The translations on this website are for practitioners, people engaged in the practice of Buddhism. I'm not translating for academics or linguists. When people use prayers or other texts in their practice, the prayers need to speak to them in their own language. Tibetan and English are very different languages in terms of grammar and the underlying concepts on which the grammar is based. Effective translations start with the acknowledgement of these differences.
Lexical translation — trying to establish explicit correspondences between the words in one language and the words in the other — is hopeless. Not only can you not map words across; the way of thinking is different, the grammatical structures are different, the way sentences are structured are different. You can't even put an ordinary Tibetan sentence into English because a Tibetan sentence just doesn't work the same way as an English sentence. Translated texts need to go into the reader's own language, so that when someone picks up the translation, he or she is reading actual English, not another language in English words.
Time and again translators try to assign one specific meaning to a Tibetan word, phrase or sentence. This reductionist approach actually limits what can be expressed. When a translation allows a range of possible interpretations, it may seem less clear, but the range allows the text to retain its richness. One person will read it and take it to mean this, another person takes it to mean that. They're both right. Allowing some ambiguity in the translation enriches the result.
A word in Tibetan will typically map to a range of words in English, and vice versa. They overlap only to an extent: there are meanings in English that are not included in the Tibetan, and meanings in the Tibetan that are not included in the English. By trying to link this word with that, you limit the meaning only to the place where the English and Tibetan overlap, and lose the richness.
It is sometimes necessary to import terms from the source language into the target language, but as soon as you do that you limit the accessibility of the translation. Consequently I avoid the use of Tibetan or Sanskrit words as much as possible.
A translator needs not only to understand Tibetan: they also have to be able to express what is being said in English. That is often much harder. When I first started translating in the 1970s a friend gave me a copy of Wittgenstein's Philosophical Investigations. This book showed me how to think about language and the use of words, so that I didn't create what Wittgenstein called "pseudo-problems" — problems that seem to be philosophical but arise only because words are misused. This can happen very easily when words aren't being used as they are in everyday language. I've incorporated this idea into how I translate: I work to make my translations read naturally.
Translations should be in English, not just anglicized Tibetan or Sanskrit, so that when you pick it up you feel like you're reading English. Natural English has a certain sound and rhythm that need to be there. You can tell whether it is natural English when you read it out loud.
English, more than many languages, has several roots. The power of English lies in its Anglo-Saxon roots. With the Normandy invasion in 1066, French became the language of the intelligentsia, e.g., words ending in -ion or -ate, such as liberation, transmission, educate or translate. These words express ideas precisely, but don't have the power or earthiness of the Anglo-Saxon roots.
The Tibetan script was originally developed for the sole purpose of translating Indian Buddhist texts into Tibetan. The style that developed is concise and powerful, a very few words in scholastic Tibetan can say a lot.
When you translate this into English, the temptation is to use very formal English academic language. But in English, academic language has no life: it doesn't move us. It is primarily intellectual; it is not passionate or emotional. Attempts to convey the sense of dignity and respect carried by the Tibetan formal expressions sound insincere and stilted. So you get translations that are very intellectual; they just sound like a bunch of words, a bunch of concepts.
As a traditional mythic culture, everything in Tibet had a mythic component. When you grow up in that culture, you know the human and you relate with the mythic. They aren't really separate. But when a rational, modern person approaches a traditional mythic culture it is very difficult for them to sort out the human and the mythic. People either relate to the human and dismiss the myth, so we lose the energy and emotion of the mythic, or, as is more common in Tibetan Buddhist circles in the West, people take the mythic literally and ignore the human side.
To balance the mythic and the human in translation is tricky, as we have largely lost the mythic component to modernism.