Sinoxenic writing and Chinese minority 
literature

The Chinese writing system has had a sweeping effect on textual traditions across the Sinosphere (nations in Southeast and East Asia that have historically been significantly influenced by China), famously being adopted to write the nearby languages of Japanese, Korean, and Vietnamese. ¹ However, China itself occupies a vast territory that contains a wide range of distinct ethnic minority groups for whom Chinese is not a native language. With what script, and in what language, do China’s minority peoples write? While some ethnic groups possess their own indigenous scripts (such as the Tibetan and Mongolian script), many groups have historically relied on borrowing “sinographs,” otherwise known as Chinese characters, a practice that is particularly prevalent among the peoples of south China. This borrowing resulted in the creation of what are known as Sinoxenic scripts, forms of writing that use Chinese characters (sometimes modified or even newly formed) to write non-Sinitic languages.

Chinese character scripts used within China (perhaps the most well-documented of which are Sawndip Zhuang and Sinoxenic Bai) developed alongside the expansion of the Chinese state that began in the Qin dynasty, and their reach extended throughout south China and into northern Vietnam. The first recorded usages of Sawndip Zhuang date to the Tang dynasty, while the earliest extant example of Sinoxenic Bai script dates back to the 12th century (although the script itself is almost certainly much older). Both were used to compose songs and stories, as well as to record historical records. Despite this long history, it is not always fully recognized within China that speakers of minority languages use sinographic character scripts to write their native languages (and indeed have been doing so for centuries). ² Consider the Bouyei (Buyi 布依) people. The Bouyei, who speak a Tai-Kadai language and have a population of some 2.5 million, primarily reside within Guizhou province, as well as some parts of Yunnan and Sichuan. In a recent article on minority languages, we learn that “Originally, the Bouyei language had no script. In the 1950s, the Chinese government facilitated the creation of the Bouyei script. However, the government failed to popularize the script. Chinese characters are now commonly used by the Bouyei people” (Zhou, 2019: 140). The first sentence is correct only in the sense that the Bouyei did not invent a script ex nihilo; there is in fact plentiful evidence of a Sinoxenic Bouyei script in use prior to the modern period, while these days, of course, written standard Chinese has become the norm. The fact that a Sinoxenic script existed prior to the creation of a Romanization system and subsequently the widespread adoption of standard Chinese makes the last sentence somewhat problematic: both scripts (Sinoxenic Bouyei and standard written Chinese) are, after all, based on Chinese characters.

Another major problem in identifying Sinoxenic writing today, concomitant with the issue of separating a writing system from a language, appears to be the lack of clearly defined textual evidence: many examples are not independent texts in their own right, but can instead be found as fragments embedded within what are ostensibly works of Chinese literature. ³ To illustrate this point I will use a few examples from Achang (阿昌) minority literature. The Achang are one of China’s 55 officially recognized ethnic minority groups, and Achang is a language spoken amongst some 80,000 Achang people who reside in China’s Yunnan province and Shan State in Myanmar. When it comes to writing, it is often stated that they do not possess their own script: “The Achang language has a strong oral tradition though Mandarin or Burmese are used as literary languages” (Minahan, 2014: 2). This would suggest that the Achang use the dominant literary language of the region in which they live (China or Myanmar). But how do they write in the vernacular? Within China, it is often assumed that, quite simply, such minority groups do not, and that “Achang literature,” for example, is composed in Chinese by Achang people. Zhou Qingsheng states unequivocally that “The Achang language has no script” (2019: 143). Another denial of writing can be seen in Yuan et al. (2019: 317), which declares that “there is no written language” for Achang. Chen Qiguang, in an appendix to his “Overview of Chinese languages and scripts,” notes that the Achang do indeed have a Romanized script that was developed during the 1980s and was still in use at the end of the previous century (1990: 310). I cannot, however, find any material written in this script, although a few online news reports do mention its existence (if not any tangible evidence of said existence). It seems to have been developed by local linguists based on Hanyu pinyin (as the majority of the Romanization systems created in the late 1950s for China’s minority languages were), but never officially implemented. Such, one assumes, is the fate of a newly invented script for an endangered language of only 30,000 speakers in China.

Discussions of language use in Achang literature (and indeed much minority literature in general) have generally been restricted to what “ethnic flavour” may be brought by the author into their Chinese writing, a kind of analysis which seems to constrain the writing within tightly controlled linguistic borders. By way of an example, the following line from Luo Han’s story Ziwu 紫雾 [Purple Mist] (as discussed by Zhang et al., 2016: 11), “只要大哥愿意, 即使把我这头割下来给大哥当尿壶, 我也没有任何的怨言” (“as long as elder brother wishes it, even if I were to cut off my own head and give it to him to use as a chamber pot, I would have no complaints”), has been said to display the Achang people’s bold and forthright nature. Such analysis merely conforms to essentializing tropes of minority literature as simplistic and “pure.” However, this is not vernacular Achang; rather, it is an (admittedly vivid) Achang turn of phrase translated into Chinese and presented to the reader as a verbal souvenir of local culture. We return to the earlier question: how, then, do the Achang write the vernacular? The answer is simple: they borrow sinographs. ⁴ When Achang author Cao Xianqiang (2015: 47) writes about traditional mountain songs in his Chinese short story Wanwan de shanlu wanwan de ge 弯弯的山路 弯弯的歌 [Meandering mountain road, meandering song], he is in fact composing in two languages, Achang and Chinese:

丹嘛俩嘛谷赫呶; 谷赫呶, 丹嘛俩嘛谷赫呶。

大田大地谷黄了; 谷黄了。大田大地谷黄了。

[“The grains in the vast fields of the earth are yellow [and ripening], the grains are yellow. The grains in the vast fields of the earth are yellow.”] ⁵

While Cao Xianqiang’s story is written primarily in standard Chinese, the first line above is clearly not Mandarin; rather, it is written in vernacular Achang using a borrowed script: “Sinoxenic Achang.” There is no way for a Mandarin speaker to make sense of the words that come before the translation, apart from one single graph, gu 谷, which is the same in both the written Achang and the Chinese translation. The sinograph gu 谷stands here in Achang for the word ku³¹, which itself is a loanword from Chinese gu, “grain” (see Yuan, 2001: 116). In this case gu is used in Achang as a “borrowed graph” that corresponds in its graphic, phonemic, and semantic totality to the borrowed word in the source language. ⁶ However, the song otherwise consists completely of phonetic borrowings. This is not just a single word or phrase of Achang language, but several lines from a song, fully written out. Cao himself notes in a preface to the story that his story is the first “romance novella” belonging to the Achang people (“这是阿昌族第一个中篇风情小说,” 2015: 3). Cao here undertakes the role of a cultural mediator and a true author of “minor literature” following Deleuze and Guattari, for his stories assume a “collective value”; he is speaking for all Achang people. Cao places the traditional songs and dances of the Achang at the heart of his story’s narrative, at once preserving and transmitting his native culture. This includes the notation and translation into Chinese of many traditional Achang songs:

嗯, 窝罗! 棒蚌剥果瓦湿果。瓦湿果, 棒蚌剥果瓦湿果。

嗯, 窝罗! 攀枝花开满寨红。满寨红, 攀枝花开满寨红。

[Oh, woluo! The kapok trees are in bloom and the whole village is red. The whole village is red, the kapok trees are in bloom and the whole village is red.] (p. 46)

In the above lyrics, Cao retains the Achang word 窝罗 (Chinese: woluo) in the translation without providing a Chinese equivalent. The 窝罗 is, we learn at the beginning of the story, the traditional name of an Achang dance: “这个叫阿昌的民族, 唱着古歌, 跳着古朴的’窝罗’ (舞蹈)” [“this people, known as the Achang, sing their ancient songs and dance the age-old ‘woluo’ (a form of dance)”] (p. 2). Given the nature of Chinese as a lingua franca – and Chinese characters as a scriptura franca – it is hardly surprising that contemporary minority literature in China is predominantly composed in Chinese. However, within these texts we can see evidence of native languages: this is to say that minority authors also use Chinese script to write what are, in many cases, non-Sinitic languages.

Now, this does not mean that any such instance of transcription or script borrowing counts as evidence of a minority “Sinoxenic script” per se. All of China’s minority peoples have borrowed sinographs in this way. Examples such as the Achang lines cited above may show the use of borrowed Chinese script to write word(s) in a different language, but they do not necessarily represent a systematic, fully developed script in its own right. Such usage can therefore be labelled Sinoxenic writing, but not necessarily Sinoxenic script. What an analysis of contemporary minority literature in China shows is the continued prevalence of this particular strategy of writing ethnic languages. This article is an attempt to historicize this phenomenon by suggesting that these are not necessarily simply haphazard phonetic transcriptions, but instead examples of minority language composition tapping into a longer tradition of Sinoxenic writing. I posit in this paper the existence of historical evidence of Sinoxenic Naxi, hitherto undocumented, which indicates that even those minorities with their own indigenous scripts developed a tradition of Sinoxenic writing that can still be found in works of contemporary minority literature. Further, I suggest that all of China’s minority authors engage in Sinoxenic writing (albeit in mostly limited forms) when they write their native languages using Chinese graphs. Put simply, minority authors in today’s China who write in Chinese often do not just write Mandarin Chinese, but also a Sinoxenic form of their native languages. In this way I am adopting a broad definition of Sinoxenic writing, following Holm’s succinct explanation of such scripts as writing “in which the Chinese character script has been borrowed” (Holm, 2013: xi), ⁷ and suggesting a split between the “old Sinoxenic writing” of the manuscript sources and the “new Sinoxenic writing” to be found in modern works of literature.

Something old, something borrowed, something new

The history of Sinoxenic writing begins with the spread of Chinese characters into China’s borderlands, where it was adopted by non-Han communities as a literary script. This was a process that began in the Qin-Han period and continued into the Tang and Song dynasties. The spread of Chinese script was the foundation for the formation of Sinoxenic scripts. Wang Feng (2003), whose own expertise lies in the study of Sinoxenic Bai, suggests one natural drawback of Chinese writing as a literary script in minority regions: the disconnect between the script and the spoken language. It was perhaps owing to this that minority peoples began to borrow Chinese graphs to write their vernacular. These borrowings can, in general terms, be divided into four categories: “borrowed graphs,” which correspond, in their phonemic and semantic totality, to the borrowed words in the source language; phonetic loans; semantic loans; and graphic loans (the borrowing of a Chinese graph that is not related to the phonetic or semantic values of the Chinese language).

Wang Feng (2003) suggests that relying on such borrowings alone made it difficult to accurately portray a minority language, and this led some Sinoxenic writing systems to modify existing Chinese graphs, or to simply create new ones. The upshot of this development is that Sinoxenic scripts employ two general methods of depicting written words: borrowing characters (jiajie Hanzi 假借汉字) and creating characters (fangzaozi 仿造字). In the first category, as we have seen, a Chinese character is taken for either its semantic properties (xundu Hanzi 训读汉字) or its pronunciation (yindu Hanzi 音读汉字), and in rare cases both. ⁸ I will illustrate the point by using a few examples from a rice-paper booklet recording a repertoire of Bai folk songs in Sinoxenic Bai script (see Fu et al., 2015, for a detailed annotation and translation of this text). The booklet was discovered in Jinquan county in 1958, and dates from before 1930. A semantic borrowing from this text would be the sinograph lao 老 “old,” which is used to mean the same thing in Bai but is read as the Bai word ku³³ (gux). A phonetic borrowing would be the sinograph nao 脑 “brain,” pronounced nɔ³³ (not), similar to its phonetic value in Chinese, which is used for the second-person pronoun “you” in Bai (Fu et al., 2015: 10–11). ⁹

While Sinoxenic Bai has been largely supplanted by a Latin alphabet Bai script developed in 1958 for the purposes of vernacular writing, some of the old phonetic borrowings persist in “new” modern-day Sinoxenic Bai. In the following line from the old Sinoxenic Bai songbook, we can see the final Bai verb, zɯ⁴⁴ tɯ⁴⁴ (seinx de) “to know”, represented by two phonetic loan characters, ren deng 认登: 阿朵佷认登/ˤa⁵⁵ to²¹ ni²¹ zɯ⁴⁴ tɯ⁴⁴, “Nobody knows the truth” (Fu et al., 2015: 24). In new Sinoxenic Bai writing, this verb has been written in a couple of different ways, ren deng 仁登 (Yang, 1994: 51) and ren deng 任登, for example “任登姆(知道吗)?” (p. 61), the Chinese ren deng (mu) being read in this case as the Bai seinx de (mox) “do you know?” These ways of writing seinx de, centuries removed, have different phonetic loans for the first syllable, seinx (i.e. homophones with different tones), but all share the same second sinograph, deng 登. We already know that Sinoxenic writing in general is not a standardized practice, 10 but the existence of common transcriptions surviving across the centuries does suggest something of a shared tradition, if not necessarily a direct inheritance.

An example of the “creating characters” category would be the graph for the Bai word, ta⁴⁴ (da), meaning “here.” In song 148 from the Bai songbook (titled “My love, you went home the other day”), the word is written by way of a newly formed character consisting of two primary elements: the radical chuo “辶” and the graph jiu “久.” Combining these two elements we get the Sinoxenic Bai graph, (Fu et al., 2015: 177). In new Sinoxenic Bai, however, the phonetic loan da 达 has been used for this word, for example in the following dialogue written by Bai author Yang Tengxiao in the short story “Yinhuaguo” 隐花果: “阿妈, 老芝梅仁登我珍阿达?” (in Romanized Bai, “At mox, not zil mel seinx de ngot zex al da?”), which is given the Chinese translation in parentheses “你咋个晓得我在这里?” (Yang, 1994: 77, English: “How did you know I was here?”). Note that the Chinese translation does not include the first two graphs, ama 阿妈, the Bai at mox (mother), but the meaning of this word would no doubt be obvious to the Chinese reader as meaning “mother,” as ama 阿妈 is also a common Chinese dialect word with the same meaning.

The modification and creation of sinographs was probably motivated by a desire to highlight the difference between the Chinese language and the minority language. An example is the old Sinoxenic Hani graph ba’la (meaning “moon”), which is the sinograph yue 月 (moon) with the addition of a circle around it (see Lu and Wang, 2020: 3). ¹¹ This character features prominently in an old Sinoxenic Hani text known as “the sun [and] the moon are born” that tells of the Hani cosmogenesis. ¹² This text was first printed in Wang Ersong’s 1994 monograph on the Hani alongside a phonetic transcription and both literal and free translation. The title in old Sinoxenic Hani is as follows (Wang, 1994: 69):


beelma/lama/divq

Sun/moon/born

The five graphs represent two disyllabic nouns, the sun, and the moon, followed by the verb “to be born.” Both nouns are poetic terms for the planetary bodies (the vernacular Hani words are naolma and ba’la). There is, however, no immediate reason why an unmodified 月 itself could not be read in Hani as lama/ba’la, the graph acting as a straight semantic loan. It would seem that the addition of the circle (very much not a traditional Chinese radical) only serves to distinguish this graph from standard written Chinese.

Following the logic of gradual separation, the final stage in Sinoxenic script development would be the creation of a phonetic alphabet or syllabary (such as Hiragana and Katakana in Japanese), representing an (albeit sometimes only partial) emancipation from sinographs. However, for minority ethnic groups within China, the reverse seems to have happened: after a period of gradual separation whereby some “new” characters were created, the Sinoxenic writing of most ethnic groups has reverted in the modern period to being predominantly comprised of phonetic loans. That is, Sinoxenic writing has by- and-large become a wholly phonetic syllabary using conventional Chinese characters. ¹³ Continuing with the Hani example used above, in a 1998 novel by Hani author Lang Que 朗确, we can see the word “moon” (Hani: ba’la) appear in new Sinoxenic Hani as two phonetic loan graphs, 巴拉 (Chinese: ba la): “巴拉、巴拉, 送我一个像你一样娇美白净的阿布吧!” (“Oh ba’la, ba’la, send me a daughter as delicate and white as you!”; Lang, 1998: 39). The difference between the two languages is still being highlighted, as “moon” appears as an untranslated Hani word within the Chinese text, but it is not as pronounced as to warrant a modified sinograph, as was the case in old Sinoxenic Hani. Lang Que writes a parallel line in reference to the sun “娜玛, 娜玛, 赐我一个像你一样活艳艳的儿子吧!” (“Oh naolma, naolma, gift me a son as bright and lively as you!” (p. 39). We can see, then, that a new Sinoxenic re-writing of the title of the old Hani cosmogenesis story could be composed entirely of phonetic borrowings, jettisoning the created graph for “moon”: “娜玛巴拉迭”, “naolma ba’la divq” (here using lower register vocabulary for the sun and moon).

By way of explaining this gulf between old and new Sinoxenic writing, Wang (2003) proposed that character-based writing in China goes through several formative stages, from borrowed characters to invented characters, then back to borrowed characters. These stages coincide with the emergence of an independent “ethnic” consciousness or indeed the degree of Sinification experienced by the group. A Sinoxenic script displaying a large number of newly created characters suggests a stage least influenced by written Chinese, while a preponderance of phonetic loans indicates a stage where Chinese language and script exerts a greater influence over society. The spread of Chinese literacy in minority regions has no doubt limited the development and usage of newly created characters in Sinoxenic writing, and research has shown that in regions and periods where contact between the centre and the periphery was more robust, some 90 percent of the characters in a text were borrowings (Fu et al., 2015: 11–12).

Sinoxenic Naxi: from semantic to phonetic borrowings

Just how many traditional Sinoxenic scripts exist? In a major study of Sinoxenic writing, Wang Feng (2003) writes that there are a total of 16 such scripts, of which nine are found in south China, a melting pot of minority groups. These nine scripts are sawndip Zhuang and Bai script, Zinan (otherwise known as the Jing script, or Zinan), ¹⁴ Dong script (leec Gaeml), ¹⁵ Bouyei script, ¹⁶ Miao script, ¹⁷ Yao script, ¹⁸ Hani script, ¹⁹ and Shui script. ²⁰ Notably, Wang also states that “as there are many ethnic groups in the south, and their usage of writing is complex, as research in this area deepens, we may discover other Sinoxenic scripts in the future” (p. 14; my translation). Indeed, even those minority groups with their own indigenous character-based writing systems have also developed sinographic writing. Two historical texts in particular may point toward the existence of a Sinoxenic script among the Naxi, 21 and both feature borrowings that can be compared to those found in the new Sinoxenic writing of contemporary literature, revealing a shift toward phonetic borrowing.

The primary evidence of an old Sinoxenic Naxi can be found in the Genealogical Chronicles of the Mu Family (Mushi huanpu 木氏宦谱), dated to 1516. This manuscript records the history of the old Mu kings (or chiefs) of the Naxi people. The botanist explorer (later turned Naxiologist) Joseph Rock copied the manuscript in Lijiang sometime between 1931 and 1932. In his 1947 anthropological work The Ancient Na-khi Kingdom of Southwest China, Rock published a translation of the Chinese text, including the section written in Sinoxenic Naxi, alongside its equivalent in Naxi dongba script (dobbaq tei’ee, an indigenous logographic script), the latter taken from a Naxi manuscript in the Mee biuq (sacrifice to heaven) ritual tradition (see Rock, 1947: 77–86). The Genealogical Chronicles begins with 11 sentences of Sinoxenic Naxi that narrate the genesis of the gods. This section is followed by an enumeration in Sinoxenic Naxi of the 12 legendary generations of Naxi ancestors. Rock originally believed these Sinoxenic sections to be incomprehensible Chinese, that is, until he translated a related manuscript written in the Naxi dongba script, a process which led to him being able to go back and decipher the Sinoxenic text. Rock noted that the traditional owners of the manuscript (the descendants of the Mu family in Lijiang) ²² could not explain the meaning of the Chinese characters to him. This is hardly surprising, as Sinoxenic texts are almost always assumed by the uninitiated to be written in indecipherable Chinese; only when read in a different language entirely can they be properly deciphered. ²³ Either this section of the Chronicles of the Mu Family was a highly idiosyncratic usage of Chinese to render Naxi writing, or evidence of the existence of a Sinoxenic Naxi script. Whether or not one agrees on the latter point, it is still, by definition, a form of Sinoxenic writing.

Rock identified the opening 11 sentences as a record of a song sung during the sacrifice to heaven ceremony of the Naxi. His realization led him to the following conclusion: “These are Na-khi sentences partly translated into Chinese, while some of the characters must be read as phonetic transcriptions of Na-khi words” (p. 76). Let us take the first sentence as an example. I’ll leave it to Rock to explain in full:

First sentence – The sentence in Chinese is as follows: 草古天能古 Ts’ao ku-t’ien-neng-ku. This translated literally would be: “Grass old heaven able old,” which, of course, would have no meaning. It must be read in Na-khi: Ts’o-gv-muan nnü-gkv. It can readily be seen that the Chinese characters Ts’ao-ku stand for Ts’o-gv, the third character t’ien (heaven) is a translation of the Na-khi muan, which means heaven, there being no such sound as muan in Chinese. The character neng is the nearest approach to the Na-khi nnü, and the last character ku is here read in Na-khi gkv. The literal translation of the sentence is: “Man egg heaven it laid.” (p. 77)

What we have here are five morphemes depicted by five sinographs, 草古天能古, read in Naxi as “Coq gv mee nee gv,” indicating the meaning “The egg of mankind was laid by heaven.” Four of the graphs are phonetic borrowings, while one is semantic.

I contend that the Chinese characters can be read not merely as transcriptions, as Rock suggested, but as Naxi words in Sinoxenic script. In his exegesis, Rock highlights a mixture of phonetic and semantic borrowing, and goes on to blame the phonetic inventory of Chinese for being unable to represent Naxi words:

In some instances the phonetic value of the Chinese characters render Na-khi words only, and their ideographic value had to be discarded, while in others the meaning of the character had to be read in Na-khi and the phonetic value rejected. The reason for this mixture of phonetic and translation rendering is the phonetic paucity of Chinese. Where the sound complex of a word could not be rendered it was translated. (p. 70).

Based upon the earlier discussion, it is possible from this description to identify the sinographs as a form of Sinoxenic writing featuring both phonetic and semantic borrowings, but Rock instead indicated that it was comprised of phonetic transcription with some additional translation. The difference between phonetic transcription and Sinoxenic writing is, I feel, an important one. Phonetic transcriptions are usually read in the transcribing language, whereas Sinoxenic writing is to be read in the native language (in this case, Naxi). For example, the Chinese cao 草 can operate as an approximate phonetic transcription of the Naxi coq (it retains its phonetic value in Chinese), whereas in Sinoxenic writing, the graph 草 is used to directly represent the Naxi coq (its phonetic value is Naxi). It is not a transcription of coq, it is coq.

There are two borrowings in this first sentence that are worthy of further discussion. First, the use of Chinese gu 古 for the Naxi gv, “egg,” has survived intact in new Sinoxenic Naxi. In a treatise on Naxi culture by He Shihua (2007: 196), we see the following terms: 古珊 (Chinese: gu shan) to be read gv zzaiq (mottled egg) and 古徐 (Chinese: gu xu) to be read gv xiuq (red egg). Second, “heaven” is today not written with a semantic borrowing (i.e. Chinese tian 天), but has become a phonetic borrowing, 美 (Chinese: mei) for Naxi mee. That is to say, the “phonetic paucity” of Chinese as Rock noted is an irrelevant issue if Naxi readings are attributed to the graphs. We can see examples of authors referring to the sacrifice to heaven as “美补,” using the graphs (standard Chinese: mei bu) to stand for the Naxi “mee biuq.” See, for example, He and Yang (1992: 188): “祭天, 纳西语称‘美补’” (“The sacrifice to heaven is called ‘mee biuq’ in Naxi”). In He Shihua’s work, we see the following saying in Sinoxenic Naxi (He, 2007: 198): “纳西美毕若, 纳西美毕敌,” which reads Naqxi mee biuq ssoq, Naqxi mee biuq ddee (“The Naxi are people who worship heaven; for the Naxi heaven worship is the most important [ceremony]”). What was a semantic borrowing in old Sinoxenic Naxi–Chinese tian, to be read mee in Naxi (both indicating “heaven”), becomes a phonetic borrowing in new Sinoxenic Naxi.

It is possible to identify a range of phonetic and semantic borrowings in Sinoxenic writing to depict cognates of “heaven” in Tibeto-Burman languages across the centuries. An early Sinoxenic reading for “sky/heaven” in a Trans-Himalayan language can be found in the transcribed songs of the Bailang 白狼 people (of which Hill, 2017, is an excellent reconstruction). The songs appear in chapter 86 of the Hou Hanshu 後漢書 (History of the Later [Eastern] Han Dynasty), compiled by Fan Ye 范曄 (395–451). These three songs from the border of Sichuan and Gansu and presented to the Chinese court in 74 AD are recorded in Chinese transcription and translation, making them perhaps the earliest discovered Sinoxenic record of a Tibeto-Burman language. In the first of the Bailang songs, the graph “冒” (old Chinese mˤus) is used to indicate the idea of tian 天 “Heaven,” perhaps related to old Tibetan, dmu དམུ་, “a type of sky god” (Hill, 2017: 400). The Naxi language is said to be distantly related to the ancient Bailang language (see Guo and He, 1999), as the Naxi are descended from ancient Qiang tribes. It is my belief that, just as created characters appear to have gradually disappeared from Sinoxenic writing within China, semantic readings have also largely been replaced by phonetic borrowings. This is one of the major differences between old Sinoxenic writing and new Sinoxenic writing.

A further piece of textual evidence for the existence of old Sinoxenic Naxi can be found on the cover of a Naxi ritual manuscript in the collection of Leiden University Library. In blue ink, beneath the Naxi graphs that adorn the centre of the cover, there are a number of sinographs. Incomprehensible in Chinese, this is in fact a rather unique example of Sinoxenic writing. Such writing appears on two manuscripts in the Leiden collection (both likely penned by the same hand). These manuscripts were acquired from a private collector in 1998, and likely date back to the late Qing or Republican period of China. The sinographs are a later addition, written in simplified Chinese; they were most probably written in the second half of the 20th century. Manuscript OR25165 has only two sinographs on the cover, while OR25166 has two lines of sinographs, 12 in total. I will examine in detail the two lines from OR25166 (see Figure 1), which form an intriguing example of Sinoxenic writing.OPEN IN VIEWER

Figure 1.

Detail from the cover of OR25166. Leiden University Libraries.

At first these two lines seem meaningless, but reference to the dongba script found above them (see Figure 2) indicates that this is in fact an example of Sinoxenic Naxi script. ²⁴ The lines read:

In English: [At] Na saq wu gv bbuq

White alcohol, black alcohol, three kinds [are spoken of?]

Now, the most notable feature of these lines are the diacritic notations. In Figure 1 we can see two graphs with ring diacritics above them: at the end of the first line, shan, mountain, and the second graph of the second line, bai, white. While the 10 other graphs are phonetic borrowings, the ring diacritic is used here to indicate a semantic borrowing. The graph 山, read in Chinese as shan (mountain), is to be read in Naxi as bbu – meaning hillside. The usage of shan (hill, mountain) for Naxi bbu (hillside) suggests that the graph being borrowed is linked to a semantic domain and not a specific morpheme. Second, the graph 日 bai (white) is to be read in Naxi as perq, also meaning white. Both graphs are used as semantic borrowings. It would, of course, be easy to substitute phonetic borrowings for these syllables: bbu could simply be bu 布, and perq, white, has been transcribed in the Naxiological literature as both pan 盘 (see Yang, 2009) and pai 排 (Sha, 2001: 2).

It is well known that phonetic marks were used in Medieval manuscripts with the purpose of clarifying the correct pronunciation of a character. Ishizuka Harumichi (1993) described how the Tang dynasty practice of marks indicating the four tones of the Chinese language developed from the earlier poyin 破音 marks used for distinguishing a character’s default and derivative readings. The most basic form of poyin notation was a red dot placed above the middle of the character, indicating that it was used in a derivative sense. The presence of the rings in OR25166 suggests a quite sophisticated level of philological complexity for Sinoxenic writing, although (to my own knowledge) no other examples of this usage have yet come to light. How widespread was this notation? Could it have simply been the work of one enterprising Naxi religious specialist, or perhaps the work of a translator? In light of this diacritic, it could also be hypothesized that the addition of the circle to the sinograph for moon in the Sinoxenic Hani graph might act as a form of diacritic, suggesting that this graph is to be taken as a semantic loan: the circle not above the graph (as with , for example), but encompassing it.

In terms of the other borrowings here, two more are of note. The use of 三 (“three”) is probably a case of both phonetic and semantic borrowing (i.e. a loanword in the strictest sense), hence there is no ring diacritic, because the Naxi for three, see, is similar in pronunciation to the Chinese san. Finally, in the Naxi dongba script title, there is what looks like a sinograph in the centre, directly above the hillside to the right of the figure holding the bamboo stick (see Figure 2). This is the Naxi graph, she (“to say, to relate”). Technically, this is a syllabic Naxi graph borrowed from Chinese shang 上 (up, above). It is a phonetic and graphic borrowing, but not a semantic one. The 上 at the end of the Sinoxenic Naxi writing can therefore be conceptualized as a phonetic and graphic loan of what was already a phonetic and graphic loan. In this particular instance, then, what began with sinographs ends with sinographs.OPEN IN VIEWER

Figure 2.

Detail from the cover of OR25166. Leiden University Libraries.

The title of OR25166 refers to a specific place, and also to the three kinds of alcohol. The location appears in Rock as “²Na-¹ssaw -²wùa-²gv-¹mbu” (Naxi pinyin: na saq wu gv bbuq), with the gloss “Name of a place where pigs were killed in the winter” (Rock, 1963: 301). ²⁵ This is a place name that appears in a funerary ceremony. For this entry, Rock lists a compound graph with the phonetics na and sa contained within a village house, wu: . On the cover of OR25166, however, the first five syllables are depicted by only one compound graph, the second syllable, sa, located inside the hillside, bbv: . This Naxi writing reflects the difference between logographic Naxi (one compound logograph of only two graphic components being read as five syllables) and the syllabic Sinoxenic writing, where five graphs each represent a single syllable. When it comes to evidence of Sinoxenic script, however, we must be vary of exaggerating the importance of the “driftwood,” and it would be wise to be alert to the fact that a few broken twigs do not indicate the existence of a forest. Nevertheless, the ring diacritic is potentially evidence of the maturity of this particular mode of hybrid writing, and perhaps other similar examples will be uncovered in the future.

Conclusion: Sinoxenic writing and contemporary minority literature

The Sinoxenic writing that we can find in Naxi literature today (in publications falling within the genre of minzu wenxue 民族文学) does not make use of such diacritics, but this is in some way to be expected, as new Sinoxenic writing appears to have become entirely phoneticized, featuring only phonetic borrowings. There is no need to highlight semantic borrowings when none exist. Some authors have written whole sentences and dialogues in new Sinoxenic Naxi (as with the examples of such renderings in Bai already shown). An example from Naxi author Yang Zhengwen’s collection of short stories would be the following line of dialogue: “‘马乃马乃! 纳罕若米底吉瓦嘛!’ (纳西语: 不谢不谢! 都是一家人罗)” (Yang, 2008: 66). Again, despite being written seemingly in Chinese, this is in fact two languages. The dialogue is spoken in Naxi and written in Sinoxenic Naxi, while Yang provides the translation in Chinese in parentheses. In Naxi, this is read: Me nee me nee, Nahaq ssomi dee jiq waq [“Don’t worry, don’t worry, we’re all one family”].

Despite the inherent idiosyncrasies of this mode of writing, we can identify some borrowings that have become somewhat unofficially standardized across the literature of certain minority groups. Taking the rather common category of kinship terminology as an example, the Hani word for grandfather, aqbol, has been written in Sinoxenic Hani as 阿波 (Chinese: abo) in Hani literature by several different authors: “那天, 在阿波家的茅草屋里, 我吃上了最好吃的糯米粑粑” (“That day, at aqbol’s thatch cottage, I ate the tastiest sticky rice cake”; Li, 2013: 227). In addition, as another example, “九个神急了: ‘怎么办呢, 阿波, 我们要死了!’” (“The nine deities became agitated: ‘What shall we do, aqbol, we’re going to die!’”; Liu and Lu, 1989: 20). Similarly, the Achang word for “grandmother”, ɑ³¹ ʑɑ³³, has been written 阿娅 (Chinese: aya, sometimes also 阿丫). In Zhao Jiajian’s story, Naliang 纳凉: “我那60岁的阿娅坐在屋前把我抱在怀里” (“My 60-year-old “aza” sat in front of her house with me in her arms,” 2015: 256). ²⁶ Cao Xianqiang also uses the same borrowed graphs in his writing: “‘阿袍’、‘阿娅’们则在那里给孩子们摆古弄经” (“The ‘ɑphɑu’ and ‘aza’ would tell stories to the children there”; Cao, 2012: 198). Such terms are likely to be commonplace borrowings; that is, they have been reached by some form of implicit consensus, no doubt by virtue of their oft-recurring nature. These consensus borrowings are noteworthy precisely because of the general lack of standardization in Sinoxenic writing, which has resulted in the same words being written using different graphs even within an individual text (Holm, 2013: 1).

Contemporary Sinoxenic writing, as has been described in this article, depicts what can be generally categorized as vernacular non-Chinese language within fiction or non-fiction texts that belongs to the genre of minority literature; it is not indicative of a standardized script used for bureaucratic or other official purposes. While Sinoxenic writing, as can be seen within Chinese minority literature, does not directly indicate the perseverance of systematic Sinoxenic writing in the same way that Holm has, for example, documented for Zhuang, these Sinoxenic readings are more appropriate in their context than Chinese language readings, and they point to the still wider possibilities of this mode of writing, perhaps even toward avenues of development for minority writing as a genre in China. For scholars and writers alike, the borders of “Sinography” have room to be widened further still.