Trisha Gupta interviews translator Daisy Rockwell for Scroll; there’s some rote stuff (how did you get into it? what about the unequal power differential between English and other languages? do you focus on accuracy or English readability?), but also some interesting details:
What do you do about dialect, or idiomatic phrases? Do you try to produce an equivalent in English? This can be a difficult thing to do… I remember in Falling Walls, you have Chetan calling his Bhai Sahib, Ramanand, the Old Codger. The nickname is remarked upon at some length, but we do not learn the original term in Hindi.
Some aspects of dialect and idiom just cannot be translated, and if they were kept in the original language in the translation, it would not be a translation anymore. There is a school of translation in India which feels that smoothing these elements out is doing violence to the original text and that translating it into English at all is doing violence really, because of the hegemony discussed above. However, if one has committed to rendering a text in English, one must bite the bullet and figure out how to get it done. If a nickname or something is particularly hilarious, I might keep it in Hindi. It’s really a case-by-case basis for me. In the case you are talking about, the nickname was baṛhaū, which is a) not that funny by contemporary Hindi standards, and b) difficult and unattractive to render in the Roman script, thus I chose to come up with something a bit old fashioned in English.
Now the big problem for a translator from Hindi and Urdu into English is that one is bound to have many readers who not only know at least a passing amount of said languages, but may actually be fully fluent in them, and literate too. Why are they reading the English? Often it’s just their habit to read in English, but they are also the most critical readers of translations, and complain of translators “over-translating”, having a preference for being able to “feel the Hindi” through the English. I have seen many reviewers say such things about translations from Hindi and Urdu (not of my books, but of others), and I must say, if they are so eager to “feel the Hindi”, they really ought to take the trouble to purchase the Hindi original, since they don’t need an English translation. [..]
Do you think there needs to be a different kind of translation for readers who are familiar with the cultural context of a work – Indians reading in English – than there is for foreign readers who have never encountered a basti or a chulha or a hakim? Are Indian publishers becoming more comfortable with and cognisant of this need?
I would say yes, and covered some of this above, as in the case of the readers who actually know Hindi but do not read in it. However, since publishers outside of India do not currently have any interest in South Asian literature in translation, I feel that we translators must attempt to create texts that can be all things to all people. I do keep many Indian words in the text, but I also tend to give a cursory gloss for terms that a non-Indian reader wouldn’t get. I don’t like having glossaries, and I do think in a long book, readers can learn certain terms from context. The trick is not to overdo it so that the non-Indian readers get overwhelmed and put the book down.
Similarly, one should not under-do it because then the Indian readers will get annoyed. One rule of thumb I use is I ask myself: “Is this word used very often in Indian English?” “Did this word make it into Hobson Jobson?” If the answer is yes, I will keep it. I might also put in a word to give the reader a hint, like “crunchy chiuda”, but never naan bread or chai tea.
Kinship terms are hands down the most difficult aspect of translation into English from South Asian languages in my opinion. Women’s writing contains way more of these terms than men’s writing, simply because there is more action inside the house than outside, generally. With these I try my best to come up with English equivalents, but also include some original terms so I won’t be accused of over-translation. The problem with the kinship terms of course, is not only are they very elaborate, but they are all context-centric, so one person’s devar is another person’s bhai sahib, is another’s chacha ji, etc. In Falling Walls, I called Chetan’s elder brother Bhai Sahib and used as I would a name, because I simply couldn’t imagine him without that title, and he was an important character. I called the mother Ma because this is perfectly understandable in English, but I didn’t call Bhai Sahib’s wife Bhabhi, because he also gives her real name, and it would get confusing with all the bhabhis in the house.
There’s some snark about Tim Parks (“although he is a translator, he refuses to disappear, and is always popping up with unnecessarily nasty critiques of other translators, particularly women who are getting more attention than him”) and an interesting point about translatability:
Hindi writers are generally speaking all about regional specificity and inhabiting their linguistic sphere fully. Hindi literature was mostly born and developed as a nationalist impulse, somewhat like modern Hebrew, and I think that Hindi authors still feel that they are forging a new idiom and a new literature. This makes Hindi extremely difficult to translate at times. Certainly way more difficult than Urdu prose, which does not have the same “newness” chip on its shoulder.
And she wants (admirably) to focus on translating women authors: “I realised suddenly that I’d only been translating men (Ashk, Bhisham Sahni, and Shrilal Shukla), and I felt fed up with the male gaze.”
Does that mean “father-in-law”? I ask because it’s straight from PIE…
< *dāiH₂u̯ēr- “husband’s brother, brother-in-law”
There’s a strange archaic Norwegian word dove m. [²du:ve] “wife’s sister’s husband” (I’ve seen it glossed as “son-in-law” dialectally). I’ve never been able to make anything of that beyond some resemblance to the IE word. It’s not attested in ON, but the most natural origin would be **dófi m. Beyond that I’m almost blank. I now wonder if it might be an early Celtic loan.
It could be an affectionate form of something like **dófr < **dōwř < **dāwz. Looking it up again just now I noticed dúfungr glossed as “nickname, lit. son of a dúfa m.?” Interestingly, there’s no dùfa m. in the dictionary. And I think **dófr might work too. Dúfr m. is on the list of dwarf’s names in the Þulur.
* Introducing ř for [R from z]
… but this is all a diversion from the topic of the post. A very good interview. It made me think, and then the next question and answer followed up, not only on the previous question but on the thoughts I’d been having when reading it. Like this:
*-āiH₂-? That’s creative.
I like that.
*-āiH₂-? That’s creative.
Yes. The whole proto-word looks like something put together of so many pieces that anything could happen in the daughter languages — including laryngeal hardening in the Germanic forms. Looking at the other kinship terms, I’d expect it to be a compound word which was reetymologised or remodeled at different times in different branches. I was thinking of a compound with *wir- parallel to *swe-sor-, but so far my only suggestions for a first element are *deH₂i- “divide” and *deik’‑ “point”. I could invent plausible-sounding semantics for both, but it’s not worth much without the phonology, and that’s too much for me.
*deH₂i- “divide”
Not really relevant to *dāiH₂u̯ēr-, but I took that from the AHD root index:
It lists deal and dole among its derivates. Other Germanic forms are e.g. Scand. (arch.) de(i)ld and Ger. Teil. But why does Germanic have *d-? Doing what I should have done before* — asking Bjorvand & Lindeman — I learn that they unite the Germanic words in a verb *dailijan- with close correspondences in Baltic and Slavic, “which likewise may reflect a proto-form with initial *dh-“.
With a stem *dail-, the proto-form should be *dʰeH₂il-, but I don’t particularly like Slavic and Germanic having a variant root of their own. Given the causative form of the verb, it ought to mean “make become a dail“. Could this **dail instead have been borrowed from Celtic as a term in property law or something?
*) I’ve been through this before but I forgot.
Or *dʰH₂eil- of course. Or an o-grade…
Back to to *dāiH₂u̯ēr. Let’s say we take it at face value. I see only two ways to break it down:
(1) *dāiH₂ + *u̯ēr
(2) *dāi + H₂u̯ēr
The first element of (1) could then be parsed as *dēH₂-iH₂, where *dēH₂- is “give” and *iH₂ is what would become the feminine ending, so “the (female) one given” = “married daughter” or — with a mere individualizing function — “individual part of the given” = “share, portion”. With “married daughter” for the first element, the second might be “man” — except it still doesn’t work phonologically. Would some form of *u̯er- “be aware, guard, watch” work without a reconstructible suffix? “Warden of a portion/married daughter”.
As for (2), the first element would still be “portion”, but I have no suggestion for *H₂u̯ēr.
Now we must call on Piotr to clean up my mess.
David M.: Or *dʰH₂eil- of course. Or an o-grade…
Admittedly. But *dail- would still look like an aberrant form.
Why are they reading the English? Often it’s just their habit to read in English, but they are also the most critical readers of translations, and complain of translators “over-translating”, having a preference for being able to “feel the Hindi” through the English.
That struck me, too. Linguistic reality (and human perceptions of language) are always more complicated than the simplistic “This is one language, and this is another one” model would have us believe.
Me: *dēH₂-iH₂, where *dēH₂- is “give” and *iH₂ is what would become the feminine ending
I was very happy when this struck me, but I don’t think it works. First, I attach an adjectival suffix to a bare verbal root, and second, *-iH₂ isn’t the same as the standard (proto-)feminine ending *-eH₂. Yesterday I thought they’d still be fine together, but that was probably just wishful thinking.
I mean, yesterday I thought dēH₂- might work as a root noun. Today I think probably not.
Eichner’s anti-law: only *e was colored by laryngeals, *ē was immune (and shows up unchanged in Hittite).
(I’ve been trying to continue the Pokorny notation of the Wiktionary form, but it’s not fun anymore.)
Mea culpa. The length shouldn’t be there before assimilation of the laryngeal. The root is *deh₂-, and it seems well established, though in the meaning “divide”, not “give”.
And mea culpa again. I confused myself to the meaning “give” when I briefly considered *deh₃- instead of *deh₂- and pondered whether the two could be related by ablaut (since ablaut plays a role in athematic root nouns, and I think it needs to be an athematic root noun to take the suffix *-ih₂).
The meaning “divide” isn’t any worse semanticaly, just more boring. A revised (but still unsatisfactory) suggestion is:
*dāih₂-wēr < *deh₂-ih₂, where *deh₂- is a root noun from “divide” and *-ih₂ is what would become the feminine ending. This would meen something like “portionee”.
Or (suddenly seeing it now) < *dh₂ei-h₂, where *dh₂ei- is an extended root also meaning “divide” and *-h₂ is the collective ending. So something like “a portion consisting of many items”.