Monday, December 10, 2012

Olivelle vs. Böhtlingk: Should we add words to a Sanskrit text if it is not understandable?

Is one allowed to add a word if the sentence just lacks sense, although no manuscript supports the proposed reading? My provisional answer, whenever I am tempted to change the text is: NO. It is much more likely that I am not understanding the text, rather than that the text is flawed.
However, we all know how unreliable manuscript traditions might be, and how extremely unreliable editions might be, especially if they are based on just a (few) manuscript(s).
An interesting example is the polemics about the first Indologists and their excessive audacy in emending texts, to which Patrick Olivelle dedicated vehement attacks in his Introduction to his edition of the Upaniṣads and in a 1998 article on the Journal of Indian Philosophy ("Unfaithful Transmitters", see here). Olivelle focusses especially on a passage of Chāndogya Upaniṣad (8.15.1), where Böhtlingk and, following him, Senart dared to add several words, because they could not make sense of the text. The text is in fact odd:

ācāryakulād vedam adhītya yathāvidhānaṃ guroḥ karmātiśeṣeṇa abhisamāvṛtya kuṭumbe śucau deśe svādhyāyam adhīyāno dhārmikān vidadhat |

The text apart from guroḥ karmātiśeṣeṇa means, at first sight:

After having studied the Veda according to the prescriptions, having returned from the teacher's home, reciting his own portion of the Veda in [one's] family, in a pure place, he should have virtuous sons.

But what could guroḥ karmātiśeṣeṇa mean? Böhtlingk in his Dictionary and in his edition of several Upaniṣads (1889a, 1889b, 1890a) postulated guroḥ karma kṛtvāviśeṣeṇa "after having performed an action for the teacher, in the undifferentiated [time]" (my translation, based on B's text). Senart followed him and even added sthitvā after kuṭumbe. Olivelle is quite against this "hubris" (his wording) in going against tradition. As partial apology of Böhtlingk, however, one might note that his first addition of kṛtvā is based on Śaṅkara's commentary (and hence, not only on his hubris):

guroḥ karma yatkartavyaṃ tatkṛtvā karmaśūnyo yo 'tiśiṣṭaḥ kālastena kālena vedamadhītyetyarthaḥ […] kuṭumbe sthitvā […]

Furthermore, I could finally understand the ChUp's wording through a later article of Böhtlingk (1897a Bemerkungen zu einigen Upanishaden BKSGW, available on, see here), where he goes back to the traditional reading and refers to Gautama Dharmasūtra 3.6: guroḥ karmaśeṣena japet. This Dharmasūtra does not add any further word and, therefore, it makes clear that guroḥ karmaśeṣa (and possibly karmātiśeṣa) had a fixed conventional meaning, possibly "the [time] remaing out of having performed one's duties towards one's teacher". It is also noteworthy that karmaśeṣa is a sāpekṣasamāsa, which needs to be connected with guroḥ.

What is your policy? When did you end up emending a text?

On my general policy, favouring the text over me as a reader, see No. 6 in this post.


Jayarava said...

An interesting example. I'm in the process of marshalling evidence for something like this, but on a smaller scale. In Conze's critical edition of the Heart Sutra we find the phrase: ārya-avalokiteśvaro bodhisattvo gambhīrāṃ prajñāpāramitā-caryāṃ caramāṇo vyavalokayati sma: panca-skandhās tāṃś ca svābhavaśūnyān paśyati sma.

I think it reads better if we drop the (superimposed) colon after "sma" and change pañcaskandhās to pañcaskandhāṃs by the addition of anusvāra; thus vyavalokayati sma becomes a transitive verb with pañcaskandha as it's object; with some other downstream consequences. The result is far more elegant and makes more sense.

In this case it is supported by manuscripts, but 60 years of scholarship have overlooked a grammatical problem that this student could not solve within the given text. I think Conze was predisposed to see vyavalokayati sma as intransitive because of his religious beliefs; and thus the confusion.

If there is a better reading one can only propose it and see what one's peers think. It's not necessary to insist on it right from the start. Such things change by consensus over a long time - perhaps beyond your lifetime. One presents the arguments for and against and waits for a reaction.

I am proposing a change to a 60 year old critical edition by the acknowledged master of the genre; but it's a simple grammatical argument that Sanskritists can evaluate and tell me if it makes sense or not.

As I recall Olivelle is particularly annoyed by *silent* changes to texts - where the text is changed without any notice or comment. I don't think this can be justified. But texts are often transmitted with errors and omissions (Buddhist texts anyway).

Is ununderstandable a word? :-) It looks like a Sanskritism.

By the way the word "hybris" is interesting. Since y and u are neighbours on the keyboard it could simply be a typo. But it might also combine "hubris" and "hybrid". A Freudian slip?

elisa freschi said...

Thank you, Jayarava, also for taking care of my English (by the way, "hybris" is due to the fact that I was thinking at the Greek version of the word, no Freudian slip –I think)!

I agree with your point: "hidden" corrections are not acceptable and the text should be as transparent as possible.

Regarding "your" colophon, how did Conze's religious beliefs interact with his understanding of vyavalok-? Is it about the fact that the skandhas cannot be directly perceived?

I agree with the fact that pañcaskandhās (with or without anusvāra) is connected with the preceding verb, on the basis of the position of ca (should be the second position in a clause) and of the presence of tāṃś (why should one need it, if there were already the corresponding noun in the same clause, given that there is nothing before to which one might want to refer as specifying that one is referring to "these" skandhas?). However, personally I try to refrain from using aesthetic adjectives such as "is more *elegant*", because I do not think I can trust my knowledge of Skt enough and because aesthetic values are by nature subjective. I prefer to refer to sense (as you do) grammatical rules, parallel occurrences in similar texts and so on.

Michael said...

My policy is: it depends on the context. Is it a popular text with many editions, living traditions, and a large manuscript base to look to? (like a major Upanisad) Then one should be more cautious in emending.

Is it a text that hasn't ever been edited and is from an essentially dead tradition? Then emending will be more necessary.

In either case we don't need to be so nervous about "doing violence" to a text. We just need to meticulously note our changes. As Jayarava and you say: no silent emendations (aside from trivial spelling variations). Then critical readers and future editors can simply disregard changes that they see to be incorrect.

Unknown said...

Dear Elisa and Jayarava,

Perhaps all would agree that translations are always interpretive in nature. If it so, then translation can be though of as a category of interpretation and interpretation a special case of translation. Thus, all exegetical enterprises are interpretive by nature. Consider for example the following verse (no. 20) from the second chapter of the Bhagavadgītā:

na jāyate mriyate vā kadācinnāyaṃ bhūtvā’abhavitā vā na bhūyaḥ/
ajo nityaḥ śāśvato’ayaṃ purāṇo na hanyate hanyamāne śarīre //

Note the portion: bhūtvā(’)bhavita. Śaṅkarācārya reads the text as भूत्वाऽभविता, thus justifying the occurrence of the avagraha (ऽ) in भूत्वाऽभविता. Thus as per Śaṅkara’s interpretation the verse would mean: “This Self is never born; It never dies either. Having been born, It never ceases to be, again. Unborn, eternal and everlasting, this ancient One is not slain when the body is slain.” (Dr. A. G. Krishna Warrier’s Translation)

However, the same verse is present in Madhusūdana Sarasvatī’s Guḍhārthadīpikā commentary with the reading भूत्वा भविता, i.e. no avagraha is present here which would have otherwise followed Śaṅkara’s reading अभविता. On the whole the verse is translated, following Madhus’dana's interpretation, like this: “It is not born, and It does not die; nor is it even that this One having been nonexistent becomes existent again. This One is birthless, eternal, undecaying, ancient; It is not killed when the body is killed.” (Swami Gambhirananda’s translation)

So when Śaṅkara says, “Having been born, It never ceases to be, again.”, he interprets the verse from a positive angle (itimukha, compare with this the positive approach to realising Brahman in the Upaniṣads in the mahāvākya, sarvaṃ khalvidaṃ brahma; astityevopalabdhavya; etc. [though this is not to say at all that Śaṅkara was ignorant of the negative method of Vedānta hermeneutics])

But when Madhusūdana says, “nor is it even that this One having been nonexistent becomes existent again”, the approach is clearly negative (netimukha, compare with this the negative approach to realising Brahman in the Upaniṣads such as neti neti, etc.)

Ultimately both Śaṅkara and Madhusūdana agree on the unborn and undying nature of the Self, but one defends the thesis from a ‘positive’ angle, while the other ‘negative’. When such authors as Śaṅkarācārya can do this (i.e. make changes within the text, without reversing the main idea [although I do condemn such silent editing]), is it not implied that such acts, far from being ‘willful’, proceed from a genuine necessity? What do you say?

Dominik Wujastyk said...

I said what I think about this topic many years ago in the introduction to my 1993 edition, Metarules of Paninian Grammar, volume 1, pp. xvi-xvii (, and in volume 2, xxxi (esp. the Housman quote, see

I would draw special attention to the statistic presented by Kenny (The Classical Text, 1974, p.62) that in "cases when it has been possible to check against subsequently discovered MSS or papyri," estimates of the number of learned conjectures that are actually correct vary from 5% to a mere 0.1%. In other words, if you make a conjecture, the statistical chance of it being right is probably between one-in-a-thousand and one in twenty. You'll most likely be wrong.

Jayarava said...

Hi Elisa

I'm an artist at heart and sometimes there is just an aesthetic pleasure for me in how something is phrased. And I'm hardly informed at all about Sanskrit aesthetics so it's entirely personal. I dare to put in a blog comment what I probably wouldn't put in a formal article (if only because its the kind of thing editors make you take out).

I think Conze saw Avalokiteśvara as a god-like being floating around on a lotus up in heaven benignly looking down on us suffering sinners (he hints at this in his commentaries) - hence he treated vyavalokayati sma as intransitive. Avalokiteśvara was just looking down on us, but not at anything in particular. It ties in with some Tibetan legends about the him as well.

But when the pañcaskandhān are the object of his gaze then we grasp that it is a story about a particular kind of meditation which has long roots in Buddhism. It has practical implications for those who would practice in order to perfect prajñā. The text seems a lot less mystical. And that also I find aesthetically pleasing.

Thanks for your comments. As a beginner I often suffer from hubris (or I would say 'diving in at the deep end of the pool', but once I realised that the sentence could not really say what Conze made it say, I was forced to come up with my own answer with what little Sanskrit I do have. I still can't believe I'm the first person to have noticed it - but I can't find anything in the literature.

elisa freschi said...

Thank you very much, Jayarava, Sudipta, Michael and Dominik. I really enjoy this sort of collective writing of blog posts (as it also happened for the post "How to translate Sanksrit in eight steps", see:

What I still do not understand about Olivelle's article and introduction is the fact that he really seems to have something personal against Böhtlingk (e.g.: "…he followed an unconventional —the less charitable but more accurate term may be outrageous— editorial practice", JIPh, p. 175). However, what are the main criticisms, apart from the fact that Böhtlingk worked with "hubris"? 1. That he did not trust enough the indigenous transmission; 2. That he put his emendations in the main text and the traditional readings in footnotes and endnotes; 3. That B's emendations violate the principle of the lectio difficilior.

As for 1, I guess that it is a matter of degree. As already pointed out, I tend to trust the text over me, at least as my first methodological approach (and Dominik expressed a similar position, although in a more articulated way). But Michael is right in suggesting that, especially in the case of "dead" traditions (and, I would add, in the case of codices unici), it is difficult to rely on the received text.

As for 2, I dislike endnotes (because no one reads it —just like I had overlooked endnote n.6 in Olivelle's article). But what about footnotes? Does anyone think that one should put the received text as the main text and one's conjecture in a footnote *even if one were convinced that the received text just does not make sense*? Before finding the parallel in Gautama Dharmasūtra, *guroḥ karmātiśeṣeṇa* seemed extremely odd. It is a sāpekṣasamāsa and it presupposes *kālena* in a way which cannot be predicted. Should not the editor have tried to offer something to the reader? Now, Olivelle is right in pointing out that Böhtlingk could have read Gautama DhSū before. But suppose it had not been published yet. Should he have still put the received text as the main text as it was? Perhaps with a wave underneath?
And don't you emend the text in the main text, if you have additional evidence (e.g., sources, parallel texts, usus scribendi, etc.), though no manuscript support?

As for 3, do we really agree that one should not violate the lectio difficilior principle (cf. Olivelle, p. 177 in JIPh: "a principle problem with many of Böhtlingk's emendations: he rejects difficult readings in favor of easy ones, violating the cardinal principle of lectio difficilior, the bedrock of textual criticism")? I would rather think of it as an ultima ratio device. Also because what appears as "difficilior" to us might not have been that difficult for the targeted audience.

What do you think?

elisa freschi said...

@ Jayarava, you are right and I forgot that you have the additional adhikāra of being a sensitive artist.

The life of a scholar of Sanskrit (and Pāli, I suppose) is full of this kind of surprises and as soon as one closer inspects a text, one is rewarded with gems never detected before.
Thank you very much for your interesting remarks about Avalokiteśvara "content-void" meditation.

Jayarava said...

Domninik - love that statistic!

Jayarava said...

One of the other things to point out is that with a text like the Heart Sutra it has historically been subjected to many silent edits by Indian, Chinese, Tibetan and Japanese editors. Even setting aside numerous scribal errors no two mss. texts are identical.

This suggests that in the history of this text almost no one passed on the text they were given - everyone altered it just a little. The trend continues into English translations.

I should add that my attempts to amend another text passed on orally to my Buddhist teacher in a mangled Tibetan version met with solid resistance. The fact that with a few very minor changes (word breaks mainly, e.g. samayam anupalaya instead of samaya manupalaya) one could produce well formed Sanskrit sentences meant far less to him than the importance of preserving the oral rendition, however corrupt. So in some contexts this issue is more complex with multiple media of transmission and sources of authority.

elisa freschi said...

@ Dominik,
it would be interesting to know, though, how many of the conjectures match the *gist* of the passage as found in manuscripts later discovered.

Unknown said...

Dear Elisa,

As for the following observation of yours:

“As for 2, I dislike endnotes (because no one reads it —just like I had overlooked endnote n.6 in Olivelle's article). But what about footnotes? Does anyone think that one should put the received text as the main text and one's conjecture in a footnote *even if one were convinced that the received text just does not make sense*?”

I would like to point out an exactly similar instance from Pt. Pañcānana Tarkavāgīśa’s Bengali edition of Jayantabhaṭṭa’s Nyāyamañjarī 2 (1941), where he says in a footnote (p. 5): “आदर्शपुस्तकस्थः ‘फलहानादिबुद्धयः’ एषः पाठो न शोभनः।”. Also in his edition of Nyāyamañjarī 1, he says at one place (first fn. to p. 3 of 2006 reprint edition): “दृष्टादृष्टभेदेन पुरुषार्थो द्विविधः, तस्य पन्था अपि द्विविधः। इति पाठः सङ्गततया प्रतिभाति मे।”.

I also observe that the problem does not exclusively concern the Western editors (using footnotes), but it is also shared by modern Indian editors, such as Pt. Pañcānana Tarkavāgīśa, who are generally thought to be respecting the text without trying to make any emendation.

windwheel said...

You are asking about guroḥ karmātiśeṣeṇa. The meaning is clear in an emic context or in your own- Guru shite- on what basis are you expecting to see grammar in this context?
Now if you wanted to genuinely expand your mental horizon you would ask what is the nature of this concept. Without knowing anything about India we can say- it's about broadcasting the Guru's nonsense. But the word chosen indicates some limits to our talking crap.
I seldom find anything genuine behind your quandaries. Apply your common sense.
It is great joke that people on this forum talk of 'sixty years of scholarship'- sixty years of stupidity is more like it.
Indologists aren't scholars they are drug traffickers or pederasts or other such worthless scum.
Tell me one thing they have 'discovered' in the last 60 years?
You take a stupid passage written in a stupid style and want to get grammar and philosophy out of it? Why?
The fact that this is Skt. gives you the clue that the guy who wrote it was just book keeping. Skt is not a sacred language- it's a shit language for everybody except the clerks who use it.
Go back to doing null morphemes- you can squeeze some Philosophy out of that but not what some shithead ancestor of mine wrote coz he thought- who the fuck cares? I wanna go home.

elisa freschi said...

@Sudipta, @Jayarava, thanks for showing that the picture is more complicated than shown by Olivelle (who might still be right in the case of the Upaniṣads). I also know of footnotes written by (excellent) Sanskrit editors and of texts altered by their transmitters, so that it is not always the case that Western editors are "unfaithful" and Indian copyists "faithful", as in Olivelle.

Licenza Creative Commons
Quest' opera è distribuita con licenza Creative Commons Attribuzione - Non commerciale - Non opere derivate 2.5 Italia.