An excerpt from Brian Murphy’s 2005 book The Root of Wild Madder, an exploration of the majestic, sometimes Delphic, and always offbeat world of Persian rugs:

The challenges of taking literature across languages are huge and thoroughly discussed. Poetry takes the chore to a higher level.

First, let me say that this statement can be generalized beyond its application to literature. It pertains, in fact, to all language, beyond that of prose or the written form. This means it applies just as well to spoken verse, which is, after all, contemporary poetry’s origin, though much popular poetry was originally based on spoken verse– think Sappho. Second, the statement seems intuitive, but is worth a little bit of discussion. I think a common statement of the Heisenberg Uncertainty Principle applies well:

The measurement of position necessarily disturbs a particle’s momentum, and vice versa.

Or, in terms of language:

The measurement of expression necessarily disturbs a statement’s meaning, and vice versa.

Scientific treatises are rarely translated because the specificity of the terms usually leaves very little ambiguity and only exist in one language (English). There’s no problem in measuring expression because the meaning usually speaks for itself. But words used in poetry are often far more ambiguous and we praise the form for its ability to prod, poke, provoke, and elicit from seemingly humble expressions. The more energy spent on translating, that is, measuring the expression, the more the actual meaning in fact changes. The closer we come to a precise measurement of electron positions in orbitals, the more the position is disturbed, meaning we will lose other types of information. We are trading one for another. In the translation of language, poetry being the hardest, the more precise measurement we obtain of the statement, binding words from one language to words of another, the more the new expression through translation necessarily differs from the other in meaning.

Of course, some translations of poetry are still wildly successful or evocative. But they are not and can never be the same, leaving alone for the time being the differences in interpretation caused by different neural structures arising from different language structures! As I mentioned in my post on Anne Carson’s book of Sappho’s poetry:

Carson uses words such as sweetbitter, honeyvoiced, mythweaver, songdelighting. These are not words that we really have in English, but their composition follows standard rules for word formation and seem to be quite intelligible. Translators should never shirk from creating new words in order to translate. We need some frame of reference to understand these terms, after all. And if these new words help us see things we already understood in new ways, like a metaphor might, then these truly expand our power of comprehension, opening our minds to possibilities that we had never before considered.

Adding new words in English is exactly what we do when there’s something new to be described that another language does better, or if the word has some sort of economy. I applaud translators capable of this feat, for word creation is the best weapon short of borrowing (which defeats the purpose of translation) for defeating the Heisenberg Uncertainty Principle in language.