Public Inscriptions and Manchu Language Reform in the Early Qianlong Reign (1740s– 1760s)1

Official signs and steles in the Qing empire were often at least bilingual, with Manchu and Chinese either side by side or on opposite sides. New institutions and renovated buildings necessitated new Manchu words and turns of phrase for the accompanying inscriptions. Under the Qianlong emperor, the text that should go on plaques and steles was carefully scrutinized, especially regarding the use of Manchu vocabulary. Public inscriptions thus involved watchful editing. The emperor was not the only driving force behind the reform. In part, individuals who moved through Beijing and Manchuria saw the words inscribed in the builtup environment around them change because officials in the field had memorialized the throne to draw attention to usages that the initial reform had missed. This paper will discuss this process.

Public Inscriptions and Manchu Language Reform in the Early Qianlong Reign (1740s-1760s) 1

Institute of Modern History, Academia Sinica
Official signs and steles in the Qing empire were often at least bilingual, with Manchu and Chinese either side by side or on opposite sides. New institutions and renovated buildings necessitated new Manchu words and turns of phrase for the accompanying inscriptions. Under the Qianlong emperor, the text that should go on plaques and steles was carefully scrutinized, especially regarding the use of Manchu vocabulary. Public inscriptions thus involved watchful editing. The emperor was not the only driving force behind the reform. In part, individuals who moved through Beijing and Manchuria saw the words inscribed in the built-up environment around them change because officials in the field had memorialized the throne to draw attention to usages that the initial reform had missed. This paper will discuss this process.

大量新創碑文則是這審慎的訂正過程之結果。需要注意的是，乾隆皇帝並非 推動滿文改革的唯一動力，而是地方官員在施政過程中發現滿文使用不規範 之處繼而上奏，使得從北京到盛京作為公共景觀的碑刻得以一體改觀。本文 即探討這一歷史過程。
In Beneath the Red Banner (Zhenghong qi xia 正紅旗下), Lao She's 老舍 (1899She's 老舍 ( -1966 unfinished novel about Manchu life in Beijing around the turn of the twentieth century, there is a famous passage that deals with one of the protagonists' language skills. Cousin Fuhai-a kindhearted housepainter with many talents-has a very poor command of Manchu. Lao She's narrator informs the reader that this particular Manchu only ever spoke and wrote Chinese. In addition, Fuhai's unfamiliarity with Manchu affected how he interacted with writing in the environment around him: When he came upon a ceremonial plaque or stone inscription written in both Manchu and Chinese, he could appreciate the elegance or boldness of the Chinese characters, but he would only glance at the Manchu words out of the corner of his eye, as if he reverently wanted to maintain his distance.

当他看到满、汉文并用的匾额或碑碣，他总是欣赏上面的汉字的秀丽或刚 劲，而对旁边的满字便只用眼角照顾一下，敬而远之。 2
Lao She's point is that a late Qing bannerman like Fuhai, without any hope of obtaining an official appointment, did not know Manchu. What is interesting for the purposes of this paper, however, is that the passage shows that Manchu inscriptions were part of the urban landscape of Beijing. If that was the case around the turn of the twentieth century, when the novel is set, it was even more so in the eighteenth, when many Manchu steles and inscribed plaques were erected or installed along with the construction or renovation of temples and official buildings. On the eve of the Boxer Rebellion, those who like Fuhai walked through the Qing capital would have chanced upon mostly older Manchu steles and inscribed tablets and plaques. A century and a half earlier, however, city dwellers could gaze at new inscriptions. Indeed, the thenreigning Qianlong emperor was erecting Manchu steles for bannermen to read. 3 As a part of the public image of the Qing government, official signs and stele were often at least bilingual, with Manchu and Chinese either side by side or on opposite sides. Since stele were often erected to commemorate recent events, recently deceased servants of the throne, or in order to mark the completion of a construction project, their inscriptions reflected the development of the Qing from a small frontier state to a great empire. New institutions and renovated buildings necessitated new Manchu words and turns of phrase for the accompanying inscriptions.
Initially, the language of steles does not appear to have posed a problem to those who wrote the inscriptions. Steles, so prominently used in China since antiquity, were initially simply referred to in Manchu using the Chinese name bei 碑 or prosaically as "stone documents" (uwehei bithe). Eventually, however, a Manchu terminology developed. Not only did the steles themselves receive the new name of "shining stones" (eldengge wehe), but the institutions commemorated through steles and the places identified with plaques received new names as well. From at least the 1740s, under the Qianlong emperor, the archival record shows that the text that should go on plaques and steles was carefully scrutinized, especially regarding their use of Manchu vocabulary. Public inscriptions thus not only involved logistical and financial considerations and technical skill, but also watchful editing.
The Qianlong emperor initiated or continued several major Manchu literary projects, including the compilation of new reference works and the translation of Buddhist and Confucian texts into Manchu from Mongolian and Chinese. In the process, the Manchu lexicon was revised, and words considered Chinese loans were removed and replaced with new coinages. Often, the new expressions mimicked the Chinese in their semantic structure, but not in their form, which appears to have been what mattered most to the reformers.
Qianlong's reform of the Manchu lexicon has attracted the attention of researchers for some time. Outside the Qing empire, lexicographers such as Ivan Ilyich Zakharov (1816Zakharov ( -1885 noted the hundreds of neologisms introduced in Qianlong's official Manchu-Chinese dictionary Han-i araha nonggime toktobuha manju gisun-i buleku bithe | Yuzhi zengding Qingwen jian 御製增訂清文鑑 (Imperially commissioned mirror of the Manchu language, expanded and emended) from 1772-1773. 4 Erwin von Zach (1872Zach ( -1942 described the structure of many of the new words, 5 which he later called "artificial creations of the emperors" (Kunstprodukte der Kaiser). 6 Peter Schmidt (1869Schmidt ( -1938 offered more context when he called these "phonetically and orthographically unjustifiable" words the products of the "ardent linguistic purism" (Sprachreinin- Until the late twentieth century, discussions of the new words were based primarily on Qianlong's Manchu-Chinese dictionary. Increased access to various kinds of archival collections changed this situation. Notably, Zhang Hong, Cheng Dakun, and Tong Yonggong published archival sources that showed how the reform unfolded over many years before the publication of Qianlong's bilingual dictionary. 8 Even though Pamela Crossley and Evelyn Rawski initially contextualized Qianlong's "language purification" as part of a greater state-building project that they argued went back to pre-conquest times, 9 Rawski later discussed it in more specific terms on the basis of Zhang, Cheng, and Tong's reprinted archival sources. 10 Mark Elliott associated it to the emperor's view of language as essential to the maintenance of Manchu identity and his attendant fear of Manchu language attrition. 11 These scholars also noted the reform's links to the aforementioned major translation projects undertaken at court. However, the reform has yet to be studied in its full extent. This paper will contribute to the literature on Qianlong's language reform by looking at how it played out in the field of public inscriptions. I will show that the emperor was not the only driving force behind the reform. Individuals who moved through Beijing and Manchuria saw the words inscribed in the built-up environment around them change in part because officials in the field had memorialized the throne to draw attention to usages that the initial reform had missed. Rather than stress the link between discussions of vocabulary and Manchu identity or imperial ideology, I will focus on the choices and difficulties that reformers faced in specific cases. The paper will thus discuss this process of, as it were, the steles' hundred visions and revisions (before the taking of a toast and tea).
First, I will give a brief overview of the erection of Manchu stele since the preconquest period. Second, I will discuss a few examples of stele and inscriptions in the context of Qianlong-era language reform. My key sources are the endorsed originals and file copies of Manchu palace memorials (Manwen zhupi zouzhe 滿文硃批奏摺 and Manwen lufu zouzhe 滿文錄副奏摺) and Manchu court letters recorded in copy books (dangbu 檔簿) and held at the First Historical Archives in Beijing, supplemented by reprinted edicts, memorials, stele rubbings, and visual sources. These sources, while informative, unfortunately do not allow me to say much about how the public Manchu inscriptions were received. What I will offer here are the stele inscriptions as discussed by the emperors and officials responsible for them. 8. I will use and cite these reprinted sources later in the paper.

Background: Manchu Steles and Public Inscriptions
Steles erected in pre-conquest Manchuria included the kinds later put up in Beijing. The early Manchu rulers recognized the role of steles as public pronouncements of policy. In 1608, Nurhaci erected a stele (uwehei bithe, "stone document"; uwehe ilibuha gisun, "words established in stone" [old spelling uwehe > new spelling wehe]) 12 on the Ming-Jurchen border that announced that trespassing from either side should be punished by death. Ming officials came from Liaodong to confirm this policy through an oath and the sacrifice of a white horse. Nurhaci considered the end to Jurchen trespassing to be one of his great achievements. Nurhaci and Hong Taiji repeatedly referenced this stele while claiming that the Ming failed to heed its words and secure the border (one of the "seven grievances" that Nurhaci directed to the Ming). The stele's intended readership was not the commoners of either side of the border, but the authorities. The stele stood as a testament-a public proof-to a cross-border agreement. 13 Yet as the early Qing state expanded, inscriptions appeared in more visible locations. The gates to the walled towns that came under Manchu rule received Manchu names that were carved on plaques. Any literate traveler would have been able to read these public inscriptions. 14 Furthermore, steles continued to be important in the early Qing state's foreign relations. In 1638-1639, after the second Manchu invasion of Korea, a trilingual (Manchu, Mongol, Chinese) stele was erected in Korea and written in the voice of the Koreans ("our small country" [meni ajige gurun]). The Qing side provided the Koreans with the Manchu original. The Chosŏn side then translated it into Chinese and repeatedly 12. Kicentai 庄声 [Zhuangsheng], Teikoku o tsukutta gengo seisaku: Daichin gurun shoki no gengo seikatsu to bunka 帝国を創 った言語政策-ダイチン과グルン初期の言語生活と文化 (Kyōto: Kyōto daigaku gakujutsu shuppankai, 2016), 61 writes that it was also referred to as a šibei following the Chinese shibei 石碑. 13. The chronicle does not mention the language in which the stele was inscribed. Quotations from the stele are similar but not identical. They can represent paraphrases of an underlying Manchu text, but I do not think it can be excluded that they constitute ad hoc translations from the Chinese, which would then, it follows, have been the language of the stele inscription.
The stele is mentioned in Guang-lu 廣祿 and Li Xuezhi 李學智, trans., Qing Taizu chao lao Manwen yuandang: Di yi ce "huang" zi lao Manwen dangce 清太祖朝老滿文原檔：第一冊荒字老滿文檔冊 (Taipei: Taiwan zhonghua shuju, 1970), 8-9 (uwehei bithe), 38-39 (where the Chinese want to redraw the border and erect a new stele, later invoked as one of the "seven grievances"), 64-66 and 79-80 (where the failure to heed the original stele is one of the "grievances"); Guang-lu 廣祿 and Li Xuezhi 李學智, trans., Qing Taizu  revised it before they were able to satisfy the Qing side. The stele was simply called bei in Manchu, using the Chinese word ("took a stone and erected a stele" [wehe be gajifi bei ilibufi]). 15 Sometimes temples and stele were erected together as elements of foreign policy. At the site of a new temple, Hong Taiji in 1630 (Tiancong 4) erected a Manchu-Chinese stele to commemorate a Tibetan cleric who died in 1621 en route to Manchuria. The Manchu text is written in the old, unreformed script. 16 In 1638 (Chongde 3), Hong Taiji dedicated a temple a few li west of Mukden and gave it the name Yargiyan etehe fucihi [fujihi?] 17 soorin (Ch. Shisheng Si 實勝寺), "Buddha throne (i.e. temple) of true victory." Two stele were erected at the site. One of them carried a Manchu text, which called the inscribed stele wehei bei [< Ch. bei 碑] de araha bithei gisun, "Words in text written on a stone stele." 18 The fact that the other languages included Mongolian and Tibetan shows that this temple too was part of the Qing policy toward the Buddhist Mongols. 19 The court continued to erect steles in the post-conquest period. A new name was introduced to refer to steles. In 1708, the preface to the normative dictionary Han-i araha manju gisun-i buleku bithe (Imperially commissioned mirror of the Manchu language) called them "shining stones" (eldengge wehe), 20 a word that caught on and became standardized as the Manchu name for steles from 1749 at the latest. 21 The preface to the dictionary was nominally composed by the Kangxi emperor. It does indeed appear that the emperor himself came up with the new coinage. A draft version of the preface used the phrase "stele stones" (bei wehe). The emperor's vermillion brush crossed out the Chinese loan bei and replaced it with eldengge "shining." 22 Often, the occasion for erecting a stele was the commemoration of deceased im-perial princes or banner servants of the throne. 23 Furthermore, officials petitioned the throne to be allowed to put up such stele. 24 Erecting steles could be a complicated process, especially in the radically enlarged post-1644 empire. When a stele was to be erected in the provinces in the emperor's handwriting, officials in the field had to communicate the size of the stone to the capital. Furthermore, the individuals on site might in reality be better placed than the emperor to actually compose the text of a stele to be erected near a local institution such as a temple. A back-and-forth with Beijing over the final wording was sometimes the result. 25 Carvers, whose job was not easy, might also have to be dispatched from the capital. The paper carrying the manuscript for the inscription could get wet from humidity, which affected the appearance of the characters written on it. Transposing them accurately to the stone was difficult. If revisions were needed, the carver might have to remain on site for more than a month. 26 Upon completion of the stele (as of 1711 still called bei in Manchu), the original manuscript (da araha bithe) might even have to be sent back to Beijing. 27 Having the text of older inscriptions on hand in Beijing was evidently useful. In 1761 (Qianlong 26/2/13, 1761/3/19), after the conquest of what is now Xinjiang, Agūi (1717-1797) reported from the field that a set of two steles erected in Ghulja (Ma. Gūlya, in Ili) had sustained damage, "erasing the inscribed text so that it is no longer legible" (ede wehe de foloho hergen mahūlabufi tuwame turibume muterakū). Agūi requested that "the draft texts for these two stele be written out again and sent over for them to be carved again onto the original stone inscription by Fukui, who did the work in the first place" (ere juwe eldengge wehe-i bithei jise be dasame arabufi unggihe manggi, da weilehe fukui de nikebufi dasame weilebuki seme wesimbuhebi,). Qianlong agreed to copy out the "quadrilingual draft" (duin hacin-i hergen jise)-probably Manchu, Mongolian, Tibetan, and Chinese (if not Oirat)-and send it out to Xinjiang for recarving. 28 23. E.g., Zhongguo di yi lishi dang'an guan, Qing chu nei guoshi yuan Manwen dang'an yibian, vol. A decision taken a few years later reveals the particular difficulties involved in carving Manchu steles. In 1764 or earlier (before Qianlong 29/11/25, 1764/12/17), it was decided that steles should be put up at government schools across China proper to celebrate the victory in Xinjiang. Similar steles had been erected in the past to celebrate military victories. Suldei 蘇爾德 (fl. 1757-1778), financial commissioner in Jiangsu, had seen that in Suzhou, the Manchu versions of those past inscriptions (of texts dating from 1729 and 1754) "contained many mistakes in the brush strokes of the Manchu characters" 所鐫清字筆畫多有訛錯. Suldei corrected the mistakes. In order to prevent similar mistakes in the new inscription, he sent a bannerman to oversee its carving. The idea was that rubbings of the new stele should then be sent to five other locations in the province where identical steles would be erected. Yet because of the sheer size of the original, it was difficult in some places to source a stone of the right size and to find a location that could support it. Suldei proposed that the layout and size of the stele should be modified according to local needs and that "banner personnel literate in Manchu" 通曉清文旗員 be dispatched to trace the Manchu text in each case, thus replicating the process he had carried out for the first of the new inscriptions. 29 The arrangement that Suldei proposed was probably costly. The court decided (Qianlong 29/12/jiashen, 1764/12/29) that since "the literati of the outer provinces have never been proficient in the dynastic script, there is also no need to order it carved on the steles" 至外省士子，本不諳習國書，碑內亦可毋庸令其鐫刻. 30 Thus the erection of Manchu steles in the south was dispensed with, and no banner official had to be sent to supervise each carving.
In the case of Chinese steles, readers of the inscriptions were not limited to those who saw the original stone in person, but it is not obvious to what extent that was the case for Manchu steles. Rubbings of some stele were indeed very widely distributed. In 1705, Kangxi's inscription on the pacification of the Northwest was reproduced in ten thousand copies, yet it is unclear in what languages. 31 Furthermore, the texts of imperial steles were recorded in local gazetteers, 32 but probably in Chinese only. Finally, when couriers who carried an imperially authorized text for a stele inscription passed through a location, local officials might greet the party in a ceremony. 33 I assume that such events drew crowds. Perhaps such assemblies of the curious got a glimpse of the Manchu text on a blueprint sent from Beijing, but I do not think that the Manchu text would have been read out aloud for them.

Public Inscriptions and Qianlong-Era Language Reform
Reproductions of Manchu steles did not have the circulation that Chinese stele rubbings and transcriptions did, but still, many Manchu steles were erected and available for viewing, especially in Beijing. Restored or newly founded temples, for instance, required a public stone inscription. Not all of them were even erected on government initiative; seven temple steles with Manchu inscriptions written by individual officials, religious organizers, or anonymous bannermen are known from Beijing. 34 The vocabulary used in Beijing temple stele differed. For example, the stele for Fahai Si 法海寺, erected in 1660 in the suburbs of Beijing, called the temple Fa hai sy juktehen, where the last element is new word for "temple" and the first three elements are transcriptions of the Chinese name. 35 The old name, "Buddha throne" (fucihi soorin), was after 1645 no longer used for what in Chinese was called si. 36 The change in name caused an asymmetry between Manchu and Chinese nomenclature. The new Manchu name juktehen was also used to translate Chinese miao 廟, which referred to a Confucian temple, thus constituting an alternative to the transliterations miyoo and miyo often used to render Chinese miao in Manchu. However, at times Manchu miyoo, from the Chinese word for Confucian temple, translated Chinese si, a word that was most often (but not always) used for Buddhist temples. An altar, tan 壇 in Chinese, was as of 1685 simply called tan in Manchu. 37 In the Qianlong period, such public inscriptions in Manchu were subjected to official scrutiny and in some cases reform.
Under Qianlong, Manchu steles continued to be erected in the same contexts as before. However, unlike for earlier periods, there is a great deal of evidence to suggest that, in this period, the Manchu wording used in public inscriptions did not represent an ad hoc problem to be solved by whoever wrote the text. Rather, officials scrutinized the wording of Manchu steles and petitioned the emperor to standardize it. Thus Qianlong-era Manchu language reform, most famously represented by the imperially authorized lexicographical collections that were published at the Beijing palace, entered the public space. Not only officials, clerks, and banner students for the translation examination were exposed to Manchu neologisms and reformed expressions, but so was anyone who took the time to look over new stele that went up around the capital's temples and altars. In the following sections, I will look at, first, the names for places and institutions and, second, at a discussion about the translation of Buddhist vocabulary as used at a temple stele.
The Names for Temples, Altars, and Gates Names for temples, altars, and gates were revised as part of the larger overhaul of the prescribed Manchu lexicon. The emperor, the Grand Council, and provincial officials were actively involved in the language reform. In several instances, officials in the field memorialized the throne on language-related issues, seeking clarification on points of vocabulary.
To be sure, deliberations over the Manchu names for public institutions or places did not begin with Qianlong. Already in 1636, Hong Taiji opposed his officials' suggestion that the capital's city gates receive Manchu names translated from their Ming counterparts, e.g., goroki be toktobure duka, "Gate of the ordering of distant places" for Chinese dingyuan men 定遠門. Hong Taiji opposed such allegedly haughty and presumptuous language. He preferred names such as jecen be tuwakiyara duka, "Gate Watching Over the Frontiers," which was the one chosen in this case. 38 Manchu government sources from the Kangxi and Yongzheng periods contain the names of many of Beijing's city and palace gates as well as of palace buildings. The city gates clearly had established Manchu translations in this period, as they are encountered in the same form in different kinds of sources, even visual ones. 39 Some palace buildings had established names too. Their names were, like their Chinese counterparts, noun phrases with words like gung (< gong 宮 "palace") or diyan (< dian 殿 "hall") as their head, e.g., amba hūwaliyambure diyan for Taihe dian 太和殿. Kangxi-period records contain a few names that are transcriptions from Chinese through and through, e.g., u ing diyan for Wuying dian 武英殿, and these become more common in Yongzhengperiod sources. 40  In the Qianlong period, a prominent goal of the revision of names was to replace such Chinese loans with words having a Manchu-looking morphology. Despite looking like old Manchu words, the replacements enacted under Qianlong were generally neologisms coined for the occasion. Surviving discussions show that the parties involved were very sensitive regarding the origin of words. Chinese was one thing, Manchu another, they thought, and preferably the two should be kept distinct. However, unlike under Hong Taiji, mimicry of the semantics of the corresponding Chinese expression was not an issue. The language could now be as flowery as the Chinese, and mean exactly what the Chinese meant, as long as the words looked Manchu.
Numerous lists with new words remain from the Qianlong period. They were sent by the Imperial Household Department in Beijing to its homolog in Mukden for promulgation there, and they are now partially accessible. The lists tend to focus on a certain theme. In one instance, that theme was public buildings.
Late in 1748 (Qianlong 13/10/22), the Imperial Household Department announced a new installment in the series of new Manchu words. This time, the list included Manchu alternatives to the Chinese loans gung and diyan. There former was fixed as gurung and the latter as deyen-words that retained echoes of the Chinese originals. 44 Less than a year later (Qianlong 14/7/17), the emperor remarked that "the Manchu text carved on the stele at Anyou 安佑 palace [at the Summer Palace northwest of Beijing] still says gung. Now, since gung . . . has been fixed as gurung, erase . . . the Manchu text gung and replace it . . . with gurung" (an io gung ni eldengge wehe de manju hergen kemuni tere gung ni bei seme folohobi, te gung be gurung . . . seme toktobuha be dahame, da foloho gung . . . sere manju hergen be nilafi encu gurung . . . seme niyeceteme folokini). 45 Qianlong was evidently a force not only in the formulation of the language reform, but also in its enactment.

Saksaha
Vol. 16 similarly replace the Chinese loanwords tan "altar" and miyoo "temple." They should now be mukdehun and juktehen (cf. the verb juktembi, "to sacrifice, to worship") in Manchu. 46 Effectuating this change in practice involved making choices for which the officials involved sought imperial guidance. The uncertainty stemmed from a provision that Qianlong made regarding the use of the new words. In early 1749 (Qianlong 13/12/12), Qianlong, presumably motivated by his aversion to mixing Manchu and Chinese elements, decided to restrict the usage of certain new words. "If Manchu translations exist for the characters that precede gurung and deyen [in the names of palaces and halls]," Qianlong said, then they should be called gurung and deyen. If no Manchu translations exist for the Chinese characters, [these places] should still be referred to using their Chinese names, calling them simply diyan and gung.
gurung, deyen-i dergi hergen ubaliyambuha manju gisun bisirengge oci, gurung, deyen seme arakini gung, diyan-i dergi nikan hergen be manju gisun-i ubaliyambuhakūngge oci, kemuni nikan hergen be dahame, an-i gung, diyan seme arakini 47 When Biyantaha, director of the court of sacrificial worship (wecen-i baita be aliha yamun-i aliha hafan, i.e. Taichang si zhengqing 太常寺正卿), considered the names of the capital's altars and temples, he found that some had names with Manchu equivalents, others not. Temples that, according to Biyantaha, had Manchu names were tian tan 天 壇, the Altar to Heaven, and lidai diwang miao 歷代帝王廟, the Temple to Past Emperors. 48 They had reportedly already been translated as abkai tan and jalan jalan-i han sei miyoo (with Chinese loans used for "altar" and "temple"). Biyantaha did not point it out, but the latter name is seen on a stele erected at the site in 1733. 49 The two institutions should now, Biyantaha proposed, receive the new names of abkai mukdehun and jalan jalan-i han sei juktehen. The latter name was indeed subsequently used on a stele erected on the site in 1764. 50 Biyantaha identified two temples without Manchu names. One was the Daoist Dongyue miao 東岳廟 (dong yo miyoo), the God of the Eastern Peak Temple, located outside the city walls to the east and benefitting from the patronage of bannermen and bannerwomen. 51 In 1704 (Kangxi 43), a bilingual stele had been erected at the temple (Biyantaha did not mention it), on which the name was written as dergi io-i miyoo, 52 using precisely the kind of Manchu-Chinese mixed language that Qianlong wanted to avoid (dergi being Manchu, io Chinese). The other institution without a Manchu name was the Altar for Asking for Grain, qigu tan 祈穀壇 (ki ku tan). What should be done with these names? In an apparent intensification of the language reform, Qianlong, in response to Biyantaha's query, ordered all of them translated (gemu ubuliyambukini). 53 A few months later (Qianlong 14/3/8, 1749/4/24), the Imperial Household Department accordingly announced new Manchu names for several more altars. 54 Finally, in 1762 (Qianlong 26), a stele was erected at Dongyue miao on which the name was given as dergi colhon-i juktehen. 55 The temples and altars where the emperor carried out various rituals received new names. Yet they were not, as it were, set in stone. More than thirty years after the institutions had been renamed (Qianlong 46/5/30, 1781/6/21), Fu[lunggan] (1743/46-1784), memorialized the throne saying that "[Buddhist] temple" (si 寺) and " [Confucian] temple" (miao 廟) were both translated as juktehen. The almost seventy-year-old emperor responded that Buddhist temples should be called juktehen and Confucian ones muktehen, 56 a word seen as a translation of miao on a stele as early as 1738 (Qianlong 3). 57 However, three decades earlier, Confucian temples had been juktehen.
As the aging Qianlong emperor's remarks to Fulanggan shows, the revision of names for public places continued beyond Biyantaha's memorials in the late 1740s. For example, in 1764 (Qianlong 29/6/4, 1764/7/2), Qianlong responded to a request by Coociowan, vice-president

Saksaha
Vol. 16 Coociowan's request was not unique. In the 1720s, a certain Maitu, who had travelled to Tibet, complained that signs along the empire's roads, borders, and city gates and passes did not carry Manchu inscriptions. 58 In Coociowan's case, there was at least Manchu text on the gate. The problem was that the Manchu simply transcribed the Chinese.
The gate whose name Coociowan wanted to change was called Weiyuan Bao 威遠 堡 "Fortress" after an old Ming-era installation (in late Qing times just a gate). Unlike the gates of Xingjing 興京 (called Yenden in Manchu), Aiyang 靉陽 (called Aiha), and that of the walled city of Fenghuang 鳳凰城 (called Fucaha), the gate at Weiyuan Bao was simply called Wei juwan beo in Manchu in a transcription of the Chinese name. The imperial decision was to give Weiyuan Bao the Manchu name goroki de horon tuwabure jase, "Gate of Far-Reaching Display of Might." The new name for Weiyuan Bao, which was a translation of the Chinese name rather than its phonetic transcription, was acceptable and considered to be properly Manchu in the context of the Qianlong reforms. Yet the choice contrasts markedly with the names of gates chosen under Hong Taiji, as discussed above. In 1636, a name similar to goroki horon tuwabure jase, to wit goroki be toktobure duka, had been rejected in favor of a more down-to-earth term.
The renaming of Weiyuan Bao is an example of how Qianlong-era language reform affected places that lay outside the capital city and carried inscriptions that would have been seen by many people-in this case, travelers in Manchuria. It is also, similarly to Bianta's proposals for imperial temples and altars, an example of an official proposing to change a name on his own initiative, which suggests that it was not just Qianlong himself who was bothered by Chinese words hiding in Manchu transcription.
In other respects, Coociowan's memorial is puzzling. Some of the place names he contrasted with Wei yuan beo, such as the pairs Xingjing/Yenden and Fenghuang/ Fucaha, were fairly straightforward. They were cases in which one name was morphologically Chinese and the other represented a local, non-Chinese tradition. In so far as those local names were used in Manchu, they can certainly be called Manchu names, but it is possible that they were older place names that had entered Manchu usage from whatever language was originally spoken in the area (a possibility that Coociowan, unsurprisingly, did not consider). Other name pairs that Coociowan mentioned, however, do not contrast with Weiyuan Bao/Wei yuan beo as clearly. Alongside Xingjing and Fenghuang, Coociowan mentioned the Manchu name Giyamcan for the palisade gate that in Chinese was called Xianchang 鹹廠. Probably speaking a form of Mandarin relatively similar to the one used today, Coociowan did not see that Giyamcan is a loan either from a local Chinese dialect or, in case it dated from a much earlier period, from Middle Chinese (γəɨmtɕ h aŋ'). 59 In southern Manchuria, speakers of Chinese and Tungusic languages had interacted for a long time and the languages shared some vocabulary; clearly, more than Coociowan was aware of or would admit. 60 I should note, moreover, that Coociowan was not bothered by cases in which there was no proper Chinese name. Alongside Xingjing and Fenghuang, he mentioned, for example, Ying'e 英額 and its Manchu counterpart Yengge. In this case, the Chinese is an obvious transcription of the Manchu, but that was no reason to change it. 61 The overhaul of the names of palace buildings and government structures thus continued for much of the eighteenth century and involved both the emperor and his officials. The names were used even into the nineteenth century, if not necessarily in inscriptions. The 1655 (Shunzhi 11/12) Manchu translation of the Classic of Poetry (Shijing 詩經) had miyoo and gung for Chinese miao and gurung. By contrast, the revised version from 1768, which circulated much more widely, used juktehen for miao and gurung for gong. 62 Furthermore, the various gurung, deyen, muktehun, muktehen, and other institutions are listed in a commercially published collection of "names of the boards and government offices" (Ma. jurgan yamun-i gebu; Ch. yashu mingmu 衙署名 目) from 1889. 63 However, the efforts invested in the project notwithstanding, when photographed in the 1920s, many palace gates and buildings in Beijing carried Manchu names that were but transcriptions of their Chinese counterparts. Qianlong's neologism gurung for "palace" was seen, but the old transcription gung was more common. 64 Similarly, when a traveler through the Weiyuan bao gate reported in in 1886 that a tablet with the name was hung on the gate, he made no mention of a Manchu inscription there. 65 Perhaps the Qianlong neologisms survived in government documents, but not always in public inscriptions.

The Inscription at Fragrance of the Teaching Temple
Revising the words used on inscriptions did not just involve the names for the institutions that the inscriptions commemorated, but also words used in the main body of longer inscriptions. I will discuss an example involving the establishment of Manchu vocabulary for Buddhist deities in the context of translation from Mongolian.
When workers on Qianlong's order erected a Temple of True Victory (Ma. yargiyan etehe juktehen; Ch. shisheng si 實勝寺) in the hills in Beijing's western suburbs, they found a dilapidated temple nearby. The emperor ordered it restored, gave it the name Fragrance of the Teaching Temple (Ma. šajingga wangga juktehen; Ch. fanxiang si 梵 香寺) and erected two steles at the site. The first carried a Manchu inscription on one side and Mongolian on the other. The second carried a Chinese inscription on one side, and, reportedly, "Sanskrit" (Fanwen 梵文) on the other. 66 Rubbings of the three first inscriptions have been published, but especially the Chinese text is damaged, showing a diagonal crack through the whole slab. Perhaps the Sanskrit (not Tibetan?) text was illegible, which might explain that no rubbing was made of it.
The restoration of the temple reflects Qing court policy. The Temple of True Victory was erected upon the conclusion of the First Jinchuan Campaign, Qianlong's war with Tibetan or Gyalrong-speaking tribes in western Sichuan. 67 That a Buddhist temple should be erected right next to one that celebrated the purported military successes of the empire indeed shows that the court continued to rely on Buddhism, so important in the Qing's new Western possessions, to mediate its relationship with the Tibetans and Mongols, and that inscribed steles played an important part in that process.
Yet the steles were also part of a larger project of producing Buddhist texts at court. As early as 1742, Qianlong's Buddhist teacher Rolpai Dorje (1717-1786) 68 headed the project to translate the Tanjur from Tibetan to Mongolian, a project that finished shortly before the Fragrance of the Teaching Temple steles were erected. 69 Translation of sutras into Manchu had evidently begun already at this time, although the most famous project of Manchu Buddhist translation, that of the Kanjur, dates to a few decades later. Buddhist terminology in Manchu was created as a result, but at the time of the erection of the Fragrance of the Teaching Temple steles, that terminology was not yet well established.
An archival document shows that before the stele was carved, officials in the Grand Council revised the wording of the Manchu text. The document implies that at least parts of the Manchu text were translated from Mongolian. The document gives a hint of the process by which multilingual stele were produced at the Qianlong court.
An official-Fuheng (1721-1770)? 70 -memorialized (Qianlong 14/11/23, 1750/1/1) the throne and remarked on a few errors that had crept into the text that was to be inscribed on the stele at Fragrance of the Teaching Temple. 71 Fuheng, if it was him, had been ordered by Qianlong to look over the text that was to be carved onto the stele. He found several problems relating to how Mongolian Buddhist vocabulary had been rendered into Manchu. The names of Buddhist deities had been carried over into Manchu with their Mongolian pronunciation nearly intact. But just as obvious Chinese loans in Manchu were a problem, so were obviously Mongolian loans. 72  . 76 It appears that the stele was only erected after this exchange took place. The text is dated to an "auspicious day" (Ma. sain inenggi; Ch. jiri 吉日) in Qianlong 14/11 (1749/12/10-1750/1/7), which must have been no more than a week after Fuheng's memorial, if the date is accurate. Interestingly, the Manchu text that was actually carved on the stele does not contain the phrase giyan-i binwat gaijara fusa, "The Bodhisattva who obtains binwat." The inscription does, however, contain the sentence: In the past, the Buddha told Śūra and Mahāsattva to practice and study the Lotus sutra. seibeni fucihi, kicen fusa, amba fusa de alame, šu ilhai enduringge nomun be urebume tacifi, 77 In the Mongolian inscription, Śūra and Mahāsattva are indeed referred to as kičiyenggüi bodisadu and mahasadu. In the Chinese, they are jingjin pusa 精進菩薩 and mohesa 摩 訶薩. 78 The full Chinese name for Śūra is not, as in the inscription, jingjin pusa, but da jingjin pusa, with da 大 meaning "great." Perhaps the full name was not used on the stele for lack of space. That the established Manchu translation for Śūra is "great Bodhisattva," however, suggests that even in the case of translation from Mongolian (mahasadu), the semantic structure of the corresponding Chinese expression (da jingjin 75 pusa) still played a part in the coining of the Manchu expression. Furthermore, as in the case of Coociowan's names for gates in the Willow Palisade, Fuheng here was unaware of-or at least not bothered by-the fact that fusa represented an early loan from the Chinese pusa (the change from p-to f-in loans from that period being regular). 79 Finally, the fact that binwat does not appear on the stele suggests that the inscription was revised further even after Fuheng's memorial. Clearly, erecting a stele even in a relatively peripheral mountainous location was not a small matter, especially if it was plurilingual and contained specialized vocabulary. At this time, Buddhist vocabulary in Manchu was not yet definitely fixed, which meant that Qianlong's officials and religious advisors had to make choices as they went along. Yet because of the attention paid to language in this period, the choices were subject to revision.
After the erection of this stele, the court eventually produced quadrilingual collections that standardized Buddhist vocabulary in Tibetan, Mongolian, Manchu, and Chinese. 80

Conclusion
The overview of the Qing government's public Manchu inscriptions from the preconquest period to the second half of the eighteenth century that I have given in this paper shows that steles and signs with Manchu text were complicated but important to the court and its servants. Manchu steles announced government policy, and the language's use on public buildings asserted the power of the dynasty. The frequent back-and-forths between the court and officials in the field give a sense of the costly and complicated logistics involved in sourcing stones of the size needed for plurilingual inscriptions, establishing and sending out the text, and allocating the skilled manpower needed to execute the work. My sources show that erecting Manchu steles, which almost always contained other languages as well, was a major undertaking.
The sources used in this paper do not show how the new steles were received. Such sources might very well exist, however. Notably, I have largely left untouched the stele erected to commemorate Manchus, for which officials petitioned the throne. Such steles and the sources surrounding them would be a good place to start for looking at interaction between the court and the communities that would have read, looked at, or otherwise interacted with Manchu steles. This work has already been done, for example, for the commemorative arches for chaste widows erected, with government sanction,

Saksaha
Vol. 16 by the local Chinese elite in Huizhou. 81 Yet even with this work still left undone, I can confidently say that the kind of public inscriptions discussed here were seen by large numbers of people and that, given the evident financial investment and advanced craftsmanship that they represented, they did not go unnoticed. Given the financial and logistical stakes involved in erecting Manchu steles, it is not surprising that there is a great deal of evidence to suggest that, at least in the eighteenth century, emperor and officials paid great attention to the appropriateness of the Manchu text that were inscribed on them. The Manchu text should not appear as a calque of either Chinese or Mongolian. In the many new contexts in which Manchu was used-as a medium for Buddhist discourse, or on buildings of a kind the Manchus had not possessed when they first wrote down their language in the early seventeenth century-therefore required the development of new terminologies. The execution of the inscriptions, carving on stone, should furthermore be free from errors. In many cases, the will to avoid infelicities and errors led to the commissioning of new inscriptions, as was, we must infer, the result of the approved name revisions discussed above. In other cases, fear of mistakes precipitated avoiding making a Manchu inscription at all, as in government steles in the south in the 1760s.
Yet at least in Beijing and parts of the Northeast, city dwellers and travelers were witnessing a changing inscribed environment of new "shining stones" and repainted signs. Perhaps they, like Fuhai in Lao She's novel, glanced on them out of the corner of their eye at the distance appropriate for such stately embodiments of the dynasty. Accordingly, when the empire weakened and fell, signs were changed, and stele faced neglect or even destruction. Today, much of the evidence remains only on paper-as rubbings, government documents, or printed books-rather than on the stone that at the time probably appeared much more durable.