Translation as Innovation in Literature: the case of a Sanskrit Buddhist poem translated into Chinese

Abstract As many other religions have done, Buddhism developed and evolved over the centuries in order to fulfil social needs and to adapt to different cultural backgrounds. Following the commercial routes in Central Asia, from around the second century BCE, Buddhist texts first arrived in China, and consequently a new compelling need to translate them for the use of new followers developed. This paper attempts to describe the many difficulties faced by the first Buddhist translators through the analysis of the translation of a particular poem, the Buddhacarita. The case study aims at pointing out how this translation process involved linguistic, religious and cultural issues.

fasting, in which assemblies of monks and devotees gathered to hear the recitation of sūtras (Hureau 2008).
As far as we know from historical records, the life of the famous translators was all but serene, especially if we consider them in their relationship with political power.Foreign monks were supposed to receive protection from powerful kings, local princes or important monasteries in order to pursue their translation work: the attention that political leaders paid to them was mainly due to reasons of authority and prestige; as a consequence foreign monks often had to live segregated and were to follow the strict rules imposed by interested forms of patronage (Chen 2004).
Nevertheless, the efforts and accomplishments of these devoted translators have been hardly dealt with by Western scholars of Translation Studies.The origin of Translation Studies is often linked with the necessity of translating the Bible into Greek and Latin (Bassnet 1999) and little mention is made about non-European scenarios.Max Deeg (2008) drew an interesting parallel between the translation of the Bible in the German languages, which took place between the Vth and the XIth century, and the process of translation of Buddhist texts into Chinese.2He also stated that while we can rely on a good deal of information about Medieval translators of the Bible (Deeg 2008: 84), we do not have the same insight on Buddhist monks' workshops, whose discussions about translation's issues "were held on an abstract level rather than dealing with technical issues of translation" (Deeg 2008: 84).
In fact, there are reasons to believe that Buddhist translators dealt with technical and theoretical issues in frequent cases.They were certainly aware of the difference among 音譯 yīnyì 'phonological rendering' (when the foreign word is borrowed into a Chinese text according to its original pronunciation, with an attempt to reproduce or imitate the sound of the foreign word by means of some Chinese syllables; the latter were considered the smallest phonetic units of the Chinese language) and 義譯 yìyì 'translation by meaning' (in this case each morpheme of the foreign word is translated with a corresponding Chinese morpheme/ character -this method is particularly tricky when translators have to deal with Sanskrit compound words, in which the relationship among different parts of the compound is not always clear).These processes exerted a sensible influence on the study of Chinese linguistics and Chinese phonology both in China and in the West.3 Ch'en (1960: 185) pointed out how, through the centuries, Buddhist translators drew several lists of words that could not be translated according to their meaning, but only rendered into Chinese using characters according to their phonographic value, which was probably due to the hieratic value attached to the recitation of words and prayers in Ancient India.These precepts are derived from religious issues yet are not simply limited to an "abstract level", since they seem to deal with technical issues and these could have influenced the work of translation.
Moreover, the difficulty of conveying the rhythm of the Sanskrit verse into Chinese constituted a very important problem for Buddhist translators.As Victor Mair and Tzu Lin Mei (1991: 375-470) pointed out, various attempts had been made to create new forms of metric patterns derived from the model of Sanskrit meters.

A case study: a peculiar translation for a particular poem
The case of 佛所行赞 Fó suǒxíng zàn (T192), the Chinese translation of the Buddhacarita, a Sanskrit poem of the 1 st -2 nd century CE written by the poet Aśvaghoṣa,4 is a peculiar example in such respect, for at least three reasons: 1. the choice of keeping the versified form of the original in the translation, thus producing the first long epic poem in the history of Chinese literature; 2. frequent use of devices meant to adapt and simplify the text for the target audience; 3. an explicit and conscious use of a form of "censorship" with regards to the content.
The Buddhacarita, or "Life of the Buddha", is a poem in twenty-eight cantos, and it tells the story of the historical Buddha, from his birth till his death.Only around fourteen cantos of the original Sanskrit text survived, the second half of the manuscript being lost, hence we have to rely upon the complete translation in Chinese and Tibetan.5The importance and, to a degree, the aesthetic value of this literary work lies in the fact that the author, the poet Aśvaghoṣa, was probably a Brahmin -i.e. a person belonging to the learned Vedic clergy -who had converted to Buddhism.The Sanskrit poem has the important function of mediating a new religion (Buddhism) to scholars of a different traditional background (brahmins).6 It should be that the authorship of the translation is still debated.The preface of the text in the digital edition of the Taishō canon7 states that the translation was the work of the monk Dharmakṣema (385-439), thus dating it back to the 4 th century CE.
However, some external evidence in the same Chinese canon contradicts the colophon.In fact the catalog created by 僧祐 Sēng Yòu, 出三藏記集 Chū sānzàng jìjí (T2145: 12a25-12a26), states that The Buddhacarita in five scrolls (written by the famous bodhisattva Aśvaghoṣa and also known as Eulogy of the origin of the Buddha's Acts, translated at the Liuheshan temple), second part, total of seven scrolls.Age of the emperor Xiaowu, Liu Song dinasty.Translated by the sage Baiyun at the Liuheshan temple.
The Xiaowu age of the Liu Song dynasty lasted from 453 to 464 AD.It has been impossible until now to find out where the oldest colophon of the work is located; the use of this material may be of great importance in order to solve the question about the authorship.
This Chinese text has been often used by western scholars as a tool for understanding the difficult passages in the Sanskrit original; Johnston (1932: xiii) explicitly criticized the attitude of the translator and the translation itself as being particularly unfaithful and inaccurate.In fact, Chinese scholars such as Hu Shi (1927: 124) and Liang Qichao (sd.: 161) demonstrated a different attitude in analyzing the text: if considered as a work of literature -and not merely as a translation -the Fó suǒxíng zàn clearly represents the first attempt in the history of Chinese literature to create a long epic in verse.8It is well known that the translation of Buddhist sūtra caused the introduction of long prosimetra as a new literary pattern in Chinese literature (Liang sd.,(161)(162).The Fó suǒxíng zàn is an even more revolutionary attempt to introduce a new pattern, in order to retain a greater adherence to the original text.A poem of more than thirty thousand pentasyllabic verses could not pass unnoticed -especially if we consider that the longest classical poems of the Chinese literature are of about one hundred verses in length (Coyaud 2009: xi).The Chinese scholar Rao Zongyi (1996: 8) assumed that the influence of this long poem extended till the point of affecting the production of a Confucian author like Han Yu (768-824), famous for having been a strenuous opponent of both Buddhism and Taoism.
As far as we know from tables of rhymes and historical records, the aesthetics of Chinese poetry was moulded by its deep relationship with traditional musical patterns (Norman 1988, 24-25); poems were meant to be sung according to modes (Bertuccioli, 1968: 146-147).We can wonder about what kind of use the Fó suǒxíng zàn might have had in this context.In fact, it could hardly be sung or learned by heart, since, though having a metric pattern, it has no clear rhyme or modal scheme.We do not have any proof of an epic tradition in the Chinese world, since classical Chinese poetry was characterized by short compositions of a few verses and therefore the Fó suǒxíng zàn is perceived as astonishingly long also by modern commentators (Liang,s.d. 162).
As its content and style suggest, the translation was probably meant for an audience of monks or lay people well versed in Buddhist teachings -specific Buddhist terms are widely used in the text.The translator did not choose to paraphrase the text and in fact it preserved the metrical pattern in the final Chinese work.This may be considered an effort to preserve the peculiarity of the original text in the translation -there are only a few examples of Buddhist epic poems in Sanskrit, and the two surviving works of Aśvaghoṣa are perhaps the best examples of this literary genre.The translator might have been aware of the originality of the text that he was dealing with; still, this supposed awareness did not prevent him from censoring some of the contents.
We thus assume that some of the most striking differences in content between the original Sanskrit text and the Chinese translation are to be attributed to deliberate choices by the translators -not depending on the fact that translators were dealing with a source text which is not identical to the one we can read today.Censored content was mainly about the detailed and very common descriptions of beautiful women and courtesans.In the third Sanskrit chapter, Aśvaghoṣa used more than forty verses to describe the beautiful ladies of the court rushing to see the young prince taking a walk outside the palace, while in the resultant translation only two verses are dedicated to the same description, with almost no mention of the beautiful ladies (we only have a cursory mention of "joyful citizens").Other examples of censorship are the comparisons between the Buddha and animals, like the one occurring in BC 8.23, where the grieving women are compared to cows abandoned by the bull (i.e. the prince) in the forest, while in the translation we only find a mention of 'cows that lost their path' (T192, 14c22).9These kinds of "adjustments" about content seem to be due to a specific aim of re-targeting the text for a new audience.This is not an easy process, if we consider that the adaptation for a new (probably religious) target audience has to deal with a shift among two very different cultural backgrounds and with religious issues that changed and developed across five or six centuries.
An aspect to be analyzed entails all the features that the translator(s) adopted to convey proper names and specific terms to the target reader.The analysis is focused on three different aspects: the translation/ transcription of proper names, the translation of Buddhist Hybrid Sanskrit terms, and the addition of external elements which are not present in the original text.

The translation/transcription of proper names
The work of translation of Buddhist texts required the cooperation of many people, and in the case of the Fó suǒxíng zàn it seems that there was someone well-versed in the field of Indian mythology, since almost all of the patronymics or secondary names of divinities and heroes had been substituted with simplified or explicative forms.The translators could also have had the possibility of reading some commentary of the Sanskrit original text that had not survived through the ages until today.We can notice examples of transcriptions of proper names as well as translations according to their meaning as well as hybrid forms (part of the name translation is a phonetic rendering and part is translated according to the meaning).
The name of the Indian author, Aśvaghoṣa, is a meaningful example: 馬鳴菩薩 mǎmíng púsà, literally 'the bodhisattva Maming', is a translation by meaning.The first part, Aśva 'horse', is translated as 馬 mǎ 'horse'; the second part ghoṣa 'sound, noise', is translated with 鳴 míng 'hoot, animal call'.This translation became the traditional way for referring to Aśvaghoṣa in Chinese.
Here are a few examples derived from an original list of about eighty proper names: The family name of the Buddha is Ikṣvāku, and it draws a link between the royal family and a mythological hero who was believed to be son of the sun.The name is translated as甘蔗 gānzhe 'sugarcane' (BC 1.1).This happens following the Sanskrit "folk etymology" of this proper name, that derives it from ikṣu, the wild sugarcane (Salomon-Baums 2007, 208).This interpretation is often found in later Chinese commentaries, where it is stated that the family of Siddhārtha Gautama was one of sugarcane farmers (cfr.

法苑珠林 Fǎyuàn zhūlín [T 2122]);10
The name of the legendary cakravāka birds, which are described as living as inseparable couples, is translated as 雙輸鳥 shuāngshūniǎo, (8.60), a single word in which the first character 雙 shuāng and the last character 鳥 niǎo mean respectively 'pair' and 'bird' while the second character means 'to transport by wheel', apparently trying to convey the meaning of the Sanskrit word cakra 'wheel';11 The name of the great sage Vasiṣṭa (1.52), a well-known mythological character mentioned many times in the text, is translated in different ways.In one case the translator chose to substitute a "matronymic" referring to the sage with his proper name.It looks like the translator knew the mythological history of Vasiṣṭa, and he wanted to provide the reader with a sort of exegesis.The Chinese characters used for the phonetic transcription changes in various examples: Similar changes can be noted for another two proper names: Antideva and Janaka.In both cases the variation of the characters does not exactly correspond to a different syntactical role of the Sanskrit word.
安低牒 āndīdié → Antidevaḥ (BC 1.52) 安低疊 āndīdié → Antidevaḥ (BC 9.70) 安提 āntí or āndī → Antidevam (BC 9.20) Another example is the name of Janaka: 闍那駒 shénàjū → Janako (BC 1.45) 闍那伽 shénàjiā → Janako (BC 12.67) 闍那 shénà → Janakaṃ (BC 9.20) These varied transcriptions of the same names are often found in different scrolls (卷 juǎn 'volumes' or 'folios'), which raises the doubt that there were problems of proof-reading in the translation team.It cannot be considered impossible that the translator was trying to convey the idea of case endings (Sanskrit is an inflectional language and Chinese is not), but we have to keep in mind that we do not know how these characters were read and pronounced in the age in which the Buddhacarita was translated (probably the Vth century CE); the transcription used here is from that of modern dictionaries.In fact, we can only notice the use of different characters, always in the last position, and in phonetic renderings.
In one case, the situation is more complicated since the translator apparently misunderstood a verse, BC 1.43, translating an adverb as a proper name.

薩羅薩仙人， 經論久斷絕， 而生婆羅婆， 續復明經論 現在知見生， 不必由先胄 (T04n0192_p0002b08, 09).
After the scriptures had long ceased, the seer Sārasvata produced Parāśara who continued the scriptural texts of the Vedas.When knowledge is produced in the present, it is not necessarily connected with the deceased (Willemen 2009: 7).12 Johnston (1932: 10) dismissed this 婆羅婆 póluópó as a mistake in translating the word pūrve.But the same pūrve seems to have been translated as 先胄 xiānzhòu 'ancestors, men of earliest time'.As 婆羅婆 póluópó is a phonetic rendering, it should be a proper name.A phonetic rendering 婆羅舍 póluóshè is used in the translation for the Sanskrit name of Parāśara (BC 4.76).The name is also translated as 毘陀波羅沙 pítuó bōluóshā, corresponding to vṛddhaś ca iva parāśaraḥ 'as the old Parāśara' (BC 12.67).The different characters at the end of the name cannot prove the translator was thinking about different proper names (since, as we have seen, they could depend on the different case ending of the original form in Sanskrit).Parāśara was in fact the name of a son of Vasiṣṭa, traditionally one of the authors of the Vedic hymns (Monier-Williams 1976).So quoting here the name of Parāśara is not completely out of place.
Other examples: Soma (2.37) -月光 yuèguāng, is a potion used during rituals of sacrifice; the king Śuddhodana is described as using it according to the precepts of the Veda.The Chinese translation means 'moonlight', probably because one of the names of the moon in Sanskrit is suma (which, to a Chinese reader, is phonetically similar to the word soma).Shortly after in the text (4.5) the name of candramā 'personified Moon' is simply translated as 月 yuè or as 月天子 yuètiānzi 'emperor of the moon'; this translation recurs also with different names of the moon, such as in BC 9.11, in which indu 'moon' is translated with the same character, 月 yuè.
Chandaka (5.68) -車匿 chēnì, This is the name of Buddha's charioteer.The first character (word) means 'chariot'; at least for this first syllable, it looks like the translator made an effort to do a phonetic transcription while explaining the role of the character (the Sanskrit name is just a proper name and contains no references to the character's trade in the play).
Śuddhodana (2.46) -淨飯 jìngfàn.It is the proper name of Buddha's father, the king of the Śākya clan.The Chinese translation sounds like 'purified rice'.In this case the translator deliberately chose to translate each part of the name (Śuddho 'pure' and dana 'food') according to their meaning.In the 後漢書 Hòu Hàn shū the dynasty chronicles of the Hou Han dynasty (25-220 AD) written in the 5 th century, the same name is translated as 屑頭邪 xiètóuyé, a phonetic rendering, to which Pulleyblank (1983: 70) proposed an hypothetical pronunciation *sɛt dow jiaʳ, suggesting an alternative phonetic rendering.

2.2
The translation of Buddhist Hybrid Sanskrit terms.Hakeda (1962: 150-163) drew a list of all the Buddhist Hybrid Sanskrit13 terms in Aśvaghoṣa's poems.Almost all the technical terms related to Buddhism in the Buddhacarita were translated according to their meaning, except for a few famous terms such as parinirvāṇa, transcribed as 涅槃 nièpán.This is an important observation since, according to some records (Ch'en 1960: 185-188), different schools of translation adopted different methods of conveying technical terms.The following list shows a few examples.14ājavaṃjavatā (12.41, 41cd) literally translated as 'the state of coming and going', or 'the rebirths cycle'; in Chinese it is translated as 生老死 shēng lǎo sǐ (27c08) 'be born, growing old, die '. upapatti (9.56, 14.4, 14.9) 'birth, rebirth'.The first occurrence of this term (BC 9.56) is when the king's ministers try to persuade the prince to leave the forest and go back to the royal palace, by stating that it is impossible to know whether there is the possibility of "being born again after death".Olivelle (2008: 264) translates the passage as "If there is continued existence beyond this, we will rejoice there in accordance with our birth".In the Chinese version upapatti is translated with 所得 suǒdé 'be obtained'.The Chinese translation according to Willemen (2009: 64) is "If there will be a later world, one will endure what will be obtained".In another two cases the Chinese version has 生死 shēngsǐ 'birth and death '. Śuddhādhivāsa (1.20, 3.26, 3.56, 13.31), literally '[gods] living in the pure sky'.It is the proper name of a class of divinities.The translator chose a literal translation: 淨居天 jìngjūtiān, in which 天 tiān, literary meaning 'sky', is also the first character in the Chinese term 天帝 tiāndì used to convey the meaning of 'god'.

The extensive use of technical Buddhist terms that are not present in the original
Various terms such as bodhisattva (菩薩 púsà) and cakravartin (轉輪王 zhuǎnlùnwáng), are very frequent in the translation but almost absent in the original text.This choice is due to a sharp change in the target audience; from a brahmanical perspective (Aśvaghoṣa apparently wrote his work for an audience of learned Brahmans) we switch to an audience which is well acquainted with mahāyāna theories but had little knowledge of brahmanical precepts.This is the reason why the Buddhacarita describes the perfect king according to a traditional brahmanical perspective (Hiltebeitel 2006), while the Fó suǒxíng zàn refers to the king with the Buddhist term of cakravartin.
In the translation, the proper name of Buddha's horse, Kaṇṭhaka, is simply translated as 白馬 bàimǎ 'white horse'.This choice is probably in analogy with the white horse that appeared in a dream to the emperor Ming of the Eastern Han (25-220 AD).After this dream the emperor chose to convert to Buddhism and started the building of the first Buddhist temple in China (白马寺 Báimǎ sì 'the white horse temple', close to the capital洛陽 Luòyáng).15 From this analysis we can infer that the translators made an effort to introduce more Buddhist terms in the translation.This is probably because the original Sanskrit text was addressed to a non-Buddhist audience; it was written with the purpose of conveying a Buddhist point of view using a style and a language proper to another literary tradition.The translation aims at relating a Buddhist tale to people that were probably used to Buddhist concepts and terminology.

The addition of external elements, not present in the original text
This alteration is, arguably, the most remarkable one, and consists of the use of specific Chinese elements where the original text presented scenes and situations that the translator felt needed to be culturally mediated.A first striking example could be the introduction of the birth date of the Buddha (四月八日sìyuè bā rì, 8 th day of the fourth month) which was a highly debated question in China (Wang-Toutain 1996: 91), while there is no mention of a specific day in the Sanskrit text.
The frequent mention of 五欲 wǔyù 'five desires' deserves attention.These are the five causes that start the process of cause and effect which binds men to the chain of worldly existence.According to the definition in the Soothill-Hodous dictionary, the five desires are related to the five senses.This Chinese expression is used five times in the second juǎn 'volume' (the Fó suǒxíng zàn consists of a total of five such volumes), but it is not present in the original.
In the translation there is the frequent mention of the 四大 sìdà 'four elements' (T192 6b02), for the Sanskrit mahābhūta 'gross elements'.But in the sāṃkhya theory according to Aśvaghoṣa's perspective, there are in fact five elements (Kent 1982: 60).
At the end of the eighth chapter of the Buddhacarita, after having sent his men into the forest to look for the prince who had fled from the royal court, the king had to accomplish his role by performing the remaining sacrificial duties.The Chinese translation passes over this reference to the religious role of the king and substitutes it by describing the king as he orders the preparation of a lavish banquet (Willemen 2010: 59), which looks to be quite a harsh way to mourn the sudden deception of a beloved son.
There are some interesting examples of different interpretations of the text.For example, in BC 3.1 the prince wants to take a walk in the royal parks.These are described as nibaddhāni 'furnished, adorned with'; both Johnston (1932: 32) and Olivelle (2008: 63) think that this means the parks were "adorned with".The Chinese translator chose a secondary meaning of the word nibaddha -'string musical instrument'.So he translated it as 弦歌 xiángē 'singing and playing string instruments', probably intending that there was music in the park.

A note on censorship
The last and most important aspect of this comparison between Aśvaghoṣa's poem and the Fó suǒxíng zàn shows that many parts of the original work were censored in the translation.The missing parts deal mainly with the description of joyful or sorrowful women, that Aśvaghoṣa provided us with a vivid and licentious richness in details.Most of the metaphors that held a comparison among humans and animals, very popular in Sanskrit poems, were censored as well.In some cases, such as in the fourth and the fifth chapters, Aśvaghoṣa gives us a very long description of the beautiful women in the royal court lasting for more than fifty verses.It looks like the poet is trying to make the reader feel like the young prince Siddhārtha felt in the end, that is to say, quite bored by the annoying and overwhelming presence of his father's courtesans, and thus ready to abandon the pleasures of the court.No trace of this literary feature survived in the translation.This is an example from the fifth canto: cala kuṇḍala cumbitānanābhir ghana niśvāsa vikampita stanībhiḥ| vanitābhir adhīralocanābhir mṛrgaśāvābhir ivābhyudīkṣyamāṇaḥ|| while young women, their faces kissed by their dangdling earrings, their breasts throbbing with deep and constant sighs, their eyes darting hither and thither, gazed up at him like young does (Olivelle 2008: 143).
The singing women all looked at him and, listening for his instruction, waited for the expression of his voice.They observed his thoughts with reverence, just as that deer in the forest (Willemen 2009: 35).

Conclusion
The Chinese philologist Ji Xianlin (1987: 163) suggested that the translators of the Buddhacarita were probably carrying out their work as specifically dealing with a text of Buddhist doctrine, rather than a literary work.16This could be true for some aspects, among which the introduction to the translation of a technical Buddhist vocabulary.On the other hand, if we consider the efforts put forward in preserving the metrical pattern of the original -in contrast with the Chinese literary and poetical tradition -and the attention paid to the transcription of names and mythological references, we are led to believe that the translators were aware of the aesthetic value of the original text, even if willing to adapt it to a new audience.In many cases the chief of the translation team was of Indian or Central Asian origin, and he was probably well-versed in the study of the original text.Bingenheimer (2010: 32-33) interestingly suggests that the model of cooperation adopted by Buddhist monks in their effort of translation could teach modern scholars a new working pattern in their task of analysis of Buddhist texts.A simple overview of the complex cultural context in which Buddhist texts were translated is sufficient to understand how translation was linked to political, religious, cultural and socio-economic issues, thus becoming part of the religious life in monasteries.The practice of translation was therefore institutionalized through rituals (in a religious perspective) and by means of translation schools and bureaus.As Bingenheimer (2010: 23-27) points out, Buddhist translators had to face a cultural and social shift in audience; in a diachronic perspective, they had to adapt their methods according to different technologies of transmission (from oral recitation and memorization to writing and print).
The analysis of a specific poem, i.e. the Buddhacarita, gave us the possibility to understand the importance of this kind of study in the field of linguistics (borrowings and phonetic renderings, issues in pronunciation of medieval Chinese, and different typology of languages in contact), in the field of history of religion (evolution of Buddhism in its shifting from the Indian context to the Chinese environment, and evolution of different schools of Buddhism testified by a new lexicon), and in the field of literary studies (the evolution of different literary genres and metrical patterns in various historical and cultural contexts).
So it is safe to assume that the study of Buddhist translation is a direct inquiry into the history of Central Asia and medieval China.If we still assume that the study of history is an issue of primary importance -as well as the study of linguistics being a crucial point for the purpose of language training -we can infer that the study of Buddhist textual translations should gain a seat in the hall of Translation Studies.By doing so, the same field of Translation Studies will gain a step towards a global and multi-cultural perspective.