The Challenge of a Faithful Translation

The Challenge of translation – Faithful yes, but not a slave

While no one disputes that a translation must be truthful, the definition of truthfulness and the ways in which translators have striven to achieve it have varied over the centuries. Word-for-word translation has given way to translation of meaning with the translated text reading as naturally in the TL as the original did in the SL. Reconciling truthfulness and beauty is one of the most important challenges faced by translators.


Much has been said and written about the notion of faithfulness (or fidelity) in translation, even the sexist comment that a translation is like a woman : if is faithful it is not beautiful and if it is beautiful it is not faithful, as if being both faithful and beautiful were mutually exclusive

Obviously, like everything else, “faithfulness” depends on how you define it – a principle of loyalty or honesty or a matter of exactness and accuracy ; or  both ; or much more that that ) – and also it depends on what you relate it to – word or meaning ; the source language or the target language ; the source text or the target text ; the author or the reader.

Faithfulness will also depend on the different choices you make and the strategies you use in different translating situations (oral or written), with different texts (literary or technical ; philosophy, poetry, logics, etc…). And accordingly, it raises different types of difficulties. Usually technical translators are envious of literary translators because they do not have technical problems to solve, and literary translators are envious of technical translators because they only have technical questions to deal with. We Dharma translators, are not envious of anybody else, because we have both : the technical problems and all the rest…

Without getting into theoretical issues about linguistic theories in translation, I would like to relate this notion of faithfulness to my personal experience as a Dharma translator and  specially to one model of translation strategy developed by Lederer (2001) at the ESIT school of translators in Paris that I find interesting and useful.  So, as this exploration of the extent of faithfulness,  has mainly given me the opportunity to reconsider my ideas about translation and my involvement in translating Dharma I am afraid that apart from being a very self-centered talk, the rest might be very familiar to you and overrun.


In the early eighties, when the director of a FPMT center in France asked me to translate orally, from English to French, the teachings of the resident gueshé on Shiné and Lhaktong, I thought he was pulling my leg. First, I did not know who Shiné and Lhaktong were and did not think that just knowing a foreign language suddenly qualified someone to be a translator or worse an interpreter. On top of that How can you translate something you do not understand ? The reason that apparently made me a translator was that I understood English and had a degree in linguistics from a Canadian university. But speaking a language and translating a Buddhist senior monk talk about meditation and philosophy are for me two different things : in one case, you think you know what you are talking about, while in the other you know you don’t.

But curiosity and temptation were stronger than I thought, so I finally went up to meet Gueshé la in his room and find out more about the subject.

After hearing all my excuses about my incompetence, Gueshé La just smiled at me and said : ” Oh don’t be so shy just say the same thing in your own language ! ”

Saying the same thing in my own language ! That was exactly what I thought I could not do, as my knowledge of the thing itself was rather a non-thing and definitely not functional.

But as you cannot resist a wise and compassionate person, a few days later, after some more encouragement by Gueshé la, convincing me that there was not any body else around who could do it, I was sitting on the hot cushion, scared as a newborn lamb, trying to convey as faithfully as I could, that is almost word by word, whatever Gueshé la was saying. Sorry, whatever the English translator was saying, as I did not know Tibetan then. This was my first experience of translating Dharma : translating a Tibetan translator translating the words of a Tibetan scholar speaking about a subject I knew nothing about. This is how Dharma teachings were introduced in France when at this time when there were no direct Tibetan-French translators available. Taking any one who came close to accomplishing the function of a merely labeled translator. In that case ME.

Everybody knows the famous expression (traduttore, traditore) : that interpreters are traitors.  And in that case we were two traitors. Although some might argue that two traitors are probably better than just one, as betraying the traitor could be one step closer to truth !?! Anyway, we both joined our efforts as best we could, trying to translate every word like a dictionary would. Isn’t a dictionary the best tool for translating ? This is when I proudly started to consider myself as being just a tool at the service of Dharma and others. A Dharma translating machine so to speak.

In those days my only strategy for translating was : as you can’t rely on your understanding, just rely on the words. What else can you do ? But you inevitably fall into the sufferings of the linguistic lower realms, as you are not translating the content (the meaning), but the language itself, (the substance), just like a dictionary, putting more or less one word at the place of another. But without a proper understanding of the thing that is being talked about, how can you even find the right words to translate. From the faces of the listeners in the audience, I could tell that they were trying to figure out what language I spoke. No wonder why many past linguists have argued that translation was an impossible dream :  two different linguistic systems will never match completely. As it is commonly accepted that there are no complete synonyms in language, there is no perfect synonymy between the words of two different languages, so equivalence in meaning cannot be provided by synonymy and a translation is not merely a linguistic operation.

Somehow I had to face the fact that a translator is not translating languages but texts, messages, in order to communicate what this text or message means, and make it meaningful to a recipient. For communication to talke place, the translator has to understand the text in order to translate it and the reader/ listener has to understand the translation. Otherwise what is the point of translating.

To use a famous example, just, suppose you were asked to translate in another languge an idiomatic expression like “it’s raining cats and dogs”. You would not think that by translating it literally you would be saying the same thing as when an English person normally says it, and that the listener would understand what you are talking about ? Saying so, you would just be saying non sense. Faithfulness to the words, therefore, is not enough. In order to translate the message, you need to interpret it. Isn’t it what an interpreter is supposed to do : to interpret, that is to give meaning to something in order to clarify it ? Although some people might say that interpreting is just making a personal comment instead of translating faithfully.

Translating literally usually does not make sense but sometimes it does. Suppose somebody wanted you to say that he feels threatened by cats and dogs every time it rains a lot, because he is obsessed by English idiomatic expressions, you, the interpreter would have to translate it literally, as this is what the speaker means. Therefore a translation faithful to the meaning depends on how the translator interprets the speaker’s intention and does not imply that one should never or always translate literally. A literal translation may be a right or a wrong strategy depending on the translator’s interpretation.

This confirms what the linguist Saussure said, that words are arbitrary and become fixed by conventions. So, I figured that if I really wanted to translate Dharman the first thing to do was to learn the conventions through which it operates. So I studied with Gueshé la for a few years, learning about the conventional thing through the conventional Tibetan words, considering that if Dharma is in the words, by translating all the words, and nothing else but these words, I will be faithful to meaning and preserve the Dharma.

Although I was reluctant to make any changes, for fear of being a traitor and that Dharma became corrupted, I had to face the fact that you cannot translate a language and still want to keep it. This is not the language that has to be kept but what the speaker means. In order to translate, one must transform the message so that it keeps the same meaning and produces the same effect on another recipient and in another language.

You probably know that in early latin, translatio meant “change” and was related to money – changing money – (when you change money you know you are not going to get the same currency, but you hope to get more or less the same value, even though you don’t) and traducere meant to transport or transfer this money. In this transportation, changes happen.

“When we translate, says Jay Garfield, we transform in many different ways : we replace terms and phrases with particular sets of resonances in their source language, with terms and phrases with very different resonances in the target language – we disambiguate ambiguous terms, and introduce new ambiguities – we interpret, or fix particular interpretations of texts in virtue of the use of theoretically loaded expressions in our target language – we take a text that is to some extent esoteric and render it exoteric simply by freeing the target language reader to approach the text without a teacher –  we shift the context in which a text is read and used.  No text survives this transformation unscathed.”

Granted that all these transformations take place, still, does the meaning change ?

So, as many linguistic and non linguistic changes happen in the translation process how could there not be a change in the language of Dharma ? And how could Dharma not be affected by changes, since the written Dharma is only words.

Like Saussure, Sakya Pandita in his Gateway to learning recognizes that due to the arbitrariness of the words and their relying on convention, the comprehension of meaning and the attainment of any goal connected to the use of language arise in the mind not due to the essential characters of the words but due to the familiarity with conventional use of terminology. It is this relative nature of language that Sapan calls upon to justify why meaning does not change as it moves through the transforming processes of translation.

And since transformations in translation change only the linguistic conventions, rather than taking these transformations to show that this is a decay of Dharma, Sapan argues that linguistic meaning does not operate through the words themselves but through the speaker’s intention. As long as this intention is preserved, the linguistic changes are insignificant. So, since it is the speaker’s intention and not the words that determines meaning, certain losses in translation should be acceptable – as long as the speaker’s intention is preserved.

How can this intention be preserved ? In the XIIIth century, as there is no central authority in charge of translation any more in Tibet, interpretive difficulties are overcome in the same way as all Buddhist knowledges are preserved : through a lineage or a  community of properly trained scholars. Therefore, in order for terms to perform their intended function, it is imperative that an appropriate cultural environment is created. It is only within such a cultural environment that selected terms used in the translation of Dharma can acquire proper fields of meaning.

Therefore in order to give their full meaning and power, Tibetan translations had to be placed in the right hands. Otherwise the differences between Sanskrit originals and their Tibetan translations could be misleading for inexperienced readers.

So in  The Dharma’s Gatekeepers, Jonathan Gold says that in a way Sapan seems to be both worried and optimistic. Worried because he recognizes that shifts in phonetic qualities, grammatical relations and etymological implications are the inevitable result of the translating process from Skt to Tib. Although translators tried their best to produce faithful translations, either in the decoding process (comprehension) or the recoding process (recreation) they have made some errors. Translators have used different strategies –  some have added clarification into Sanskrit words instead of making a direct substitution, (like adding an extra word in order to clarify the meaning),

  • others have used compounded words instead of uncompounded ones in order to facilitate understanding,
  • others have sacrificed ease of comprehension in Tibetan in order to stay close to the SKT,
  • for ease of understanding, others have used honorific forms in accordance with proper Tibetan usage.

According to Sapan, the greatest treason to avoid is when, based on ignorance, the translator sacrifices the meaning in Sanskrit in order to make comprehension easier in the target language.

But in another way, Sapan is optimistic, as for him these recoding strategies are not problematic, and he would rather criticize translators when they make errors of interpretation (in the decoding phase).

In Dharma, faithfulness is related to authenticity, as the translator’s task is to provide the conditions for the preservation of the Buddhist teachings helped by a community of experts.

Interpreting the speaker’s intention thus plays a crucial part as a translation cannot remain obscure but must clarify. That may be the reason why a translation is sometimes less pungent than the original.

In a recent experience, the question of interpretation and choosing the proper recoding strategy became a practical issue.

Last year, I was asked to translate a public talk given by the young Kalu Rinpoché on karma (and a Chenrezi initiation). As I was familiar with the subject and had been translating Dharma for about 30 years, I felt quite confident and accepted without any reserve. Still, as I had not done any oral translation for quite a while, I just hoped that the same automatism would come back. Things started very nicely, in automatic mode. As usual, I was writing and saying all the words that Rinpoché said as faithfully as I could. But Rinpoché uses a lot of words and speaks quite fast. And as time passed, he talked longer and longer, so it was more and more difficult for me to write down every word he said. I suddenly felt as if I was drowning in an ocean of words. It reached a point when he talked for about 30 minutes without a pause, not letting one bit of space to fill in with my translation. When he finally stopped after noticing my desperate waving, I had written something like 25 pages both sides. Rinpoché turned to me and asked in a very practical manner  :”How long will it take you to say what I said ?” Roughly the same time, I answered/ “Then Just say whatever makes sense!”. As I looked at my paper, some words were there but they did not make much sense as they looked like hieroglyphs. I somehow could not read my handwriting anymore. Without the possibility of relying on the false security of a word translation,  I was prompted to do a “meaning translation”, and to jump immediately from one extreme to the other, from the safe ground of words to the empty space of my own interpretation. This was very scary to let go of the words, but I had to make that jump quickly, as hundreds of people were waiting for my “act of speech”. How did I make this jump ? I don’t know. Some psycho-linguists might say that I finally relied on the threefold process of a typical interpretive translation comprehension, deverbalization and reverbalization (or putting it simply, after understanding the message, letting go of the original phrasing of the source language and recreating it in the target language in order to communicate the message). But, basically, I was lucky to know about the subject and that it was somehow a very personal talk, so I had a lot of freedom of expression in order to recreate the message by relying on my interpretation. In that case, my strategy was exactly the opposite as it was with Guéshé-la: as you can’t rely on the words, just rely on  you understanding.

After I finished translating, he question that was on the listeners mind was : Did Rinpoché really say that ? In other words, was it a faithful translation ?

If you apply the strategy of a literal translation faithful to the words, to produce a free translation, are you really faithful to the meaning ? Not sure ! What are the conditions for a translation to be faithful to meaning ? According to a model developed by the ESIT in Paris, two basic equations must be there :

  1. The meaning as understood by the translator must be the same as what the speaker means. This is the purpose of the comprehension (or decoding) phase. In order to give meaning to a message, the translator has to rely on and interpret the speaker’s intention. In order to do this, the translator must have some linguistic competence (like knowing what such and such words mean) and some extra-linguistic knowledge (the translator must have the cognitive requirements necessary to recognize the different clues to really understand the meaning).
  2. The meaning understood by the reader/hearer of the translation must be the same as the meaning understood by  the recipient in the source language. This is the purpose of the recreating (recoding) phase which involves finding equivalent words, expressions, etc., in your own language. In order to be faithful to the meaning, the translator must also be faithful to the target language and to the recipient of the target language.

So, faithfulness to the meaning involves three types of allegiances , to the…

  1. Speaker’s intention. If the translation is not faithful to what the speaker means, it is not faithful to the meaning. This implies that the translator should have enough linguistic and non-linguistic knowledge to understand the message.
  2. Target language. If the translation contains language errors, it is not faithful to the meaning. The translator must accommodate to target linguistic conventions. This implies that the translator should have complete mastery of the target language and finds proper equivalents. If the expression and the words are strange (not only foreign), it will be difficult to understand
  3. Recipient of the translated text. If the translation is not clear to the recipient, it is not faithful to the meaning. This implies that, if necessary the translator might add extra information to adjust to the reader with a different background, so that the translation might be pleasant to hear or to read. Therefore, a translation is faithful to the meaning when it is meaningful to the reader and produces the same effect

In order to be faithful to the meaning, it is necessary to be faithful to all these three parameters. If you keep only one and betray the others, you are not faithful to meaning.

In order to be faithful to meaning you have to betray the words.

As translation means comprehension and recreation : you cannot recreate the message if you don’t understand what the speaker means. And the recipient cannot understand the translation if the translator cannot recreate the message.

So, translation fails when there is trouble in understanding and when there is zero communication.

Following this model, then neither a literal translation nor a free translation are real translations. Why ?  Because they are not faithful to the meaning.

As a literal translation is only focusing on the language of the original text, it is not faithful to the speaker’s intention (what he/she means to say), it is not faithful to the target language (as it is trying to keep the source language) and  it is not faithful to the recipient of the translation (as it does not try to make understanding easier). If the translator is enslaved by the words, he cannot faithful to meaning.

Although a free translation is interpretive (involving comprehension, deverbalization and reverbalization), it disturbs the balance of the above parameters : the translator is giving a free interpretation of the speaker’s intention, and is going beyond the limits of the choices imposed by the target language and its recipient. Although it is necessary to interpret the message in order to translate it, one has to stay between the limits of the meaning conveyed by this message. It is also necessary to find the proper equivalence in the target language and to fulfill the needs of the recipient who wants to understand what the speaker really said and receive the same effect the speaker wanted to produce. The three parameters impose some limits and constraints that should not be transgressed in order not to fall into the extreme of a free translation.

The translator is free in his relation with the speaker because he is recreating a new text, but he is also his servant because he must convey the same meaning and  the same effect. Between these two extremes of translation, there is a gap and this is precisely within this gap that a faithful interpretive translation can operate. As long as the translator pursues the same goal as the author and preserves the latter’s intention, the risks of a free interpretation are minimized

In brief, there are three parameters to take into consideration in order to be faithful to meaning, but there is only one mental process to translate meaning : comprehension, deverbalization and reverbalization. I think that these findings of translatology are helpful to translate and to explain why sometimes translation fails. They can also apply to all the different types of translation, (literary or technical, written or oral), and to translation in any language. The notion of deverbalization is quite important as it creates a link between comprehension and recreation (either partial or total ) and it frees the translator from the straightjacket of the original expression used in the source text,. It is somehow easier to deverbalize when languages are far apart and more difficult when the two languages are closely related. Concretely this phase of deverbalization is what happens when one forces oneself to take one’s nose out the text, or to close the book, in order to translate in one’s own words.

One can say that, paradoxically, it is when one gets free from the source language that one is the most faithful to meaning.

The space of fidelity to meaning

As it is important to chose the appropriate strategy to reach the desired goal,  the act of translation requires competence and method. . Therefore, according to the translators’ skills and methods there should be different ways of being faithful to the meaning.

That subjectivity plays a very important part in translation is undeniable, as you can tell when you read different translations of the original text and notice their differences. These differences are simply due to the fact that the translations have been written by different individuals, with different linguistic skills and different extra-linguistic knowledge, and also who have different was of being faithful to the original. Different translating methods and strategies also bring different results. Some translators will be more faithful to the language itself (its words, its syntax) while others will be more faithful to the overall meaning and their understanding of the whole text. But an excess of subjectivity can also alter the interpretation of the original and too much freedom can lead to mistakes even if the translator understands what the author means. Therefore, it is important not to go beyond certain limits

So, there are many ways of being faithful to a text depending on how the translator concretely applies the three parameters and depending on how he/she associates his/her linguistic and non-linguistic knowledge. This also shows that faithfulness is more a question of degrees than of nature.

As each historical period produces its own translation of past texts, the concrete applications of these three parameters will of course vary as the translator is submitted to different types of constraints – linguistic and non- linguistic, depending on time and environment factors, the target language and the type of translation (oral, written, simultaneous, consecutive).

The application of faithfulness to the meaning is also related to time and to the history of the contacts between the 2 languages involved. Each new translation bridges a gap between two languages and cultures and paves the way for more communication and comprehension.

To some extent, one can say that there is not only one type of interpretation possible but a space opened to different types of interpretation

To quote Umberto Eco about the interpretation of meaning : “Once the message or the text is separated from the speaker’s intention and from the concrete circumstances of its utterance, it somehow floats around in the wide open space of an infinite number of interpretations (as limitless as the possibilities of human thought). Therefore no text can be interpreted according to a unique and utopist authorized meaning”

Subjectivity is essential in the translation of poetry

A poetical text is different from another genre as in a poetical text, the author decides to focus essentially on the text itself and not to present the information in a direct and transparent manner. The author then uses or creates special effects in order to move or inspire the reader, with a particular style, with ornaments and expressive power. From the language they use, these texts carry different semantic fields from ordinary language. How is it possible to create the same semantic and stylistic richness of the source text in another language ? If the translator wants to be faithful to meaning, he should be faithful to whatever conveys meaning, including the substance (the form) of the poem. If we let go of the rimes and meters that the author has deliberately chosen to use to express what he means, are we really faithful to meaning ?

How can one be faithful to form when form is part of meaning ?

Some people have argued against the translatability of poetry as a real translation of a poem should be another poem. Then translation cannot be literal but must be a complete recreation. The translator must try to see things the way the author saw them and render the whole process of the poem.The translator needs to recreate a new poem that is faithful to the meaning and the effect but not to the means of the original poem. In order to reproduce the same effect in the target language reader, the translator must take distances from the poetical means of the source language but must be faithful to the inspiring force of the poem and to the esthetic principles he/she has chosen to follow.

In a poem/song, the words are just a starting point, the translator must try to see things the way the author saw them and render the whole process of the poem. The meaning of the poem is the sum of all the elements it is made of. So, despite the difficulties, the translator should not sacrifice anything, he should be faithful to all the elements, although compensations have to be made.

In order to convey the spirit of the text, the translator, here again, has to betray the words and recreate the whole poem. The challenge of faithfulness is then how to reveal the same content with another linguistic substance, by using or creating different means in his own language. In that sense, it is really a creative recreation. Two strategies give the most successful results  :

  1. either the translator imitates the original without recreating it. The poem does not read as a translation but as if it was an original creation so that the reader who doesn’t understand the Source Language and has never read the author in its original tongue may experience the same pleasure as the SL reader
  2. Or the translator recreates the whole (substance and content) while keeping the structure of the original (using rhymes, meters, rythm).  This is regarded as the actual way of translating poetry

A poem being like a living body where each element plays a vital role, it is difficult to be faithful to all of them. This was my main problem, when I translated the songs of the Kagyu Gurtso. A faithful translation of a song should produce the same effect but with different means : In this case, achieving equivalence and faithfulness meant taking the songs out of their original metric husk and inject a free flowing rhythm. This is how I felt these songs would appeal both to Dharma readers in general and particularly to mahamudra practitioners. They should not be enriched either to make them more palatable to the target audience The less one manipulates them, the better. Since these vajra verses can be regarded as direct oral instructions that are imparted simultaneously to reading them, will it still carry the same enlightening potential if the translator manipulates it ? Instead of enriching it, will it not destroy it ? One must pay attention to the connotative value of the words and not introduce new connotations. The language of the translator should not introduce effects that were not in the original text.

I finally chose to make an informative translation, without any artistic claim, rather than a complete recreation using meters and rimes. Faithfulness was therefore limited to rendering the information conveyed by the songs. And although I tried to use alliterations and to focus on some of the expressions of the content but not all, the overall translation did not reflect any deliberate aesthetic or poetic pattern. Therefore, I do not consider the translation of these songs as a real poetic translation, but, let’s say as a poetic paraphrase, so I would not claim that they produce the same effect on the target reader than on the original songs. So it’s not complete faithfulness.

Chögyam Trungpa Rinpoché said that he wept and laughed while reading Milarepa’s songs. Reciting these vajra songs is a way to connect oneself to the Kagyu lineage, through the life examples of our forefathers. “These songs should be regarded as the best of butter which has been churned from the ocean of milk of the Buddha’s teachings” …”these songs should not be regarded as ordinary poetry, as a purely literary endeavor. They are the insight of our forefathers, conceived, described and proclaimed… Even reading one passage is better than going to a psychiatrist or taking a dose of aspirin. They provide a staircase to liberation… Many of these songs act to clear obstacles and generate exertion in practice… The lineage songs are genuine and precise, and due to these, they are powerful and helpful.”


In a wide open space of a large number of interpretations, the translator has to fill a gap between truthfulness and treason, negotiating each solution step by step, according to the goal he/she wants to reach.

While insisting on faithfulness to the meaning, I do not imply that a good translation must necessarily be faithful to the meaning. It depends essentially on the translator’s goal. Both a literal translation or an adaptation can be of very good quality even though they are not faithful to the meaning. To assess the quality of a translation, one should take into account 1) the means (the method), 2) the result achieved and 3) the goal. Still, I would argue that the assessment of a faithful translation does not depend so much on the opinion of a bilingual assessor but on the reaction of a monolingual reader when he/she reads the translated text.

As a perfect translation is an impossible dream – saying exactly the same thing in another way is not possible – a translation is faithful when it says almost the same thing (as Umberto Eco would say). Indeed, this small but crucial adverb – almost – contains all the questions, the doubts, the polemics and the suspicions related to this issue. Because of this adverb, there have been different types of approach in translating literary texts. Because of this adverb some texts are considered untranslatable and because of this adverb, a translation is never completely finished.