Vernacular Language and the Wu Dialect in the Formation of a Chan Koine and the Rise of Chan/Zen Philology: The Seventh to Seventeenth Centuries

Abstract

Chan monks had a language problem. They needed to deal with at least four registers; the language of the street in their district, the Chan koine with its colloquialisms and argot, the guanhua or official language, and the elite formal language in Literary Sinitic that was packed with allusions. Zen monks in Japan had to deal with registers of Japanese, plus the Chan koine and the formal Literary Chinese. One response was philological, glosses on vocabulary that were likely to be misunderstood, such as dialect words and words that had changed meaning over time. Notably, most of these responses originated in or were connected to the greater Hangzhou region and its Wu language. After discussing whether there was a “standard” or common language used by elites throughout China, this article contends that the awareness of language differences between the “standard” or Mandarin Chinese and the Wu language by native Wu speakers contributed to the rise of Chan philology and then Zen philology. A few key examples of specialized Chan philological texts will be examined, but it should be kept in mind that examples of Chan philology may be embedded in other Chan works.

1. Introduction

Philology (xiaoxue, 小學), in the sense of an examination of changes in meaning and their nuances over time in different historical contexts, was long part of the scholarly apparatus of learned Chinese monks. This was primarily because the foundational Buddhist scriptures had been translated from various Indic languages over a long time frame during which the Chinese literary language itself had changed. This philology included the creation of text-critical editions by collation, glossing of meanings of obscure words, the selection of “correct” characters, the explanation of dialect vocabulary, and phonology. Chinese monks also needed to demonstrate that Buddhism was scholarly and that it could match the academic levels of the native elite Confucian study of the Classics (jingxue 經學). Jingxue was closely associated with the creation of philology.1 Buddhist monks adopted this because they attempted to proselytize the ruling class, who were educated in the Confucian classics. To do so, the monks needed to be versed in the Confucian classics and how to read and interpret them, in other words, philology. This was a means of understanding the culture of the elite as a basis for explaining Buddhism in Chinese terms. The monks also needed to use appropriate language in dealing with the bureaucracy in such matters as the management of the monasteries and their residents (Cao 1994, pp. 18, 114).

As time passed, Chan monks also faced the problems created by a rapidly expanding Chan literature, mostly circulating in manuscript copies that were susceptible to transcription errors due to being composed over centuries by authors coming from nearly all corners of the empire (and some, such as Koreans from beyond that empire) who may have had a wide range of Sinitic languages as their native tongues. This meant that texts were often corrupt and that correct meaning could not be ascertained because the meanings of words had changed over time or had been used in different contexts or, in fact, were “dialect” (regional language) forms. By the early Song, there had been nearly four centuries of Chan texts, some no longer easily readable and so forgotten (such as some “Northern Chan” texts and some of the corpus of Shenhui 神會 of the eighth century) and only retrieved in the early twentieth century from Dunhuang. The language of some of them was so divergent from the “common” or standard language of the Northern Song court that attempts were made not only by Chan monks but also by Song courtiers to standardize a Chan language, a koine or common dialect, that was close to the common language (but was possibly largely restricted to writing). These forces, combined with an awareness of the difference between the “common” or Mandarin language and the Wu language, which was spoken in one of the main centers of Chan culture and publication and which had connections with Song-court culture, inspired the rise of Chan philology.

This article is divided into an initial investigation of how to characterize the language used by Chan and how it was related to the state-sanctioned standard language (Section 1, Section 2, Section 3, Section 4 and Section 5). This is followed by descriptions and analyses of Chan texts that are basically philological, linking most of them to the region where the Wu language was native. It concludes with some observations about how this Chan philology was received in Japan.

2. Chan and Regional Languages

Language has considerable importance in Chan and Zen, not just in philosophy but also in social networks. Recent research has indicated that regional networks in Chinese Chan were possibly more important than lineage connections.2 Such regional networks, at least in modern times, are largely based on one’s native “dialect” or rather language.3 Therefore, one’s native dialect was a very important element of a monk’s everyday life because it often facilitated associations with other monks and laypersons who used the same dialect and made a monk aware of the differences between his native language and that used in the Chan texts and from the official language (guanhua 官話) used in government offices and by administrators.4 This question of Chan monks speaking different languages or attempting to communicate in an artificial common language has been largely ignored.

Despite state restrictions on travel, which was tightly regulated (Benn 2002, pp. 51, 184), Chan monks traveled extensively to consult with Chan teachers all over the country (biancan zhufang 遍參諸方). This pilgrimage in search of an enlightening Chan master was modeled on the quest by Sudhana for enlightenment from fifty-two teachers as described in the last part of the Avataṃsaka Sūtra (Huayan jing 華嚴經). For example, Wenxi 文喜 (821–900) of Longquan Cloister in Hangzhou asked, “Why don’t you learn the pilgrimage consultation of Sudhana?”5 Such travels in search of a spiritual guide had come to be expected, for Xuansha Shibei 玄沙師備 (835–908) was asked, “Ascetic, why don’t you travel in search [of a spiritual master]?”6 This search is sometimes expressed in terms of “consultation and questioning [of teachers] in various regions” 參學諸方.7 The famous Zongmi 宗密 (780–841) declared that “I had consulted and questioned [teachers] in the various regions”.8 In the Chuanxin fayao 傳心法要 (Dharma Essentials of the Transmission of the Mind) dated 857, Huangbo Xiyun 黄檗希運 (d. 850) concluded that “If you wish to practice, question the lineage teachers of the various regions”.9 Therefore, most Chan monks traveled widely across China in search of an enlightening master and, in so doing, encountered a number of regional languages and dialects.

This movement across linguistic boundary lines was furthered by rulers and eminent officials ordering Chan monks, among others, to attend court or to reside in specified monasteries, often far from their home monasteries. For example, Shenhui 神會 (684–758) was ordered by the emperor to be banished from Luoyang to Nanyang in 720.10 Later, in 743, he was ordered to live in Kaiyuan Monastery in Jingzhou.11 Weikuan 惟寛 (755–817), a native of Quzhou in Zhejiang, studied under Mazu Daoyi 馬祖道一 in Jiangxi, then taught throughout Min (Fujian) and Yue (northern Zhejiang). In 809, he was ordered by the emperor to reside in Anguo Monastery in distant Chang’an.12

This need for travel brought monks into contact with different languages and dialects. An early example of such travel and study of different languages, or at least different pronunciations, can be found in the life of Shenxiu 神秀 (606–706), leader of the so-called “Northern Chan”. An epitaph stele inscription for Shenxiu written by Zhang Yue 張説 (d. 730) outlined Shenxiu’s early career:

As a youth he was a Confucian student and he traveled to question [teachers] in Jiangbiao (Jingzhou region). In regard to the profound teachings of the Daoist philosophers, the general meaning of the Documents and Changes [two of the Confucian classics], the sutras and treatises of the Three Vehicles [of Buddhism] and the Dharmaguptika Vinaya, his explanations were conversant with gloss explications (xungu) and his pronunciations referred to the Wu and Jin [pronunciation schemes].13

Here, Wu probably refers to the pronunciation or old standard dialect that prevailed in Chang’an up to the end of the seventh century based on the Qieyun 切韻 rhyme schemes that were compiled in 601. It reflected the “Mandarin” spoken in the former southern capitals.14 Nevertheless, this example of Shenxiu demonstrates that even early Chan monks were aware of language differences and of philology, of which xungu was a part.15

3. Common Languages

In the Chinese states, languages differed according to social status and region. The language now called Literary Sinitic (or more commonly Classical Chinese), a written language used by the literati, the court, and mainstream bureaucrats as a cosmopolitan written language, was hegemonic. It was grounded in the Confucian classics and was increasingly divergent from the spoken language, even of the court elite. Emperor Gaozu of the Sui (r. 618–626) recognized this when he said, “Even though I do not understand the written language, I also know that the [spoken] language of my native place is not inferior to this language”.16

While Literary Sinitic could be used to record conversations, this record was a “translation” from the standard language. In contrast, the spoken language, even when written down, was more able to “transcribe these subtle nuances of the speech acts” (Anderl 2012, p. 22). The problem was that in Tang times, “Since there did not yet exist any specific Chinese characters for transcribing items of the spoken language many function words were recorded phonetically by ‘loaning’ the pronunciation of other characters”(Anderl 2012, p. 18). This means that the transcription of speech by people such as Chan monks was relatively unstable as it appears they adopted loan words according to their own predilections and language background, such as native dialect. On the other hand, Literary Sinitic was relatively stable because it was largely based on the Confucian classics and jingxue. The spoken language of the court was less fixed as the composition of the ruling class in different dynasties changed. A “new standard based on Chang’an” emerged by the end of the seventh century under the Tang dynasty and, at least phonologically, replaced pre-Tang dialects. Edwin Pulleyblank thinks the change of the Tang capital, Chang’an, to the Northern Song capital, Bianjing (Kaifeng), “must have meant a further change in the linguistic standard”, but the evidence for this is slim. However, when the Northern Song fell, and Southern Song shifted their capital to Hangzhou due to the loss of North China, it “did not, apparently, lead to a new linguistic standard. The northern standard of Kaifeng was continued by the refugee officials from Kaifeng, a situation that is reflected in the present dialect of Hangzhou which has Mandarin-like features that distinguish it from other Wu dialects”.17

3.1. A Common Language?

Although there was definitely a court or official language (guanhua 官話), which after the seventh century under the Tang dynasty was most probably based on the Guanzhong or north-western dialect of the greater Chang’an region, sometimes called Qin秦, the question still remains as to whether there was a standard language “based on the speech of the capital” that “spread as a koine over the whole country”.18 The capital-based elite and the central mainstream bureaucrats who had qualified via state-run examinations on the Confucian classics and related topics may have spoken this koine, in other words, a prestigious “common dialect” that is supra-regional, throughout the country. However, at least in the distant regions of the empire, it was not spoken by local petty officials with whom most people had dealings. Posted in 819 to southern Guangdong, Han Yu 韓愈 (768–824) found that he could not communicate with the locals and complained that even the local officials “speak like birds and look like barbarians”.19 In a Buddhist work, the Yiqiejing yinyi (Pronunciations and Meanings [of Characters] In All the Scriptures) by Huilin from the period 783–807 that glossed difficult characters for pronunciation and morphology and explained unusual terms, there are mentions of a “common language of north and south”,20 “a common language of east and west”,21 and a “common language of the Qin and Jin”,22 as well as a “common language of all directions”.23 All of this suggests that there was more than one “common language” and not just that of the earlier southern courts that were being supplanted by that of the northern court and the Chang’an district around the end of the seventh century. This depends on what “common language” (tongyu 通語) meant. In the Fangyan 方言 (Regional Languages) by Yang Xiong 揚雄 (53 B.C.E.—18 C.E.), it meant a word shared across all the regional boundaries and was contrasted to a word belonging to only one region (See Zhou 2000, p. 257). In the Yiqiejing yinyi it means a “common language” in the sense of a shared language, but with some boundaries, as is hinted at where it writes of “a common language from the passes [east of Chang’an] to the east between Chen and Song”, 24 which means it may not have been spoken across the whole empire. This observation is confirmed by the earlier Qieyun, the preface of which refers to the pronunciations and tones used by the Wu and Chu (the south, Jiangnan) as being

occasionally light and shallow and that of Yan and Zhao (the northeast) as being mostly thick, with Qin and Long (the northwest) replacing the departing tone with the entering tone and Liang and Yi (Sichuan) having the even tone resembling the departing tone.

(Quoted in Huang 1998, p. 226)

Although the word guanhua (official language) would seem to imply only one standard language, it is only mentioned occasionally in Chan texts from the Ming dynasty (1368–1644). Rather, as we have seen, the term tongyu 通語or “common language” as distinct from fangyan 方言 or “dialect” appears frequently in Tang texts, especially in the Yiqiejing yinyi. Huilin’s text differentiated Qin 秦 and Jin 晋 (Shaanxi) “common language” from dialect,25 or Jiangnan and central country “common language”,26 and even the “common language of all directions”.27 He also mentioned “dialects” and “the current vernacular common language” 今俗通語, as well as dialects and the common language of the north or of north and south.28 Some of this was written, as Huilin refers to a term as “common popular text, of the common language” 通俗文通語.29 He even characterizes one term, zhituo [0881b22] 擲碢, as “The written language of the south of the Yangzi and Wu-Yue [regions] and not the common language of the classics and histories”.30 This suggests that there were a number of common or shared languages, some of them written, though these may not have been the language of the mainstream bureaucrats.

However, if we exclude evidence from Chan texts, we know little of what this common language or vernacular during the Tang and Song was like beyond individual words and terms, for hardly any sentences, if at all, survive. All we have are fragments, mostly found in poems and a few stories. For example, “Release this crazy man” 放此風漢 is found in the late Tang collection of stories from the late eighth to the early ninth century, the Yuquanzi 玉泉子 and “This barbarian has misled me” 這胡誤我 reportedly said during the An Lushan rebellion (755–763).31 The poetry of Wang Fanzhi 王梵志, probably of the late seventh to the early eighth century, likely from Weizhou in northern Henan, contains several examples of what might be called the “common language”.32

Others laugh at my poverty. My poverty is very enjoyable. I have no oxen and no horses. I do not worry that thieves will steal them. You are rich, and the house tax is high. The differential tax uses up all you have. I have nothing to summons. I eat my fill and always stretch out my legs… 他家笑吾貧、吾貧極快楽。無牛亦無馬、不愁賊抄掠。你富戸役高、差科並用却。吾無呼喚處、飽喫長展脚。

Again, “I see that fellow has died, His stomach burning like fire. It is not that I sympathize with him, I am afraid that it will also come to me”. 我見那漢死、肚裏熱如火。不是惜那、恐畏還到我。33 Similar language can be found in the Hanshan 寒山 (Cold Mountain) poems.

In other cases, conversations have survived because the wit or punning used needed a common language. For instance, the general Wu Yizong 武懿宗was sent to conquer the Khitan, but he threw away his armor and fled. A wit at court, Zhang Yuanyi, ridiculed him for “Riding pigs directly south in flight”. 騎豬正南掾. Empress Wu (Zetian) asked, “Yizong has horses. Why does he ride a pig (zhu)?” 懿宗有馬、何因騎豬. Zhang said, “Riding a pig is running squeezing/clamping onto a pig/privy (shi, 豕 = shi 屎, shit, i.e., he is running shit-scared)”. 騎豬，來（夾）豕走也.34

Given the paucity of evidence for what the standard court language was like, or even whether there were regional variants of the official spoken language, we have no proof that the language found in early Chan and other forms of Chinese Buddhism was the same as a “common” or standard language. The Yuanming lun 圓明論 (On Perfect Enlightenment), a text probably belonging to the Shelun School, was connected to, but not the same as, early Chan (Dongshan famen),35 and which was directed at a doctrinally well-informed audience, uses the metaphor of a stupid husband and wife making liquor for the projection of the insentient from the sentient mind. They look at the clear liquor in the vat and, seeing their reflections in it, accuse the other partner of hiding a lover in the vat. The husband hit the wife, and she asked, “’What have I done?’ He said, ‘Why have you hidden a man in the vat?’ …. At that, the wife got even angrier and said, ‘There is a man bringing a woman.’ They hit each other again”. 我有何事。其夫即言、你何故将一男人藏着甕中。。。。爾時其婦更瞋口云、有一男人送一箇女婦來。又復相打36 The language has similarities to Northern Chan Dasheng wusheng fangbian lun 大乗五生方便論 (On the Five Productive Skillful Means of Mahāyāna) and Dasheng wufangbian 大乗五方便論 (On the Five Skillful Means of Mahāyāna) associated with Shenxiu and the later corpus of Shenhui, such as the Nanyang Heshang wenda zazhengyi 南陽和尚問答雜徴義 (Miscellaneous Proofs in the Dialogues of the Reverend of Nanyang) of post-732, but added to until 792,37 which contains supposed dialogues with literati and monks. One could argue that they are similar because this reflected the “common language”, but one could also argue that they differ from the “common language” used by the mainstream bureaucrats of the court and administration because they use different characters for what are evidently the same words (the “what” 是没物/勿 of Shenhui versus 甚谟物 of the Northern Chan Sanbao wenda 三寶問答 (Dialogue on the Three Jewels of Buddhism), possibly reflecting Dunhuang’s north-western dialect，the “where” 阿没處 of Northern Chan versus more standard “where” 何處 of Shenhui).38 A “common language” of the court would more than likely follow a common standard and not use different characters with possibly slightly different pronunciations for the same words. However, differences in language occur over time, place, and genre, so determining if the language of Shenhui, for example, was the same as that used by the mainstream bureaucrats and literati will be difficult. Moreover, in the language of Shenhui, we often find usages typical of Buddhist Hybrid Sinitic, such as yunhe 云何 and ruhe 如何, alongside words that appear to be colloquial, such as shimeiwushi 是没物/勿是, with the same or a similar meaning.39

3.2. Liwen

There was another level of language, the documentary liwen, the written language of the petty clerks of the local magistrate office and probably of the hamlet and village functionaries.40 Evidence for it first appears in the Tang period in the biography of Pei Yanling 裴延齡 (728–796), a rapacious finance official, in the Xin Tangshu 新唐書 (New History of the Tang Dynasty). It says that in 805, in a discussion of his financial legacy, “There was no real benefit and yet there was the trouble of liwen”.41 Buddhist records alluding to this documentary language date back even earlier. The Mingbao ji 冥報記 (Records of Retributions in Hell) by Tang Lin 唐臨 (active 650s) tells the tale of Wang Shu, who died a violent death and came back to life two days later. At the tribunal in the underworld, there was an official acting as a secretary at the bottom of the stairs有吏文案.42 Thus, liwen was used for recording tax payments, legal documents, and depositions, as well as orders to petty local officials. Although they have not been specifically identified as such, liwen was the language of the documents unearthed at Dunhuang and Turfan that dealt with tax, contracts, and legal proceedings. The language of these Tang-dynasty socio-economic documents, along with that of the Buddhist sutras, is sometimes called “colloquial plain language” 口語白話 as distinct from that of the literati officials文人官方.43 However, there seem to be only marginal differences between the language of these socio-economic documents and the Buddhist scriptural language or Buddhist Hybrid Sinitic, and between the Buddhist Hybrid Sinitic and Literary Sinitic (wenyan文言).44 There is rather a gradual simplification and slight vernacularization as one moves from the elite Literary Sinitic down to the writings of the petty clerks and village headmen.

Buddhist monks had to be conversant with this liwen because they had to deal with petty bureaucrats and navigate the bureaucracy of what was a surveillance state in Tang times. They had to apply for ordination certificates, permission to reside in specified monasteries, obtain travel permits, and, sometimes, an exemption from taxation.45 These reports and applications went from the local functionaries at the village level in a bureaucratic chain up to the prefecture level, even when concerning foreign monks, such as Ennin 圓仁 in 839.46 In fact, monks, especially abbots and monastery secretaries, had to know the elite Literary Sinitic in order to communicate with the literati and senior officials, Buddhist Hybrid Sinitic to read and preach the sutras, liwen to deal with everyday activities such as taxation, application for monk certificates, contracts, and travel permits.47 In addition, they needed to speak the language of the streets and, if they were Chan monks, to read the colloquial-influenced plain language that many Chan monks had adopted.

4. The Wu-Dialect Area

In addition to written Literary Sinitic, the spoken standard or common language and the documentary liwen for bureaucratic purposes, if Chan monks, especially abbots and eminent teachers, lived in a “dialect”-speaking area, they needed some familiarity with that dialect also. One example is the Wu “dialect”, a distinct language, which had a long history. Some claim that there was a distinct Wu-Yue culture, a subset of a wider Chinese culture, that was marked by a water-borne focus and belief in “shamans” (Zhang 2004, p. 3). In contrast to North China, where transport was primarily by land, the Wu-Yue region was linked by watercourses such as the Yangzi River, the Grand Canal that had been constructed during the Sui dynasty (581–618) and which terminated in Hangzhou, and the ocean. This promoted economic development and an increase in population from the mid-eighth century.48 However, geographically, the Wu-Yue region differed greatly, with a broad low floodplain stretching from Hangzhou north towards the Yangzi River delta, plus a narrow plain skirting the southern coast of Hangzhou Bay that runs from Hangzhou towards Ningbo. These lowlands were very fertile and amenable to water transport. To the west and south, there were hills backed by ranges. These uplands were less productive. Most of the monasteries were located along the hillsides.

What probably unified the region was language (Suzuki 1985, pp. 115–16). However, in the city of Hangzhou, there is a dialect that some scholars have classified as Mandarin rather than the Wu language. This dialect, restricted to Hangzhou and its immediate vicinity, resulted from large-scale immigration of the elite of the Northern Song capital of Kaifeng after its fall in 1126. The northern elite moved to Hangzhou from 1128, and they became the majority of the population in the city, outnumbering the natives (Simmons 1992, pp. 8–10). “The northern quality of the Song Hangzhou dialect joined with the status Hangzhou gained as the dynasty’s capital … allowed the dialect implicitly to serve as one of the models for guanhua 官話, the traditional koine spoken by members of the official class” (Simmons 1992, p. 2). Because this influx from the north came after Chan philology was established, the existence of this dialect does not seem to have had any influence on Chan philology, but it may have influenced the Zen monks who came to Hangzhou from Japan during the Southern Song period.

The Wu “dialect” was noticed in medieval times. Emperor Yang of the Sui is said to have loved the Wu dialect, even using it, saying, “There are many externally who plot against (or counsel) me”. 外間大有人圖儂. The last word, nong, a first-person pronoun, was used in the Wu dialect (Cen 1960, p. 272).

Chan spread into the Wu region, especially Hangzhou, Yuezhou, and Taizhou (and not so much into the lowlands stretching to the north of Hangzhou), and developed there during the Tang period, slowly until the 790s49. Then, it rapidly expanded over the next century. After the Tang ended, the region was controlled by the Qian clan of the Wu-Yue Kingdom (907–978), who were enthusiastic devotees and sponsors of Buddhism. Many Chan monasteries were founded from the late Tang to the end of the Wu-Yue (Jia 2010, pp. 363–72). Yet we find very few, if any, traces of Wu dialect in the records about these Chan monks from the period because most of those records had been subjected to a standardization of the language imposed by the new Song dynasty. Typical of this is the Jingde chuandeng lu 景德傳燈録 (Record of the Transmission of the Lamplight Compiled in the Jingde Era) and subsequent “lamplight histories” sanctioned by the Song court.

5. A Standard Chan Koine

If the language of the late Tang Chan texts was like the court language or guanhua that has been characterized as a koine and was standardized,50 as was the case with the Jingde chuandeng lu that was standardized by members of the early Song court, that Song-dynasty Chan language should also be called a koine. A koine is defined as a language that

Results from the mixing and subsequent levelling of features of varieties [of dialects] which are similar enough to be mutually intelligible, such as regional or social dialects. This occurs in the context of increased interaction among speakers of these varieties.

(Siegel 2001, p. 175)

As a koine, colloquial dialect forms should be found in Chan texts. The Zutang ji 祖堂集 (Collection From the Halls of the Patriarchs), the initial core of which was compiled in 952 in Quanzhou, now in the heartland of the Min dialect (actually a number of separate languages now) (Ramsey 1987, p. 108, see 16 on “dialects”), arguably has minor traces of the Min and Wu dialects of the tenth century.51 This was probably not acceptable to states that wished to centralize control and dampen diversity or to Chan authors who wanted their texts available to all literate Chinese. There may already have been a growing consensus on how to represent colloquial speech in the tenth century, for Christoph Anderl observes that the Zutang ji is relatively homogenous in its use of colloquial function words (Anderl 2012, p. 18). However, the first known attempt at standardization of the Chan language was in the Jingde chuandeng lu of 1009 or rather in its original form, the Fozu tongcan ji 佛祖同參集 (Collection of the Joint Consultations of the Buddha and Patriarchs) of 1004. The Jingde chuandeng lu was sponsored by the Song dynasty court, and its chief editor, the academician Yang Yi 楊億 (974–1020), who hailed from the Min region (Fujian), wrote of the original, the Fozu tongcan ji by the Chan monk Daoyuan 道原 (probably from the Wu-Yue region), that “in some cases the ordering of the words was confusing and in some cases the language used was coarse—all of this we deliberately removed in order to make it of imperial quality”.52

Daoyuan seems to have started this standardization. Yang Yi, in a preface to the Fozu tongcan ji, refers to this when he wrote of what Daoyuan had done when he compiled the text from earlier records:

In cases where the principle was not clear, like Zichan of Dongli, he polished the words and where the phrasing was not applicable, like the style of Confucius in the Spring and Autumn Annals, he added to them or removed them. In cases where it was only a name and there was disdain for the facts, he dealt with them as the lacunae were handled in the Shiji (Records of the Grand Historian).53

This passage is full of references to the Confucian classics, with the reference to Zichan being to the Analects (Lunyu, XIV chapter 9) on the polishing of documents and to the theory that Confucius collated and edited the Confucian classics, and was in that sense the father of philology.54 The reference to the phrasing being not suitable and to Daoyuan’s removal of some phrasing suggests that Daoyuan had in mind a standard language as a criterion for changing his source texts. Moreover, his practice seems to have echoed the practices of jingxue and its ally, philology.

There probably was a need for standardization of the language. Chan records that developed from the late Tang onward came to be called yulu 語録 (recorded sayings). These yulu mostly recorded the dialogues between masters and students but were allegedly written down by members of the audience in secret, though they were subject to later amendments by third parties, and many may, in fact, be fictional, at least in part (Yanagida 1985, pp. 227–29; Welter 2008, pp. 46–47, 175 and Note 19). Whether records of actual conversations or fictional creations, there must have been a need to make these records intelligible across the Chinese empire because leading Chan monks came from all corners of China. Some, such as Shitou Xiqian 石頭希遷 (700–790), were probably not even native speakers of a Chinese language. Xiqian probably spoke a language like modern Yao.55 Mazu Daoyi 馬祖道一 (709–788) originally came from Sichuan, and he and many of his pupils resided in what is now Jiangxi Province, an area where the dominant language was Gan, a non-Mandarin language. The influential teachers Linji Yixuan (d. 867) and Zhaozhou Congshen 趙州從諗 (778?–897) both came from Shandong and later lived in Hebei Province in the northeast. Thus, there was probably a tendency by these monks to imitate the standard spoken language of the Tang court, which was adopted to forge some mutual intelligibility. However, this process is difficult to detect because most of the yulu of the mid-to-late Tang were edited later in the Song dynasty and may have been subjected to a Song-dynasty standardization, possibly imposed by the court.56

It appears then that “many Chan texts from the Tang were unacceptable for the educated Song Dynasty readership”(Anderl 2012, p. 25). Yang, a leading author in both poetry and prose, was part of the Song court’s attempt to champion Chan and record it in a language different from that of the preceding Tang and Wu-Yue dynasties.57 This process has been labeled ‘textual sanitation’, in which the language was, to a certain degree, adapted and homogenized, including orthography and grammatical markers. Phonetic loans were removed or reduced, the rhetorical structure became more stringent, and dialect influences were removed”.58 This was a considerable undertaking, for Yang Yi and his associates undertook the compilation or standardization of the earliest of the extant yulu, and not just the Jingde chuandeng lu (Welter 2008, pp. 68–71). Yet this standardization was incomplete, for there may be traces of the languages (Min and Wu) of the former Min and Wu-Yue regions, respectively, (and not just of the dialects of the standard language) in some Song-dynasty Chan compilations (Lu 1998, pp. 91–94), and possibly even in the Jingde chuandeng lu, which uses nongjia 儂家, meaning I, that probably derives from a general South China (Jiangnan) use.59

Yang Yi and his associates checked dates against calendars and alleged events against historical records and against the Song gaoseng zhuan and Zongmi’s Chanyuan zhuquan ji.60 But exactly what Yang Yi meant by “coarse” 猥俗 or “vulgar” language and making the language that of “imperial quality” is uncertain, for comparisons of the language of the Jingde chuandeng lu and that of earlier or contemporary Chan texts are extremely problematic. Firstly, changes were made over time to the Jingde chuandeng lu after 1009 (Welter 2006, pp. 116–17), and the chief possible source for a pre-Song Chan language, the Zutang ji, was expanded several times from a possible one fascicle of 952 to a ten-fascicle version of post 990. Then it was expanded and reorganized according to lineages by Kwangjun 匡儁 in Koryŏ Korea in 1245.61 These layers in the Zutang ji are difficult to differentiate with certainty. There are virtually no texts after the Shenhui corpus of the 730s to 790s, with the exception of the Dawei jingce 大潙警策 (The Cautions of Dawei, Pelliot 4638) of ca. 980 (manuscript not available to me),62 to be compared with the Jingde chuandeng lu to see what Yang Yi meant by changing coarse language into language of imperial quality 綸貫, namely, that which is consistent and excellent, a process which was a component of philology. What is available for comparison, and with all the above caveats, is the 1245 print of the Zutang ji.63

Could “coarse language” mean scatological language? The Zutang ji contains a number of examples. Chan texts often have scatological references that rarely, if ever, entered the Literary Sinitic texts.64 For example, a native of Shaozhou, Dongsi Ruhui東寺如會 (744–823), a pupil of Mazu, was visited by Cui, a Vice Director of the Department of State Affairs, who “saw sparrows defecating on the head of [the image of the] Buddha [in the Buddha Hall]. He asked, ‘Does this sparrow have the Buddha-nature?’ The master said, ‘It does.’ The minister said, ‘Since it has it, why does it shit on the head of the Buddha?’ The master said, ‘If it did not, why doesn’t it shit on the head of the chicken?’”65 This does not appear in the Jingde chuandeng lu but does appear in the Liandeng huiyao 聯燈會要 of 1183.66 Again, Yaoshan 夭山 (884–960), a native of Fujian, quoted a dialogue, “’It is your body that has a Buddha in it; do you recognize it?’ The chief monk said, ‘Where did you get such crap to piss on the Buddha?’”67 This also does not appear in the Jingde chuandeng lu. This vulgarity has a peasant earthiness about it, as Daan 大安 (793–883), a native of Fuzhou in Fujian, reflects in his formal sermon, “Therefore I spent thirty years on Guishan eating Guishan’s rice and shat Guishan’s crap. I did not study Guishan’s Chan, I just usually looked after a water buffalo. When it went off the path into the grass I pulled it back, when it encroached on people’s seedling crops, I would whip it”.68 This does appear in the Jingde chuandeng lu.69 Thus, Yang Yi did not remove all scatological language.

Did Yang Yi remove abusive language?70 Again, the Jingde chuandeng lu contains only a few of the abusive words found in the Zutang ji. However, “Drag that person into the mud” 推人向泥著 is only in Zutang ji.71 “Do not shit here”莫向這裏 is only in Zutang ji,72 as is “Huh! What a butcher?” 咄、這屠兒.73 Likewise, “You are just a lunatic wearing a robe” 汝只是狂被袈裟74 and “What a garrulous novice!”75 only appear in the Zutang ji. Also, “What demon made you renounce home?” 那個魔魅教你出家76 and “These [deeds] are the dregs of hell” 是地狱柤/渣滓77 only occur in the Zutang ji. Other insults or put-downs are only used in the Zutang ji.78 Some slightly less-insulting addresses, such as “This garrulous monk”, 多口阿師 appear in both the Zutang ji and Jingde chuandeng lu.79 Some language probably could not be avoided, such as “The master then spat into his face”, 師便驀面唾, which appears in both compilations plus the Yunmen Guangzhen Chanshi guanglu 雲門與匡真禪師廣録 (Extensive Records of Chan Master Guangzhen of Yunmen),80 because it was integral to the story. A final example suggests something else Yang Yi and his co-editors did. “All of you are total drunkards [lit., wine-dreg drinkers]. If you carry on like this, you will be so ridiculed you will leave”. 盡是一隊喫酒糟漢。與摩行脚、笑殺而去. This dialogue is in the Zutang ji and Jingde chuandeng lu, but the last problematic phrase was reworded as 行脚取笑於人 “carry on [on pilgrimage], being laughed at by people”.81 This substitution of phrases hints that Yang Yi removed language that was not well understood and replaced it with less ambiguous wording.

Finally, there is the question of dialect. A number of Chinese linguists have studied this topic, arguing that there are elements of Min 閩, specifically the Southern Min dialect, in the Zutang ji (Mei 1997; Zhang 1996). However, Kinugawa Kenji has refuted much of this, suggesting that some of the elements detected were typical of the broader Southern region, such as Wu and possibly Guangdong or Jiangxi, or were older forms of Northern dialects introduced by migrants into the region from the North (Kinugawa 2013, pp. 144, 148, 162–64, 203). He says there is some evidence of pronunciations, based on rhymes in verses, that some influence was exercised on the verses, at least by the Min pronunciation, and that there are one or two grammatical structures that are Southern in origin (Kinugawa 2013, pp. 165, 203). In fact, there is no evidence that Yang Yi removed dialect forms.

It seems then that Yang Yi only corrected what he saw as grammatical mistakes and toned down some of the scatological and abusive language, possibly because Daoyuan had already standardized much of the text. Moreover, the Jingde chuandeng lu and the Zutang ji shared the common language that was the official language of the North, but it was probably the Jingde chuandeng lu that provided a model for what became the Chan koine. This was because the Jingde chuandeng lu gained the imperial imprimatur, entered into the Tripitaka in 1101, and was imitated and summarized. It also reflected the interests of the Song state and literati. Moreover, it marks a shift away from the Xuefeng-lineage-centered Zutang ji towards the Wu-Yue Fayan lineage in the Jingde chuandeng lu and the subsequent shift toward the Linji lineage.82 Moreover, it was more comprehensive than the Zutang ji (either the one-fascicle version or the assumed ten-fascicle version) and more accurate in dating and other aspects than the Baolin zhuan 寶林傳 (Records of Baolin Monastery) of 801, and according to some scholars, more in conformity with the Northern common colloquial. This meant it was repeatedly included in the Song to modern-day Tripitaka printings, and according to the poet Lou Yue 樓鑰 (1137–1213), coincidentally a native of Ningbo in the Wu region, “Recent generations of Song literati do not make as much effort as previous generations and just appear to study Buddhism. They can only skim through the Lengyan, Yuanjue, and Jingming (Vimalakīrti) sutras, and the recorded sayings of the [Jingde] chuandeng [lu] as material for discussion”.83 Therefore the Jingde chuandeng lu became the standard for the Chan language, perhaps not yet a fixed koine, but moving towards one.

Though we have little evidence for how the language of the Jingde chuandeng lu and the early Song-dynasty yulu was edited by Yang Yi and his associates, we do know that later Chan used the Jingde chuandeng lu as a model. It has been described as epoch-making, for it fixed the form of Chan historiography into the lamplight genre (Huang 2008, p. 155). It attracted some notes in later redactions besides those made by Daoyuan and Yang Yi. These include extra notes for the Song imprint of the Tripitaka and the following Tripitakas, and for the individual printings of the text, with the most in the 1316 imprint (Huang 2008, p. 124). The Wudeng huiyuan 五燈會元 (The Five Lamplight [Lineages] Returning to the Source) of 1252 took most of its material for its first two fascicles from the Jingde chuandeng lu and much of the rest of it is a combination of the biographical elements from the Jingde chuandeng lu with the recorded sayings and gong’an elements from the Zongmen liandeng huiyao 宗門聯燈會要 (Essentials of the Linked Lamplight of the Chan School) of 1183 (Huang 2008, pp. 24–25, 197–98). Moreover, it was one of the main sources for the Xuedou songgu 雪竇頌古 (Praises of Old Cases by Xuedou) by Xuedou Chongxian 雪竇重顯 (980–1052) from Mt. Xuedou in the Wu region, which in turn was a basis for the Biyan lu 碧巖録 (Blue Cliff Record) (Yanagida 1981, pp. 281–82, 296, 298–99). All of this incorporation of material from the Jingde chuandeng lu and the examples of yulu edited by Yang Yi as models probably led to what had been a reflection of the commonly spoken language or guanhua to be increasingly frozen into a Chan koine, increasingly divorced from the shared written languages of Literary Sinitic, the documentary liwen of the petty functionaries, and the spoken language of the Song court and then the Yuan. Thus, it became another “literary language” of a “specialised cleric elite”, one maintained until the last century (Anderl 2004, xxvi). In that sense, there was little change in the language of the Chan koine from Northern Song until recent times.

6. Zuting Shiyuan 祖庭事苑 (Anthology of Allusions from the Courtyard of the Patriarchs) and “Dialect”

The Zuting shiyuan of 1108, the first text solely devoted to philological concerns of the Chan School, was compiled in order to explain difficult terms and obscure references and allusions in a series of Chan works. This occurrence of unusual or easily misunderstood terms indicates that some Chan authors may not have immediately adopted all the norms of a koine based on the Jingde chuandeng lu or a court-sanctioned language.

Given the nature of a koine as reflecting a shared language that had evolved out of the meeting of mutually comprehensible dialects and of Chan texts being written in regions of China where other languages prevailed, at least among the native populations, it is no surprise that some “dialect” or popular colloquial language appeared in the Chan koine. Incidents such as the following mentioned in the Wudeng huiyuan show that there was a problem with dialect. Purong Daoping 普融道平 (d. 1127), a native of Guangdong, would, “Whenever a local monk came to visit, chant a rustic song in Min pronunciation: ‘At the start of the letter it instructed the maid rest from her labors, at the letter’s end it instructed the maid not to nod off,’ ‘Now say, what did it say in between?’”84 Of course, it is an issue as to whether or not dialect vocabulary and grammar were used and whether or not the specific item of vocabulary was simply part of the common language but has only survived in modern dialects. Nevertheless, Muan Shanqing’s 睦庵善卿 Zuting shiyuan of 1108 indicates that dialect words had made their way into the Chan texts, possibly even in the Jingde chuandeng lu (Hanqing Lei 2010, p. 177). Lei and others have identified a number of vocabulary items from the Wu dialect (8), Min dialect (5), north-western (18), and south-western (6) dialects of Mandarin and so on in Chan texts, including some from the Jingde chuandeng lu.85

Muan Shanqing (fl. 1088–1108) was a native of Dongyue, the southern Wu region in Zhejiang Province, and he said his hometown was Muzhou (mod. Jiande). He first became a monk in Yuezhou (Shaoxing). He then traveled widely, consulting Chan masters before returning to Muzhou sometime between 1098 to 1100 to look after his aged mother. He called his residence Muan (Mu Hermitage) as he admired Chen Zunsu 陳尊宿 (aka Daozong, 792?–895), who had woven sandals to support his mother.86 In 1108, in the Song capital, Bianjing (Kaifeng), in North China, Shanqing met the Yunmen-lineage monk Faying法英 (fl. 1050–1131), who was a native of the Ningbo and an abbot of the nearby Mt. Siming. On reading Shanqing’s notes on the editions of the Yunmen lu雲門録 (Records of Yunmen), Faying realized the value of Shanqing’s work, and so he arranged to have the Zuting shiyuan printed (Huang 2006, pp. 131–33).

The Zuting shiyuan mainly deals with Yunmen-lineage texts. Of the eight fascicles, the first five fascicles are Yunmen texts (Yunmen lu, works of Xuedou Chongxian, a third-generation heir from Yunmen Wenyan, in fascicles one to four, and the works of Xuedou’s pupil Tianyi Yihuai 天衣義懐 993–1064, the compiler of the works of Yunmen). These, plus the Yongjia zhengdao ge 永嘉證道歌 (Yongjia’s Songs on the Realization of the Way) that is in fascicle seven, plus the authors of the Fengxue zhonghou ji 風穴衆吼集 (Collection of the Roars of the Assembly of Fengxue) (Fengxue Yanshao, 896–973), the Fayan lu 法眼録 (Record of Fayan) (Fayan Wenyi, 885–958) of fascicle six, and the Lianhuafeng ji 蓮華峰集 (Collection of Lianhua peak) (Tiantai Deshao, 891–972) of fascicle seven (Zheng 2021, pp. 2–4), were all connected to the Wu-speaking area of Zhejiang.

Yunmen was born in Jiaxing and ordained in Changzhou near Suzhou, and then visited Muzhou Daozong (Chen Zunsu) in Muzhou shortly before Daozong died. These districts are all in the Wu-language region. The modern scholar Urs App thinks Daozong had considerable influence on Yunmen’s written language. Yunmen went onto Min to study with Xuefeng Yicun, then returned occasionally to the Wu region before finally settling in northern Guangdong (App 1989, pp. 2–5). The issue of the alleged author of the Zhengdao ge, Xuanjue 玄覺, is complex (See Jorgensen 2009, pp. 78–87), but by late Tang times, its author was said to have come from Wenzhou, and the Yongjia of the title is apparently in or near Wenzhou.87 Xuedou Chongxian was born in Sichuan, but after he was enlightened, he spent three years at Lingyin Monastery in Hangzhou, then went to Suzhou, and a short time later, he went to be abbot at Zisheng Monastery on Mt. Xuedou, just south-west of Ningbo. Tianyi Yihuai was born in Wenzhou to a fisher’s family, was a student of Chongxian, and then spent most of his life on Mt. Tianyi in Yuezhou before retiring to Chizhou in Anhui Province. Fengxue Yanshao was a native of Yuhang who studied for a while in Yuezhou before leaving and eventually settling in Ruzhou in the north. Fayan Wenyi was also born in Yuhang and studied for a time in Zhejiang before going to Fujian (Min); in his later years, he moved to Jiangxi and then Nanjing. Tiantai Deshao was born in Chuzhou in Zhejiang Province but left for Jiangxi when he was eighteen. After being enlightened by Wenyi, he then went to Mt. Tiantai in Zhejiang and then to Hangzhou. Thus, with few exceptions, most of the monks whose works Shanqing glossed were either native to the Wu-speaking regions or came and spent their most productive years in that region, which may suggest that language was part of Shanqing’s motive for writing the Zuting shiyuan.

When asked by Faying about his motives, Shanqing said he had visited and consulted at monasteries throughout the country and found that “when lecturing on or reading the records of Yunmen or Xuedou”, even great lineage masters, when questioned, could not provide accurate responses,

and the students sometimes got the offal and the weeds, the fringes and the leftovers,88 always transmitting them on. In that [instruction], they quote Buddhist stories and the incidents of the Confucian books, often not knowing their sources, erroneously forming opinionated theories. How can one be a knower of this who merely ridicules them, for their errors will rope in future students?89

This was not the only problem, for some Chan followers thought that Chan was solely about the transmission from mind to mind outside of the doctrinal teachings, but “many have lost the sense of non-reliance on written words, frequently saying [one should] remove written words, as silent sitting [in meditation] is Chan. These really are the dumb sheep of our school”.90

Shanqing’s response was to spend twenty years combing through the Buddhist scriptures and Chinese classics, questioning knowledgeable teachers, and collecting and comparing the various copies or redactions of the old records of the Chan masters, making notes. He quoted 247 non-Buddhist texts, ranging from the Confucian classics to character and rhyme dictionaries, and 85 Buddhist texts.91 He created over 2400 entries, which can be classified into the forms of characters, meant to correct the use of “incorrect characters”, the pronunciations of characters, the explanation of the meanings of characters, quotations providing evidence for the explanation, and personal notes.92

The most common entry was the correction of characters, using “dictionaries” such as the Shuowen jiezi 説文解字(Explanations of Characters), Boya 博雅 (Expanded Elegance), and Sancang guxun 三倉詁訓 (The Three Types of Gloss Explanations of Cang Jie) of the early Han as support for the change. Some of the characters used by Chan texts were evidently incorrect, not being applicable to the intended meaning of the term or phrase. However, in a few cases, such as zhe 這 and zhe 遮, used widely in Chan texts for “this”, Shanqing says that these should be zhe 者 (Huang 2006, pp. 146–47; Kinugawa 2013, p. 261), which he contends, following the Sancang guxun and Shuowen, comes from shi 適.93 This differs from what appears in the Jingde chuandeng lu and the Zutang ji (377 times for zhe 這 versus 42 times for zhe 者), which possibly reflects a Southern usage (Lu 1998, pp. 59, 62, 71–72, 74). Mao Huang 毛晃 (fl. 1131–1162), a philologist from Quzhou (mod. Quxian in Zhejiang), wrote of a poem by Wang Yan (r. 919–925), King of Shu, where Wang used zhebian 者邊, that zhe 者 as meaning “this” is vulgarly written zhe 這 (Liu 1954, pp. 164–65). It is clear from this and other examples, such as for the function words renmo and shenmo, that Shanqing was trying to make Chan works conform more strictly to the orthodox philological mainstream than even Yang Yi had done. He was trying to replace “vulgar” or “popular” characters with correct characters, saying, “It should be written” or “should follow” (Zheng 2021, p. 44). This demonstrates that some Chan authors had not strictly followed the model of the language in the Jingde chuandeng lu and the early yulu as edited by Yang Yi and his colleagues.

The evidence from pronunciation also demonstrates that Shanqing aimed at bringing the Chan language or koine as close as possible to the literati-sanctioned common language. His suggested pronunciations largely agree with court-sanctioned rhyme dictionaries such as the Guangyun 廣韻 (Extended Rhymes) of 1008, its revision, the Jiyun 集韻(Collected Rhymes) of 1043 and the Yupian 玉篇 (Jade Compilation) of 1013 et cetera. Three hundred and forty-four of the Zuting shiyuan entries follow the Guangyun. As Zheng Lijun concludes, “Shanqing definitely did not base himself on one rhyme dictionary or character dictionary for pronunciation, but based it on the contemporary common language pronunciation used in reading” (Zheng 2021, pp. 31–32).

Furthermore, in his explanations of meanings and citing evidence for allusions, or explaining Sanskrit terms, Shanqing was not only assisting readers in understanding the Chan text, but he was also attempting to make Chan masters and students, be more like the literati: learned and widely read. He cited the Confucian classics, Laozi, Buddhist sutras, and commentaries by the likes of the Huayan scholiast Chengguan 澄觀 (737–838).

Shanqing also dealt with dialect, with nearly all entries applying to the Yunmen lu and the works of Xuedou. I count about eleven entries, but most without reference to region. Yunmen probably used dialect more than other masters because he was teaching in areas where the languages were totally incomprehensible to Northerners and mainstream bureaucrats. Yunmen was teaching in a region where many hill tribes, such as the Gelao 猲獠 (perhaps the ancestors of the modern Yao who still live in the vicinity of Yunmen-si), lived. The local Chinese probably spoke a variety of Cantonese.94 Thus, outsiders like Yunmen had to use as much of the colloquial, perhaps even a creole, to teach local people. Thus, when Yunmen was asked, “What is the intention of the teaching?” he replied, “The tongue of the chiliao has again brought a question”.95 Shanqing glossed this as jiliao.

The latter [character] is pronounced liao. In Northern dialect, it combines the sounds into one character [like klao] … It can also be that much speech is [said to be] jiliao. In Lingnan (Guangdong) there is a bird like a mynah (Gracula religiosa). When raised in a cage for a long time it can speak. Southerners call it a jiliao…. Its nature is not the same as that of parrots and mynahs. Yunmen lived in Lingnan, so it is likely that he used it in this sense.96

Again, Shanqing glossed an original wanluan 䴷圝 as a dialect word meaning complete,97 and later redactions of the Yunmen lu have followed Shanqing.98

Another example is where a saying of one region of China or an earlier period is no longer understood. For example, Xuedou wrote, “Three to five, past midday, urgently leaping down the road. His straw sandals trodden through, his home town far. He wears as shade a burning bell of ten thousand catties”.99 Shanqing explained,

Wearing as shade a burning bell: Some in the assembly propose that this [means] that he wore a fire-plate/fiery plates [of armor] on his head to ward off (?) the conditions of non-Buddhists/the external road. The meaning is definitely of this type. I have heard a Sichuanese monk say, “This is a saying of Sichuan”. When ridiculing a person for their ignorance, the people of Sichuan sometimes say “A burning bell is covering your head” and they often say this of the increasing heat of the start of summer. After all, Xuedou was a native of Sichuan.100

This suggests that Shanqing was made aware of the language issue because he had traveled from his native Wu-speaking region throughout China, including to the capital. As a story about an eminent official from the Min-speaking region having language difficulties while in the Northern Song capital in the north China heartlands demonstrates, it was not just monks who were faced with communication problems. The story concerns Liu Changyan 劉昌言, a native of Quanzhou in the Min-speaking region.

When Liu was the imperial diarist, he spent three continuous days dealing with the imperial consultations and proposals. He did not even have time to eat. There was the humor of ready wit in his responses to proposals, and the emperor (Taizong, r. 976–997) was pleased with his skillful oratory. Therefore, he took over the management of the Bureau of Military Affairs as Grand Master of Remonstrance. However, an objection arose among court officials [to the effect that] meetings between ruler and ministers can run from the cordial to the cold, but do not continue for long. A favor can go cold in a morning. The emperor then said to the attending ministers, “In some responses to proposals, Liu had a strong southern accent, and I did not understand anything”. So, the opponents could not say anything, and the [position] was permitted.101

If a high official could face such prejudices because of his accent, surely some monks, especially those traveling around consulting Chan masters, would have also faced similar prejudices, making them aware of language problems. I suspect Shanqing had such experiences, encouraging his philological studies to aid in the standardization of the Chan koine and to provide pronunciation guides, in addition to the scholarly knowledge of allusions to Buddhist and non-Buddhist texts and literature.

Clearly, the Zuting shiyuan met a need, for after its first print in 1108, another print was made in 1154 (Huang 2006, p. 136). It was referred to in the Biyan lu of 1125, the Chanyuan mengqiu 禪苑蒙求 (Elementary Investigations into Chan) of 1225, the Chixiu Baizhang qinggui 勅修百丈清规 (Imperially Revised Pure Regulations of Baizhang) of 1343 (T48.1131a2-5), the Fanyi mingyi ji翻譯名義集 (Collections of Names and Meanings of Buddhist Translations), a Sanskrit-Chinese dictionary of 1143, and the Famen chugui 法門鋤宄 (Elimination of Traitors in the Dharma Gate) of 1667 (Huang 2006, p. 127; Zheng 2021, p. 17), and beyond that in Korea and Japan.102 Moreover, the rise of lettered Chan (wenzi 文字) Chan and the recording of the sayings and sermons of leading Chan monks prompted a number of cautions. In a preface to the Dunwu rudao yaomen 頓悟入道要門 (Essential Entrances into the Way via Sudden Enlightenment) by Dazhu Huihai (ca. 800), the following caution was issued:

It is essential that one follows the master’s [Huihai’s?] verbal instructions, for if not, then one may have a misunderstanding and the use of the text would be ineffective. If one is divergent even in the slightest, then one misses it by a thousand miles. These are not erroneous words.103

In other words, even a text requires that one follows the master’s instructions, and that implies that if there is an error in the text, the results will not be conducive to enlightenment. This requires the utmost attention to the fidelity to the original text, a prime concern of philology. Shanqing took this lesson to heart, and because of the popularity of his Zuting shiyuan, he contributed to the fixing of the Chan koine into a largely frozen form that endured through to the start of the twentieth century.

7. Other Philological Works: Chanlin Baoxun 禅林寶訓 (Valuable Glosses on the Chan Monasteries) and Its Commentaries

Even though the Zuting shiyuan had become an influential philological guide, the texts it covered were limited. Beyond the works the Zuting shiyuan commented on, more yulu had been compiled both before and after the Zuting shiyuan was circulated and published. Moreover, compilations of gong’an 公案 (Jap. kōan, public cases) had been made by Fenyang Shanzhao 汾陽善昭 (947–1024) and Xuedou Chongxian (980–1052), but they led to problems beyond the paradoxes and seemingly illogical content they contained, for they also incorporated literary and colloquial components, some of it dated, as they were quotes of dialogues and exchanges between earlier masters and pupils from most regions of China. They also had introductions and commentaries in the most erudite Literary Sinitic, packed with allusions to the huge treasure trove of Chinese literature and history. The Zuting shiyuan helped address some of these issues. However, this was not enough, and Dahui Zonggao attacked the “tendency to equate knowledge about gong’an with insight into them and thus enlightenment itself”.104 Zonggao was attacking the practice of memorizing gong’an and attempting to explain them logically or in terms of Buddhist doctrine. It was not enough, in his estimation, to recognize the allusions they contained or understand the difficult words they sometimes used. There had been a trend towards compiling gong’an, with students reciting the Biyan lu of 1125 “day and night and [they] regarded it as the ultimate of learning”.105 Dahui infamously, at least according to legend, burnt the blocks of the Biyan lu, but the text survived. The language of the exchanges about it was characterized by Wan’an Daoyan as

tasteless [not conducive to enlightenment], like gruel made up of boiled shavings or rice composed of crooked iron nails. This then became food for future generations’ gestation. Such was the beginning of niangu (‘picking up the words of the men of the past’). Fenyang was the originator of the tradition, and Xuedou continued its and made it as boundless as the ocean.106

Perhaps this implied that one had to go beyond language and that philology on its own was not enough to bring one on the path toward enlightenment.

7.1. Chanlin Baoxun

The problems of understanding texts extended beyond the gong’an and the dialogues of the yulu. Other texts, such as letters, epitaphs, encomia, and biographies, contained allusions and vocabulary that needed explanations and philological analysis. Thus, despite Dahui opposing the entanglements of words in the lettered Chan of gong’an, he also contributed to the linguistic difficulties for readers in his own works, not so much in the yulu and Dahui shu 大慧書 that largely conformed to the Chan koine, but in a book he jointly compiled with Zhu’an Shigui 竹庵士珪 (1083–1146), the Chanlin baoxun. This work is in Literary Sinitic, with the rare touch of the colloquial, and came to require commentaries of its own. Compiled in 1130, soon after the defeat of the Northern Song court, it was written near Yunju Monastery in northwestern Jiangxi (Li 2004, pp. 2–3). However, it remained in manuscript form there and ended up tattered, worm-eaten, and in disorder. In the Junxi era (1174–1189), a monk from Dongwu (i.e., Zhejiang), Jingshan 淨善, obtained it from Yunju monastery (the last date in the text is 1179) and spent ten years collating it by checking against its sources (Li 2004, pp. 3–4), and so the revised collection may date from ca. 1200 (Yü 1989, p. 72 note 40).

This collection of advice draws heavily on many letters from masters to pupils, both monastic and lay, and draws upon records, biographies, texts engraved in stone, and real records (shilu 實録, veritable records), but not yulu 語録 (only two references). It includes advice on teaching, the running of large monasteries, warning against violations of regulations, insincere practice, the pursuit of fame or luxury, and other malpractices in what they felt was a period of the end of the Dharma. The advice is mainly aimed at abbots, and is informed by fears of decline, the end period of the Dharma, the dangers of luxury, the need for morality, and the rejection of lay entanglements, and for practitioners to concentrate on attaining enlightenment by respecting the Dharma.107 It was first printed in 1362 in Jiangsu and then in 1378 in Koryŏ (Shiina 1993, pp. 80–81). It was then printed in the Ming in 1443 (Ono 1932–1936, 6: 411).

According to a preface dated 1644, “This book had circulated greatly in Jiangbei (Jiangsu and Anhui) and greatly brought forth in Wu, but eight or nine out of ten masters in Min and Guangdong had not heard of it”.108 This means that this text was widely used in the Zhejiang region for the reform of Chan monasteries, which some had thought had become overly secularized, possibly even effete, in part because of the prosperous commercial developments in south-eastern China. Later, in the very last years of the Ming, the predations of the Qing meant that the problems were due to warfare and the impoverishment of the people, in addition to a baneful influence from branches of Wang Yangming Confucianism.109 Symbolic of this corruption was Xuelang Hong’en 雪浪洪恩 (1545–1608), an important scholar in the revival of Huayan and Yogācāra. The strong advocate of vinaya and regulations for the then lax and even corrupt monkhood, Yunqi Zhuhong 雲棲袾宏 (1535–1615), accused Hong’en of indulging in banquets, plays, and music (Liao 1998, pp. 37, 40). One of Hong’en’s poetry students, the famous Qian Qianyi 錢謙益 remarked that “it seems he had lost the genuine nature of a monk”, and Zibo Zhenke 紫柏真可 (1543–1603) made a similar comment. Hong’en ignored criticism and apparently succumbed to the secularization of the monkhood (Liao 1998, p. 38).

Dajian 大建, an heir of Yunqi Zhuhong, wrote the Chanlin baoxun yinyi 禪林寶訓音義 (Pronunciations and Meanings of the Valuable Glosses on Chan Monasteries) in 1635. It was a collection of phonetic and semantic glosses for the masterwork, and is in the tradition of the Tang-dynasty Yiqiejing yinyi. Noticeably, in his preface, Dajian said that in 1608, the year Hong’en died, “I farewelled Hong’en and took a teacher in Wangting in Suzhou [on the banks of Taihu, just over the border from Zhejiang in Anhui], and studied it (the Chanlin baoxun) in Chang’an (the northern capital, Beijing?), and I consulted various masters”.110 He found that while many lectured on it and studied it, the texts were brief and rich in meaning, but beginning students could not understand it because of the errors of manuscript transcription over the ages. Therefore, he collated the best of the copies and added pronunciation guides and explanations, including identifications of people, places, terms, and allusions. He also explained the meaning of entire passages. This was done at Yunqi Monastery just outside of Hangzhou.111 Dajian clearly used the techniques of text-critical collation (jiaochou 校讎) to establish his reliable text (shanben 善本), and by broad consultation, used the principles of evidential scholarship (kaozheng 考證), which were just then coming into vogue.112

Yet this was clearly insufficient, as three more commentaries, all using Dajian’s work as a basis, were produced. The first was the Chanlin baoxun hezhu 禪林寶訓合註 (Combined Annotations on the Valuable Glosses on Chan Monasteries) by Zhang Wenjia 張文嘉 of Wulin (Hangzhou) in 1650. He incorporated the original text but removed some repetition in Dajian’s work and added supplementary notes.113 In 1683, Deyu 德玉 from Sichuan compiled the Chanlin baoxun shunzhu 禪林寶訓順硃 (Concordant Corrections to the Valuable Glosses on Chan Monasteries). He requested an explanation by a Qixia Shanren (probably from near Nanjing), which left Deyu dissatisfied, so he wrote his own, with brief summaries of the biographies and the themes of each entry.114 It was printed in Shaoxing, northern Zhejiang (Ono 1932–1936, 6: 412b). Another commentary, the Chanlin baoxun bishuo 禪林寶訓筆説 (Notes on the Valuable Glosses on Chan Monasteries) of 1706, was written by Zhixiang 智祥, south of Changsha in Hunan.

The Chanlin baoxun, incorporated into the Tripitaka in 1584, then attracted the attention of monks, mostly from the greater Hangzhou region, who used philological methods to make it more useful for the reform of Chan monasteries. But Chan monks faced other problems with language in the late Ming. Within a brief span in the late Ming and early Qing, a number of new reference works appeared. The first was to assist the secretaries of Chan monasteries in the composing of their correspondence, in particular the shuyu 疏語 or praises of the virtues of Buddhas and eminent monks that were read out before them or their representatives on formal occasions such as the birthday of the emperor or the Buddha or similar ceremonial occasions.115 As these ritual occasions increased, so did the onerous tasks for the secretaries and the increased need for models of composition. Finally, Yongjue Yuanxian 永覺元賢 (1578–1657), who had been in the role of monastic secretary, gathered a number of these documents into the Chanlin shuyu kaozheng 禪林疏語考證 (Evidential Proofs on the Praises Used in Chan Monasteries), and his pupil Zhaoran Daoguo 超然道果 wrote a commentary on the meaning and the sources. According to Yuanxian’s preface, these shuyu were a Chinese innovation in Buddhism meant to attract believers through the praises in the text. The post of secretary, who composed these works, was an especially valued position in Tang and Song monasteries, and only the most erudite were chosen. By the Ming, there were very few monks capable of filling this role, and so they simply followed stale materials, wrote in poor language, and merely entertained the guests. Coming from a key Buddhist district of this time, Wulin, in other words, Hangzhou, Yuanxian’s work was printed by friends.116 Interestingly, Daoguo went on to found Sofukuji in Nagasaki (Yanagida 1974, p. 512). Daoguo examined the text to verify it, and when it was to be printed, he tried to correct any possible lacunae or errors.117 The shuyu covered many occasions, such as the emperor’s birthday, Buddha’s birthday, Buddha’s day of enlightenment, the induction of a new abbot, prayers for peace, opening the eyes of a new statue, the release of fish, prayers for spring rain or the warding off of fire, protection of the fetus of a believer, prayers for the safety of a baby, vows by a wife on behalf of a trading husband who has not returned, prayers for recovery from various illnesses, a master’s birthday, the ceremony of listening to readings from sutras et cetera. As the language was often difficult, the vocabulary had to be explained and the sources given. Notably, Shanqing’s Zuting shiyuan was often cited as an authority, usually in the form of a gloss. Certainly, the difficulty of the language of such encomia was probably well beyond the average monk, for it required an extensive knowledge of Buddhist terminology and events. But that was insufficient, as knowledge of secular texts seems also to have been another prerequisite.

7.2. Chanlin Leiju 禪林類聚 (Encyclopedia of Chan Monasteries)

There were other works that were closer than the Chanlin baoxun to what we would call a dictionary or encyclopedia. The next Chan work of this ilk after the Zuting shiyuan was the Chanlin leiju of 1307. This joint compilation by Shanjun 善俊, Zhijing 智境, Daotai 道泰, and others from Wanshou Monastery of Tianning 天寧萬壽寺 (probably that in Jiaxing, just northeast of Hangzhou), is a “Categorized Assemblage from the Chan Monasteries”. It is made up of gong’an and commenting verses. It contains a total of 5272 cases selected from works such as the Wudeng lu and the various yulu. They are categorized in order of emperors, chief ministers, Confucian scholars, Buddhas and patriarchs, Dharmakaya, Buddhist statues, monasteries, halls, et cetera, ending with plants and animals. These 102 sections (men) were to facilitate referencing (Yanagida 1974, p. 512). It was immediately printed in Yuan China and again in Japan in 1367 (Shiina 1993, pp. 96, 576). Sometimes, the case and comment were brief; some are commented on rather in verse. Sometimes the entry is very long.

Significantly, this text has borrowed the term leiju for its title from the ‘encyclopedias’ such as the Yiwen leiju 藝文類聚 (Encyclopedia of Artistic Works) compiled by Ouyang Xun (557–641) et al. as a sourcebook for the writing of essays or the Gujin shiwen leiju 古今事文類聚 (Encyclopedia of Facts and Literature Past and Present) of ca. 1246. The Chanlin leiju was probably modeled on the former given the order of its classifications, which are similar.118 As Teng and Biggerstaff commented, “Chinese encyclopedias (lei-shu) differ from Western encyclopedias in that they consist almost entirely of selected quotations from earlier writings, the name encyclopedia having been applied to them because they embrace the whole realm of knowledge” (Teng and Biggerstaff 1971, p. 83). In this case, the “encyclopedia” is organized into topics for cases or gong’an used only in Chan monasteries, but the gong’an had become so numerous that a taxonomy was required. It was like a handbook that could be used in philology applied to gong’an.

8. Japan

Chan was not limited to China. It spread to Korea, Japan, and Vietnam. Because of the direct ocean route from southern Japan to the ports serving Hangzhou, Japanese monks came in considerable numbers, certainly more so than monks from Korea and Vietnam. There was also considerable trade between Hangzhou and Japan.

When Chan was introduced into Kamakura Japan, it was introduced from Southern Song and not from the Khitan empire that controlled North China. Virtually all the pioneer Zen monks traveled to Mingzhou to the east of Hangzhou, doing so on trading vessels that launched from Hakata. The number of Japanese monks arriving rose over time, and the names of over two hundred Japanese monks are known. Most of the Zen monks studied in the Hangzhou region, but they came in groups rather than alone.119 All of this took place in a short time frame of about a century, from 1168 to ca. 1277. To what extent the Japanese monks learned colloquial Chinese is unclear. While many of the Chan monks in the region had probably fled from the North and spoke the “common language”, the natives spoke the Wu dialect, and there was a dialect particular to Hangzhou that resulted from the meeting of Northern Mandarin and the Wu language. Doubtless, the Japanese monks would have attempted to master the Chan koine, but it seems it was no longer a colloquial language.

There is indirect evidence that the Japanese Zen monks had not mastered the Chan koine and possibly even the common Chinese language. Yōsai 栄西 (1141–1215) rarely wrote in the koine, and Dōgen 道元, who returned from Southern Song in 1227, may have communicated with written conversation 筆談 while in China, but some of his earliest works, dating from 1233, contain no Chan koine, but by the 1240s he wrote in an occasional hodge-podge of Japanese, Buddhist Hybrid Sinitic, and the Chan koine, which may reflect how he had communicated while in Song China.120 In a record of a conversation with a Chinese monk, Dōgen reported his own speech in Japanese as follows: “ima (or nyokon) tarehito ka kore o taiji seru (Now who has it?” and he reported the reply of the monk in colloquial Chinese, followed by Japanese: “tantou laohan, nali yu xiangsi. Nochi ni shinshutsu nengoro niseba, sadamete mi-suru koto aran (paraphrase: The old abbot seems to have it. Later, if you kindly request its divulgence, you will probably get to see it”.121 Again, he quoted a Chinese master, saying to him in a mixture of Chinese and Japanese: “wu nali (Jap. go nari) ni ichi-juku no koseki ari, renmo cidi nari, yu laoxiong kan (I there have a scroll of old writings, and when there is such an occasion, I will show it to you.)”122 This precedes a quote in pure Literary Sinitic.123 It is unclear whether this is a reasonable representation of the actual conversation, the Chinese speaking in colloquial Chinese, Dōgen struggling with a mixture of Japanese and broken Chinese, or was a literary artifact, a compromise reflecting the partial back translation from Literary Sinitic that the Shōbōgenzō was initially written in (the Manaji Shōbōgenzō 真字正法眼藏, The Storehouse that Perceives the Correct Dharma in Chinese Characters). Elsewhere in the same book, in a chapter dated 1241, Dōgen takes a quote from a passage in the Chan koine, such as one about where the Buddha-nature is located when a worm is cut in half, but then analyses and comments in Japanese, suggesting he understood the koine.124

As time went on, it is clear that some Japanese monks, such as Enni Ben’en 圓爾辯圓 (1201–1280), could read and use the Chan koine, and some Zen monasteries, especially those headed by Chinese abbots, may have been miniature bilingual societies (Jorgensen 2022, p. 24; Protass 2022, pp. 130, 142, 147). A monk like Nanpo Shōmyō 南浦紹明, in Song China from 1259 to ca. 1267, was even made a monastic official there for a period, meaning he had requisite linguistic skills (Satō 2006, p. 27). However, other Japanese monks, such as Myōken Dōyū 妙見道祐 (1201–1256), who stayed in Song China ca. 1240–1245, may not have been able to effectively communicate orally, at least initially.125 Such linguistic obstacles were particularly marked in the practice of gong’an, in which there was a pivotal role in the sharp repartee between master and student, often in metaphorical and poetic language, as we find in “capping phrases” (jakugo 著語or agyo 下語). These “literary games” required mastery of allusions and nuances in the Chan koine (Hori 2003, pp. 43–45, 54–56). As only elite, educated Chinese of extraordinary abilities were able to participate, the difficulty would have been multiplied many times for the Japanese.

Although earlier Japanese and Chinese monks had brought Chan into Japan, they had often compromised with the influential established sects such as Tendai and Shingon, thereby diluting much of the very Chineseness of Chan, tending to Literary Sinitic composition rather than to the colloquial language, for example. The first to introduce “undiluted Sung monastic practice” was Lanxi Daolong 蘭溪道隆 (1213–1278), who arrived in 1246. However, he tried to have his students avoid scholarly indulgences. “The practice of Zen (sanzen bendō) does not lie in the study of four-and-six-character parallel prose”, he stated.126 In other words, Zen practice is not about being able to use the most difficult of elite Literary Sinitic, even though most Japanese Zen monks thought that Chinese poetry in Literary Sinitic and Zen had a close relationship (Protass 2022, p. 154).

Later Chinese monks arriving in Japan realized the obstacles to communication. Wuxue Zuyuan 無學祖元, who arrived in Japan in 1279, immediately after the Mongol Yuan forces had taken Southern Song, was born near Mingzhou in Zhejiang and became a monk in Hangzhou. He had a dialogue with someone who understood Chinese:

The student continued, “I ask in this way and you, Reverend, reply in this way. Our minds know each other, each sentence knows the other. But as with Japanese who do not understand [our] language, how will you teach them to obtain this jewel?”

The master said, “I have a skillful means”.127

Thus, Chan teachers had on many occasions to resort to non-verbal communication, and while some learned Japanese, others did not (Protass 2022, p. 147–48).

Hence, even after nearly a century of some exposure to the Chan koine and a half-century during which Chinese Chan teachers operated in Japan, many Zen monks who had mastered Literary Sinitic could not understand even the simplest Chan koine. This became evident when Yishan Yining 一山一寧 arrived from Yuan China in 1299. According to his biography as recorded by Kokan Shiren 虎關師錬 (1278–1346), who knew Yining personally:

The master sat alone on a couch and it was not necessary to go through an intermediary to meet him. Whether one had just arrived or were from far off, it made no difference. And so, people consulted and questioned him. In the Chan books there was to be no seeking of secret meanings, and only [the use] of the Zuting shiyuan. There were many who went to excess in putting in corrections and the world [of monks] suffered from this. And so, the master [with respect to] the highest principles put aside doubts, but when the words could not be comprehended, he lectured on the texts character by character and line by line.128

Yining’s openness led to his popularity, and he attracted so many students he had to limit their numbers via entrance exams on Chan-style verse (Dumoulin 1990, p. 36; Colcutt 1981, p. 74). Note he advocated the use of the Zuting shiyuan as the only guide, which would have introduced Japanese students to Chan philology. However, he found that many Japanese monks did not even understand the most ordinary of the colloquial words such as shenme 什么 (what) in the koine of standard Chan texts such as the Jingde chuandeng lu and the Wudeng huiyuan that he wrote Zokugokai 俗語解 (Explanations of Colloquial Words). This compiled his answers to their questions. Yet according to Mujaku and modern scholars, Yining was sometimes wrong (Iriya 1992), probably because the Yuan dynasty colloquial differed in some aspects from that of the Song or possibly because his own native Wu dialect (he was born in Linhai or Taizhou in south-eastern Zhejiang) differed from the standard. Perhaps this explains why Musō Soseki 夢窓疏石 (1275–1351), who won in the aforementioned verse contest, after some years of study, said that he and Yining “had failed to communicate in spoken language” and mentioned parenthetically that Yining was from Taizhou台州 (Protass 2022, p. 152). Perhaps Soseki was blaming Yining’s pronunciation or dialect (in modern Wu dialects, the Taizhou dialect differs from that of northern Zhejiang).

Other texts in the same vein as the Zokugokai (no longer extant), such as the Shūmon hōgo 宗門方語 (Proper Words of the Zen School, n.d., quotes Yining and Zuting shiyuan often), and another by the Chinese monk Leiyin 雷音 (d.u.) were produced, and then Mujaku Dōchū 無著道忠 (1653–1745) made a series of studies of the Chan koine. Like Yining, all used the Zuting shiyuan.

After about two decades of interrupted exchanges between Japan and China, there were new obstacles to communications on top of suspicions that some monks were spies.129 The numbers of Japanese monks going to Yuan were much greater than previously, and most were Zen monks.130 However, they faced a new linguistic challenge, for although the vast majority traveled and studied in Zhejiang, some went further, such as to Nanjing and even up to Beijing. But a creole, made up of a Chinese base with features of northern languages such as Khitan and Jurchen, had emerged as a common working language for the people living in the Khitan empire.131 The Mongols, who did not value education in Literary Sinitic for administration, preferred the documentary language such as liwen and possibly merged this with the existing creole, but with more Mongolian elements (See Liang 2010, pp. 183–184). Whether or not the language, often called Yuan baihua 元白話 was spoken or whether it was a form of direct translation and an official language of edicts from the Mongol khans and elites is contested (Miya 2006, p. 188), but it is clear there was some overlap between the two (Miya 2006, p. 194).

When or if this creole was spoken, it rapidly caused problems for the people in the south of the Mongol empire who spoke the Wu and other southern dialects (Miya 2006, pp. 215, 258, note 112). This may have troubled Japanese Zen monks, such as Betsugen Enshi 別源圓旨 (1294–1364), who wrote in a poem that “The southern sounds and the northern language alarm one with their divergence” (Quoted in Dongchu 1970, p. 555). When the creole was written, it also affected Zen monks, who had to deal with the Yuan authorities. This language appears in accounts of debates between religions in the Bianwei lu 辯偽録 (Records of Debates Over False Religions) by the Yuan-dynasty monk Ruyi Xiangmai 如意祥邁 of 1291,132 and is repeated in the Fozu lidai tongzai佛祖歴代通載 (A Comprehensive Registry of the Successive Ages of the Buddha and the Patriarchs).133 The most important document that includes this Mongolian-inflected Chinese is an imperial directive and a response by Dongyang Dehui 東陽德輝, the Chan master then abbot at Baizhang-shan, dated 1342, ordering the composition of the Chixiu Baizhang qinggui 勅修百丈清規. This was done in Nanjing.134 This text reached Japan in 1356, almost immediately after it was printed in Yuan in 1343, and many prints and commentaries were made thereafter in Japan. Dehui’s pupil Chūgan Engetsu 中巖圓月 (1300–1375), who had been Dehui’s secretary before he returned to Japan in 1332, may have understood some of this official language. However, it is clear from Mujaku Dōchū’s commentary in the text, the Chokushū Hyakujōshingi sakei, 勅修百丈清規左觽 (Commentarial Keys to the Imperially Revised Pure Regulations of Baizhang) that Mujaku and the preceding Japanese commentators struggled with this language.135

In Japan, the philological texts that had been produced in the Wu-language-speaking region of Zhejiang, beginning with the Zuting shiyuan, were printed and used extensively, especially as the literary tendencies of Gozan Zen grew. For example, the Zuting shiyuan was printed in Japan, probably in the Muromachi period (1338–1573), then in a movable type print in the period between 1624 and 1644 in Kyoto, followed by woodblock prints in 1647 and 1758. A kunten version was made in 1647 (Takamatsu 1993, p. 42). The great Zen philologist Mujaku Dōchū used the Zuting shiyuan extensively in his Kattōgosen 葛藤語箋 (Dictionary of Entangling Words) (55 times), Zenrin shōkisen 禪林象器箋 (Dictionary of the Images and Implements of Zen Monasteries) (over 60 times), and Wan’un reiu 盌雲靈雨 (Numinous Rain of the Buddha’s Cloud Canopy) (29 times) (Zheng 2021, pp. 17, 25–26). The Shoroku zokugokai 諸録俗語解 (Explanations of the Colloquialisms of the Zen Records), a Japanese counterpart to the Zuting shiyuan, begun by Keishū Dōrin 桂洲道倫 (1714–1794) and completed by others in 1804 (Yoshizawa 1999, pp. 1, 414), which explains terms from Chinese Chan texts in Japanese, quotes the Zuting shiyuan twenty-eight times.

The Chanlin baoxun was probably introduced to Japan in 1287 and kept as blocks in Kamakura and reprinted (?) in 1631. It was printed in Kyoto in 1681 and several times thereafter. Mujaku’s Zenrin shōkisen quotes it twelve times, and the Shoroku zokugokai quotes it six times. The Chanlin baoxun yinyi was printed in 1635 and in 1663 in Japan and was quoted twice in the Shoroku zokugokai. The Chanlin baoxun shunzhu printed in Jiaxing in 1683 was also quoted four times in the Shoroku zokugokai. There was a Japanese “commentary”, the Zenrin hōkun shūi 禪林寶訓拾遺 (Unrecorded Information on the Valuable Glosses on Chan Monasteries), printed in 1721 and quoted in the Shoroku zokugokai four times (Ono 1932–1936, 6: 411–412; and index in Yoshizawa 1999).

The Chanlin leiju was “summarized” by Shanjun during the Yuan as the Chanlin ba leiju 禪林抜類聚 (Selections from the Encyclopedia of Chan Monasteries). Both texts were popular in Japan. The former was printed in the period 1345–1349 and again in Kyoto in 1367, then in 1649, 1659, 1677, and again in 1711. It was commented on by Ban’an Eishu 萬安英種 (1591–1654) in the shōmono 抄物 (Japanese colloquial commentaries) genre and Shiban 師蠻 (1626–1710) added kunten reading marks to it in 1675. There is also a Zenrin ruijū satsuyōshō 禪林類聚撮要抄 (Selected Extracts from the Encyclopedia of Chan Monasteries) produced in 1650 (Ono 1932–1936, 6: 412–413; Takamatsu 1993, pp. 42–47). The Shoroku zokugokai quotes the Chanlin leiju twenty-one times.

Others, such as Giyō Hōshū 岐陽方秀 (1361–1424), wrote linguistically-informed commentaries on works like the Biyan lu, which came to function as dictionaries. For the formal language, Kokan Shiren (1278–1346) compiled the Zengi gemonshū 禪儀外文集 (Collection of External Texts of Zen Ritual), rather like the Chanlin shuyu kaozheng, and it was commented on by Mujaku, who also wrote exemplars and explanations of this genre. The arrival of the Mt. Huangbo (Ōbaku) Chan monks with their Min (Fujian) “dialect” and the rise of the Kogaku branch of Japanese Confucianism further promoted the philological study of colloquial Chinese. It was in this environment that the greatest Zen philologist, Mujaku Dōchū, revived Rinzai scholarship and wrote many works on the Chan koine, informed in part by the Chan philology from the Greater Hangzhou region.

9. Conclusions

Dialect and language are unavoidable topics in Chan and Zen. There were, of course, language barriers between Chinese and Japanese, but there were also “dialect” barriers between Chinese and problems of changed meanings in the colloquial that created difficulties for later generations. The wide range of texts and registers of language meant that, eventually, philological studies were required. It is curious, however, how many of these originated in the Wu-speaking area of Zhejiang. This cannot be explained solely on the basis of the economic prosperity and political clout of the region, for the first truly philological work, the Zuting shiyuan, was produced before Hangzhou became a capital and experienced unprecedented economic growth in the Southern Song. Zhejiang was the smallest province in China, and from the tenth century on, wielded an influence in Buddhism well out of proportion. I think the “conflict of dialects”, with waves of Northern Mandarin-speaking immigrants, both economic and political, into the region, was a factor. Thus, native Wu-speaking Chan monks were more aware of language problems, and so they contributed to the creation of a Chan philology. Because Zhejiang was the chief destination for Japanese monks wishing to learn Chan, this influence flowed on into Japan and Korea.

This article began by linking the rise of philology in Chan to the posited existence of a standard language supported by the court. During the Tang dynasty, there was no evidence of state or literati intervention in the standardization of the language written down by Chan monks, unlike in the early Song. Moreover, although it is widely agreed that there was a standard court-based language in the Tang, I argue that there is evidence that it was not spoken or understood in every corner of the Tang Empire. In fact, there may have been a number of “common languages”. Therefore, there is a possibility that some Chan texts did not conform to the norms of the court language and that dialect influences and differences in preferences in the use of certain characters to represent vernacular grammatical markers changed over time and space. Moreover, the comparison of the Chan language of the Tang with a posited standard language is difficult because of a paucity of evidence for what the standard court language was like.

However, there was an increasing tendency to standardize the Chan language under the influence of the new Song regime’s desire to differentiate itself from the Tang dynasty via a modified court language and court-supported literary compilations. While it seems there were moves by some Chan monks to adhere to a written language that was more consistent and standardized, influential courtiers such as Yang Yi intervened to edit and amend Chan texts so they would conform to the court-sanctioned language of “imperial quality”. As there is a dearth of information about that court language from the Northern Song, only indirect conclusions can be made as to what these amendments involved from limited comparisons between texts like the Jingde chuandeng lu and the problematic Zutang ji or from comparisons of the Tang and Song versions of the Platform Sutra.

Despite such attempts at linguistic standardization, it is evident that not all Chan authors adhered strictly to these models, for according to Muan Shanqing, they used inappropriate characters and dialect or dated vocabulary. Shanqing was often prescriptive, saying character A should be written as B, basing this on court-sanctioned rhyme texts and other works of philology. While Shanqing was trying to encourage authors to use the normative standards established in the Jingde chuandeng lu and early yulu, his awareness of the issue of language was probably heightened because he was a native of the Greater Hangzhou region where the Wu language was spoken. He had traveled to the Northern Song capital, where the court-based standard language was spoken. Most of the Chan figures Shanqing admired were also from the Wu-speaking area or had connections with it, and the majority of the works he commented on in the Zuting shiyuan were linked to the Wu-speaking area. The Zuting shiyuan is a work of philological scholarship, and it became a guide for monks in China, Japan, and Korea.

Many of the later works of Chan philology, though limited in number, were written or published in the Wu-speaking region. Even when they were written by monks who did not come from this region, they were written in adjacent areas of South China (Jiangnan), where Mandarin was not the native language. In addition, the first Chan “encyclopedia” was compiled in the Wu-speaking region.

While the reason for this concentration of Chan philological works originating in the Wu-speaking area is unclear beyond the awareness of language inspired by the contrast of Wu language and the court-based Mandarin, there is room for investigation of the role printing, commerce, and high concentration of population in Hangzhou had on the production and circulation of philological texts.

The Greater Hangzhou region during the Southern Song and Yuan dynasties was also where Japanese monks came to study, and it is noteworthy that these Japanese took back and used works of Chan philology, such as the Zuting shiyuan, as guides to the texts they were studying. Eventually, Japanese Zen monks wrote their own philological works, often based in part on texts like the Zuting shiyuan.

Further investigations of the Chan language need to be conducted by historical linguists armed with a deep knowledge of the provenance and history of Chan texts from the manuscripts of Dunhuang to the Song dynasty yulu. This article, moreover, is only a preliminary investigation into the philology of Chan and Zen Buddhism, and it is likely that other contributions to philology may be buried in the vast Chan and Zen literature.