Four Challenges Faced by Early Chinese Buddhist Translators: A Case Study of Zhi Qian’s Chinese Translation of Dhammapada

Abstract

This study focuses on the translation of Buddhist scriptures into Chinese, specifically the Faju jing, a Chinese version of the Dhammapada completed in the third century CE. It reveals that the Faju jing is not a straightforward translation but a combination of different sources. The translator, Zhi Qian, faced challenges in integrating multiple translation practices, dealing with diverse original Indian languages, incorporating pre-existing phrases from other translators’ work, and managing divergent opinions within the translation team regarding the translation style. This multi-layered process of translation, involving the participation of multiple translators, also likely occurred in other early translations. These challenges extended beyond mere comprehension of the Indian text, resulting in potential errors and deviations from straightforward translations. It is possible that some mistranslations were a consequence of integrating multiple traditions within the source text, making it difficult for translators to maintain a consistent linguistic framework and leading to errors. Furthermore, this study highlights the remarkable efforts of Chinese translators who collaborated with foreign monks in translation groups. It emphasizes the important role of Chinese translators in integrating diverse translation processes and refining the language to suit Chinese readers. They incorporated earlier translations and modified the language to align with Chinese forms. Overall, this case study sheds light on the complexity of early Chinese Buddhist translations, influenced by the integration of multiple traditions and the localization of the texts. It underscores the significance of Chinese translators in the translation process and their contributions to the development of Chinese Buddhist literature.

1. Introduction

The translation of Buddhist scriptures into Chinese has long been a subject of scholarly study.1 It provides valuable insights into how early Chinese society assimilated foreign cultural traditions. However, the limited historical records from this era make it difficult to establish specific details. Fortunately, the Faju jing 法句經 (T no. 210), a Chinese version of the Dhammapada completed around 224 CE, serves as a unique translation that encompasses complex translation processes.

The Dhammapada is an essential Buddhist text available in several languages and has a global influence.2 And the Chinese version, the Faju jing, has been extensively quoted3 and has significantly contributed to the dissemination of Chinese Buddhism. In-depth analyses of this classic Chinese translation have been conducted by numerous scholars, such as Willemen (1974), Mizuno (1981), Su (2018), Dhammajoti (1995, 2009), and Nattier (2023). It is worth noting that Ji (1993, p. 203) has suggested that the Faju jing can help understand other languages, such as Tocharian. Dhammajoti (1995, pp. 91–93) suggests that by studying the Faju jing, we can investigate the transmission of other language versions and determine which is the more authentic or which has undergone subsequent modifications. Thus, the Faju jing is not only a significant text in the history of Chinese Buddhism but also plays an essential role in the transmission of Buddhism as a whole. Moreover, since the existing studies have addressed the complex situation of the Faju jing from multiple perspectives, it is now necessary to integrate these studies and explore the translation process of the text in a new way.

It is worth noting that Zhi Qian (支謙, fl. ca. 222–254 CE), the primary translator of the Faju jing, is a renowned figure in the field of Chinese Buddhist translation and has been extensively studied by scholars.4 Zhi Qian lived during the Three Kingdoms period and showed an early interest in learning various languages as well as studying various classics. His teacher was Zhi Liang (支亮, fl. ca. early third century), whose teacher was the famous translator of the Later Han Dynasty, Zhi Loujiachen 支婁迦讖 (Lokaṣema, fl. ca. 170–190 CE). Zhi Qian’s experience provided him with a solid foundation for translating Buddhist scriptures. He was responsible for translating many early Mahayana scriptures, including the earliest extant version of Larger Sukhāvatīvyūha-sūtra, the Da amituo jing (大阿彌陀經, T no. 362);5 the earliest extant version of Buddhāvataṃsaka-sūtra, the Pusa benye jing (菩薩本業經, T no. 281);6 and the earliest extant version of Vimalakīrti-nirdeśa-sūtra, Weimojie jing (維摩詰經, T no. 474).7 According to Zürcher (2007, p. 50), “he (Zhi Qian) was, in fact, the only important translator in Southern China before the late fourth century”. Thus, studying the life and works of such an important figure in Chinese Buddhist translation can provide valuable insights into the translation practices and habits prevalent in Chinese Buddhist history.

In summary, the in-depth study of Zhi Qian’s Faju jing can provide a reliable and authentic reference for comprehending the challenges encountered by Chinese Buddhist translators.

2. The First Challenge: Integrating Diverse Translation Practices

The Preface of Faju jing provides a comprehensive account of the entire translation process. However, previous scholarship has primarily focused on the discussion of translation style in the Preface, overlooking the translator’s integration of diverse translation practices.8

The author of the Preface is the translator Zhi Qian himself,9 and the specific content is as follows:

曇鉢偈者，眾經之要義。曇之言法，鉢者句也。而《法句經》別有數部，有九百偈，或七百偈及五百偈……近世葛氏傳七百偈，偈義致深。譯人出之，頗使其渾漫……

始者維祇難出自天竺，以黃武三年來適武昌。僕從受此五百偈本，請其同道竺將炎10為譯。將炎雖善天竺語，未備曉漢，其所傳言或得胡語，或以義出音，近於質直。僕初嫌其辭不雅。11維祇難曰：“佛言：‘依其義不用飾，取其法不以嚴。’其傳經者，當令易曉，勿失厥義，是則為善。”座中咸曰：“老氏稱：‘美言不信，信言不美。”仲尼亦云：“書不盡言，言不盡意。”明聖人意深邃無極。今傳胡義，實宜經達。

是以自竭12受譯人口，因循本旨，不加文飾。譯所不解13，則闕不傳。故有脫失，多不出者。……昔傳此時有所不出，會將炎來，更從諮問，受此偈等，重得十三品。并挍往故，有所增定，第其品目，合為一部三十九篇，大凡偈七百五十二章。庶有補益，共廣聞焉。

(CBETA, T55, no. 2145, pp. 49c20–50a28)14

The verses of Tanbo (曇鉢, Dhammapada) is the essence of all sutras. Tan (曇) means teaching (Dhamma), while Bo (鉢) means sentences (Pada). There are various versions of the Dhammapada, such as the 900-verse, the 700-verses, and the 500-verse versions…In recent time, Ge Shi (葛氏) has transmitted the 700-verse version. The meaning of the verses is too profound. The translator rendered it in a mixed and disorderly manner…

In the beginning, Weiqinan (維祇難) came from India and arrived in Wuchang in the third year of the Huangwu period (224). I received this text of five hundred verses and asked his fellow monk Zhu Jiangyan (竺將炎) to translate it. Although Zhu Jiangyan was proficient in Indian languages, he was not familiar with Chinese, and his translations sometimes transliterated Indian words, sometimes translating the phonetic words directly,15 being rather straight forward and unadorned.

At first, I disliked his language for its inelegance. Then, Weiqinan said, “The Buddha said, ‘Relying on the meaning, without any need of adornments; taking the teaching, with any need of decoration’. The one who transmits the scripture must make it easy to understand and not lose its meaning. This is what is considered good”. Everyone in the assembly said, “Laozi (老子) said, ‘Trustworthy words is not beautiful, and beautiful words is not trustworthy’”. Confucius also said: ‘The Scriptures does not thoroughly express the speech, nor does the speech thoroughly express the thoughts’. We should know that the thoughts of the saints are limitless profound”. Now, as the conveying the meaning of the Indian text, it is appropriate to translate it directly and faithfully.

Therefore, I carefully followed the original text from the translator without adding any embellishments. If there were any passages that Zhu Jiangyan did not understand, they were omitted from the translation. As a result, some parts were lost and not translated…

Earlier on, some parts of the text were missing. As Zhu Jiangyan arrived, I consulted with him again and received verses, resulting in a total of thirteen additional chapters. Further additions and revisions were made, and the text was compiled into thirty-nine chapters with a total of seven hundred and fifty-two verses. I hope that this text will be beneficial and widely circulated.

16

The Preface elucidates four crucial points concerning the text’s translation process.

Firstly, it emphasizes that the text underwent two separate translation processes before taking its present form. During the first process, Weiqinan brought a text comprising 500 verses, which was subsequently translated by Zhu Jiangyan. The second process entailed Zhi Qian reconsulting with Zhu Jiangyan and retranslating the text with some revisions and adjustments. Secondly, the Preface explicates the intricacy of the original language of the Faju jing. The challenge faced in comprehending certain verses during the first translation process implies that the obstacle was not in the content, which does not involve profound philosophical analysis, but in the language.17 Thirdly, the Preface emphasizes that the Faju jing is a composite text comprised of different traditions.18

In addition, the Preface to Faju jing reveals that Zhi Qian was not only the translator but also responsible for the final editing and organizing of the text. The Preface also suggests that multiple translators contributed to the translation effort. In the first translation, Zhu Jiangyan was the actual translator, while Zhi Qian’s role was to record it in Chinese.19 In the second translation, the Preface did not identify the real translator, and it should be considered a joint effort by Zhu Jiangyan and Zhi Qian, with Zhi Qian responsible for the final editing. Additionally, the Preface suggests that a team of translators may have worked on the project, with input from various individuals beyond Wei Qinan, Zhu Jiangyan, and Zhi Qian. The Preface employs phrases such as zuo zhong xian yue 座中咸曰 (all those present said) and implies that other individuals’ opinions may have been involved.20

In brief, during the translation of this sutra, Zhi Qian had to integrate diverse translation practices from different translators in various processes.

3. The Second Challenge: Dealing with Multilingual Source Materials

One of the most difficult tasks in early Chinese Buddhist translation is dealing with different kinds of languages. The strong relationship between Faju jing and the Pāḷi Dhammapada (hereafter Dhp) is widely acknowledged among scholars, with 26 chapters of the former believed to have been translated from the latter (see Lü 1991, p. 644; Shi 2011, pp. 647–53). Despite this similarity, scholars have also identified differences between the two versions, leading to the suggestion that these variations may have been influenced by other linguistic traditions, such as the Gāndhārī Dharmapada (hereafter GDhp), Patna Dharmapada (hereafter PDhp), and Udānavarga (hereafter Udv).21 Through a comparative analysis of these different linguistic traditions, we can determine their relationship to the Faju jing.

Firstly, there are instances where Faju jing is identical to the Udv.

Faju jing 17.9d: 漸盈大器 (gradually filling a large container)22;

Dhp 121d: udakumbho pi pūrati (water pot is filled);

PDhp 193d: udakumbho pi pūrati (water pot is filled);

GDhp 209d: udakubho va puyadi (water pot is filled);

Udv 17.5d: mahākumbho ’pi pūryate (large pot is filled).

Faju jing 17.9d uses the term daqi 大器 (large container), which is different from the udakumbho (water bottle or water jug) in Dhp 121d, PDhp 193d, and GDhp 209d, but it is the same as Skt. mahākumbho (large pot, large jar) in Udv 17.5d. Other examples include Faju jing 36.2c, which reads 厚為最友 (loyalty is the best friend) and uses the term you 友 (friend), which is different from ñātī (relatives) in Dhp 204c and PDhp 76c, but the same as mitra (friend) in Udv 26.6c and GDhp 162c. Additionally, Faju jing 35.16c reads 自覺出塹 (self-wakened and out of the moat), and 35.32b reads 墮塹受苦 (suffering in the moat), both use the term qian 塹 (moat), which is different from palighaṃ (latch, obstacle) in Dhp 398c and palipathaṃ (obstacle, mud, dangerous road) in Dhp 414a, but the same as parikhaṃ (ditch) in Udv 33.58c and Udv 33.41a.23 These examples provide evidence for the relationship between Faju jing and the Udv:

Secondly, there are instances where Faju jing is only identical to the PDhp.

Faju jing 33.2c: 愚為此害賢 (the foolish damages the wise);

Dhp 72c: hanti bālassa sukkaṃsaṃ (it destroys the fool’s good happiness);

Udv 13.2c: hanti bālasya śuklāṃśaṃ (it destroys the fool’s good happiness);

PDhp 177c: hanti bālassa śukrāṅggaṃ (it destroys the fool’s good quality).

In Faju jing 33.2c, it is the xian 賢 (wise) that is damaged, which is different from the sukkaṃsaṃ/śuklāṃśaṃ (brightness or happiness)24 that is damaged in Dhp and Udv. However, xian 賢 is closer to śukrāṅggaṃ (good qualities)25 in PDhp 177c. This demonstrates the connection between the Faju jing and the PDhp:

Faju jing 1.17b: 亦非父兄 (neither father nor brother);

Dhp 288b: na pitā na pi bandhavā (nor fathers, not even relatives);

Udv 1.40b: na pitā nāpi bāndhavāḥ (nor fathers, not even relatives);

GDhp 261b: na bhoa na vi banava (nor fathers, not even relatives);

PDhp 366b: na pitā no pi bhātaro (nor fathers, not even brothers).

In Faju jing 1.17b, the character xiong 兄 (brother) diverges from bandhavā (relatives) in Dhp 288b,26 bāndhavāḥ in Udv 1.40b, and banava in GDhp 261b, but corresponds with bh-ātaro (brothers) in PDhp 366b.

Other analogous instances can be found. For example, Faju jing 17.20a uses the character duo 墮 (fall) in 有識墮胞胎 (sentient being falls into the womb), which differs from upapajjanti (are born, arise) in Dhp 126a27, but aligns with okraṃmanti (enter, fall into) in PDhp 274a. Additionally, in Faju jing 31.12d, the character e 惡 (evil) appears in 眾惡不犯安 (it is safe to refrain from doing evils), which contrasts with dukkhassa (of suffering) in Dhp 331d and duḥkhasya (of suffering) in Udv 30.34d, but agrees with pāpassa (of evil) in PDhp 65d.

Moreover, as there are several verses in the Faju jing that are not found in any other surviving Indian texts, it is only reasonable to suggest that they have originated from unknown sources. The terms indrakīla in Śarīrārthagāthā and indakhīla in Dhp 95b both denote “a bar or bolt for a gate or door”. However, the corresponding verse in Faju jing 15.7b (不動如山) employs the term shan 山 (mountain), which carries a distinct meaning. Nevertheless, Monier-Williams et al. (1999, p. 166) indicate that indrakīla assumes the meaning of “mountain” in the Mahābhārata. Additionally, the Chuyao jing (出曜經, T no. 212), extensively referencing Faju jing, utilizes the term Anming 安明 in the phrase 不動如安明, which indeed designates a mountain. This particular instance suggests that Faju jing draws from an unidentified or unique source.

Nonetheless, there are also instances that indicate the Chinese translation combines different traditions. One such instance is found in Faju jing 18. 5, where the phrase 死入地獄，如是為十 (die into hell; this is the tenth situation) appears. The Chinese translation includes the phrase ru diyu 入地獄 and the numeral shi 十. However, in Dhp 140d, the phrase nirayaṃ so upapajjati (he is reborn in hell) is used, while Udv 28.29d employs daśamāṃ durgatiṃ vrajet (the tenth situation is to go to evil destinies).28 These terms differ from the Chinese translation. None of the surviving versions match the Chinese translation, suggesting that the translator may have merged different traditions and made modifications.

These examples demonstrate that the relationship between the Faju jing and other various language versions is not limited to specific chapters but can be found throughout the entire text of the Faju jing. This implies that the translation of the Faju jing was a complex process that involved various sources, and it can be difficult to discern their individual contributions.

4. The Third Challenge: Incorporating and Refining Previous Translations

There are so many Buddhist scriptures that have been translated into Chinese throughout history. It is inevitable that some of them have almost the same content but were translated by different translators. How does the latest translator deal with the work of the previous translators? One of the methods is to revise the work of the previous translators into a new one. Zhi Mindu (支敏度, fl. first half of the fourth century) in He shoulengyangjing ji 合首楞嚴經記 (Note on the Combination of the Śūraṃgama-sūtra) pointed out that Zhi Qian might have modified the works of Zhi loujiachen, who was active during the Han Dynasty.29 There have been many academic discussions on the phenomenon of Zhi Qian’s revision of previous translations, including works by Zürcher (2007, p. 50), Okayama (1980, p. 735), Matsuda (1988, pp. 484–85), Harrison (1998, p. 557), and Harrison et al. (2002, p. 180). Nattier (2008, pp. 177–78) suggested that nine of the 24 core texts of Zhi Qian are revisions of previous translations.

4.1. Predecessors’ Translations

The use of predecessors’ works in the Faju jing adds another layer of complexity. Nineteen verses within the Faju jing display significant similarity to the Zhong benqi jing (中本起經, T no. 196), a text translated by Kang Mengxiang 康孟詳 in CE. 196–220.30 However, with the exception of three verses, the remaining sixteen exhibit varying degrees of modification.

Three types of modifications can be identified in the Faju jing. The first type involves changes based on different Indian traditions.

Zhong benqi jing: 持戒終老安 信正所止善 智慧最安身 眾惡不犯安.

31

Faju jing 31.14: 持戒終老安 信正所正善 智慧最安身 不犯惡最安.

Dhp 333: sukhaṃ yāva jarā sīlaṃ sukhā saddhā patiṭṭhitā sukho paññāya paṭilābho pāpānaṃ akaraṇaṃ sukhaṃ.

32

The instance reveals that the verse in Faju jing 31.14 bears a striking resemblance to the verse in Zhong benqi jing. The two verses are nearly identical, except for the final sentence, where the plural meaning of evil (zhonge 眾惡) in Zhong benqi jing is replaced with the singular evil (e 惡) in Faju jing. Comparison of various Indic-language versions demonstrates that the term pāpānaṃ in Dhp 333d is plural, whereas pāpasya in Udv 30.20d and pāpassa in PDhp 82d are singular. Thus, Zhi Qian appears to have copied the verse from the Zhong benqi jing and subsequently modified it based on the Indic-language versions similar to Udv and PDhp.

Several additional examples demonstrate similar phenomena. For instance, the verse 快哉福報 所願皆成 敏於上寂 自致泥洹 at Faju jing 22.13 largely follows the verse 快哉福報 所願皆成 上寂大人 自見泥洹 in Zhong benqi jing. However, the third line of the verse in the Zhong benqi jing reads上寂大人 (great being of nirvana), while in the Faju jing, it is changed to 敏於上寂 (diligent for nirvana). The meaning of daren 大人 (great being) in the context is unclear and lacks any corresponding word in other language traditions. By contrast, the use of min 敏 (quick, diligent) in the Faju jing makes the reference clearer and corresponds to kṣipraṃ (quickly) in Udv 30.11c.

The Zhong benqi jing employs the verse 雖壽千年 亦死過去 (one may live for a thousand years and yet die), which is modified in Faju jing 1.12 by replacing the term qiannian 千年 (thousand years) with bainian 百年 (a hundred years). The substitution aligns more closely with the Indic-language version Skt. varṣaśataṃ in Udv 1.30a. Additionally, the sentence 不如尊正諦 (it is better to respect righteousness) in the Zhong benqi jing is altered in Faju jing 2.18 by replacing zunzhengdi 尊正諦 with xuyujing 須臾敬 (respect in a moment). This change corresponds to the phrase muhuttampi pūjaye in Dhp 107d.

The last example of modification found in the comparison of the Zhong benqi jing and Faju jing is particularly interesting. In the former, the phrase 至道無往返 (ultimate truth leads to no return) is employed, whereas in the latter, Faju jing 36.21 modifies this by replacing zhidao 至道 (ultimate truth) with wo 我 (I). As a result of this change, the sentence now means, “I have already achieved a state of no return”. The modification of the subject is a topic of interest. The phrase in Udv 26.25a is naivāgatir na ca gatir (neither coming nor going), while Ud 8.1 reads neva āgatiṃ vadāmi, na gatiṃ with the addition of vadāmi (I say), indicating that the Buddha himself is speaking the contents of the verse. It is possible that when Zhi Qian translated the Dhammapada, he had access to a text similar to Ud 8.1 and included vadāmi in his translation, treating “I” as the subject being described.

It is notable that in some cases, the Faju jing only makes adjustments to specific words and phrases in the Zhong benqi jing without altering the overall meaning of the text. For instance, in the Zhong benqi jing, the phrase 誰能致不死 (who can achieve immortality) is used, while in the Faju jing 37.1d, the character shui 誰 is changed to shu 孰. Interestingly, both shui and shu have been used by Zhi Qian in his other translations, with Faju jing utilizing both shui (one instance) and shu (five instances). It should be noted that the interrogative pronoun shu had already replaced shui during the Eastern Han Dynasty.33 Therefore, Zhi Qian’s choice of the older interrogative pronoun reflects the translator’s personal preference.

In the Zhong benqi jing, the phrase 是處為泥洹 (this is the place of nirvana) is used, while in the Faju jing 36.21d, the word chu 處 (place) is changed to ji 際 (boundary).34 The reason for this change can be traced to Zhi Qian’s personal preference for translation. Zhi Qian consistently used the character ji 際 to translate Skt. anta or other similar Indian words. In his translation of the Faju jing, for instance, the Indian word Skt. duhkhāntaḥ (the boundary of suffering) is rendered as 苦際:

Faju jing 6.8ab: 諸念、生、死棄 為能作苦際.

35

Udv 15.6: samyojanaṃ jāti jarāṃ ca hitvehaiva duḥkhasya karoti so ’ntam (By abandoning attachment to bond, birth, and death, one can attain the end of suffering).

Faju jing 36.8ef: 無所樂 為苦際.

36

Udv 26.14：tṛṣṇākṣayaṃ paśyato hi duḥkhasyānto nirucyate (For one who sees the cessation of craving, the end of suffering is declared).

The term ji 際 can also convey the meaning of “the end or limit of something” in the Faju jing. For example, in Faju jing 1.12cd, the phrase 為老所壓 病條至際37 (when old age comes and illness strikes, life comes to an end) corresponds to the sentences anu hy enaṃ jarā hanti vyādhir vā yadi vāntakaḥ (he will be destroyed to death by aging or illness) of Udv 1.30d. In this context, ji 際 corresponds to the Skt. antaka and indicates “the end”.

In general, in the Zhong benqi jing, the phrase 是處為泥洹 can be understood as “this state/situation/context is Nirvana”, while in Faju jing 36.21d, the phrase 是際為泥洹 emphasizes that “this end/limit (of suffering/samsara) is nirvana”. This change indicates Zhi Qian’s careful consideration of the nuances of the Indic language sources and the flexibility of the Chinese language. This demonstrates Zhi Qian’s scholarly rigor and his commitment to accurately conveying the teachings of Buddhism to Chinese readers.

Another special situation is shown in

Table 1. The corresponding relationships of the Zhong benqi jing and the Faju jing in different Indian texts.

Zhong benqi jing	Faju jing 1.12	Sn 589	Udv 1.30
雖壽千年 亦死過去	雖壽百歲 亦死過去 為老所壓 病條至際	api ce vassasataṃ jīve bhiyyo vā pana mānavo ñātisaṃghā vinā hoti jahāti idha jīvitaṃ	yo ’pi varṣaśataṃ jīvet so ’pi mṛtyuparāyaṇaḥ anu hy enaṃ jarā hanti vyādhir vā yadi vāntakaḥ38
	Faju jing 1.8		Udv 1.22
合會有離	常者皆盡 高者亦墮 合會有離 生者有死		sarve kṣayāntā nicayāḥ patanāntāḥ samucchrayāḥ samyogā viprayogāntā maraṇāntaṃ hi jīvitam39
	Faju jing 1.17	Dhp 288
無親可恃	非有子恃 亦非父兄 為死所迫 無親可怙	na santi puttā ttāṇāya na pitā no pi bhātaro antakenādhibhūtassa nāsti ñātīsu ttāṇatā40

.

Table 1 illustrates that the four-line verse found in the singular form in the Zhong benqi jing is distributed across three verses of the first chapter of Faju jing. It is evident that the content and sequence of the verse in the Zhong benqi jing deviates from all known Indic language versions. Nevertheless, all three verses in the Faju jing can be traced back to the Indic language versions, and the content and sequence display a higher degree of similarity. They belong to the same system as Udv 1.30, Udv 1.22, and PDhp 366. Notably, examples of splitting the verses of Zhong benqi jing and placing them in different verses of Faju jing can also be found in two other places.41

As evident from the discussion above, the Faju jing not only preserved the content of the Zhong benqi jing but also made certain modifications. These changes range from individual word changes to entire sentence modifications and even scattering complete verses in different sections. Some of the alterations in the Faju jing can be traced back to the Indiac language versions. However, for certain changes, the reason behind them remains unknown. It is evident from the differences in expression that the translator consulted an unknown tradition during the translation process.

4.2. Zhi Qian’s Other Translations

Zhi Qian not only incorporated verses from his predecessors but also from his own translations. Fourteen verses in the Faju jing can be found in other texts translated by Zhi Qian, such as the Taizi ruiying benqi jing 太子瑞應本起經 (T no. 185), Liaoben shengsi jing 了本生死經 (T no. 708), and Beijing chao 孛經抄 (T no. 790). With the dating of Zhi Qian’s translations being unclear, it becomes challenging to ascertain whether the Faju jing derived from his other translations or vice versa. Nevertheless, through a comparative analysis of the similarities between them, one can discern the translator’s diverse considerations during the process of translation.

The connection between the Faju jing and the Taizi ruiying benqi jing is evident in verse 貪婬致老 瞋恚致病 愚癡致死 除此三者 乃可得道 (Greed and lust cause aging, anger causes illness, foolishness and ignorance cause death; only by eliminating these three can one attain the way). In Faju jing 28.12, the last two sentences of the verse are condensed into a single sentence 除三得道 (eliminating three to attain the way).42

Another example list below:

Taizi ruiying benqi jing: 無病第一利 知足第一富 善友第一厚 無為第一安.

43

Faju jing 36.2: 無病最利 知足最富 厚為最友 泥洹最快.

44

Dhp 204: ārogyaparamā lābhā, santuṭṭhiparamaṃ dhanaṃ, vissāsaparamā ñātī, nibbānaṃ paramaṃ sukhaṃ.

45

Udv 26.6: ārogyaparamā lābhā, saṃtuṣṭiparamaṃ dhanam, viśvāsaparamaṃ mitraṃ, nirvāṇaparamaṃ sukham.

46

In this instance, a distinction can be observed between the Taizi ruiying benqi jing, which employs a five-character-verse sutra, and the Faju jing, which uses a four-character-verse. This difference is attributed to the use of different words to correspond to the Pa. parama (highest, supreme). Specifically, the Taizi ruiying benqi jing utilizes the disyllabic word diyi 第一 (first), whereas the Faju jing, uses the monosyllabic zui 最 (most).

It is worth noting that both the Faju jing and the Indiac language versions share a similar structure of linking an abstract quality to a worldly benefit through four lines. However, the phrase 厚為最友 (loyalty is the best friend) in the Faju jing diverges from 善友第一厚 (good friends are most trustworthy) in the Taizi ruiying benqi jing. The former highlights the abstract quality of hou 厚 (loyalty), which corresponds to Pa. vissāsa and Skt. viśvāsa (closeness, trust), while the latter underscores the significance of “good friends” and their trustworthiness. The latter form of content is different from all the other traditions.47

In general, three verses in the Faju jing exhibit similarities with those in the Taizi ruiying benqi jing, with one verse abbreviated and one modified. This suggests that the translation of the Taizi ruiying benqi jing preceded that of the Faju jing, and Zhi Qian drew from and adapted the former when translating the latter.

The second point of comparison concerns the relationship between the Faju jing and the Liaoben shengsi jing 了本生死經 (T no. 708).48 Notably, they differ in the wording of the second sentence.

Faju jing 37.15: 性癡淨常想 樂身想疑想 嫌望非上要 佛說是不明.

Liaoben shengsi jing: 性癡淨常想 樂想身想疑 嫌妄非上要 佛說是不明.

49

Further scrutiny reveals that the syntax of le shen xiang 樂身想 (the concepts of happiness and body) in the Faju jing ’s second sentence follows a pattern of “object 1 (樂), object 2 (身) + verb (想)”, which is consistent with the phrasing of the first sentence, jing chang xiang 淨常想 (the concepts of purity and eternal constancy). On the other hand, the phrase 樂想身想疑 in the second sentence of the Liaoben shengsi jing is unclear in meaning. Overall, the phrasing of the Faju jing ’s verse appears to be more reasonable.50

Finally, the relationship between the Faju jing and the Beijing chao 孛經抄 (T no. 790)51 can be examined. The former contains ten verses that are highly correlated with the latter but with some modifications to the wording.52 It is evident that the Faju jing has made efforts to conform to the Indic language tradition.

The sentence 快心放意 (unrestrained and indulgence of the mind) in Beijing chao uses the phrase fangyi 放意 (indulgence), which does not have a direct equivalent in the Indiac language tradition. However, the sentence 快心作惡 (unrestrained and committing evil actions) in Faju jing 13.9 utilizes e 惡 (evil), which can be associated with the term pāpakaṃ (evil) in Dhp 66c. The sentence 有佛興快 (having the Buddha on earth is something to be happy about) in Beijing chao uses fo 佛, which is in the singular form and corresponds to buddhasya in Udv 30.22a. In contrast, the sentence 諸佛興快 (having the Buddhas on earth is something to be happy about) in Faju jing 22.21 uses zhufo 諸佛, which is in the plural form and corresponds to buddhānam in Dhp 194a.

In Beijing chao, the phrase 消毀其形 (destroy its form) has no exact corresponding term in the Sanskrit text. However, in Faju jing 26. 4d, the phrase 反食其身 (eat its body) is used. The character shi 食 (eat) corresponds to the term khādati (eat) in Dhp 240d and Udv 9.19d. Furthermore, while the phrase 攝意從正 (control the mind to be correct) is used in Beijing chao, the phrase 制根從止 (control the faculty to be calm)53 is used in Faju jing 15.5. Here, yi 意 (mind) and gen 根 (faculty) correspond to Pa. indriyāni (faculties of sense) in Dhp 94a. Strictly speaking, gen 根 is closer to indriyāni than yi 意, which is just one of the six faculties.

The comparisons made between the Faju jing and Beijing chao suggest that they contain similar concepts but differ in their specific phrasing. The Faju jing exhibits a deliberate effort to adhere to the Indian text. It is possible that when translating the Faju jing, Zhi Qian incorporated the content of Beijing chao but made modifications based on the original Indian text.

5. The Fourth Challenge: Adapting to Diverse Translation Style

Translators inevitably have their own preferences when translating, but since translation is a complex process that requires translators to balance different kinds of situations, the translation style may have changed a lot. As a Chinese translator, Zhi Qian was known for his high cultural level and preference for an elegant translation style.54 This preference is evident in his revisions of previous works, where he made modifications to reflect his preferred style.

One such example is Zhi Qian’s translation of the Da mingdu jing 大明度經 (T no. 225), where he largely followed the Daoxing banruo jing 道行般若經 (T no. 224) translated by the Eastern Han dynasty translator Zhi Loujiachen. Scholars have compared these two texts and concluded that Zhi Qian’s modifications made the text more concise, elegant, and neat compared to the original version, which was long, disorganized, and plain (see Nattier 2010, pp. 309–11; Hu 2012, pp. 225–26; Ji 2013, pp. 129–31). However, it should be noted that the rules summarized by scholars mainly focus on the Da mingdu jing. When examining the translation of the Taizi ruiying benqi jing, which contains several passages rewritten from Zhong benqi jing, different rules can be seen. For example, the Taizi ruiying benqi jing supplements the sentence components of the Zhong benqi jing and breaks the neat four-character format used by the latter from a grammatical perspective (see Li 2020a, pp. 262–66).

Furthermore, a crucial discussion can be found in the Preface of the Faju jing. Initially, Zhi Qian believed that the expression of Zhu Jiangyan’s original translation was not elegant (其辭不雅), but after being persuaded by Weiqinan and others, Zhi Qian adopted the translation policy of following the original text’s meaning and not overly embellishing it (因循本旨，不加文飾). The examples of translation revisions in this article also demonstrate the translator’s emphasis on the source text, indicating that Zhi Qian considered how to conform to the original meaning to a great extent during the translation process.

In other words, Zhi Qian’s preference for an elegant translation style was a feature during a certain period of his long translation process and cannot represent his entire preference.55 Before translating the Faju jing, his preferred translation style was likely more elegant. However, it is highly likely that during the process of translating the Faju jing, he reconsidered his approach and opted for a simpler translation style that was closely aligned with the Indian text in order to meet the expectations of the translation team.56

6. Rethinking the Factors behind “Mistranslation”

Based on the aforementioned research, the four challenges faced by early Chinese Buddhist translators can be clearly identified. These findings greatly contribute to our understanding of Chinese translations, particularly shedding light on the notable phenomenon of “mistranslation”.

Scholars have previously observed that early translators of Indian texts into Chinese committed errors, which could be ascribed to the confusion between Middle Indic languages and Sanskrit or the translators’ intentional misinterpretation resulting from their cultural context.57 For instance, Faju jing 28.4a translates sabbe saṅkhārā dukkhā in Dhp 278a as 生死非常苦 (rebirth is extremely painful). Su (2016, p. 126) suggests that the mistranslation of Pa. saṅkhārā (all formations) as shengsi 生死 (birth and death, Skt. saṃsāra) may stem from confusion arising from the occurrence of -kh-/-s- in the original Gandhari language.

However, as demonstrated above, the Chinese rendition of the Dhammapada faced four distinctive challenges in its translation process. It is important to note that these challenges potentially played a significant role in the production of a final product that deviated from a straightforward translation. Therefore, some translation errors were probably due to the integration of multiple traditions within the source text, posing difficulty for translators to maintain a consistent linguistic framework in their translations and making them susceptible to errors.

For instance, in Dhp 59b, the sentence andhabhūte puthujjane (amongst ignorant ordinary people) is translated in the Faju jing 12.17b as 凡夫處邊 (normal people at the edge). This could be due to confusion between Skt. andha (blind, dark) and Skt. anta (boundary), as both can be pronounced similarly to anda in Gandhari.58 However, according to Su (2016, p. 127), in the Faju jing 19.1c, the term andhakārena (darkness) in Dhp 146c was accurately translated as youming 幽冥, indicating that the translator had a proper understanding of the word andha (blind, dark, ignorant).

Moreover, Faju jing’s source texts often contain colloquial words, which can have various interpretations, necessitating the translators to randomly select one. For example, Faju jing 1.8a reads 常者皆盡 (all that is constant will cease), corresponding to sarve kṣayāntā nicayāḥ (all that is accumulated will eventually be destroyed) in Udv 1.22a. The term changzhe 常者 (constant) in Faju jing corresponds to Skt. nicayāḥ (accumulation, heap) in Udv 1.22a. However, these two terms have different meanings. In fact, the term nicaya (accumulation, heap) may drop the final syllable and become *nica in Prakrit. Moreover, Skt. nitya (constant, eternal) may also become *nica.59 Therefore, when presented with *nica in the original text, Zhi Qian faced the task of making a decision and translated it as 常者 (constant), which differs from the intended meaning of the extant Indian text.60

Additionally, errors may arise if the quotes in the Faju jing do not match their source texts exactly. For example, one edition of Zhong benqi jing includes the following passage: 我行無師保 志獨無伴侶 積一行作佛 從是通聖道 (I walk without a teacher, with determination and no companions. I practice one path to become a Buddha and follow the holy way). However, some editions of the text replace the word banlü 伴侶 (companions) with denglü 等侶 (equals). The corresponding verse, Udv 21.4, reads: ācāryo me na vai kaścit sadṛśas ca na vidyate eko ’smin loke sambuddhaḥ prāptaḥ sambodhim uttamām (No one is my teacher, and no one is my equal. I am alone in this world, and I have attained supreme enlightenment). The term banlü 伴侶 or denglü 等侶 corresponds to the Skt. sadṛśa (similar, same, matching), so 等侶 is considered the correct word. Furthermore, the Zhong benqi jing describes the story of Śākyamuni’s first sermon to his five original companions, which means “no companions” is not true. Therefore, denglü 等侶 is more appropriate in this context. However, the Faju jing inherited the wrong editions and utilized banlü 伴侶.

Another instance of a translation error can be observed in the Faju jing 15.7c, wherein the phrase 真人無垢 (true man without defilement) appears, corresponding to Udv 17.12c and Dhp 95c. The term zhenren 真人 (true man)61 carries a distinct meaning, differing from hrada (lake) in Udv 17.12c and rahada (lake, pond, water) in Dhp 95c. It is evident that the translators may have made a mistake, but how did it happen? It is noteworthy that Skt. arhant (Arhat) can take the form of *rahada in Gandhari.62 Therefore, when faced with the word *rahada, which can be interpreted as Skt. arhant or Pa. rahada, the translator made a choice and believed it referred to Arhat, so he translated it as “true man”.63 However, it should be noted that the content of Dhp 95 appears not only in Faju jing 15 but also in Faju jing 36. The latter translates the verse as 淨如水無垢 (pure like water, free from impurities) and correctly renders the meaning of the Pa. rahada.64

In fact, the case of the translation error in the Faju jing is more intricate than previously discussed. The sentence 真人無垢, which appears in the Faju jing, was directly inherited from the Zhong benqi jing, indicating that the Faju jing inherited the errors from its source. The Chinese translations encompassed not only multiple Indian traditions from the original text but also integrated various other traditions due to the utilization of works by other translators. These factors contributed to a higher likelihood of errors in the translation process.

This multi-layered process of translation, involving the participation of multiple translators, is likely to have also been a common occurrence in other early translations. An example illustrating this is the Shisong lü 十誦律 (Ten Recitations Vinaya, T no. 1435), originally translated by Furuoduoluo 弗若多羅 and Jiumoluoshi 鳩摩羅什 (Kumārajīva). However, the translation was only partially completed due to Furuoduoluo’s death. Subsequently, Tanmonanti 曇摩難提, the Indian monks, collaborated with Kumārajīva and completed the remaining sections, resulting in a total of 58 volumes. Later, Beimoluocha 卑摩羅叉 revised the text, bringing the total number of volumes to 61.65

Comprehending the practice of integration in translation can provide fresh insights into certain unique phenomena. A notable example of such a phenomenon is the occurrence of overlapping translations and double translations, as identified by Nattier (2004, pp. 8–9). Overlapping translations occur when a portion of an expression, typically one syllable, is interpreted as belonging to two distinct words and is consequently translated twice. For example, the Chinese term duwuji 度無極 corresponds to the Skt. pāramitā, representing an overlapping translation of pāramitā (度) and amita (無極). Similarly, the Chinese term xinjie 信解 corresponds to the Skt. abhimukti, constituting an overlapping translation of abhimukti (信) and mukti (解). Double translations, on the other hand, refer to the practice of assigning two different interpretations to a single term. For instance, in Zhi Qian’s translations, the Chinese term yuanyijue 緣一覺 corresponds to the Skt. pratyekabuddha. This represents a double translation, as it encompasses two similar terms in Prakrit: pratyeka (一) and pratyaya (緣).

These linguistic phenomena were not uncommon in early Chinese translations of Buddhist scriptures.66 Two main explanations for their occurrence have been suggested in existing research: either the translators lacked a proper understanding of the source text’s true meaning, or they intended to convey multiple meanings simultaneously.67 However, these phenomena can be understood within the framework of the translation process, which involves the layering of different versions rather than being a singular, isolated event. The collectors documented the outcomes of various translation processes, which occasionally resulted in double translations. In the similar case of Faju jing, a notable outcome is the presence of 35 sets of verses with identical content but varying wording (see Mizuno 1981, pp. 314–37; Li 2015, pp. 39–52; Su 2014). The translations of these verses often align with either the Pāli version or the Sanskrit/Gandhari language versions, indicating that the Chinese translations have preserved content from different traditions.

In the seventh century, a similar phenomenon can be seen in the translation work of Xuanzang 玄奘 in the Shuo wugoucheng jing 說無垢稱經 (T, no. 476). Xuanzang used the term yaoshi要施 to translate Skt. nimantraya in the Vimalakīrtinirdeśa, which means “to invite” (要) or “offer anything to” (施). It appears that Xuanzang employed a double translation. However, upon examining previous corresponding Chinese versions, it is evident that both Zhi Qian and Kumārajīva used the character 要. This indicates that Xuanzang inherited the 要 and found that it did not adequately convey the true meaning of “offer anything to” in this particular context.68 As a result, the addition of the character 施 in the translation of 要施 serves as another example of the integration of different traditions within the translation process.

Drawing upon the preceding research, it becomes evident that the challenges faced by translators extended beyond their proficiency in comprehending the Indian text. In the early stages of translation history, they were confronted with the task of not only adapting to a single Indian text but also choosing a method of combining multiple complex sources.

7. Unveiling the Role of Native Chinese Translators

Zhi Qian, a renowned figure in Chinese Buddhist history, was a native of China and played a significant role as an early Chinese translator in translation groups. Based on the four types of challenges mentioned above, we can gain a deeper understanding of his contributions to the translation process. This understanding can, in turn, shed light on the roles of other Chinese translators within the translation groups led by foreign monks.

Although previous research has highlighted the importance of Chinese translators’ language skills, there is limited concrete evidence regarding their specific translation practices. As a matter of fact, the biographies featured in Gaoseng zhuan 高僧傳 (T no. 2059) provide concrete examples of the Chinese translators’ actual practices in these translation groups:

時有清信士聶承遠，明解有才，篤志務法，護公出經，多參正文句……又有竺法首、陳士倫、孫伯虎、虞世雅等，皆共承護旨，執筆詳校

(CBETA, T50, no. 2059, p. 327a1-7)

At that time, there was a male lay disciple named Nie Chengyuan 聶承遠, who had a thorough understanding of the principles and was talented, with a firm commitment to Buddhism. He often helped to review and correct the scriptures translated by Zhu Fahu 竺法護. Zhu Fashou 竺法首, Chen Shilun 陳士倫, Sun Bohu 孫伯虎, Yu Shiya 虞世雅 and others also followed Zhu Fahu’s intentions and carried out detailed corrections.

跋澄又齎《婆須蜜》梵本自隨，明年，趙正復請出之，跋澄乃與曇摩難提及僧伽提婆三人共執梵本，秦沙門佛念宣譯，慧嵩筆受，安公、法和對共校定。

(CBETA, T50, no. 2059, p. 328b12-14)

Sengjiabacheng 僧伽跋澄 carried the Indian version of the Poxumi 婆須密 with him. The following year, Zhao Zheng 趙正invited him to translate it. Sengjiabacheng 僧伽跋澄, Tanmonanti 曇摩難提, and Sengjiatipo 僧伽提婆 worked together with the Indian text. Chinese monk Zhu Fonian 竺佛念 preaches and translates the text. Huisong 慧嵩 made a record in Chinese, and Dao’an道安 and Fahe 法和 helped to review and correct this text.

請跋摩譯焉，泰即筆受，沙門慧嵩、道朗與義學僧三百餘人，考正文義，再周方訖，凡一百卷，沙門道挻為之作序

(CBETA, T50, no. 2059, p. 339a16-25)

Sengjiabamo 僧伽跋摩 was invited to undertake the translation, with Daotai 道泰 entrusted to record. Alongside Huisong 慧嵩, Daolang 道朗, and over three hundred monks who specialized in Buddhist teachings and doctrines, they collaborated to verify the accuracy and meaning of the text. This process took two years to complete, resulting in the translation of one hundred volumes. Monk Daoshan 道挻 wrote the Preface for the scripture.

As we can see above, numerous Chinese individuals participated in the initial Buddhist translation process. Their roles are commonly described as canzheng wenju 參正文句, xiangjiao 詳校, jiaoding 校訂, or kaozheng wenyi 考正文義, which involved editing the final text. However, the specific details of these roles and their associated practices remain unclear.

The case study of Faju jing illustrates the crucial role of Chinese translators in integrating diverse translation processes and determining the final style of translation. In the course of their work, they also refined the language and produced the final text. Notably, one of the major responsibilities of Chinese translators was to modify the language to conform to established forms in China, facilitating comprehension by Chinese readers of the newly translated texts. To achieve this, Chinese translators employed two primary methods for refining their translations.

The first method involved utilizing the accomplishments of earlier translators as a foundation and then replacing certain portions of text based on Indic sources or their own aesthetic criteria. For example, Zhi Qian’s Da mingdu jing was translated based on Zhi Loujiachen’s Daoxing banruo jing,69 and Zhu Fahu 竺法護 (Dharmarakṣa)’s Puyao jing 普曜經 (Lalitavistara) was translated based on Zhi Qian’s Taizi ruiying benqi jing.70 Although utilizing existing translations as a foundation for new works is a clever technique, it poses practical difficulties. The most significant challenge is striking a balance between preserving the integrity of the previous translations and adapting to the new Indian text. If a translator seeks to blend these two traditions into a new one, the process can be intricate and daunting. As evidenced in the revision of Taizi ruiying benqi jing, the translator combined elements of both Indian text and pre-existing translations, utilizing the former to frame the beginning and end of the story and the latter as a basis for the main portion (see Li 2020a, pp. 245–46). Such an approach necessitated consulting and integrating two distinct texts simultaneously, a task requiring a deep comprehension of the language and cultural context of each source.

Another method is to use the original text in an Indic language as the basis for translation and selectively incorporate existing phrases and sentences from previous translations. For instance, the Faju jing preserves 19 verses from the Zhong benqi jing, while Zengyi ahan 增壹阿含 (Ekottarikāgamaḥ, T no. 125) includes multiple verses from the Faju jing. It is noteworthy that the Faju jing splits one verse from the Zhong benqi jing and distributes it across four distinct verses, indicating the translator’s familiarity with prior translations. This type of approach necessitates a comprehensive understanding of existing Chinese Buddhist literature, and only those with extensive knowledge can flexibly integrate the works of their predecessors.71 Consequently, this strategy reflects the translator’s expertise and competence.

This line of inquiry can be extended to shed light on the roles of other Chinese translators in their respective teams. For instance, just as mentioned above, the translation of the Wufen lü 五分律 (Five Part Vinaya, T no. 1421) in the fifth century involved Faxian’s 法顯 provision of the Indian scripture and the participation of two translators from the Western Regions, Zhisheng 智昇 and Fotuoshi 佛馱什. Additionally, the team included two prominent Chinese exegetical monks, Daosheng 道生 and Huiyan 慧嚴, who were responsible for zhibi canzheng 執筆參正 (verifying the correct version).72 The specific contributions of these two monks are not well documented, but it is apparent that some of the expressions in the Wufen lü are consistent with those in earlier vinaya texts, such as the sentence 欲得好心莫放逸，聖人善法當勤學 (To attain a good heart, do not indulge in idleness; diligent study of the virtuous dharma is what the saints do), which is also found in the previously translated Mohe sengqi lü 摩訶僧祇律 (Mahāsāṃghika-vinaya, T no. 1425).73 It can be concluded that it is Daosheng and Huiyan that polished Wufen lü and made some of the text more familiar to the readers.74

8. Conclusions

The process of translating Buddhist scriptures into Chinese has been the subject of extensive scholarly investigation. In this study, a detailed examination of the Faju jing, including its Preface, original language, and rewriting phenomenon, reveals that it is not a simple translation but a fusion of different sources. Zhi Qian, the translator, faced four major challenges: harmonizing diverse translation practices, integrating the work of previous translators into a cohesive text, navigating various original languages, refining the translation by incorporating pre-existing phrases from earlier works, and dealing with differing opinions on translation style within the translation team.

It becomes apparent that the challenges encountered by the translators went beyond their ability to understand the Indian text. They were faced with the task of not only adapting to a single Indian text but also choosing a method to blend multiple complex sources. Consequently, some translation errors may have arisen from the integration of diverse traditions within the source text, making it difficult for translators to maintain a consistent linguistic framework for translating, making their work susceptible to mistakes.

Moreover, this case study highlights the critical role played by native Chinese translators in integrating diverse translation processes and determining the final style of the translation. Throughout their work, they refined the language and produced the final text. Specifically, Chinese translators were entrusted with the responsibility of adapting the language to conform to established forms in China and to make it easier for Chinese readers to understand the newly translated texts. They employed two primary approaches to achieve this objective.

The first approach was to use the achievements of previous translators and selectively replace parts of the text based on Indian sources or one’s own aesthetic criteria. The second approach was to use the original Indian text as the basis for the translation while selectively incorporating existing phrases and sentences from previous translations. These strategies made the final work more comprehensible to Chinese readers. All of this work reflects the significant contributions made by Chinese translators within translation groups led by foreign monks.

In conclusion, this study sheds light on the complex nature of the translation process and the role of Chinese translators in integrating diverse sources and refining translations. It underscores the challenges faced by translators and the multi-layered nature of early Chinese Buddhist translations. The findings contribute to a deeper understanding of the translation practices employed during this period and highlight the importance of considering the role of Chinese translators in the production of early Buddhist translations.