, and then by 1780 a trilingual dictionary was completed, now including Chinese. Building on these lexicographical successes, two more dictionaries were compiled, one that added Tibetan, and finally one that added Uighur, resulting in the Yu zhi Wu ti Qing wen jian 御製五體清文鑑 (Imperially Commissioned Qing Pentaglot Dictionary) of the eighteenth century⁴³ (see Figure 15.9.)

One of the most remarkable of the linguistic and orthographic features of this dictionary is the fact that for every Tibetan and Uighur word, that is, those parts of the dictionary that are written in a script not immediately familiar in a Manchu / Mongol / Chinese context, the entry gives the pronunciation in a Manchu transcription. In the case of Tibetan, an orthography notoriously conservative relative to the pronunciation of the language in the eighteenth century (and later), the dictionary actually gives both a letter-for-letter Manchu transliteration and a representation in Manchu script of the actual contemporaneous pronunciation of the Tibetan word (see Figure 15.10).

The top item is the Manchu entry jakūn “eight” and the bottom is the Chinese. The second from the top is the Tibetan word for ‘eight,’ viz., brgyad, followed by two smaller size Manchu entries. The first of these, (a) in Figure 15.10, is brgyad, written in the standard Manchu script, representing the Tibetan orthography letter-for-letter, and the second, (b) in the illustration, is jad, i.e., /džad/, reproducing the actual eighteenth-century pronunciation of the Tibetan word for “eight” as closely as possible in the Manchu script. This same distinction between an orthographically accurate Manchu transliteration and an actual pronunciation given in Manchu script is maintained for every Tibetan word in the dictionary.⁴⁴ This practice reveals a surprisingly keen awareness of the relation (and non-commensurability) between language and script, something not directly in evidence in any comparable Chinese materials.

Wolfgang Behr, writing about the history of linguistic and translation concerns in pre-Qing China, concluded his survey by saying:

Looking at the development of translation and interpretation as reflected in non-Buddhist Ancient and Mediaeval Chinese texts, the most striking observation to be made is the enormous paucity of available evidence and references. It is painfully obvious that throughout most of China’s pre-Qing history, there was a deeply engrained lack of interest in foreign languages [...]. (Behr 2010, 201)

Precisely the same thing can be said about the Chinese interest in foreign writing systems. Only when the state was in the hands of non-Chinese rulers do we find any real attention paid to those scripts that reached China’s northern and western borders from across Central Asia, scripts that had evolved and were adapted in a multitude of ways to serve the goal of writing a multitude of languages, but that still, when carefully analyzed, reflect their shared Aramaic origin. None of these scripts ever rose to the level of orthographic predominance vis-à-vis Chinese writing, even in the powerful Mongol or Manchu periods. For as long a period as the historical record in China testifies to, there was no lingua franca, there was only a lingua sinica with its highly prized cultural treasure of grammata serica.

Acknowledgements

A modest version of this paper was presented at the “Crossing Boundaries: Multilingualism, Lingua Franca and Lingua Sacra” TOPOI conference at the Max Planck Institute for the History of Science in Berlin in November 2010, where I benefitted greatly from comments and suggestions offered by the workshop participants and attendees. In preparing this fuller version I have been fortunate to have received very helpful advice, suggestions and corrections from Wolfgang Behr, Ian Chapman, Mårten Söderblom Saarela, Steve Wadley and the late Jerry Norman. For all of this I am very grateful; what remains wrong or confused (or both) is of course my own responsibility.

References

Footnotes