notes on machine translation
Does my use of machine translation mean that the use is representational? What are you using the technologies of machine translation for? Personally i am translating the Textbucher 1–6 of Helmet Heisenbuttel. The Heisenbuttel translation experiment has been in my view remarkably successful. It is inevitable that it had to be, would be, because I was interested in translating texts that only about fifty people in the world would be interested in reading. I am not interested in reading the original texts in German either, as reading them in English interests me more. It is the act of translation of thought that interests me in this estrangement of language practice…. A forgone conclusion you might say… The machine translation technology has freed us from having to learn to speak and read in multiple languages. We are not as dependent on the skills of human professionals, at whatever level they can translate at. We can certainly learn to speak and read the important everyday things, order coffee, talk with shop assistants, without having to learn to read German High Modernism, which after all is better done in English, an experiment, it is after all an experimental prose. I started by translating enough material from the German and comparing it with Micheal Hamburger’s original 1970s translations, compared and lived with the experimental results and am reasonably confident that the machine translation results work. Though what “work” means is a difficult concept when comparing machine translation to nothing. Are they comparable to the original translations? Is the experiment valuable? are they the machine translations the same (quality) as the Hamburger translations? Questions of value seem unnecessary as the question of what the untranslated text translated by the machine translator reads like, has no meaning.
I realize that i am breaking the laws of copyright, (buying a new German language paperback, scanning, machine translation… ) but since as far as I can imagine there are probably only fifty English language readers in the world interested in Heissenbuttels extreme modernism, i think it is inevitable that the effort will produce intriguing results. The Heissenbuttel machine translation experiment has been (in my view) successful, i am planning to do the same to his untranslated “Uber Benjamin” and the schizo-analytically named “Oedipus complex in Germany.” Is the machine less trustworthy than non-existent human translators? Or is it that it’s as good but different than a person translating the text?
a personal remark then: I was never afraid of Helmut Heissenbüttel . What I liked and admired about him, because of his theoretical as well as practical-poetic statements, was the “frightening” level of the kind of intelligence that impressed me when I was younger; the ability to understand (artistic and scientific) modernity, including its ideas, programs , repetitions, differences, surveying developments and drawing conclusions, summing up and prognosticating, unorthodox for literature as a whole, strictly conceptual for one’s own writing. Such thinking, such thought would welcome the difference that machine translation enables…
