Hostname: page-component-78c5997874-dh8gc Total loading time: 0 Render date: 2024-11-10T08:21:49.115Z Has data issue: true hasContentIssue false
  • English
  • Français

Chinese primers along the north-western frontier

Published online by Cambridge University Press:  19 September 2024

Imre Galambos
Imre Galambos*
Affiliation:
School of Literature, Zhejiang University, Hangzhou, P. R. China
*
Email: iig21@cam.ac.uk
Rights & Permissions [Opens in a new window]

Abstract

By the second half of the first millennium CE, a substantial body of texts written in Literary Sinitic developed into a shared repertoire of writings throughout Central and East Asia. In no small part, this was the result of the spread of Buddhism, which, in many regions, was adopted in its Sinitic form and relied on Chinese versions of the scriptures. As part of the means to cope with Sinitic Buddhist texts, most states also adopted a range of auxiliary texts, including primers and dictionaries. While modern scholarship has directed substantial attention at the participation of Japan, Korea, and Vietnam in the world of Sinitic texts, it has placed much less emphasis on the western and north-western regions of what is now China. This article attempts to redress the imbalance and demonstrate not only that, at one point, Inner Asian states actively participated in the Sinographic world, but also that they did this through a similar process of adaptation to that documented elsewhere. Primers played a special role in this, in that they had a more immediate connection with the core competency of reading and writing.

Keywords

Chinese primerstranslationOld UyghurTibetanTangut
Type
Article
Information
Journal of the Royal Asiatic Society , First View , pp. 1 - 18
Check for updates
Copyright
Copyright © The Author(s), 2024. Published by Cambridge University Press on behalf of The Royal Asiatic Society

By the second half of the first millennium CE, a substantial body of texts written in Literary Sinitic developed into a shared repertoire of writings throughout Central and East Asia. In no small part, this was the result of the spread of Buddhism, which, in many regions, was adopted in its Sinitic form and relied on Chinese versions of the scriptures. In some cases, the reception of Buddhist literature occurred in parallel with the adoption of the Chinese script, while, in others, a new script was invented for the purpose of writing down the scriptures in the local language. As part of the means to cope with Sinitic Buddhist texts, most states also adopted a range of auxiliary texts, including primers and dictionaries. In regions such as Korea, Vietnam, and Japan, Chinese primers initially remained untranslated, just as was the case with Buddhist literature. But states such as the Tangut kingdom developed their own script and translated Chinese primers into the local language, as a result of which localised versions of those were somewhat paradoxically used for teaching students to read and write in another language. This process of adapting primers to a new linguistic environment happened repeatedly in geographically distant regions. When we look at these cases together, the process seems to have been an inevitable step in the development of vernacular written cultures.

Even though the adaptation and translation of Sinitic texts continued for over a millennium, this article is concerned with the period prior to the Mongol conquest of Northern China—that is, up to about the mid-thirteenth century.Footnote 1 This was a formative period in the development of local scripts in Central Asia and the translation of Sinitic texts into vernacular languages. With regard to the geographical framework of this phenomenon, while modern scholarship has directed substantial attention at the participation of Japan, Korea, and Vietnam in the world of Sinitic texts,Footnote 2 it has placed much less emphasis on the western and north-western regions of what is now China. Presumably, this is largely due the lack of modern successor states that would have claimed the politico-cultural legacy of these regions as their own.Footnote 3 This article attempts to redress the imbalance and demonstrate not only that, at one point, Inner Asian states actively participated in the Sinographic world, but also that they did this through a similar process of adaptation to that documented in Japan, Korea, and Vietnam. Primers played a special role in this, in that they had a more immediate connection with the core competency of reading and writing.

Chinese primers

Primers are part of the technology of writing. They are tools that facilitate the acquisition of literacy and impart basic cultural information. Accordingly, they feature prominently in the earliest narratives of the history of writing in China. The famous account of the invention of writing in the ‘Yiwenzhi’ 藝文志 (‘Treatise on literature’) chapter of the Hanshu 漢書 (Book of the Han) from the end of the first century CE describes the compilation of three new primers for the sake of teaching the reformed Qin 秦 script following the Qin unification (221 BCE).Footnote 4 Of the primers from the Qin and Han periods, archaeological discoveries have yielded fragments of the Jijiu pian 急就篇 (Quick Mastery Chapters) and the Cangjie pian 倉頡篇 (Cangjie's Chapters).Footnote 5 Most of the early fragments are on slips of wood or bamboo but there are also several examples on more durable writing support, such as stone or terracotta.Footnote 6 Even though these two primers were designed for children, the acquisition of literacy skills in general was seen as the initial step towards one day becoming a professional scribe.Footnote 7

Moving forward to the time of paper manuscripts, we have a considerable number of surviving manuscripts and fragments, evidencing the consistent popularity of Chinese primers in scribal cultures throughout East and Central Asia. Perhaps because of the favourable conditions of the desert climate for the preservation of paper, most paper manuscripts, including primers, survived in Western China. Among the early examples are fragments of the Jijiu pian from Loulan 樓蘭 (British Library Or.8212/480 through Or.8212/484), possibly from the fourth century CE. These disjointed fragments belong to the same sheet of paper and contain the beginning portion of the text. One side has the text in parallel columns, in both regular and cursive scripts, possibly with the aim of serving as a template for copying, whereas the other side presents the same text in a less practised regular hand of a student. There are also fragments of the same text among the Turfan manuscripts—some pre-dating the Tang Dynasty (618–907).Footnote 8

Medieval manuscripts reveal that, from the sixth century onward, the most popular primer was the Qianziwen 千字文 (Thousand Character Text)—a text traditionally attributed to the scholar Zhou Xingsi 周興嗣 (d. 521).Footnote 9 The authorship and dating of this work is not unproblematic but the Turfan and Dunhuang manuscripts attest to its popularity. They also commonly include the name of Zhou Xingsi at the beginning of the text, showing that he was considered the author. There are at least a dozen copies from Turfan and almost three times as many from the Dunhuang library cave.Footnote 10 The earliest dated Dunhuang copy is Pelliot chinois 3561, in which the colophon says: ‘Record made by Jiang Shanjin of having finished tracing this copy in the seventh month of the fifteenth year of the Zhenguan era (641)’ 貞觀十五年七月臨出此本蔣善進記.Footnote 11 In contrast to other colophons, in which the verb for copying texts (e.g. Buddhist sutras) is xie 寫 (‘to copy’), this colophon uses the word lin 臨, which signifies ‘tracing’—that is, making a faithful reproduction. The implication is that whoever copied the text intended to replicate not only the textual content of the source manuscript, but also its layout and calligraphy. The manuscript is written in a skilled hand and the paper is also of good quality, suggesting that this was not a personal copy made for the sake of practice, but an exemplar of some import. Accordingly, the function of this copy—and possibly other copies of the Qianziwen in this period—was to practise calligraphy, rather than teaching schoolchildren to read and write. This, of course, fits well with the legend according to which Zhou Xingsi composed the text from individual characters of Wang Xizhi's 王羲之 (303–361) calligraphy that had been copied onto individual sheets of paper.Footnote 12

The Dunhuang copies of the Qianziwen from the ninth and tenth centuries are different and most of them were copied by students as writing exercises.Footnote 13 One such example is scroll Or.8210/S.3835, the recto of which has the primers Taigong jiajiao 太公家教 (Family Instructions of the Grand Duke), the Qianziwen, and the Bainiaoming 百鳥名 (Names of the Hundred Birds), copied in succession. A colophon at the end states that the texts were copied in the twelfth month of the gengyin 庚寅 year (possibly 930) by a certain Suo Buzi 索不子.Footnote 14 Similar (sometimes unfinished) copies of primers are not unusual among the manuscripts; in some cases, we only have a couple of lines from the beginning of the text or just the title. The examples, regardless of how fragmentary they are, provide evidence of the popularity of the Qianziwen in Dunhuang during the ninth to tenth centuries, when the Gansu Corridor was no longer part of Tang China and its successor states.

The Qianziwen had diverse functions and it would be an oversimplification to categorise it as merely a list of 1,000 common characters that schoolchildren had to acquire. As mentioned above, one of its functions was to practise calligraphy—an application that may have been related to the origin of the text. More importantly, along with some other primers such as the Mengqiu 蒙求 (What Beginners Seek), the Qianziwen represented a manageable collection of traditional tropes, quotes, and allusions, functioning as ‘a basic mnemonic structure in which a large volume of information is encoded into a small number of words, referencing a wider body of cultural knowledge’.Footnote 15 Essentially, the text served as a tool that facilitated the storing and retrieval of such knowledge.

As a by-product of its mnemonic application, the Qianziwen also served as a numbering and ordering sequence, similarly to the way in which the alphabet is employed in Western cultures for marking items in a list. The text was so popular and so intimately connected with elementary education that all literate people would have known it by heart, which made it ideal for numbering longer lists, for which the traditional Heavenly Stems 天干 and Earthly Branches 地支 sequences, or even their combinations, were not enough. One such concrete necessity was the organisation of the Buddhist Canon in which the bundles (zhi 帙) were marked with characters from the Qianziwen as a way of organising them. This is evidenced on the sutra wrappers preserved at Dunhuang—the Qianziwen character is still legible on some of them.Footnote 16

Besides the Qianziwen, among the more common primers identified as having been copied by students was the Taigong jiajiao mentioned above.Footnote 17 This was similarly arranged into four-character segments but was more literary in terms of its content. The text was rediscovered only among the Dunhuang manuscripts and, since then, later versions have also been identified in Vietnam, showing that it circulated throughout East Asia.Footnote 18 There are several Dunhuang scrolls in which the Taigong jiajiao concludes with a colophon explicitly naming the copyist as a student. For example, manuscript Or.8210/S.1163 with this text has a colophon at the end, stating that it was personally copied by a lay student of the Eternal Peace monastery 永安寺 on the seventeenth day of the twelfth month of the gengxu 庚戌 year, which in this context probably referred to 950.

Student colophons help us to identify a group of about 150 manuscripts that were produced in an educational setting at monasteries in Dunhuang.Footnote 19 Based on the information in the colophons, we can roughly date this activity from the second half of the ninth until the end of the tenth centuries. We can see that, within the same context, several texts not normally considered primers also feature as copying exercise for the same group of students. Most prominent among these were the Lunyu 論語 (Confucian Analects), the Xiaojing 孝經 (Classic of Filial Piety), and poems such as the ‘Laments of Lady Qin’ 秦婦吟 and the ‘Rhapsody of the swallow’ 燕子賦.Footnote 20 If we examine all Dunhuang copies of the Xiaojing, it is clear that many of them share a series of codicological characteristics with those bearing lay students’ colophons. The analysis of the codicological features of manuscripts with students’ colophons helps us to identify many more manuscripts that likely belong to the same group, even though they bear no colophons to this effect.Footnote 21

Dunhuang was certainly not the only place where manuscripts of primers, or other texts that functioned as primers, surfaced. Copies of the Qianziwen were found in other locations with significant manuscript finds, most notably the region of Turfan. Tombs at Turfan yielded fragments of students’ copies of the Lunyu and Xiaojing.Footnote 22 The Mengqiu—another common Tang-Dynasty primer—was found at several geographically distant sites, including Dunhuang, Khara-khoto, and a Liao stupa in Ying county 應縣. Even though, in later times, this text essentially fell out of circulation in China, it became immensely popular in Japan, as a handful of relatively early manuscripts demonstrate.Footnote 23 In fact, many primers that were popular in China were also in common circulation in Japan.Footnote 24

Before the Tang period

The history of the transmission of Chinese primers beyond the Chinese border was closely related to the spread of the script. When neighbouring states adopted Sinitic Buddhist texts, along with the scriptures came Literary Sinitic and the Chinese script. Because linguistic knowledge of Chinese would have inevitably been limited among the general population, a range of reading/translating techniques developed with the aim to enable access to Sinitic texts. There is evidence, chiefly from Japan and Korea, but later also from Inner Asia, that texts written in Literary Sinitic were often vocalised in the vernacular language, according to its word order and phonetic peculiarities.Footnote 25

Some states, however, chose to have their own script and translated Sinitic texts into their respective languages. Among the earliest examples on record is the Northern Wei 北魏 (386–535) version of the Xiaojing, which is mentioned in the bibliographic treatise of the Suishu 隋書 (Book of Sui):Footnote 26

又云魏氏遷洛，未達華語，孝文帝命侯伏侯可悉陵，以夷言譯孝經之旨，教于國人，謂之國語孝經。

They also say that when the Wei moved [the capital] to Luoyang, they have not understood Chinese yet, so Emperor Xiaowen (r. 471–499) ordered Houfuhou Kexiling to translate the key points of the Xiaojing into barbarian language and teach it to the people of the state. This was called the Vernacular Xiaojing.

The language of the Tuoba Wei 拓跋魏 was Xianbei 鮮卑, which was probably a para-Mongolic language.Footnote 27 The emperor entrusted the translation to Houfuhou Kexiling (EMC: ɣəw-buwk-ɣəw kha’-sit-liŋ), who was also the author of at least two other works in Xianbei, namely the Guoyu wuming 國語物名 (Names of Things in Vernacular) and the Guoyu zawu ming 國語雜物名 (Names of Miscellaneous Things in Vernacular).Footnote 28 These are listed in the Suishu among a total of 13 Xianbei titles, including such works such as the Guoyu zawen 國語雜文 (Miscellaneous Writings in Vernacular), Guoyu shiba zhuan 國語十八傳 (Eighteen Biographies in Vernacular), and Xianbei yu 鮮卑語 (Xianbei Language).Footnote 29 It seems that most of the titles were either related to the teaching of the Xianbei language or represented literary collections in Xianbei. Some of them seem to have been lists of objects (e.g. Guoyu wuming and Guoyu zawu ming)—that is, primers composed expressly for teaching the vernacular language.Footnote 30

In the list, the Guoyu Xiaojing seems to be the only work that we can identify and, since it appears in the xiaoxue 小學 (‘small learning’) category, there can be little doubt that it functioned as a primer.Footnote 31 The question arises as to the script in which these works were written, especially since there is no record of a Xianbei or Tuoba script. There is a record in the Weishu 魏書 (Book of Wei) concerning the creation of about ‘a thousand new characters’ in 425 but these seem to be variants of existing Chinese characters, which are indeed extremely common in Wei-Dynasty inscriptions.Footnote 32 It has been proposed that the Xianbei texts were written using the Chinese script but there is no hard evidence to substantiate or refute this theory. None of the ‘vernacular’ books listed in the Suishu catalogue has come down to us, and there are no identifiable traces of texts written in the Xianbei language. It is clear, however, that the Guoyu Xiaojing was a written translation into a different language.

By this time, Buddhism had gained a strong hold in Northern China and Buddhist works were also translated from Chinese into other languages. One such case is Emperor Houzhu 後主 (r. 565–577) of the Northern Qi 北齊 (550–577) commissioning Liu Shiqing 劉世清 (d.u.), who was fluent in four foreign languages, to translate the Nirvāṇa sūtra into the Tujue 突厥 language. The scholar Li Delin 李德林 (532–591) wrote a preface and the court presented the complete work to the Tujue qaghan.Footnote 33 Several scholars have noted that there are no records in Old Turkic from this period and thus Tujue in this place may actually signify Sogdian, which was the administrative language of the First Turk empire.Footnote 34 Regardless of the target language, this episode proves the growing significance of Sinitic Buddhism in East and Central Asia. In addition, it also shows that Sinitic Buddhist texts spread, whether in translation or original language, beyond the boundaries of the Chinese states.

The Tang and later periods

By the first half of the eighth century, Sinitic written culture was at the height of its prestige and many of the neighbouring regions integrated Sinitic texts into their local traditions. From about this time, Buddhist texts started to be translated into Sogdian, to the extent that most known Buddhist texts in that language are in fact translations from Chinese.Footnote 35 Most of the archaeologically excavated translations of primers, however, are in Tibetan, Old Uyghur, and Tangut, which represent only three of several major written languages that were around in China during the period in question. The Tibetans derived their alphabet from the Brāhmī script, the Uyghurs from the Sogdian and––ultimately––Aramaic scripts, whereas the Tanguts developed their own script with Chinese inspiration.Footnote 36 The origin of a script typically reflects a state's cultural and religious orientation at the time of the introduction of writing. The survival of larger bodies of text in these three languages is likely a matter of archaeological coincidence and, in the future, there may be discoveries of manuscripts written in other languages. Nevertheless, manuscripts in each of these languages present an interesting scenario for the ways in which Chinese primers were adopted into another linguistic environment.

Tibetan

Most of the early Tibetan manuscripts come from Dunhuang, where the Tibetans remained a strong cultural force even after the end of their political control of the region in 848. The surviving manuscripts reveal that that the two cultures impacted each other in a variety of ways, including scribal practices and Buddhist rituals. Although most of the texts translated from Chinese into Tibetan were of Buddhist content, there was also a demand for specifically non-Buddhist literature. For example, the Jiu Tangshu 舊唐書 (Old Book of Tang) records that, in 730, the Tibetans asked for copies of the Shijing 詩經 (Classic of Odes), Liji 禮記 (Records of Rites), Zuozhuan 左傳 (The Zuo Commentary), and the literary collection Wenxuan 文選 (Literary Anthology), ascribing the request to the Tang Princess Jincheng 金城公主 (698–739), who had been given in marriage to the Tibetan emperor as part of the policy of amicable relations.Footnote 37 An edict ordered the copies to be made and delivered to the Tibetans, even though there was also opposition to this, claiming that the texts would give the Tibetans undesirable military and administrative expertise. According to the tradition, the princess played an important role in the introduction of Sinitic texts into Tibet but, as Matthew T. Kapstein cautions us, the surviving copies of such works come from Dunhuang—a region far removed from the centre of the Tibetan empire—and from about a century after the event.Footnote 38

Tibetan translations of secular texts are relatively few in the surviving body of manuscripts. These include a Tibetan version of the Shangshu 尚書 (Revered Documents) (Pelliot tibétain 986), which was apparently not a verbatim translation.Footnote 39 Another interesting text is known from manuscript Pelliot tibétain 1291, which Yoshiro Imaeda first identified as a translation of the Zhanguoce 戰國策 (Stratagems of the Warring States). Later research showed, however, that this was a translation of the Chunqiu houyu 春秋後語 (Later Comments on the Spring and Autumn Annals).Footnote 40 Yet another historical text is a passage from the Shiji 史記 (Records of the Historian) incorporated into Chapter 4 of the Old Tibetan Chronicle (Pelliot tibétain 1287, 173–246).Footnote 41

As for texts that can be positively identified as primers, there are three extant copies of the Kongzi Xiang Tuo xiangwen shu 孔子項託相問書 (Dialogue between Confucius and Xiang Tuo) about the encounter between Confucius and the wise lad Xiang Tuo 項橐, the Chinese original, nearly 20 copies of which survived in Dunhuang.Footnote 42 A student's colophon dates one of the Chinese copies (i.e. Or.8210/S.395) to 943 and this provides the approximate time frame during which the text was in circulation in Dunhuang. The composition of the text, however, goes back to the early or mid-Tang period.Footnote 43 In his inventory of dated Chinese manuscripts in the Stein collection, Lionel Giles pointed out that the text was transmitted into the modern age under the title Xiao'er lun 小兒論 (Discussions with a Child), which appeared in the Ming-Dynasty booklet Dongyuan zazi 東園雜字 (Miscellaneous Characters from the Eastern Garden).Footnote 44 The Tibetan manuscripts were studied by Michel Soymié, who compared them with versions in other languages, including literary and vernacular Chinese, Thai, and Mongolian. He traced the origin of some of the motifs in earlier Chinese literature and made a number of insightful observations regarding the development of the plot over time.Footnote 45 From the 1980s onward, scholars from Mainland China also made important contributions to the study of the text.Footnote 46

Another related text that survives in two copies (Pelliot tibétain 987 and 988) is an untitled work that R. A. Stein called ‘Maximes confucianistes’.Footnote 47 This is a collection of short allusions to Chinese literary and historical sources in a manner that is reminiscent of Chinese primers, such as the Mengqiu. Based on R. A. Stein's initial attempt to decipher some of the references, Nie Hongyin was able to identify several others and pointed out that the maxims, including their sequence, were similar—but not identical—to the Chinese primer Taigong jiajiao. Based on this, he concluded that the Tibetan text was a translation of a similar type of text.Footnote 48 More recently, Saren Gaowa drew attention to a Tibetan manuscript in the collection of the Nakamura Museum of Calligraphy in Tokyo, in which the second of two texts was titled Old Instructions of Taigong. This text was very similar to, and to a significant degree overlapped with, the text in Pelliot tibétain 987 and 988. Both texts had an apparent connection with the Chinese Taigong jiajiao, yet also contained quotes that were not from that text, but could be matched with content in didactic texts such as the Baixing zhang 百行章 (Hundred Conducts) and the Xinji wenci jiujing chao 新集文詞九經抄 (Newly Collected Excerpts from the Nine Classics).Footnote 49 Therefore, Saerji and Saren suggested that the Tibetan text was not a direct translation of the Chinese Taigong jiajiao, but rather a collection of material from several primers. They transcribed the Tibetan text and matched most of it against the Chinese version, pointing out that the Tibetan text seemed to be more of an annotated version of the references, rather than a collection of maxims.Footnote 50 This may also be the reason for the discrepancy between the Tibetan and Chinese titles.

In addition to translations, another fascinating group of texts that alerts us to the complex linguistic situation in Dunhuang are Tibetan transcriptions of Sinitic texts. Extant manuscript evidence proves that such texts were by no means uncommon and were likely used by the Chinese speakers of the region.Footnote 51 A manuscript that received quite a bit of scholarly attention from early on is a copy of the Qianziwen with a Tibetan transcription of each Chinese character.Footnote 52 It was first published by Haneda Tōru in 1923 and has since then been studied by several scholars.Footnote 53 The Chinese text is complete but the transcription stops 32 four-character segments (i.e. 128 characters) before the end. The manuscript may have been produced for the sake of someone who could not read Chinese characters, or as a study tool to ensure the correct pronunciation of Chinese words.

Two other Sinitic texts transcribed phonetically using the Tibetan alphabet are the Zachao 雜抄 (Miscellaneous Jottings) (Pelliot tibétain 1238) and the Jiujiu biao 九九表 (Nine Times Nine Table) (Pelliot tibétain 1256)—both copied sometime after the end of the Tibetan rule of Dunhuang.Footnote 54 Although a tiny fragment, the surviving bit of the Zachao is very close to the beginning of the Chinese Zachao from Dunhuang, even if there are also some discrepancies.Footnote 55 The Jiujiu biao is an even shorter fragment likely dating to the tenth century, as is implied by a reference to a Khotanese envoy.Footnote 56 Takata Tokio points out that such Tibetan transcriptions were probably intended as auxiliary material for Chinese students to help them during the early stages of learning by enabling them to read texts in a phonetic script.Footnote 57 It is also possible, however, that the users of transcriptions of Sinitic Buddhist texts were Tibetan speakers who engaged in a Chinese form of Buddhism and thus recited Sinitic texts as part of their liturgical practices but were not necessarily literate in Chinese.Footnote 58 This is less likely in the case of primers, although it is also possible that we are dealing with more than one scenario.

In either case, the above examples make it clear that Tibetan speakers were relying on Chinese primers both in original language and in translation. It is also clear that the primers functioned as an entryway into interacting with Sinitic texts, in settings largely removed from native Tibetan literacy. Naturally, some of this activity may have been the result of unique local conditions at Dunhuang, where both languages were common during the ninth and tenth centuries. At the same time, the need to interact with Sinitic written culture would have surely been a concern well beyond the Gansu Corridor and for a significantly longer period of time.

Old Uyghur

The surviving body of Old Uyghur texts includes many that were translated from Chinese. These include Mahāyāna sūtras and apocryphal scriptures, but also a limited number of non-Buddhist works.Footnote 59 Transmitted histories record that, already in the second half of the sixth century, the people in Gaochang 高昌 used Chinese writing alongside their own script and read classical works such as the Mao Shi 毛詩 (Mao Version of the Odes), Lunyu, and Xiaojing, pronouncing these in their own language.Footnote 60 Although reading these texts in the vernacular would have technically been equivalent to translating them, the manuscripts themselves remained in Literary Sinitic. Indeed, numerous Chinese-language copies of texts such as the Lunyu (with and without commentary) and the Xiaojing have been excavated from the Turfan region, substantiating the popularity of these works. Some copies carry teachers’ corrections or assessment, revealing that they were produced in an educational context.

There are only a few surviving fragments of Chinese secular works in Old Uyghur. In a study of Chinese classical works in the Uyghur tradition, Peter Zieme draws attention to a series of quotes from texts such as the Xiaojing, Shijing, and Yijing 易經 (Classic of Changes), which appear in Old Uyghur translations of Buddhist texts, most notably in the Biography of Xuanzang.Footnote 61 These are merely quotes cited in other works and there are only a few texts translated in their own right. The most important of these is the Qianziwen, about half of which is preserved in multiple fragments of at least five different manuscripts from the Turfan region.Footnote 62 The Old Uyghur Qianziwen has been studied in a series of articles by Shōgaito Masahiro, who described the variety of ways in which the text was arranged. For example, several fragments from the Krotkov collection in St. Petersburg begin each four-character segment by writing out the first Chinese character of that segment but substituting the other three with horizontal lines, before adding the Uyghur translation of the entire segment below.Footnote 63 The fact that the Sinitic text is represented by the first character of each tetrasyllabic segment suggests either that the students had already committed the text to memory or, perhaps more likely, that the Chinese characters were merely inserted as reference anchors and the students primarily handled the text in Uyghur.Footnote 64

In other versions, a phonetic transcription in the Uyghur script of the four-character segments, and in some cases the Chinese characters themselves, precede the translation.Footnote 65 In such cases, when the transcription comes before the translation, the Uyghurs probably first vocalised the Sinitic text using their ‘received’ pronunciation and then also read it in Uyghur.Footnote 66 This reading technique is known to us from Japan (i.e. monzen-yomi 文選読み), where it appeared as early as the tenth century but had been largely discontinued by the sixteenth; the Qianziwen is the only text that continued to be read in this manner.Footnote 67 The Uyghur translation itself essentially meant a type of reading similar to the Japanese practice of kundoku 訓読.Footnote 68

Why did the Uyghurs need the Qianziwen? It is, after all, a text that arranges 1,000 Chinese characters into a mnemonic sequence, used by Chinese speakers to teach basic literacy skills. The text itself is not particularly engaging and contains a multitude of allusions that would not have been easily palatable to Uyghur students who were less familiar with Chinese history and culture. As there are surviving fragments in both Chinese and Uyghur, it is clear that they studied the text in both languages. It seems that Sinitic texts played an important role in Uyghur literacy and the Qianziwen was part of the technical apparatus accompanying the script. The Uyghurs primarily relied on Chinese characters for reading Buddhist texts, and the Qianziwen offered a tried-and-tested tool for facilitating the acquisition process. Indeed, the Uyghur Qianziwen fragments typically have Sinitic Buddhist texts (e.g. Lotus sūtra, Mahāprajñāpāramitā sūtra) on the other side, confirming that the principal application of the primer may have been in connection with reading Buddhist texts.Footnote 69 Perhaps less importantly, the allusions and references provided access to a wider body of literary knowledge, as was the case with students in Central China.

In a recent article, Abdurishid Yakup reinterprets a colophon written by a certain Šingtsün Tutong (Shengquan Dutong 勝泉都統), arguing that, in addition to its educational function, the Qianziwen may have also been considered as having apotropaic qualities. The manuscript in question is a small booklet from Turfan (Ch 3716), featuring the Qianziwen in an unusually clumsy hand. Although the colophon is written in Chinese characters, the word order suggests that the underlying language was Old Uyghur.Footnote 70

Similarly located on the verso of a scroll with a Buddhist text, namely the Säkiz yükmäk yaruk sudur (Ch. Tiandi bayang shenzhou jing 天地八陽神呪經, Scripture of the Divine Spell of the Eight Yang of Heaven and Earth), is the Uyghur translation of the Chinese primer Kaimeng yaoxun 開蒙要訓 (Essential Teachings for Beginners). This was a text containing a total of 1,400 characters divided into 350 segments, with a rhyme at the end of every second segment. It enjoyed popularity in both China and beyond, and its title also occurs in the ninth-century Japanese catalogue Nihonkoku genzai sho mokuroku 日本國見在書目錄 (Catalogue of Books Present in Japan). Although it was lost in China in around the Yuan period, about 40 copies survived among the Dunhuang manuscripts.Footnote 71 The Uyghur version, identified only very recently, does not include the Chinese characters, but marks their place with slanting strokes, thereby helping Uyghur readers to find their place in the text.Footnote 72

Finally, a translation of the story of the filial son Shunzi 舜子 (i.e. Shun the Son) survives on three separate fragments, which were originally part of a single manuscript. In Chinese, the story is documented among the Dunhuang manuscripts both as part of a non-Buddhist collection of stories about filial sons and as a ‘transformation text’ (bianwen 變文) of Buddhist inspiration. Although the Uyghur version is too fragmentary to allow a definite conclusion, it is likely that it was a translation of the transformation text.Footnote 73 In either case, the Chinese versions seem to have functioned within an educational context, and this was also likely true of the Uyghur translation.

The Uyghurs used Chinese primers in many ways that were similar to how the Tibetans did. The archaeological record suggests that at the core of such activities was the desire to engage with Sinitic Buddhist texts, either in their original language or via translation. Primers were technical tools that facilitated access to Sinitic texts and, through them, to Sinitic Buddhism. They seem to have had little functionality with regard to native Uyghur texts or those translated from other languages.

Tangut translations

As in the case of the Tibetan and Old Uyghur translations, the overall majority of surviving Tangut texts are Buddhist in nature, yet there is a smaller group of non-Buddhist works translated from Chinese, such as Confucian Classics, military treatises, medical works, and, not least, primers. We know from historical sources that, when the Tangut script was invented in 1036, the Xiaojing, Erya 爾雅 (Approaching Elegance) and Siyan zazi 四言雜字 (Miscellaneous Characters in Four Syllables) were the first texts to be translated into Tangut.Footnote 74 Of these, the last text must have been a primer, probably similar to the Chinese Zazi. The Erya, which is essentially a thesaurus, may strike us as an unexpected choice to translate into a foreign language, but it is clear that it was closely linked to the early stages of literacy and, in this respect, was considered less language-dependent than it may seem to us today. Similarly, the Xiaojing functioned as an educational text in which the didactic content together with a relatively uncomplicated vocabulary provided ideal material for students to copy and study.Footnote 75 Paradoxically, the adaptation of Chinese primers played a role in the introduction of the native Tangut script.

Excavations at the ruins of Khara-khoto yielded a large body of handwritten and printed material in Tangut, Chinese, Tibetan, as well a smaller number in other languages. The material includes a large number of Sinitic Buddhist texts and there are also a few texts that we can identify as primers, such as the Qianziwen, Mengqiu, and Zazi. These attest to the use of these works in their original language in the Tangut state. This is unsurprising in view of the large number of Sinitic texts found at the site, which point to the existence of a robust scribal culture that would have certainly included a process of training.

Among the Tangut texts translated from Chinese is the translation of the Xiaojing with the lost commentary and preface of Lü Huiqing 呂惠卿 (1032–1111),Footnote 76 although the full cursive hand, the multitude of corrections, and the presence of a scholarly commentary indicate that this particular copy would not have been used in primary education. It nonetheless confirms the existence of a Tangut translation of this work. There is also a surviving Tangut text with the title Miscellaneous Characters of the Three Realms (Ch. *Sancai zazi 三才雜字).Footnote 77 Four copies of three different printed editions in the St. Petersburg collection and a number of additional fragments in other collections evidence the popularity of this book among the Tanguts. Although the content of the text corroborates that it was composed in the Tangut state, it emulates Chinese zazi-like primers, which list two- or four-character segments in a thematical arrangement.Footnote 78 This was, therefore, a Tangut primer composed after a Chinese model.

Another primer is a Tangut text tentatively named by modern scholars as Excerpts from the Classics and Histories (Ch. *Jingshi zachao 經史雜抄).Footnote 79 The text contains brief quotes from classical texts such as the Shangshu, Zhouyi 周易 (Zhou Changes), Chuci 楚辭 (Chu Lyrics), Guanzi 管子, Laozi 老子, Huainanzi 淮南子, and Lunheng 論衡 (Discourses on Balance). Interestingly, it also contains quotes from the Taigong jiajiao, which was certainly not part of the elite literary canon. Indeed, the Tangut text itself is conspicuously crude and has a multitude of glaring mistakes.Footnote 80 Scholars had suspected that the text must have been a translation from a Chinese original, but only recently succeeded in linking it with the Chinese primer Xinji wenci jiujing chao, found in the Dunhuang library cave in over a dozen copies.Footnote 81 Although the Chinese and Tangut texts are not fully identical, they share 170 matching versus 57 unmatching sections. In addition, they share some mistakes, which proves the existence of a direct textual link between them.Footnote 82 Even though we are certain that the Tangut text is related to the Chinese version of the Xinji wenci jiujing chao, it is difficult to say whether the discrepancies are due to the Tangut translator's having relied on a different version or having deliberately modified the source text.Footnote 83

A collection of anecdotes from the Warring States period is a Tangut printed book that bears the title Twelve States, which was early on recognised as a translation of a lost Chinese text.Footnote 84 Recently, the Tangut text was successfully identified as a translation of the Shi'erguo shi 十二國史 (History of the Twelve States) by Sun Yu 孫昱 (d.u.), the title of which appears in the bibliographic treatise of the Songshi 宋史 (History of the Song).Footnote 85 Although the Sinitic text did not come down to us, enough of it survived in quotations to allow a positive identification. As is the case with several other Tangut texts, the translation is the earliest—in this particular instance, the only—surviving version of the Sinitic text.

There are also some bilingual works, some of which would have been used in primary education. Perhaps the most noteworthy among these is the glossary Tangut-Chinese Timely Pearl in the Palm (Ch. Fan-Han heshi zhang zhong zhu 番漢合時掌中珠), which consists of thematically grouped Tangut and Chinese word pairs, aligned side by side, including the pronunciation in both languages.Footnote 86 It is clear that this was an interlingual learning tool that worked in either direction, depending on the language of the user. It was not a translation of a Sinitic text, but a glossary tailored for local needs in the Tangut state.

The extant material shows that the Tanguts were highly invested in translating not only Sinitic Buddhist texts and even Confucian Classics, but also works associated with elementary learning. Although it is undeniable that, as was the case among the Tibetans and Uyghurs, some of these texts would have been utilised as part of the interaction with Sinitic writings, many of them were linked with acquiring literacy in Tangut. This relates to the aforementioned record that the invention of the Tangut script was accompanied by the translation of the Xiaojing, Erya, and Siyan zazi, none of which would have been intended for the study of Chinese texts in that context. Instead, the translations of these three works were clearly meant to provide practical tools for learning the new script and enhancing the learners’ Tangut literacy skills. Thinking about the possible reasons for the differences in how the Tibetans and Uyghurs used Chinese primers, we are reminded that the Tangut script was not borrowed from elsewhere. Instead, it was a new invention only partially inspired by the Chinese example, which means that, structurally (and even visually), it shared a number of characteristics with the Chinese script, while also being completely unintelligible for a Chinese reader. Nonetheless, as a result of the typological similarities between the two scripts, translations of Chinese primers seem to have been thought to be appropriate for teaching Tangut.

Other languages

The number of texts that survive in Jurchen and Khitan is extremely limited, which is why it is hard to form an accurate picture of the translation activities in these two languages.Footnote 87 Interestingly, a Chinese printed copy of the primer Mengqiu survives in a Liao-period stupa in Ying county but we have no physical evidence of translated versions of primers or other Sinitic texts.Footnote 88 Just as was the case with the Tanguts, there is a mention of a Jurchen Xiaojing but no copy of it is extant today.Footnote 89

Even though this article does not aim to go beyond the Mongol conquest of Northern China, it is apparent that translations of primers—and other Sinitic texts—became more common in later centuries.Footnote 90 Among the most interesting cases is the Xiaojing translated into pre-classical Mongolian. As we have seen, the text has been among the most popular works even in non-Chinese states, with translations made in Xianbei, Tangut, Khitan, and Jurchen, and later in further languages.Footnote 91 The pre-classical Mongolian version is arranged in a manner that is similar to the Uyghur Qianziwen in that the Chinese segments precede their Mongolian translation and thus the two languages follow each other consecutively on the same line.Footnote 92 The similarities of presentation on the one hand evidence a continuity between different book cultures around China and, on the other, their indebtedness to Chinese precedents.

Conclusions

By all accounts, the Sui–Tang unification of China at the beginning of the seventh century was of major significance for the political structure of East and Central Asia. It created a dynamic empire that, for the first time in centuries, managed to live up to the legacy of the Han Dynasty and (re)gain control over the subcelestial realm. Although, from the mid-eighth century onward, the Tang empire lost its original power base, including its Central Asian territories, Sinitic texts remained influential throughout East and Central Asia. Archaeological evidence reveals that most of the translations of primers and non-Buddhist Chinese works date to the later, and thus weaker, half of the Tang Dynasty and to the following period in which China was ruled by competing states. This was a time of growth for non-Chinese states, which needed scripts and texts in their quest for political legitimation and bureaucratic governance.

This article aimed to show that Chinese primers were actively used by a variety of cultures and peoples speaking other languages in Inner Asia. These peoples interacted with the primers in different ways, ranging from reading original Sinitic texts and vocalising them using specific techniques to creating written translations in local languages. Far from being unique to particular cultures, this kind of interaction is observable wherever significant numbers of manuscripts are excavated. In other words, this situation was the norm, rather than an exception. While such engagement with primers and Sinitic texts in general is relatively well documented in modern scholarship with regard to Japan, Korea, and Vietnam, similar processes were at play along China's northern and western borders.

The adoption and translation of Chinese primers by Tibetan speakers were linked to the study of Sinitic texts. Since early Tibetan manuscripts primarily come from Dunhuang, it is difficult to assess how much of this interest reflected the multicultural background of the local population and was specific to that particular region. Some of the transcriptions of Chinese primers may have been written by Chinese individuals who used the phonetic alphabet merely as an auxiliary means to record pronunciation at the early stages of learning. In parallel with this, it is also possible that Tibetan speakers who were practising forms of Sinitic Buddhism also produced phonetic transcriptions of Buddhist texts in order to be able to recite them in a liturgical setting. The Uyghurs used a variety of Chinese primers, in both translation and original language. The main motivation behind using these would have been to facilitate access to Sinitic Buddhist texts. Among the most noteworthy primers was the Qianziwen, which survived in multiple fragments, attesting to the popularity of the text in the Turfan region. It is also possible that the Uyghurs attributed apotropaic qualities to the text, copying and reciting it to avoid harm. In the case of the Tanguts, the translation of primers was closely connected to the introduction of the Tangut script, which in turn was an integral part of their nation-building efforts. They may have been following Jurchen and Khitan precedents, although little concrete evidence survives to corroborate this hypothesis. Unlike the Tibetans and Uyghurs, their translations of Chinese primers were also used in a monolingual setting to teach Tangut students to read and write their own language.

Acknowledgements

I would like to express my gratitude to Sam van Schaik, Françoise Bottéro, Yukiyo Kasai, Zhang Daying, Jonathan A. Silk, and Junfu Wong for their comments on draft versions of this article. I am also grateful for the valuable suggestions that I received anonymously as part of the review process.

Conflicts of interest

None.

References

1 The rise of the Mongols marks the advent of an entirely new era and some scholars argue that it represents the beginning of the early modern period; e.g. Biran, M., ‘The Mongol transformation: from the steppe to Eurasian empire’, Medieval Encounters 10.1–3 (2004), p. 339Google Scholar.

2 E.g. Handel, Z. J., Sinography: The Borrowing and Adaptation of the Chinese Script (Leiden and Boston, 2019)Google Scholar.

3 For a more inclusive view of East Asian written cultures that goes beyond the typical CJKV model, see Kornicki, P. F., Languages, Scripts, and Chinese Texts in East Asia (Oxford, 2018)Google Scholar. For an overview of Chinese linguistic and literary influences specifically in the region that is now China's Northwest, see Chin, T. T., ‘Colonization, sinicization, and the polyscriptic northwest’, in The Oxford Handbook of Classical Chinese Literature, (eds.) W. Denecke, W.-Y. Li, and X. Tian (Oxford, 2017), pp. 477–493Google Scholar.

4 Hanshu 漢書 30, p. 1721. Dynastic histories are cited in this article on the basis of the Zhonghua shuju editions digitised as part of the Scripta Sinica online database.

5 On excavated fragments of the Cangjie pian, see C. J. Foster, ‘Study of the Cang Jie Pian: Past and Present’ (unpublished PhD dissertation, Harvard University, 2017); and Foster, C. J., ‘The spread of scribal literacy in Han China: all along the watchtowers’, in Power from Below in Premodern Societies: The Dynamics of Political Complexity in the Archaeological Record, (eds.) T. L. Thurston and M. Fernández-Götz (Cambridge, 2021), pp. 175–201Google Scholar. Foster's work also shows the remarkably wide distribution of the text far beyond the confines of Central China. Thus, there are fragments from sites in Gansu Province, Inner Mongolia, and even the ruins of Niya 尼雅 on the southern edge of the Taklamakan (ibid., pp. 61–108). For fragments from the Khotan region, see also Chen Lifang 陳麗芳, ‘Tangdai Yutian de tongmeng jiaoyu: Yi Zhongguo Renmin daxue bowuguan cang Hetian xizi wenshu wei zhongxin’ 唐代于闐的童蒙教育——以中國人民大學博物館藏和田習字文書為中心, Xiyu yanjiu 西域研究 1 (2014), pp. 39–40. For fragments coming from China's north-western region, see Chō Narei [Zhang Nali] 張娜麗, Saiiki monjo no kisoteki kenkyū: Chūgoku kodai ni okeru shōgakusho dōmōsho no shosō 西域出土文書の基礎的研究 : 中国古代における小学書・童蒙書の諸相 (Tokyo, 2006), pp. 85–132.

6 For example, fragments of the Jiujiu pian (under the title Jijiu zhang 急就章) were discovered in Luoyang 洛陽 on steles from the Wei 魏 period and, on another occasion, on an Eastern Han brick shard; Zhang Xinpeng 張新朋 ‘Dagu wenshu zhong de Jijiu pian canpian kao’ 大谷文書中的《急就篇》殘片考, Xinan minzu daxue xuebao (Renwen shehui kexue ban) 西南民族大學學報 (人文社會科學版) 11 (2016), p. 193.

7 F. Bottéro, ‘Les “manuels de caractères” à l’époque des Han Occidentaux’, in Éducation et instruction en Chine, I. L’éducation élémentaire, (eds.) C. Nguyen Try and C. Despeux (Paris, 2003), p. 120; cf. F. Bottéro, ‘Les manuels de caractères de l’époque Qin-Han’, in La fabrique du lisible: La mise en texte des manuscrits de la Chine ancienne et médiévale, (eds.) J.-P. Drège and C. Moretti (Paris, 2014), pp. 55–59.

8 Zhou Zumo 周祖謨 ‘Ji Tulufan chutu Jijiupian zhu’ 記吐魯番出土急就篇注 in Dunhuang Tulufan wenxian yanjiu lunji (di'er ji) 敦煌吐魯番文獻研究論集（第二輯）, (ed.) Beijing daxue Zhongguo zhonggu wenxian yanjiu zhongxin 北京大學中國中古文獻研究中心 (Beijing, 1983), pp. 309–314.

9 This was also true for Japan, where both the Lunyu 論語 (Analects of Confucius) and the Qianziwen were commonly cited on wooden documents (mokkan 木簡) during the seventh and eight centuries; see Lurie, D. B., Realms of Literacy: Early Japan and the History of Writing (Cambridge, MA and London, 2011), p. 112Google Scholar; and J. Guest, ‘Primers, Commentaries, and Kanbun Literacy in Japanese Literary Culture, 950-1250CE’ (unpublished PhD dissertation, Columbia University, 2013), pp. 71–72.

10 For fragments from Khotan, see Chen Lifang 陳麗芳, ‘Tangdai Yutian de tongmeng jiaoyu’, pp. 40–41.

11 Soymié, M., Drège, J.-P., Eliasberg, D., Magnin, P., Schneider, R., and Trombert, E., Catalogue des manuscrits chinois de Touen-houang: Fonds Pelliot chinois de la Bibliothèque nationale (Paris, 1991), : Volume IV, Nos 3501–4000, p. 52Google Scholar.

12 Wang Xizhi's calligraphy was commonly copied as a way of practising calligraphy. Among the most popular works was the ‘Lanting xu’ 蘭亭序 (Preface to the Orchid Pavilion), which survives in numerous fragments from sites in Western China as far as Khotan; see Rong, X., ‘The Lanting xu in the Western Regions’, in The Silk Road and Cultural Exchanges between East and West, (ed.) X. Rong (Leiden and Boston, 2022), pp. 293–315Google Scholar; and X. Rong, ‘The transmission of Wang Xizhi's Shang xiang Huang Qi tie in the western regions’, in The Silk Road, (ed.) Rong, pp. 316–330.

13 On manuscripts copied by students in Dunhuang, see Galambos, I., Dunhuang Manuscript Culture: End of the First Millennium (Berlin, 2020), pp. 85–138Google Scholar.

14 Mair, V. H., ‘Lay students and the making of written vernacular narrative: an inventory of Tun-huang manuscripts’, Chinoperl Papers 10 (1981), p. 57Google Scholar.

15 Nugent, C. M. B., Textual Practices of Literary Training in Medieval China Evidence from Dunhuang Manuscripts (Leiden and Boston, 2024), p. 82Google Scholar.

16 X. Rong, Eighteen Lectures on Dunhuang, (trans.) I. Galambos (Leiden and Boston, 2013), pp. 112–113.

17 For a French translation of the text, see P. Demiéville, L'oeuvre de Wang le zélateur (Wang Fan-tche): Suivi des Instructions domestiques de l'aïeul (T'ai-kong kia-kiao): poèmes populaires des T'ang (VIIIe-Xe siècles) (Paris, 1982), pp. 611–835; for a Japanese translation, see Itō Mieko 伊藤美重子, Tonkō monjo ni miru gakkō kyōiku 敦煌文書にみる学校教育 (Tokyo, 2008), pp. 150–180.

18 On Vietnamese versions of the Taigong jiajiao, see Wang Xiaodun 王小盾, ‘Cong Yuenan suwenxue wenxian kan Dunhuang wenxue yanjiu he wenti yanjiu de qianjing’ 從越南俗文學文獻看敦煌文學研究和文體研究的前景, Zhongguo shehui kexue 中國社會科學 1 (2003), p. 172.

19 For an inventory of manuscripts and colophons, see Li Zhengyu 李正宇, ‘Dunhuang xuelang tiji jizhu’ 敦煌學郎題記輯注, Dunhuangxue jikan 敦煌學輯刊 1 (1987), pp. 26–40; Itō, Tonkō monjo ni miru gakkō kyōiku, pp. 44–68.

20 For an overview of texts most commonly copied by students in Dunhuang, see Drège, J.-P., ‘La lecture et l’écriture en Chine et la xylographie’, Études chinoises 10.1–2 (1991), pp. 85–86Google Scholar. For the Laments of Lady Qin manuscripts, see C. M. B. Nugent, Manifest in Words, Written on Paper: Producing and Circulating Poetry in Tang Dynasty China (Cambridge, MA, 2010). Both Hans van Ess and Valerie Lavoix note the similarities between the Rhapsody of the Swallow and the Han-Dynasty Rhapsody of the Divine Crow 神烏傅(賦) excavated in 1993 in Yinwan 尹灣, Jiangsu Province; see H. van Ess, ‘An interpretation of the Shenwu fu of Tomb No. 6, Yinwan’, Monumenta Serica 51 (2003), p. 611; and V. Lavoix, ‘Un poème des Han – L’“Exposition du corbeau prodigieux” de Yinwan’, in La fabrique du lisible, (eds.) Drège and Moretti, p. 182.

21 Mair, ‘Lay students’, pp. 5–96, catalogues nearly 600 such manuscripts.

22 On a Turfan copy of the Lunyu Zheng shi zhu 論語鄭氏注 (The Zheng Commentary to the Analects of Confucius), copied by a 12-year-old student, see Galambos, Dunhuang Manuscript Culture, pp. 86–87.

23 See Ikeda Toshio 池田利夫, Mōgyū kochū shūsei 蒙求古注集成 (Tokyo, 1988–1990), which is a four-volume facsimile collection of early annotated copies of the Mengqiu preserved in Japan.

24 Guest, ‘Primers, Commentaries, and Kanbun Literacy’, pp. 39–44.

25 On reading and glossing texts in Literary Sinitic in Japan and Korea, see Whitman, J., ‘The ubiquity of the gloss’, Scripta 3 (2011), pp. 95–121Google Scholar; Kornicki, Languages, pp. 157–186; Fraleigh, M., ‘Rearranging the figures on the tapestry: what Japanese direct translation of European texts can tell us about kanbun kundoku’, Japan Forum 31 (2019), pp. 4–32Google Scholar; and Bunkyō, Kin, Literary Sinitic and East Asia: A Cultural Sphere of Vernacular Reading (Leiden and Boston, 2021)Google Scholar. For examples from Gaochang 高昌, see Zhoushu 周書 50, p. 915; from among the Khitans, see Yijian zhi 夷堅志, bing 丙 18, p. 514.

26 Suishu 隋書 32, p. 935.

27 Vovin, A., ‘Once again on the Tabγač language’, Mongolian Studies XXIX (2007), pp. 191–206Google Scholar; see also L. Ligeti, ‘Tabghatch, un dialecte de la langue Sien-pi’, in Mongolian Studies, (ed.) L. Ligeti (Budapest, 1970), pp. 265–308; and E. G. Pulleyblank, ‘The Chinese and their neighbors in prehistoric and early historic times’, in The Origins of Chinese Civilization, (ed.) D. N. Keightley (Berkeley, 1983), p. 453.

28 Houfuhou is sometimes interpreted as ‘Lord of Hou-fu’, yet it is clear that hou 侯 in this place is not a title, but part of the Xianbei surname. The Zhoushu (Book of Zhou) (29, p. 514) records, for example, that, in 535, the general Hou Zhi 侯植 (fl. 530–555) was awarded the surname Houfuhou; see also A. E. Dien, ‘The bestowal of surnames under the Western Wei-Northern Chou: a case of counter acculturation’, T'oung Pao (Second Series) 63.2/3 (1977), pp. 173–174; Yao Weiyuan 姚薇元, Beichao huxing kao 北朝胡姓考 (Beijing, 1962), pp. 82–83 argues that the surname Houfuhou is merely a different transliteration of the tribal name Hufohou 護佛侯.

29 Suishu 32, p. 945; see also G. Schreiber, Japanese Morphography: Deconstructing Hentai Kanbun (Leiden and Boston, 2022), pp. 387–388.

30 Cf. Shimunek, A., Languages of Ancient Southern Mongolia and North China: A Historical-Comparative Study of the Serbi or Xianbei Branch of the Serbi-Mongolic Language Family, with an Analysis of Northeastern Frontier Chinese and Old Tibetan Phonology (Wiesbaden, 2017), pp. 121–122Google Scholar.

31 The choice to translate specifically the Xiaojing may have been linked to the significance of the theme of filial piety during the Northern Wei and the pre-Tang period in general. There are numerous examples of illustrations of the stories of filial children in Northern Wei art; see, for example, Bradford, R. E., The Guyuan Sarcophagus: Motifs from All Asia (Saarbrücken, 2011)Google Scholar; P. E. Karetzky and A. C. Soper, ‘A Northern Wei painted coffin’, Artibus Asiae 51½ (1991), pp. 5–28.

32 Weishu 魏書 4A, p. 70.

33 Bei Qishu 北齊書 20, p. 267; see also Liu, C., Selected Papers from the Hall of Harmonious Wind (Leiden, 1976), p. 14Google Scholar, esp. n. 36. For a translation of this passage, see F. Thierry, ‘La monétarisation de la société türke (VIe -IXe siècle), Influence chinoise, influence sogdienne’, in Les Sogdiens en Chine, (eds.) E. Vaissière and E. Trombert (Paris, 2005), p. 403.

34 X. Tremblay, ‘The spread of Buddhism in Serindia: Buddhism among Iranians, Tocharians, and Turks before the 13th century’, in The Spread of Buddhism, (eds.) A. Heirman and S. P. Bumbacher (Leiden and Boston, 2007), p. 108; J. Wilkens, ‘Buddhism in the West Uyghur Kingdom and beyond’, in Transfer of Buddhism Across Central Asian Networks, (ed.) C. Meinert (Leiden and Boston, 2016), p. 193; and Thierry, ‘La monétarisation’, p. 403.

35 For a brief overview of Buddhist texts in Sogdian, see Yoshida, Y., ‘A handlist of Buddhist Sogdian texts’, Memoirs of the Faculty of Letters, Kyoto University 54 (2015), pp. 167–180Google Scholar. The only dated Sogdian translation of a Buddhist text discovered at Dunhuang is from 728 and there is no indication that the others were ‘significantly younger or older’ than that (ibid., p. 168). The latest ones, with Uyghur elements in the colophons, may come from the tenth century.

36 The surviving Uyghur manuscripts are in fact written in a variety of different scripts, depending on the tradition from which the texts were adopted. Thus, Manichaean texts were typically written using the Manichaean script, Christian ones in the Syriac, and Buddhist ones in the Brāhmī or Chinese scripts. See Y. Kasai, ‘Multiscripturality in Old Uyghur: relationship between scripts and religions’, in The Silk Roads: Critical Concepts in Asian Studies, (ed.) B. Meisterernst (London, 2016), pp. 193–201.

37 Jiu Tangshu 舊唐書 196, p. 5232.

38 M. T. Kapstein, ‘The Tibetan Yulanpen jing’, in Contributions to the Cultural History of Early Tibet, (eds.) M. T. Kapstein and B. Dotson (Leiden and Boston, 2007), pp. 212–213.

39 W. South Coblin calls it a paraphrase; see Coblin, W. S., ‘A study of the old Tibetan Shangshu paraphrase, Part I’, Journal of the American Oriental Society 111.2 (1991), pp. 303–322Google Scholar.

40 Y. Imaeda, ‘L'identification de l'original chinois du Pelliot tibétain 1291—traduction tibétaine du Zhangguoce’, Acta Orientalia Hung 34.1–3 (1980), pp. 53–68. For the identification of the text as a translation of the Chunqiu houyu, see Ma Mingda 馬明達, ‘P.T.1291 hao Dunhuang Zangwen wenshu yijie dingwu’ P.T.1291號敦煌藏文文書譯解訂誤. Dunhuangxue jikan 敦煌學輯刊 2 (1984), pp. 14–24.

41 T. Takeuchi, ‘A passage from the Shi Chi in the Old Tibetan Chronicle’, in Soundings in Tibetan Civilization, (eds.) B. N. Aziz and M. T. Kapstein (New Delhi, 1985), pp. 135–146.

42 O. Weingarten, ‘The unorthodox master: the serious and the playful in depictions of Confucius’, in A Concise Companion to Confucius, (ed.) P. R. Goldin (Hoboken and Chichester, 2017), pp. 63–65; for the Tibetan text, see M. T. Kapstein, ‘Confucius and the Marvelous Lad’, in Sources of Tibetan Tradition, (eds.) K. R. Schaeffer, M. T. Kapstein, and G. Tuttle (New York, 2013), pp. 96–100; for the Chinese manuscripts, see Zheng Acai 鄭阿財, ‘Dunhuang xieben Kongzi Xiang Tuo wenshu chutan’ 敦煌寫本《孔子項托相問書》初探 in Dunhuang wenxian yu wenxue 敦煌文獻與文學, (ed.) Zheng Acai (Taibei, 1993), pp. 395–436. This story seems to have enjoyed considerable popularity in later centuries, as closely related versions survived in Turfan and, from a later period, in Southern China and Vietnam. For the Vietnamese manuscripts, see Zhu Fengyu 朱鳳玉, ‘Cong Yuenan hanwen xiaoshuo kan zhengqi wenxue zai hanzi wenhuaquan de fazhan’ 從越南漢文小說看爭奇文學在漢字文化圈的發展, Chengda zhongwen xuebao 成大中文學報 38 (2012), pp. 73–74. For a close comparison of extant copies of the text, including those in Tibetan, see Wang Xiaodun 王小盾 and He Qiannian 何仟年, ‘Yuenan ben Kongzi Xiang Tuo wenda shu jianlun’ 越南本《孔子項橐問答書》譾論 in Xin shiji Dunhuangxue lunji 新世紀敦煌學論集, (eds.) Xiang Chu 項楚 and Zheng Acai (Chengdu, 2003), pp. 239–257.

43 Zheng, Dunhuang wenxian yu wenxue, pp. 407–408.

44 Giles, L., ‘Dated Chinese manuscripts in the Stein collection’, Bulletin of the School of Oriental Studies, University of London 10.2 (1940), pp. 341–342Google Scholar. The Xiao'er lun had apparently also been translated into Jurchen, which served as the basis for a Manchu version later on; see Min Yŏng-gyu 閔泳珪, ‘Manjucha Soaron kwa Tonhwang ŭi Hang T'ak pyŏnmun’ 滿州字小兒論과敦煌의項託變文, in Yi Sang-paek paksa hoe'gap kinyŏn nonch'ong 李相佰博士回甲紀念論叢, (ed.) Sangbaek Yi Sang-baek paksa hoegap kinyŏm nonch'ong p'yŏnjip wiwŏnhoe 想白李相伯博士回甲紀念論叢編輯委員會 (Seoul, 1964), pp. 321–332. A Japanese version of the story also made it into the Konjaku monogatari shū 今昔物語集 (Collection of New and Old Stories)—an anthology of 1,040 tales probably written down in the early twelfth century. For a comparison of the Dunhuang version of the story with that in the Konjaku monogatari shū, see Itō Mieko 伊藤美重子, ‘Tonkō monjo to Nihon no kakawari: Tonkō shahon Kōshi Kō Taku sōmonsho to Konjaku monogatari shū’ 敦煌文書と日本のかかわり―敦煌写本「孔子項託相問書」と『今昔物語集』, Daigakuin kyōiku kaikaku shien puroguramu (Nihon bunka kenkyū no kokusaiteki jōhō dentatsu sukiru no ikusei) katsudō hōkokusho 大学院教育改革支援プログラム「日本文化研究の国際的情報伝達スキルの育成」活動報告書, v. 20 (2008), pp. 358–362.

45 M. Soymié, ‘L'entrevue de Confucius et de Hiang T'o’, Journal Asiatique CCXLII (1954), pp. 311–391.

46 See, for example, Zhang Hongxun 張鴻勳, ‘Dunhuang ben Kongzi Xiang Tuo xiangwen shu’ 敦煌本《孔子項託相問書》研究, Dunhuang yanjiu 敦煌研究 2 (1985), pp. 99–110.

47 R. A. Stein, ‘Tibetica Antique VI: Maximes confucianistes dans deux manuscrits de Touen-houang’, Bulletin de l’École Française d'Extrême-Orient 79.1 (1992), pp. 9–17; for an English version, see R. A. Stein, ‘Confucian maxims in two Dunhuang manuscripts’, in Rolf Stein's Tibetica Antiqua, (eds.) R. A. Stein and A. P. McKeown (Leiden and Boston, 2002), pp. 273–283.

48 Nie Hongyin 聶鴻音, ‘Dunhuang P.988 hao Zangwen xiejuan kaobu’ 敦煌 P. 988 號藏文寫卷考補, Minzu yanjiu 民族研究 3 (2005), pp. 78–84.

49 Saerji 薩爾吉 and Saren Gaowa 薩仁高娃, ‘Dunhuang Zangwen rujia geyan duwu yanjiu: Yi Zhongcun Buzhe jiucangben Gu Taigong jiao wei zhongxin’ 敦煌藏文儒家格言讀物研究——以中村不折舊藏本《古太公教》為中心, Zhongguo Zangxue 中國藏學 1 (2017), pp. 39–59.

50 Ibid., p. 58.

51 Takata, T., ‘Multilingualism in Tun-huang’, Acta Asiatica, Bulletin of the Institute of Eastern Culture 78 (2000), pp. 58–59Google Scholar. The so-called ‘Long scroll’ (IOL Tib J 1772, originally Ch. 9.II.17) in London contains a series of Sinitic Buddhist texts transcribed using the Tibetan alphabet. For the restoration of Chinese characters based on Tibetan transcription, see Takata Tokio 高田時雄 ‘Chibetto moji shosha Chōkan no kenkyū (honbun hen)’ チベット文字書寫『長卷』の研究（本文編）, Tōhō gakuhō 東方學報 (Kyoto) 65 (1993), pp. 313–380, 14pl.; cf. W. Simon, ‘A note on Chinese texts in Tibetan transcription’, Bulletin of the School of Oriental and African Studies 21.1/3 (1958), pp. 334–343.

52 Originally catalogued as Pelliot chinois 3419, this manuscript is now Pelliot tibétain 1046.

53 Haneda Tōru 羽田亨, ‘Kan-Ban taion Senjimon no dankan’ 漢蕃對音千字文の斷簡, Tōyō gakuhō 東洋學報 13.3 (1923), pp. 84–104; reprinted in Haneda hakushi shigaku ronbunshū, gekan: gengo, shūkyō-hen 羽田博士史学論文集下巻 言語・宗教篇 (Kyoto, 1958), pp. 396–419. See also B. Csongor, ‘Some Chinese texts in Tibetan script from Tun-huang’, Acta Orientalia Hung. 10.2 (1960), pp. 99–100.

54 Takata Tokio, ‘Zasshō to kukuhyō: Tonkō ni okeru Chibetto moji shiyō no ichimen’ 雜抄と九九表― 敦煌に於けるチベット文字使用の一面, Insha ronsō 均社論叢 14 (1983), pp. 1–4. For a Chinese version of this article, see Gaotian Shixiong 高田時雄, ‘Zachao yu jiujiubiao’《雜抄》與九九表, in Dunhuang, minzu, yuyan 敦煌⋅民族⋅語言, (ed.) Gaotian Shixiong (Beijing, 2005), pp. 79–85.

55 The Tibetan transcription of the Zachao was first published in Simon, ‘Note on Chinese texts in Tibetan transcription’, p. 342, but it was Takata Tokio who later managed to identify the text.

56 The tenth century was the period of close contacts between the rulers of Khotan and Dunhuang, who also became related through marriage.

57 Tokio, ‘Zasshō to kukuhyō’. The reasoning behind this view is that learning the Tibetan alphabet would have been much easier and quicker than acquiring the skills to read and write Chinese, and so Chinese students in this heavily Tibetanised region would have probably learned to read the Tibetan alphabet before they achieved full literacy in Chinese.

58 I owe this insight to Jonathan A. Silk.

59 Wilkens, ‘Buddhism in the West Uyghur Kingdom and beyond’, p. 207. For an overview of Buddhist literature in Old Uyghur, see J. Elverskog, Uygur Buddhist Literature (Turnhout, 1997). An early classic on the topic is B. Laufer, Berthold, ‘Zur buddhistischen Litteratur der Uiguren’, T'oung Pao (Second Series) 8.3 (1907), pp. 391–409.

60 Zhoushu 50, 915.

61 P. Zieme, ‘Chinese classical works in Uighur tradition’, in Journal of the Turfan Studies: Essays on the Third International Conference on Turfan Studies; The Origins and Migrations of Eurasian Nomadic Peoples, (ed.) Academia Turfanica (Shanghai, 2010), pp. 459–471.

62 Umemura, H. and Zieme, P., ‘A further fragment of the Old Uighur Qianziwen’, Written Monuments of the Orient 2 (2015), p. 4Google Scholar.

63 Shōgaito, M. and Yakup, A., ‘Four Uyghur fragments of Qian-zi-wen Thousand Character Essay’, Turkic Languages 5 (2004), 3–28Google Scholar, 313–317, p. 4.

64 This method was not limited to the Qianziwen, as it is also documented in Buddhist literature. For example, in the Old Uyghur translation of the Chinese version of the Abhidharmakośa-kārikā (Apidamo jushe lun bensong 阿毗達磨俱舍論本頌, T.1560), the Uyghur segments of text are normally preceded by two or more Chinese characters, marking their location in the Sinitic text; see Kudara, K., ‘A fragment of an Uigur version of the Abhidharmakośakārikā’, Journal Asiatique CCLXIX/1–2 (1981), p. 328Google Scholar. A comparable practice can also be seen in monolingual Chinese commentaries to Buddhist texts, in which the commentaries sometimes reference the main text by citing only the first and last characters of the relevant passage; see C. Moretti, ‘L'organisation du texte et du commentaire dans le Traité de la grande vertu de sagesse’, in La fabrique du lisible, (eds.) Drège and Moretti, pp. 252–253.

65 Shōgaito Masahiro 庄垣内正弘, ‘Roshia shozō uigurugo danpen no kenkyū 2: Agonkyō, Senjimon, Abidatsumaku sharon jitsugisho (Chūgoku shozō)’ ロシア所蔵ウイグル語断片の研究2 : 『阿含経』『千字文』『阿毘達磨倶舎論実義疏』(中国所蔵), Kyōto daigaku gengogaku kenkyū 京都大学言語学研究 19 (2000), pp. 164–188.

66 On the Uyghur inherited pronunciation of Chinese, see M. Shōgaito, ‘How were Chinese characters read in Uighur?’, in Turfan Revisited: The First Century of Research into the Arts and Cultures of the Silk Road, (eds.) D. Durkin-Meisterernst et al. (Berlin, 2004), pp. 321–324; and Shōgaito, M., Fujishiro, S., Ohsaki, N., Sugahara, M., and Yakup, A., The Berlin Chinese Text U 5335 Written in Uighur Script: A Reconstruction of the Inherited Uighur Pronunciation of Chinese (Turnhout, 2015)Google Scholar.

67 Kornicki, Languages, p. 178.

68 On this, see Shōgaito Masahiro, ‘Roshia shozō uigurugo danpen no kenkyū 4: Senjimon (zokuhen), Dajō hōon girinjō’ ロシア所蔵ウイグル語断片の研究4 : 『千字文』(続編)『大乗法苑義林章』(続編), Kyōto daigaku gengogaku kenkyū 23 (2004), pp. 191–209; Shōgaito, ‘How were Chinese characters read in Uighur?’, pp. 321–324; and Shōgaito et al., Berlin Chinese Text U 5335, esp. pp. 157–168.

69 Shōgaito, ‘Roshia shozō uigurugo danpen no kenkyū 2’; Shōgaito, ‘Roshia shozō uigurugo danpen no kenkyū 4’.

70 Abdurishid Yakup, ‘An example of Old Uyghur's use of Chinese: interpretation of a colophon to the Qianziwen in the Berlin Turfan collection’, Acta Orientalia Hung. 76.4 (2023), pp. 529–542. The point about the apotropaic nature of the Qianziwen was initially raised by Nishiwaki Tsuneki 西脇常記, who mentioned a similar case in Chinese; see Nishiwaki Tsuneki 西脇常記, Doitsu shōrai no Torufan Kango monjo ドイツ將來のトルファン漢語文書 (Kyoto, 2002), pp. 81–82.

71 Yakup, ‘Example of Old Uyghur's use of Chinese’, pp. 536–537, also cites colophon-like scribbles on wooden objects that have a similar context and may be referencing the apotropaic function of the Qianziwen. The scribbles on wooden objects are described in S.-C. Raschmann, ‘Uygur scribbles on a wooden object’, in The Ruins of Kocho: Traces of Wooden Architecture on the Ancient Silk Road, (eds.) L. Russell-Smith and I. Konczak-Nagel (Berlin, 2016), pp. 42–48.

72 Zheng Acai and Zhu Fengyu, Dunhuang mengshu yanjiu 敦煌蒙書研究 (Lanzhou, 2002), p. 51, counts 37 manuscripts, but additional fragments have been identified since then. See also Nugent, Textual Practices, pp. 26–81.

73 Zieme, P., ‘An Old Uyghur translation of the 開蒙要訓 Kaimeng yaoxun’, Written Monuments of the Orient 7.1 (2021), pp. 71–99Google Scholar.

74 Kitsudō, K. and Galambos, I., ‘The story of Shunzi in Old Uyghur’, Acta Orientalia Hung. 73.3 (2020), pp. 451–466Google Scholar.

75 Songshi 宋史 485, p. 13995. Manuscript fragments of the Erya survived at both Dunhuang and Turfan; see Chō Narei [Zhang Nali], ‘Torufan bon Jiga chū ni tsuite’ 吐魯番本『爾雅注』について, in Tonkō, Torufan chutsudo kanbun monjo no shin kenkyū 敦煌・吐魯番出土漢文文書の新研究, (ed.) Dohi Yoshikazu 土肥義和 (Tokyo, 2009), pp. 365–389.

76 On Tangut translations of the Chinese classics, including the Xiaojing, see Kolokolov, V. S. and Kychanov, E. I., Kitajskaja Klassika v Tangutskom Perevode (Lun’ juj, Men dzy, Sjao tzyn) (Moscow, 1966)Google Scholar.

77 Grinstead, E. D., Analysis of the Tangut Script (Lund, 1972), pp. 277–376Google Scholar; Hu Ruofei 胡若飛, ‘E cang Xixiawen caoshu Xiaojing zhuan zhengwen yikao’ 俄藏西夏文草書《孝經傳》正文譯考, Ningxia daxue xuebao (Renwen shehui kexue ban) 寧夏大學學報(人文社會科學版) 28 (2006), pp. 14–17; and Nie Hongyin, ‘Lü zhu Xiaojing kao’ 呂注《孝經》考, Zhonghua wenshi luncong 中華文史論叢 2 (2007), pp. 285–306.

78 Nie Hongyin and Shi Jinbo 史金波, ‘Xixiawen Sancai zazi kao’ 西夏文《三才雜字》考, Zhongyang minzu daxue xuebao 中央民族大學學報 6 (1995), pp. 81–88. A full translation into Russian is available in A. P. Terent'ev-Katanskij and M. V. Sofronov, Smeshannye Znaki Trekh Chastej Mirozdanija (Moscow, 2002).

79 A related Chinese primer from Khara-khoto (initially erroneously catalogued as being from Dunhuang) is Дх-2822, which also seems to be a local composition. For its relationship with the Tangut text, see Wang Shizhen 王使臻, ‘Ecang wenxian Дх.2822 zishu de laiyuan ji xiangguan wenti’ 俄藏文獻 Дх.2822’字書’的來源及相關問題, Xixiaxue 西夏學 5 (2010), pp. 116–125.

80 For a more detailed description of this text, see I. Galambos, Translating Chinese Tradition and Teaching Tangut Culture: Manuscripts and Printed Books from Khara-Khoto (Berlin, 2015), pp. 156–161.

81 Reflecting upon its deficiencies, Nie Hongyin calls it the most unsophisticated text among Tangut texts of this type; see Nie Hongyin, ‘Xixiaben Jingshi zachao chutan’ 西夏本《經史雜抄》初探, Ningxia shehui kexue 寧夏社會科學 5 (2002), pp. 85–86.

82 Huang Yanjun 黃延軍, ‘Xixiawen Jingshi zachao kaoyuan’ 西夏文《經史雜抄》考源, Minzu yanjiu 民族研究 2 (2009), pp. 97–103.

83 Ibid., pp. 98–99.

84 On the different possibilities concerning the relationship between the Chinese and Tangut texts, see Galambos, Translating Chinese Tradition, pp. 160–161.

85 For an edition with Russian translation, see Solonin, K. J., Dvenadtsat’ Tsarstv (St. Petersburg, 1995)Google Scholar.

86 Songshi 203, 5097. For the identification of the text, see Sun Yingxin 孫穎新, ‘Shi'erguo de Xixiawen yiben’《十二國》的西夏文譯本, Minzu yuwen 民族语文 6 (2003), pp. 13–21.

87 Kwanten, L., The Timely Pearl: A 12th Century Tangut-Chinese Glossary (Bloomington, 1982)Google Scholar.

88 For a fascinating new discovery of a Khitan translation of a Chinese historical work and the current state of its decipherment, see V. P. Zaytsev, ‘Identifikatsija kidan'skogo istoricheskogo sochinenija v sostave rukopisnoj knigokodeksa Nova N 176 iz kollektsii IVR RAN i soputstvujushchie problemy’, Acta Linguistica Petropolitana XI.3 (2015), pp. 167–208.

89 At most, bits and pieces of text, some of which may come from didactic primers, can be identified as quotes and allusions in extant inscriptions; see Ōtake Masami 大竹昌巳, ‘Kittan shōji bunken shoin no Kanbun kotenseki’ 契丹小字文献所引の漢文古典籍, Kotonoha 152 (2015), pp. 1–19; and Ōtake Masami ‘Kittan shōji bunken shoin no Kanjin tenko’ 契丹小字文献所引の漢人典故, Kotonoha 160 (2016), 1–18.

90 Jinshi 金史 8, p. 184. H. Franke and D. Twitchett, ‘Introduction’, in The Cambridge History of China, (eds.) H. Franke and D. Twitchett (Cambridge, 1994), vol. 6, pp. 33–36, offers a general overview of the translation efforts by the Khitans, Jurchens, and Tanguts.

91 On written translations into other languages in East Asia, and the similar techniques employed in such projects, see Kornicki, Languages, pp. 187–213.

92 A Mongolian translation written in the ‘Phags-pa script is mentioned in the Yuanshi 元史 (22, p. 486) as having been presented to the throne in 1307.

The pre-classical Mongolian version of the Xiaojing inspired a rich body of scholarship, most of which is listed in de Rachewiltz, I., ‘The missing first page of the preclassical Mongolian version of the Hsiao-ching: a tentative reconstruction’, East Asian History 27 (2004), p. 51, n. 1Google Scholar; see also de Rachewiltz, I., ‘The preclassical Mongolian version of the Hsiao-ching’, Zentralasiatische Studien 16 (1982), pp. 7–109Google Scholar.