How hard should we try? – Questions of detail in literary translation

Ketaki Kushari Dyson





We have all been sent a sheet outlining the questions that you would like us to address, and I would like to take off from the first two sets of questions posed there. To me, ‘What is the purpose of translation?’ is the crucial question. What we decide to call ‘a good translation’ depends on the answer we give to that first question. Whenever I undertake any task of translation, I ask myself: what is the purpose of this particular task that I am taking on? There could indeed be a plurality of purposes in any single task, so the idea of what is a good translation needs to be broad and accommodating rather than narrow and rigid. I certainly think that the quality of literary vitality can be conveyed in translation, and I believe that our attitude to form needs to be flexible. The question of mistakes is an intriguing one. Some mistakes may be of the straightforward kind (say, a word or phrase inadvertently missed out, or a word misunderstood) regarding which we can reach an immediate agreement; with other mistakes, it may be necessary to have quite a long discussion before any such consensus can be reached; and sometimes slightly different interpretations are entirely possible, so that translators (and scholars) will have to agree to differ. A few mistakes do not invalidate the whole work, and shifts of meaning are inevitable when a text moves from the terrain of one language to the terrain of another. In that case, how hard should we try? I think we have to try our best without getting wound up about it. I would like to illustrate this with examples, using some English translations of the Bengali poet Vijaya Mukhopadhyay which have been published as a booklet. I hope to focus on concrete examples in the spirit of a workshop. Through such focusing we can raise our awareness of the practical issues involved in the craft of literary translation and improve our skills.



[What follows is a slightly revised version of a talk I gave with the help of an overhead projector at a translation workshop held at the School of Oriental and African Studies, University of London, on 19 September 2002. Although called a workshop, the gathering was really in the nature of a mini-conference. It was a two-day event, over 19-20 September 2002, called ‘Workshop Three: The Art of Translation’, within a series entitled ‘Translations and Translation Theories East and West’, organized jointly by Dr William Radice of the School of Oriental and African Studies and Prof Theo Hermans of University College London, with the assistance of Dr Ross Forman of the SOAS. I am delighted that they have agreed that this workshop paper can be published in Parabaas. Interestingly, prior to this publication some fruitful discussion took place between the editor of Parabaas and myself which was itself in the nature of a workshop, and I have made some amendments in the text accordingly. My sincere thanks to the editor of Parabaas.]



Since time is limited, I shall not recap the basic points which underlie my approach: they are already in the Abstract, and you can look them up. At the end of the day, literary translation is a creative and imaginative art and a practical craft in its own right, requiring very special writing and problem-solving skills. I think most practising translators would agree that nowadays there is far too much preoccupation with theory in academic circles and not enough understanding of the nitty-grity of the actual task. I would like my contribution to be as in a real workshop, a hard-nosed dive into details. It will be in the nature of notes.


I shall take all my examples from a booklet of translations from the poetry of the contemporary Bengali poet Vijaya Mukhopadhyay (b. 1937), entitled No symbol, No Prayer, published by Cambridge India (Educational Publishers), Calcutta, in 2001. The translators are Carolyne Wright in conjunction with Paramita Banerjee and Sunil B. Ray, and in collaboration with the author. I have chosen this booklet as the basis of my presentation as it is a serious and on the whole reasonably competent effort, and yet certain ‘problematic’ areas remain in it, even after - or is it partly because of? - such a collaborative effort, and despite the close consultation with the poet which is supposed to have taken place. Vijaya is an articulate woman and I would have thought quite capable of discussing fine points with her translators; nevertheless, the residual ‘problems’ indicate how easy it is for gaps to develop when, as in this case, two or more parties with different cultural backgrounds and professional trainings, and varying levels of ability in the two relevant languages, are trying to communicate.


Certainly, what I am referring to as ‘problems’ do not invalidate the whole work, but if the main purpose of issuing a book like this is to generate interest in the work of a particular poet amongst those who cannot read his/her original texts, to capture a new readership for that poet, then is an even finer attention to detail required to do justice to the poetry and to capture the attention of new readers in an age when fewer and fewer people are reading poetry in the first place? Twentieth-century poetry is frequently dense and concentrated, with a special reliance on images and oblique innuendoes, and more often than not there is no story-line to carry the reader through. Modern poetry in translation is therefore particularly vulnerable to flagging reader-attention: if the translated texts lack vitality and vibrancy, if they lack the accent of poetry and sound prosaic, readers soon go to sleep. There is no one way, or perfect way, to translate a poem, but I believe that we can sometimes see better ways to do this or that, and that translation skills, like any other craft skills, can be polished and improved.


Vijaya, a qualified Sanskritist, is a poet who uses words extremely carefully. Her poetry is lean and taut, characterized by precision, economy, irony, and acerbity. A rigorous and masterly approach to language is needed to capture the distinctive flavour of her poetry.


Of necessity I shall focus on some selected small-sized examples, since looking before and after is quite impossible unless all of you are furnished with the original and translated poems in front of your eyes. This may suggest that I am just interested in an exercise of spotting the problems for its own sake. Not at all. My goal is to improve our understanding of the nature of the task, to see how we may tackle problems on the ground and refine our techniques. We can never hope to improve our techniques unless we learn to focus on details. And it is the overall competence of the translators that makes such a focus all the more rewarding and educative. My queries and suggestions are offered with humility, and with due respect to all the translators. Looking at smallish samples is dictated by the format of this workshop and the time-limit. So let me plunge into the job in medias res. My first bundle of notes is called:


Some simple examples

on your forehead put a dot of “bindi” powder

made from burnt postal cards,

put a sprig of jasmine in your hair (‘That’s not for Puti’)


I am not sure that ‘sandhyamalati’ is jasmine. I suspect it is a local name for a completely different flower. As far as I know, it is a small bush with purplish flowers. If Vijaya Mukhopadhyay had really meant jasmine, would she not have written ‘jui’? The editor of Parabaas and myself have had a lot of discussion about the identity of the ‘sandhyamalati’.
Mirabilis jalapa, sent to us as
sandhyamalati by the photographer,
Arunangshu Sinha
What some of us know as ‘sandhyamalati’ is called ‘sandhyamoni’ or ‘krishnakali’ (the Mirabilis jalapa) by some others. This flower does blossom in the evening (‘sandhya’). There is also considerable diversity in the interpretation of the name ‘malati’. If it is interpreted as ‘jasmine’, as it sometimes is, then ‘sandhyamalati’ could indeed be interpreted as ‘evening-flowering jasmine’, a sprig of which would be appropriate on a woman’s hair. However, I remember doing a great deal of research into plant-names when compiling the glossary for my translations of Tagore, looking up books compiled by botanists and horticulturists, pestering students of botany in the Santiniketan campus, getting them to consult their teachers and prepare lists for me, getting people to show me the actual plants, and as a result of those investigations, the ‘juthika’ or ‘jui’ was identified as the Jasminum auriculatum, the ‘bel’ as the Jasminum sambac, the ‘chameli’ as the Jasminum grandiflorum, all of the Oleaceae family, whereas the ‘malati’ was identified as the Aganosma dichotoma (or the Aganosma caryophyllata of earlier classifications) of the Apocynaceae family. I was shown the ‘malati’ plant, a climbing shrub with fragrant white flowers. Does all this matter? That will depend on the translator’s overall approach to local details. Personally, I like to carry over as much detail as can be accommodated in the target language without upsetting the poetical balance, and in this case I would have gone for the simple option of retaining the original name and adding a note. I am intrigued why this was not an option for these translators when they took much greater trouble in the immediately preceding line, resulting in a somewhat heavy-footed line, where six words have become fourteen, and the poetry has been compromised by a cumbersome explanation. The Hindi word ‘bindi’ itself requires an explanation, as does the process of ‘tip’-making referred to, neither of which is provided. To attract new readers, the translation of poetry needs to be ‘sharp’ and rhythmic, not bland or tired or stale. One could have written:


put a dot of burnt-paper powder on your forehead,

stick a sprig of sandhyamalati in your coiled hair ....


and added appropriate notes.


What about the following example? -

I have removed all my trinkets,

have lifted off my veil, my gold tiara

and hair ornaments. (‘To Be Worthy’)


I feel there is a question of interpretation here: the philosophical force of ‘bahulya’ is not conveyed by the word ‘trinkets’. What the poetic persona is saying is closer to: Look, I am divesting myself of all superfluities. As in the original text, a period would be appropriate after the first line, which is making a general statement, after which come some specific details of ongoing action in the second line. And ‘sinthimour’ is one detail the precision of which needed to be respected. It is the sola crown worn by bride and groom. The phrase ‘hair ornaments’, in the plural, is too vague and fudges the issue. After all, ‘tiara’ is also a hair ornament. The difference between the two needs to be indicated - for the sake of the poetry, because some of the poetry resides in the collocation of such details. We need to keep the end always in view - which, in this type of edition, is to recreate the poetry and recruit new readers for the poet.


A similar example is this one:

Sometimes it’s better

mentally to let go. (‘Better let go’)


I am puzzled by this rendering; ‘nijeke arpon kora’ is ‘offering oneself’, not just ‘letting go’. Shouldn’t one at least write: ‘Sometimes it’s better/ mentally to let go of yourself’?


Here are examples of some off-the-mark renderings of single words, of the kind that do matter in intensely terse and reticent poetry such as Vijaya’s.

Mere words, empty words, meaningless. (‘Words’)


Here the interesting adjective ‘parinamheen’, meaning ‘leading nowhere, without a future’, has been translated by the more bland adjective ‘meaningless’. If Vijaya wanted to say ‘meaningless’, she would have surely written ‘arthaheen’.


‘Duallir moth’ in ‘Kolkatar Karakamale’ is rendered as ‘the monument at Dualli’ (‘Homage to Calcutta’): but ‘moth’ is a monastery or similar meeting-place of sannyasis; ‘monument’ has architectural associations, but no necessary religious associations.


‘E bishal yajnashala’(‘Kon Flore’) is rendered as ‘this colossal edifice’ (‘Which Floor’): however, ‘yajnashala’ implies the hectic activity or busy-ness of a religious ritual, whereas ‘edifice’ merely refers to a big building. In the same poem the phrase ‘koi, kon flore’ is rendered as ‘where,/ which floor’. I feel that ‘which floor’ should be ‘on which floor’; omitting the locative causes the question to limp.


‘Boithaki moutat’ (‘Punarlikhito Kabita’) is rendered as ‘opium-smoke haze of the drawing-room’ (‘Rewritten Poem’): this narrows the meaning unnecessarily; ‘moutat’ could be the haze of any kind of intoxication, most likely, from the context of the poem, from tobacco or alcohol. Opium is unlikely.


‘Pradipe kotu tel chhilo’ (‘Deepavriksha o Holud’) is rendered as ‘There was mustard oil in the earthen lamp’ (‘The Lampstand and its Yellow Flame’): I haven’t understood why the lamp is earthen when a few lines down a brass artefact is referred to. Is it a brass lampstand with an earthen lamp at the top? If so, didn’t it need to be explained in a note?


In the phrase ‘the heels’ pale rhythm’ (‘Henna’), ‘pale’ is not the right translation of the word ‘mlaan’: amongst the various nuances of this adjective are ‘tired, exhausted, languid’; hence in the context of human gait ‘flagging’ or ‘halting’would have been more appropriate. Vijaya has used this adjective here in that special sense.



Let us now look at some more complex cases, where, in my opinion, ‘trying harder’ might have yielded better results.

The cat jumps up on the corrugated iron roof (‘Film’)


As far as conveying accurate information goes, the translation is absolutely correct. I have checked this with people knowledgeable in materials: what is called corrugated tin in common parlance is actually corrugated iron, but in this line is there perhaps an oblique reference to the film Cat on a Hot Tin Roof (based on a Tennessee Williams play)? After all, the poem is entitled ‘Film’. Wasn’t some annotation necessary? This book provides no notes at all.


The following example, the last lines of the poem in question, deserves our close attention:

In the half-darkness

the upright lampstand and its yellow flame -

no symbol, no prayer. (‘The Lampstand and Its Yellow Flame’)


The way the lines have been translated, with the addition of a dash where Vijaya herself does not punctuate at all, suggests that the last line has been taken to mean: ‘Here there is no symbol, no prayer’. This is OK as far as it goes. But with a little effort the translation could have been given a tilt towards what I reckon is the poet’s intended syntax: ‘The upright lampstand with its yellow flame is neither a symbol nor a prayer (though one is tempted to take it as such).’ The present indicative of the verb ‘to be’ is implicit, not explicit in Bengali, and its hidden presence can be sensed between lines 2 and 3 of the Bengali excerpt. I also think that the last line needed to be repeated, as in the original poem. The original poem is a strong philosophical statement, which is reinforced by the repetition of the last line. Since the English volume gets its name from the last line of this poem, the repetition would have been helpful. The English lines could be rewritten thus:


In the half-darkness

the upright lampstand with its yellow flame

is neither a symbol nor a prayer.


Neither a symbol nor a prayer.


The other point worth considering is the fact that the word ‘holud’ (‘yellow’) has been used in a rather uncommon, ambiguous, starkly modernist way in the original poem. Strictly speaking, in the four original lines given above and in the title of the poem the word ‘holud’ is not an adjective, but a noun, the quality of yellow-ness, the name of a colour. The title should really be translated as ‘The Lampstand and its Yellow’. The word for ‘flame’, ‘shikha’, does occur in the original poem twice, and its yellow colour is explicitly acknowledged, but interestingly the word is not used either at this point or in the title. The word ‘flame’ has been kind of interpolated by the translators at this point and in the title, presumably for the sake of fluency. So while they are not wrong to have translated as they have done, the meaning has been somewhat narrowed and the modernist use of language has been toned down. In the semi-darkness the brass lampstand itself has a yellow sheen, to which is added the yellow tint of the flame: there are two yellows, so to speak, one big, the other small, and together they make up a special image, a tall, dominant splash of yellow. In view of this, the second line of the extract given above could have been ‘the upright lampstand and its yellow’ or ‘the upright lampstand with its yellow’, and that would have been stronger and sharper, and nearer to the original effect.


The following is an interesting example:

I came back home and found my home wasn’t there. (‘No Home There’)


We have been given a somewhat matter-of-fact statement; the play of words which is the heart of the original line isn’t quite there. There the word for ‘home’ is repeated three times, and the poetry is in the paradox of the statement. To me the line implies: ‘The house is still standing, but there is no home for me in it. It has ceased to be my home.’ To bring this over a creative gesture is called for. What about this? - just a suggestion! - ‘I come back home, and find my home’s not at home.’ Further down in the translation of the same poem we encounter the phrase ‘Like a stranger on a wooden chair’: in the original poem, it is not a stranger but a tenant who is sitting on the chair. The transformation of the tenant into a stranger is mysterious. I would have said that the original line suggests a different transformation: the owner of the house has now become a tenant, waiting nervously on a hard chair for the real owner to turn up. This meaning could have been carried over with no effort at all. Why it wasn’t I don’t know. Were the translators seduced by a Camus-like idea of ‘the stranger’? True, it all boils down to an image of alienation, but personally I would have tried to retain the image of the uncomfortable tenant, as it touches a Bengali social reality - the ambiguous attitude of tenant to landlord.


And here is another one:

O fire

you know, this is my lonely ritual of Sati. (‘Are You Not Fire’)


The point is that the woman is burning on her own, whereas in the ritual of sati a woman usually burns with the dead body of her husband, or some emblem to stand for him. The point to grasp is that the poem is not really about sati, but about death itself, its inevitable loneliness. The poet visualizes her own death and the cremation of her body, which is earlier described as ‘nirbandhab’, ‘friendless’ (as translated by these translators), but more radically, following Sanskrit, ‘without kinsmen’. The function of the image of sati is to sharpen the essential loneliness of death. This poet knows that she will not even have the dubious comfort of company in death. And the word for ‘fire’ used in the original is ‘paavak’, the root-meaning of which is ‘purifer’ and which I am quite sure has been used here deliberately by the Sanskritist poet. The title, echoing one of the lines in the poem, is not quite saying, simplistically, ‘Are you not fire?’, as in this translation, but something much more like: ‘Are you not meant to be a purifier?’ So, incorporating the root-meaning of the word ‘paavak’, and redistributing the words between the lines a little bit for the sake of rhythm, one could rewrite the lines as:


O Fire, Purifier, you know

this is my act of sati, where I burn alone.

Disdainful city

do you fly like dust over my head,

cling to my feet like mud, ... (‘Rock of Resolve’)


I would say the word ‘like’ needs to be replaced with ‘as’. The city is dust, is mud. These are parts of its being.

You can have only five abortions,

the advertisement says so. (‘Advertisement’)


The context of the poem indicates that the first line needs to be translated as: ‘So you can have up to five abortions’.

First you put your hands on the system, then on the strings. (‘Sandhiprakash Raga’)


I wonder why ‘ganit’ has been translated as ‘system’ and not as ‘mathematics’, the clear meaning. Isn’t ‘system’ too broad? Isn’t the poet referring to the maths of music?

The heart-rending cries of beggars

keep exploding in your ears -

who can say if this beggary is a commercial racket

or real indigence? (‘Climb up the Rope’)


The last two lines have moved away a little from the original. What the poet is saying may be paraphrased thus: ‘Looking at all these beggars and listening to their cries, who can say the begging of which one amongst them is simply a ploy to make money, and whose is real helplessness?’ In other words, some of it is genuine and some fake. The translation could have easily indicated this.




The lack of annotation in this book needs to be mentioned separately as an issue. I cannot see how in translating across cultural boundaries we can opt out of annotation altogether, as has been done in this volume.

What do you understand of the ULF? (‘That’s not for Puti’)


Well, none of us can understand, unless the acronym is explained! Is it the United Leftist Front? Can we blame people if they think it is the United Liberation Force?


The publisher’s blurb says: ‘The anthology is worth international acclaim.’ But I don’t see how an international readership can be expected to understand the meanings/associations of terms and phrases such as - and I cull examples from various poems - ‘the Devisukta’, ‘Sunilmadhab’s confusion of colours’, ‘Sandhiprakash Raga’, ‘Nirmal Hriday’, or ‘nine doors in your virile body’ without the assistance of notes. I want you to look at the extract which has the first two phrases:

starting with the mastication of a few lines

from the Devisukta or King Lear

and ending with Sunilmadhab’s confusion of colors

or the universal scope of Esperanto. (‘Climb up the Rope’)


I have been puzzled by the phrase ‘confusion of colours’ for ‘varnavibhram’. ‘Varna’ can mean both ‘colour’ and ‘a letter of the alphabet’. The other items in the context - the Devisukta, King Lear, Esperanto - all have a connection with language and literature, so at first I wondered if Sunilmadhab’s confusion was some kind of alphabetical or script-related chaos! I wished I knew who this guy Sunilmadhab was! - some linguist, philologist, or grammarian? BUT the editor of Parabaas tells me that there is indeed a painter called Sunilmadhab Sen, so a reference to the way he uses colours is entirely possible. Such a reference does, however, need elucidation. I am not even sure which verses of which text the Devisukta refers to. So even with my reasonable knowledge of Indian material I need notes, and none are provided! How will the ‘overseas’ reader fare?


The poem ‘Graase, snaanajale’, translated as ‘In Devourings, in Bathwater’, surely has a link with the image of ‘graas’ as eclipse? But there is nothing in the translation to indicate it and no note to elucidate it.


What is one to make of this one - is this a cock-up or a considered decision?

the solemn, myopic boy walks by,

the unyielding principle of tradition,

cold naiveté

glued to his back like a stone. (‘What You Gradually Realize’)


But ‘Mugdhabodh’ is the name of a famous grammatical text, by Vopa-deva, the root meaning of the term being ‘that by means of which even the dumbest students can be taught, made to understand’, i.e. ‘Sanskrit grammar made easy’! Should the name of such a famous text have been translated as ‘naiveté’, without any note? Shouldn’t it at least be ‘grammar-made-easy’ or some such formula? This example makes me think that the translators did not have their entire work vetted by the poet, but consulted her only when they thought it necessary, and this little problem slipped through the net because none of the translators realized that it was a problem!


Missing phrase or line/ wrong connections


Missing something out is a hazard of translation, most often due to the sheer fatigue of maintaining the focus of the eyes on two separate texts, the original one and the one that is in the process of being built. We can ask a friend to check things for us, but if the work is substantial, he or she will face the same problem, noting some problems maybe, but missing others. Translation is a labour-intensive process carrying very little remuneration for the time and labour spent on it.

in the exclusive shops of which cities are these terylene shirts collected for droughts seen hanging, and is that robust girl’s name Love, ...... (‘The Magic Mirror’)


The phrase in the original saying ‘and why I still don’t know how to knit wool’ has been missed out. Of course, this could be a case of printing error rather than a translator’s slip. But what about this one? -

The debate is over.

At any moment an earthquake could occur

just as a friend might drop in at another friend’s house,

from darkness an even thicker darkness flicks off.

An uncomplaining silence sits with lowered head

close to the fringes of flyaway hair.

Sleep, inebriation, memories float away, fragments

in the undertow.

The no’s - scattered like a sackful of nails on the floor,

a sackful of witchlike pitch-black no’s. (‘Twelve Fifty A.M.’)


You can see here how not paying attention to the original layout is accompanied by a string of errors. The phrase ‘bhese jaay prajnaa o prithivi’ has been swept off; the gap between the stanzas has been closed; and the last line of the first stanza has been wrongly joined up with the first line of the second stanza. Interestingly, there are two further errors: ‘bhrangsha’ is not really ‘fragments’ but ‘slippage’; and ‘ekguchchha’ can never be ‘a sackful’- it is not more than ‘a handful’.


I would therefore say that punctuation and layout are also details to which we have to pay as much attention as we can. In this volume, space between stanzas is observed in an erratic fashion. It is sometimes there, sometimes not. Sometimes gaps have been sacrificed just in order to fit the poem on the page. I don’t think that is acceptable. We have to try harder. A smaller font might have done the trick.

I never know who gets into my blood,

my veins, nerves, and brain go numb. (‘Let It Come Down the Highway’)


Even though there is no punctuation after the first line, the Bengali lines are quite clear, because of the word-order. But reading the English on its own, I read ‘my veins, nerves, and brain’ as a continuation of a list beginning with ‘my blood’, until I hit the verb-phrase ‘go numb’, which indicated otherwise. Then I looked up the original to check what was going on. This kind of thing is what often puts readers off modern poetry in the first place, and is a further irritant when reading poetry in translation. A semi-colon rather than a comma would have been better after the first line.


Here is an interesting example of a wrong connection, the kind I think we could eliminate by just a little more attention to syntax:

Thus sufficiently wearied

even when you make a spectacle

of turning your back

we shall hail your “glorious retreat!” (‘Man’)


Let us for the moment ignore the question of whether the irony-charged image of ‘darshaniyo pith’ could have been rendered better. Indeed, the present rendering is quite imaginative. However, I think the syntax should have been: ‘Thus, even when, sufficiently wearied,/ you make a spectacle of turning your back’; otherwise the reader will link the phrase ‘sufficiently wearied’ with the last line’s ‘we’, which is what I did until I checked the lines against the original. I think the translator needs to make it quite clear who is wearied - the viewers or the viewed. Note also that the title of the poem (‘Purush’) has not been adequately translated. ‘Man’, so often used generically for all humanity, is too general in the context of the poem, which requires a clear statement of gender such as ‘The Male’.


These are just a few selected examples of ‘problems’. More can be gathered from the book, but I have no time for more examples. I end these notes with reiterating my basic position: I have attempted this exercise in the spirit of a workshop not with the intention of showing others’ errors - all of us, but all of us, make some mistakes in every task we perform - but with the hope of showing how as professional translators we could be more conscientious and try harder, for without excluding the elements of play and pleasure, anything that is worth doing is surely worth doing well, with a clear understanding of our objectives, and with an attention to the craft skills appropriate to the task. I hope some translator somewhere can learn something from this exercise. When a poem is transferred from one language to another, some slippage of ‘meaning’ is unavoidable at times; and sometimes, inevitably, new meanings are generated, which is all-right too, and can be fun, I think. When reading translated poetry, we are not in search of pale replicas of inaccessible texts: we want the joy of reading good poetry while savouring a slight taste of ‘otherness’. But skills in any discipline can be improved by training. If we as translators are prepared to go the extra mile, our re-creations can have that oomph which can kindle flames in new readers.



Published July 25, 2003

Ketaki Kushari Dyson [ketakI kushaarI Daaisan*] was born in Calcutta in 1940 and educated at Calcutta and Oxford. She has been based in England since ... (more)

Photograph of Mirabilis jalapa taken by Arunangshu Sinha who is a Department Head and Professor at Panshkura Banamali College in West Bengal.

Click here to send your feedback