Principles of Translation: Lord Woodhouselee

In case you're not familiar with this name, Lord Woodhouselee, properly known as Alexander Fraser Tytler (1747-1813), is the author of Essay on the Principles of Translation, which despite being originally published in 1791, is still relevant today. He sets down principles for translation and gives guidelines for quality assessment that are strikingly modern.
Woodhouselee's three principles are as follows:
1. A translation should give a complete transcript of the ideas of the original work;
2. The style and manner of writing should be of the same character as that of the original;
3. A translation should have all the ease of original composition.
None of these would surprise a modern translator, yet they are nonetheless often touchy issues. So let's look at each a bit.
The first principle seems obvious enough. It means no errors or omissions, but also includes the idea that a translation is not a copy or a reproduction, but an original work in and of itself, even if its content is based entirely on another's work. The translator is not supposed to make additions, amendments, or annotations, except possibly when working on esoteric literary texts. The translator should not play the game of rewriting an original text so that it says "what the author meant" or "what the author should have said."
The second principle is similarly obvious, though difficult to execute in practice. It means that the translator must not only have a complete command of writing in the target language (the translator's native language, in almost all cases), but also must be able to perceive stylistic touches and understand their meaning in the original text. A simple example from the realm of patent translation should suffice: "means" is the term of choice in a U.S. patent when explaining how the invention operates; by custom, it takes no article, which in any other document would be grammatically and stylistically peculiar, but in a patent is what we do. A patent translator working into English must know this, otherwise the character of the translation will be flawed.
The third principle is the hardest to achieve, because it harks back to the Russian axiom that states that if a translation is beautiful, it is not faithful, and if it is faithful, it is not beautiful. There is a delicate balance to achieve here, in other words, and translators must aspire to be good writers in their native language, and must know all the finer points of writing in the subject and language they are working in so as to produce a translation with "all the ease of original composition." In other words, a translation should not sound like a translation.
Of course, these principles are open to debate, and in some obvious cases are deliberately ignored or even ridiculed. The Bible, for instance, would not sound at all pleasing if written in modern American English; rather the various modern versions based on the King James translation seem to be what people want, and anything else seems flawed or fake. Literary translation, particularly poetry, often cannot capture the manner of writing of the original, or as Robert Frost said: "poetry is what is lost in translation." How, after all, do you recreate the meter of the Roman poet Horace in a language like English (which is metrically very different), or in Japanese (which doesn't allow for meter as the Romans used it)?
Further, most translators at present are hired guns; they function on a "work for hire" basis and therefore have to please their clients, or if in-house, satisfy their team leaders and adhere to whatever corporate guidelines or protocols exist. Many times I've had clients tell me to ignore this or change that, or use a term I know to be incorrect or irregular, or insist of major formatting or stylistic changes. All of this might send Woodhouselee spinning in his grave, but he was writing about literary translation, and not thinking of modern freelance translators working with clients in different time zones, communicating by email, researching terminology and subject material knowledge on the Web, all while under very tight deadlines.
Woodhouselee has a lot to offer. The fact that his book is still available and worth reading is significant in and of itself. But as with all suggestions regarding how to translate, we must remain vigilant to the realities of our own particular situations, which can vary depending on the client, agency, or company, and even vary within a single document.