Wednesday, November 21, 2012

How to translate śāstric Sanskrit in 8 steps UPDATED

How do you start translating a Sanskrit passage? What do you do if you cannot make sense of it? The following are my usual steps:

  1. 1. Identify the structure of the sentence (verb--»agent--»cause, usually in the form "A is B because of C"). A 10-lines long sentence might be scary and its translation will be time-consuming if you just start translating the first word and then the second and so on. By contrast, holding immediately grasp of the whole structure makes it handable. 
  2. 1a. In this regard Sudipta Munsi (see comments below) highlight the importance of remembering that Sanskrit is a SOV language (subject-object-verb, unlike English, whose structure is SVO). This means that the verb comes at the end and the object, most likely, immediately after it. Thus: start from the end!
  3. 2. Check words you might not know. I recommend using Āpte's dictionary or —if you know enough German— the Petersburger Wörterbuch (PW). Monier Williams' one (MW) can be used, but while being aware of the fact that it is often a translation of the PW. Āpte has the advantage of being more aware of indigeneous lexicographies. As for on-line dictionaries: beware the ones which only give you a summary of the MW, without any information about sources, because you might end up trying to adapt a Vedic meaning into a much later śāstra. 
  4. (2a. I am assuming you know Sanskrit grammar, but there might be some difficult forms you would have forgotten. Thus, check a good grammar for them.)
  5. 3. Know the context. If you cannot make sense of the passage, chances are there that you would understand it, if only you knew the context better. In most cases, my equation about originality applies: one only learns a little bit of fresh information (say 5%) in every cognitive act. Thus: you only understand a passage if you already know most of what it means. In other words: don't start diving into Madhusūdana Sarasvatī's Advaitasiddhi unless you know already something about Advaita Vedānta (and Nāvya Nyāya).
  6. 4. Read more similar passages, by the same author, within the same school or in coeval schools. It is very likely, that similar passages will throw light on yours. If you collect Sanskrit works, you will more probably be able to search through them efficiently. (I do not do it in such a broad way, but just because I have not enough searchable texts.) 
  7. 4a. Read more Sanskrit. Vidya (see comments below) suggests that checking similar structures in another text might be helpful. This is not the same as my no. 4, because formally similar passages can be found also in texts whose content has nothing to do with the one you are reading.
  8. 5. Connected with the above: read further. If you do not understand the passage, but feel you know enough about the school or the context, just keep on reading. Chances are there that once you have read 10 pages more, that initial trouble you encountered will be solved.
  9. 6. Read with someone. Discussing problems is in itself a suitable way to solve most of them. In this sense, it is worth reading also with junior colleagues or students. Which, by the way, is also a way to help them, thus a perfect case of win-win.
  10. 7. Ask for help. You surely have senior colleagues or teachers who can help you. Or, you can use mailing lists such as, or ask a question on Or ask scholars you have never met per email. Or ask me (if I do not know the answer, I will tell you whom you might ask to).
  11. 8. In case you have not done it after no. 2 and/or after nos. 5--7: Look at the manuscript(s). As aptly suggested by Jason (see comments below) it is more than possible that what does not make sense to you is the result of a misprint or of a misinterpretation by the editor. It is not sure that the same does not apply to the manuscripts you will read, though.

Am I missing some further important point? What works for you? And: do these steps apply to other languages?

For some basic Sanskrit syntax see this post.


Vidya said...

The only thing I would add to this is that has worked for me is:
Think of a similar usage in another similar / related text.

Unknown said...
This comment has been removed by the author.
Unknown said...

Here I would like to put forward another method, which the late Bimal Krishna Matilal suggested in one of his writings (I don't exactly remember the source):

Matilal said that during his college days when he was reading Bāṇabhaṭṭa’s Kādambarī, he was advised by his teacher to start from the end of the extremely long sentences, made up of lengthy compounds. That is to say, if we start from the beginning of such lengthy sentences (as begin with multiple and connected series of adjectives for a single noun, and which spread across pages), it will not be easy for us to get a clear idea about the subject, to which such adjectives are applied - a position which is reversed if we "begin" from the "end".

The credit of making this point only goes to the late Bimal Krishna Matilal.

Anonymous said...

If you are using a printed edition and think there are errors, try to consult a manuscript.. In fact, it never hurts to consult a manuscript particularly if its unknown to previous editors, as variations may be worth consideration. Even it the manuscript has been consulted for an edition one may still uncover errors in the transcription and so on.

elisa freschi said...

Thanks to you all for these insightful comments. I will integrate them in the post.

Dominik Wujastyk said...
This comment has been removed by the author.
Dominik Wujastyk said...

Regarding your point 2, I don't know why it is so persistently fashionable amongst indologists to deprecate Monier-Williams' dictionary as a tr. of PW. It isn't, although it does contain supplementary materials from PW. This was rigorously examined long ago in an article by Ladislav Zgusta ("Copying in Lexicography. Monier-Williams' Sanskrit Dictionary and Other Cases (Dvaikosyam)", Lexicographica, 4 (1988), 145-164). The result of Zgusta's careful comparison was that MW was indeed based almost entirely on original research, but that as fascicles of the PW appeared between the first and second editions of MW, MW did start to include some materials from PW. But this was never done blindly, and such incorporations were done in a critical spirit. In short, the relationship between these dictionaries is slightly complicated and certainly not adequately captured by the idea that MW is "more often than not" a tr. of PW. It takes some care to use MW well, but it remains a magnificent dictionary that has unique virtues.

elisa freschi said...

@Dominik, many thanks for this informed comment and for the reference to the article, which I really enjoyed (also because it has to do with one of my other projects, i.e., the one on re-use of texts and originality). You are right, the issue deserves a post on its own. Nonetheless, I would not say that the idea of the *material* dependence of the MW on the PW is baseless. In fact, Monier Williams himself is quoted in the article you mention to claim that his *plan* is new and original (pp.153--154): "The extant of the indebtedness to the great seven-volumed Sanskrit-German Thesaurus…was fully acknowledged by me in the Preface…Having regard, however, to the entire originality of the *plan* of my own work, I did not venture to describe it as based on the great Sanskrit-German Wörterbuch… The words and the meanings of words of a Dictionary can scarcely be proved by its compilers to belong exclusively to themselves. It is not the mere aggregation of words and meanings, but the method of dealing with them and arranging them, which gives a Dictionary the best right to be called an original production".

Long story short, I changed "more often than not a translation of PW" into "often a translation of PW".

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