[INDOLOGY] Accuracy in translations

Matthew Kapstein mkapstei at uchicago.edu
Sat Jun 9 21:29:57 UTC 2018

Dear David and Dan,

To my way of thinking the emphasis on "literalness" in translation is misplaced.

Tibetan translations give an impression of great literalness because the Tibetans adhered rather closely -- sometimes too closely -- to a rigid scheme of lexical equivalents. But translation is something more than this; it requires sensitivity to syntax, idiom, nuance and more. The great 13th c. Tibetan translator Byang-chub-rtse-mo clearly recognized this and is quoted in his biography as saying that only poor translators follow the MahAvyutpatti with complete faithfulness.

It is not difficult to find instances of Tibetan translations that are, to all intents and purposes, equally "literal," but that one is a fine, elegant work and the other unreadable hash. A case in point is Vasubandhu's ADhKBh, a great achievement, vs. Sthiramati's commentary on the ADhk, a pretty dismal exercise. And it would not be difficult to multiply the examples.

I could go on, but my point is concise: in judging the value of a translation, literalness is but one value that must be considered alongside a range of others. And while I have great respect for Tibetan transmission lineages, it would be naive to imagine that sometimes grotesque misunderstandings have never crept into tradition. But I'll leave it at that.

best regards,


Matthew Kapstein
Directeur d'études,
Ecole Pratique des Hautes Etudes

Numata Visiting Professor of Buddhist Studies,
The University of Chicago
From: INDOLOGY <indology-bounces at list.indology.info> on behalf of David and Nancy Reigle via INDOLOGY <indology at list.indology.info>
Sent: Saturday, June 9, 2018 2:08:22 PM
To: Dan Lusthaus
Cc: Indology
Subject: Re: [INDOLOGY] Accuracy in translations

Dear Dan,

Among the additional observations you have provided, I thought your last paragraph was especially relevant for providing a proper perspective on the Tibetan translations of Sanskrit texts.

First, it is only fair to say that I am unable to regard as a myth the idea that the Tibetan translations are more literal than the Chinese translations, and to concisely state the reasons why. In my attempt to gather all printed editions of Sanskrit Buddhist texts I have usually read the introductions. These, when they spoke of the Tibetan and Chinese translations, uniformly spoke of the Tibetan translations as being more literal than the Chinese translations. I do not recall a single exception. This is equally true of introductions to English translations of Sanskrit Buddhist texts. Additionally, the many critical reviews by J. W. de Jong and others said the same. My own forays into the Tibetan canonical texts in comparison with the Sanskrit texts had also impressed me with their literalness. More often than not they even followed the Sanskrit word order, only moving the verb to the end when it was not already there, such as in Sanskrit verse. The two examples I cited from Gadjin Nagao are particularly instructive, not only because his wide familiarity with Buddhist texts is undoubted, but also because they are from word-indexes that obviously took full and careful account of the words in the texts. Thus, I cannot regard the idea that the Tibetan translations are more literal than the Chinese translations as a myth.

Your last paragraph points to what can well be regarded as a myth: the idea that the Tibetan translations are so accurate that nothing else is needed, and hence they can be accurately translated without reference to the Sanskrit. I, too, have been disappointed in the quality of the many translations made only from the Tibetan translations, when Sanskrit editions exist. Why would a translator purposely choose to translate a translation when an original is available? Leaving aside the many cases where we no longer have the Sanskrit, the rationale for translating a Tibetan translation rather than the Sanskrit is the transmission lineage. There is a widespread belief that a text cannot be properly understood unless there is an unbroken lineage of transmission of the text from the time it was written up to the present. The Tibetans have retained the transmission lineages for their translations, while the transmission lineages have been lost for the Sanskrit texts when these fell out of use in India. So we have the strange situation that the Tibetans highly respect the Sanskrit texts as their sourceworks, but do not wish to use them.

As a personal aside, of little or no relevance to this discussion, Xuanzang is one of my heroes. I have the greatest respect for his incredible journey to India in search of Sanskrit texts to bring back to China. Then back there, he translated more Sanskrit texts than anyone has ever done in known history. As É<https://en.wiktionary.org/wiki/%C3%89>tienne Lamotte wrote: “Hsüan-tsang, who spent the last nineteen years of his life translating some 75 Sūtras and Śāstras, was the greatest expert in Buddhist literature of all time.” (The Teaching of Vimalakīrti, 1976, p. xxxv)

Best regards,

David Reigle
Colorado, U.S.A.

On Fri, Jun 8, 2018 at 11:33 AM, Dan Lusthaus <prajnapti at gmail.com<mailto:prajnapti at gmail.com>> wrote:

Your analogy to modern translations is apt.

While you continue to document the pervasiveness of the myth, perhaps we should let the other shoe drop. While it is true that there are differences between the various Chinese translations of the Mahāyānasamgraha, a more complete account would mention that there are actually seven Chinese translations (including bhāṣya and upanibandhana), by four translators.

Mahāyānasaṃgraha (Asaṅga)
(1) (T. 1592) Translation by Buddhaśanta in 531.

(2) (T. 1593) Translation by Paramārtha in 563.

(3) (T. 1594) Translation by Xuanzang in 649.

(4) (T. 1595) Mahāyānasaṃgraha-bhāṣya (Vasubandhu’s comm.)
Translation by Paramārtha in 563.

(5) (T. 1596) Mahāyānasaṅgraha-bhāṣya
Translation by Dharmagupta, sometime between 605-616.

(6) (T. 1597) Mahāyānasaṅgraha-bhāṣya
Translation by Xuanzang in 648.

(7) (T. 1598) Mahāyānasaṃgrahopanibandhana (*Asvabhāva’s sub-comm.)
Translation by Xuanzang between 647-649.

The two versions that dominated E. Asian attention were those by Paramārtha (with -bhāṣya) and the three versions by Xuanzang. Dharmagupta’s translation, done while Xuanzang was a youngster, never rivaled Paramārtha’s version which was the hot item in that period, totally eclipsing Buddhaśanta’s earlier effort. Study groups and temples all devoted inordinate attention to Paramārtha’s She lun; everyone studied it, masters lectured on it, it was on the reading list for all Buddhist sects in China, including early Chan. It is precisely due to its defects — ideological distortions, not just translation errors — that led Xuanzang to put not only the Mahāyānasaṃgraha on his to-do list when he returned from India in 645, but to retranslate Vasubandhu’s comm. and introduce *Asvabhāva’s subcommentary, just to make sure the correct reading of the text would become available in China. When Griffiths, Keenan, et al., worked on the 10th chapter with Hakamaya, they found that the Tibetan of Asvabhāva’s subcommentary basically concurred with Xuanzang’s text — though, if you read their introduction, they are somewhat struggling with that realization, since another important Japanese scholar, Ui Hakuju, had championed Paramārtha’s texts over Xuanzang’s (another ideological myth with some live embers in Asia) — and the evidence of the texts they were reading was indicating otherwise.

There are only three Tibetan versions: one for the main text, one for Vasubandhu’s comm., and one for the Asvabhāva subcommentary. The first and third are attributed to basically the same translators, with a redactor homogenizing the third as well. So “uniformity” is not surprising.

(1) Mahāyānasaṃgraha (theg pa chen po bsdus pa)
translation ascribed to Jinamitra, Śīlendrabodhi, Ye-shes-sde

(2) Mahāyānasaṃgraha-bhāṣya (theg pa chen po bsdus pa'i 'grel pa)
translation ascribed to Ātīśa Dīpaṃkaraśrījñāna, Nag-tsho Tshul-khrims-rgyal-ba

(3) Mahāyānasaṃgrahopanibandhana (theg pa chen po bsdus pa'i bshad sbyar)
translation ascribed to Jinamitra, Śīlendrabodhi, Ye-shes-sde
with revisions ascribed to R. Dpal brtsegs

That they agree with Xuanzang’s text would suggest either that Xuanzang and the Tibetans were good at capturing the Sanskrit, or (!) that the Tibetan translations and/or redactions were done with one eye on Xuanzang’s rendition. The Sanskrit has not survived, so we can’t rule either possibility out.

But a bigger issue is the role of redaction, revision, on what becomes the received texts. Even apart from known variants in different editions of the Chinese and Tibetan canons, there were wholesale ‘revisions’ of many texts, some multiple times.

But more germane to the Chinese vs Tibetan translation issues:

There are two Tib. versions of the Trisvabhāvanirdeśa — one designating Vasubandhu as the original author and the other naming Nāgārjuna as the author. (I don’t think it is possible that either of them had anything to do with it). There are serious discrepancies between these two translations.

Or, to take a more influential text, there are two divergent translations of Dignāga’s Pramāṇasamuccaya, both of which, while unavoidable since the Skt was lost, were commonly panned as bad translations. With the recovery of the Skt of Jinendrabuddhi’s -ṭīkā, which retains perhaps 70% of Dignāga’s original text, we can now see more clearly what the Sanskrit underlying the two Tib translations was, and suddenly some are beginning to say the Tib translations weren’t so bad after all. I suspect that reappraisal is a result of knowing what the Tibetan is *supposed* to indicate, and then finding it there. Reading in the other direction is more fraught.

One could make a similar case for some Chinese translations as well. Some years back a reading group of grad students was reading the Yogācārabhūmi with me. Our procedure was to read a passage in Skt, translate it orally, and then look at Xuanzang’s Chinese and translate that. We often looked at the Tibetan as well, esp. for thornier passages. After some weeks, the students, most of whom were Chinese, remarked how impressed they were with Xuanzang’s translation, declaring him great and accurate. I cautioned it only looks that way because they had already worked on the Sanskrit and knew what to look for. They disagreed. So the next session, I had them read Xuanzang’s Chinese version first, and translate that. They cobbled together a plausible reading. Then I had them read the Sanskrit. They immediately recognized the mistakes they had just made. Since they were native readers of Chinese, trained in reading classical and Buddhist Chinese as well as Sanskrit (two had studied Sanskrit and Pali for years in India and Sri Lanka, and one was a monk), the mistakes they made were not ’student’ mistakes but natural misreadings that can occur in the absence of the underlying Sanskrit. (Incidentally, on the few occasions when Xuanzang and the Tib diverged, it was almost always because each was offering a legitimate interpretation of what the Sanskrit said, but their interpretations differed.)

One can find better and worse translations in Chinese and Tibetan. Some years back I complained at a conference on Buddhist translations that many of the translations of classic Buddhist texts coming out in English from Tibetan versions of Indian texts were insipid, lacking in conceptual distinctness and clarity (of course, there are many exceptions to that, but that does identify a certain genre of English representations, which I will leave unidentified). One of the conference participants, an important scholar, in defense of those translations, replied: “To be fair, that’s how they read in Tibetan, too."

As we all know, translation is never as easy as it looks — and it doesn’t even look easy.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://list.indology.info/pipermail/indology/attachments/20180609/00615286/attachment.htm>

More information about the INDOLOGY mailing list