On Translation: Exhibit 1

Translating Sanskrit poetry into English presents unique difficulties. To be sure, translation is always tricky. Passing to a different language invariably loses some nuances and overtones. What can be naturally expressed in one language may require more effort in another.

With Sanskrit, though, even essential features are often untranslatable to a native English audience.

[Disclaimer: Before going further, I must point out that I am an amateur. Everything below is probably wrong, they are banal and pointless observations, anyway, and I amaze myself by my ability to take something interesting and make it boring. I thought I had something to say, but it took writing it out to realise I didn’t.]

Firstly, there are the little matters about language and style considered acceptable. Take the minor point of “elegant variation”.¹ This is considered distracting in English (and it is), but good style in Sanskrit and a few other languages including French. Thus in the Bhagavad Gita in Sanskrit, it is perfectly natural that Arjuna is addressed variously as Pārtha, Kaunteya, Dhanañjaya, Parantapa, and several other names, but translations that use “O son of Pritha”, “O son of Kunti”, “O conqueror of wealth”, “O subduer of enemies” draw too much attention away from the main point, while using the Sanskrit names would be confusing to those unfamiliar with them.

The reason elegant variation is possible at all is because of compounds. Sanskrit has endless possibilities for creating compounds (samāsa), and a lot can be expressed in one word without sounding odd — in some “difficult” works, words spanning multiple lines are commonplace.² This is not possible in English, and translating a compound into a phrase of several of words makes it clunky. (German famously has some facility for long compounds, and — to make a silly, wild connection — the enthusiasm of Germans like Goethe³ for Sanskrit may not have been a coincidence.)

The compounds give Sanskrit a large vocabulary (potentially infinite, in theory), and this allows picking words for their sound in addition to the meaning. [A story is told about a dying poet (some versions have Bāṇabhaṭṭa) who summoned his two sons to decide who would complete his unfinished work. He pointed to a dry tree in front of them and asked them to describe the fact, and one said, harshly, “शुष्को वृक्षस्तिष्ठत्यग्रे” (śuṣko vṛkṣastiṣṭhatyagre), while the other said, sweetly, “नीरस-तरुरिह निवसति पुरतः” (nīrasa-taruriha nivasati purataḥ).^m The latter got the job, but for a different goal, it might as well have been the former.] Barbara Stoler Miller laments in her preface to her translation of Bhartṛhari⁴ that it is impossible to reproduce in English the effect of phrases like karṇe kokila-kāminī-kala-ravaḥ.

The “sound” of the poem depends on, besides its individual syllables, its metre. Picking an apt metre is done to some extent in all languages, (or was done until recent centuries, when “free verse” became fashionable) but Sanskrit has a substantially richer repertoire of metre than in any other language. Thus for instance, the Meghadūta would not be what it is if it weren’t in the distinctive mandākrāntā metre, which some say was invented by Kālidāsa for that purpose (and was used by later sandeśa kāvyas⁵). The metre is usually strictly adhered to, unlike English where occasional deviation — e.g. in an iambic metre, trochaic inversions, pyrrhic substitutions — is considered desirable to break the monotony. In English poetry, the meaning dictates how the poem is to be recited, with a little help from the metre. In Sanskrit, it is the other way around. This gives poems a musical quality, and they can be said to be not so much recited as chanted or sung.

Further, Sanskrit poets evolved a rich set of highly stylised conventions: specific seasons, plants and birds evoked distinct associations. See this series of posts (starting with this one) by Venetia Ansell at her Sanskrit Literature blog. As Andrew Schelling says in his essay on translating Sanskrit poetry:⁶

The mere hint of fragrance off a nearby forested hill told not only in what calendrical moment of what season the poem was located, but also evoked a constellation of human relationships, a precise mood, and vivid moments echoed in other poems.

There is a “deep, ancient regard shown for the natural world”, and the poets are “minutely familiar with nature”, but unlike English or Chinese poetry, they are not written by recluses, but by urbane and cosmopolitan city-dwellers (often courtiers or courtesans): the approach to nature is not as if it were wild — “there are few brooding mountain escarpments, few unvisited gorges along thundering rivers” — but rather:

the human and nonhuman orders seem linked in unsensational daily intimacy. Local villages with birds in flowering trees. The whiff of odors from a nearby forest grove. Farmland crops or native grasses in fertile alluvial soil. Sweet-smelling blossoms along a village path. To put it another way: what flowering creeper shares the details of your life because you walk past it every day to fetch water? What pliant reed did you collect one spring night to weave a mat for your lover?

Thus many poems were refined and sophisticated (some would say indulgent), written for a similarly sophisticated audience, one which no longer exists. (This does not necessarily mean we cannot enjoy them today…)

One important difference between some Sanskrit poetry (kāvya, as opposed to the epics etc.) and much English poetry is that Sanskrit verses are generally free-standing and intended to be read alone. There may occasionally be continuity between verses (at least in theme), but none of them depends on others for its enjoyment. Thus the effect produced from savouring each verse individually is different from, say, reading a ballad. This is actually not so much a difficulty with translation, as a requirement to instruct readers “how to” read.

Perhaps the best way to summarise all of this would be to say that Sanskrit kāvya possesses a high degree of “artificiality”, not in a negative sense but in that they are constructed, cultivated, refined, polished. Each verse stands alone, a finely crafted gem to be examined in all lights before proceeding. Figures of speech and other embellishments (alaṅkāra, lit. ornaments/decoration) are highly prized, and Kālidāsa is so highly revered in part for his upamā, similes. Readers of English poetry, where poets strive to conceal their artifice, may consider some of these to be excesses, but there is no such notion in Sanskrit.

In light of all this, it is no exaggeration to say that Sanskrit poetry is one of the hardest to successfully translate into English. Even when the effort succeeds in producing something beautiful — as, most recently, in Andrew Schelling’s translation of the Amaru Shataka,⁷ the result is something rather far from the Sanskrit original. Indeed, some Frenchman once said: “Translations are like women: if they are beautiful they are not faithful, and if they are faithful they are not beautiful”. Jakobson went so far as to say that poetry by definition is untranslatable, and the Italians say “Traduttore traditore”: “translator, traitor”.

The one notable feature of all Sanskrit translations into English is how bad they are.

---

Despite all this, I love translations.

For one thing, I enjoy them. There is something unique about the experience of reading the thoughts of one language expressed in another — Sanskrit sentiments expressed in English, so to speak. (Mention R K Narayan here.) Moreover (and perhaps I say this more as a lover of the English verse form than one of Sanskrit poetry), comparing them is great fun. It is instructive — well, at least interesting — to see what choices translators have made, of metre and rhyme, and what inevitable changes in meaning have been forced upon them.

[Note: Translations in prose, or ones that are meant to accompany the Sanskrit originals, do not count, as they are not required to be beautiful, and rarely are. Free verse can be beautiful occasionally, but since it does not operate under the constraints of metre, it has no excuse for a bad translation: If you’re not even following the rules, you’d better be bloody good.]

More importantly, I think translations are highly valuable. While it would be nice to read Sanskrit poetry in the original, our time is limited, and very few people have the interest to do so. Sanskrit is not dead, but is quite probably dying, if ongoing attempts at its resuscitation do not prove sufficiently successful. To make matters worse, the few “scholars” who take interest in Sanskit are often not the sort to appreciate poetry. To quote from Schelling again:

Few Sanskrit scholars appear to like the poetry. The standard reference books—British scholars compiled them during colonial times—treat it with dismay or a sneer. […] If the professionals don’t like it, readers will be indifferent. Predictably, the books get harder to find. […] many volumes are hopelessly scarce: they went out of print in Bombay or Poona a hundred years ago.

As AND Haksar argues,⁸ English is the new Sanskrit. For better or for worse, its role parallels that of Sanskrit 800–1000 years ago: a near-universal link language of the educated (“educated elite”), the standard language for technical works, and a way of ensuring the widest possible audience for a work.⁹ More Indians know English than ever knew Sanskrit, and translation is the only way of bringing this rich literature to them. For this to succeed, though, the translations have to be gripping and readable on their own.

---

Enough nonsense. Here is an example, from Bhartṛhari’s Nīti-śataka. It’s not an especially fine example, but it’s a simple one. (It even occurs in some editions of the Panchatantra.) Here it is:

   आरम्भगुर्वी क्षयिणी क्रमेण
   लघ्वी पुरा वृद्धिमती च पश्चात्
   दिनस्य पुर्वार्धपरार्धभिन्ना
   छायेव मैत्री खलसज्जनानाम्

   ārambha-gurvī kṣayiṇī krameṇa
   laghvī purā vṛddhimatī ca paścāt
   dinasya purvārdha-parārdha-bhinnā
   chāyeva maitrī khalasajjanānām

My literal translation:

   Initially long/strong and diminishing subsequently,
   Or small at first and growing thereafter:
   As in the day's first and second halves the shadow [progresses differently],
   Thus [does] the friendship of bad and good men.

The point of the verse is clear enough — an ingenious analogy is being made, between friendships and shadows, their growth or decline. I do not know if you find such analogies delightful — it is an acquired taste. And in our postmodern world, even making such statements about bad and good men is unfashionable in some circles, but nevermind that. (And what does it say about us and our lost friendships?)

Focusing on the trivial, I think one of the aspects that makes this verse successful, at least on a first reading, is the pleasure of recognition on figuring out the analogy. You read the first three lines, without knowing what they are about, and then in the last line, in consecutive words, you discover almost simultaneously that it’s referring to shadows and to friendships. The joy is similar to that of hearing a good pun, or working out a simple puzzle.

Note that the translation must suffer from vagueness: while in Sanskrit the terms gurvī and laghvī (heavy and light, among other meanings) are applied to both shadows and friendships, the parallel does not easily work in English: we speak of long and short shadows, but strong and weak friendships. And the literal translation of “day’s first and second halves” was a mistake: even though sometimes “midday” can mean noon, we do not nowadays think of the day’s first half as the period before noon.

The translation must also lose the sound. The verse is in a kind of upajāti metre (शाला, says the Sanskrit metre recognizer), and can be musical. (If you are not familiar with the metre, letting the verse play on your lips for a few hours should do the trick.)

Look at what compromises translators have chosen.

First, included just because it’s the most recent, even though it’s not in metre, Greg Bailey’s translation from the Clay Sanskrit Library version:

   Heavy to begin with, weakening steadily;
   Slight at first, then later expansive;
   Different as the first half and the second half of the day,
   The friendship of rogues and that of good men and
       are like shadows. [sic]

Next, the verse from Barbara Stoler Miller’s acclaimed 1967 translation:

   One bursts forth expansive, dwindling the course,
   The other, scant in prelude, finally grows great.
   As shadows of afternoon differ from those of morning,
   So differs the friendship of evil and good men.

These put the analogy first and the things being compared in the last line. Three old rhyming translations (1877, 1899, 1930s) take the opposite approach:

Tawney

   The kindness of the bad at first
       Is great, and then doth wane;
   The good man's love, at th' outset small,
       Slowly doth bulk attain;
   Such difference between these two
       In nature doth abide,
   As 'twixt the shadow of the morn
       And that of eventide.

Too antiquated even for 1877, perhaps, but we can live with it. The wane/attain rhyme is ok, but this was the first I‘d seen ’eventide’.

Paul Elmer More

   Like as the shadows of the twilight hour
   Differ from those at morn,
   So doth a good man's friendship in its power
   From that of evil born : —
   One small at first still stronger, deeper grows,
   One shortens to the close.

Better.

Arthur W. Ryder:

   TWO KINDS OF FRIENDSHIP
   The friendship of the rogue or saint,
   Like shade at dawn or shade at noon,
   Starts large and slowly grows more faint,
   Or starting faint, grows larger soon.

Simple and sweet.

Again, I’m ignoring prose translations and commentaries, though, just for fun, this:

50. Initio magna, sed paullatim sese dissolvens; brevis antea, sed postea crescens, similis umbrae diei antemeridianae et post meridiem, est amicitia cum malis bonisque hominibus.

---

1. Elegant variation is the excessive use of synonyms to avoid repeating the same word. Fowler coined it, and gave examples like the use of “the emperor”, “His Majesty”, and “the monarch” in the same paragraph. He says

   “When the choice lies between monotonous repetition on the one hand and clumsy variation on the other, it may fairly be laid down that of two undesirable alternatives the natural is to be preferred to the artificial.”

   ↩

2. See What is the longest word of Sanskrit?, posted today on the Ἡλληνιστεύκοντος blog: it has an example of a “word” spanning 428 letters in Roman transliteration, which is one of 28 adjectives qualifying the sentence “On the way, he passed through the Tuṇḍīra country”, which is also a chapter — and takes 130 lines of English translation. ↩

3. Goethe on Shakuntala:

     Willst du die Blüthe des frühen, die Früchte des späteren Jahres,
     Willst du, was reizt and entzückt, willst du was sättigt and nährt,
     Willst du den Himmel, die Erde, mit Einem Namen begreifen;
     Nenn’ ich, Sakuntala, Dich, and so ist Alles gesagt.
   
   

   Translated by a certain Eastwick:

     Wouldst thou the young year's blossoms and the fruits of its decline
     And all by which the soul is charmed, enraptured, feasted, fed,
     Wouldst thou the earth and heaven itself in one sole name combine?
     I name thee, O Sakuntala! and all at once is said.
   
   

   ↩

4. Barbara Stoler Miller, Bhartṛhari: Poems (1967), Columbia University Press ↩

5. A sandeśa kāvya or “messenger poem” is the genre of poetry originated by the Meghadūta. See e.g. Wikipedia on Hamsa-Sandesha. ↩

6. Andrew Schelling, Manuscript Fragments and Eco-Guardians: Translating Sanskrit Poetry, Manoa 11.2 (1999) 106–115. ↩

7. Andrew Schelling, Erotic love poems from India: a translation of the Amarushataka (2004). See some sample poems here, but the Google Books preview reveals more. The common translation of śṛṅgāra as “erotic” is unfortunate, but more on that some other time. ↩

8. An interview with AND Haksar ↩

9. In the best days of Sanskrit, this was surprisingly large in extent — including, for example, Sri Lanka, parts of Central and Western Asia (incl. China), and the far reaches of Southeast Asia. See e.g. Sheldon Pollock’s discussion of The Sanskrit Cosmopolis, in his work of that name and in Language of the Gods in the World of Men: Sanskrit, Culture, and Power in Premodern India. ↩

10. There’s more to this; see comments below.↩