In this article it is brought under the attention that there appears to be no adequate theory for translation. The problem of translation is theoretical in the strict sense, being a problem in and of theory, not just the right theory but first and foremost the right kind of theory, which turns out to be the only real kind. The author attempts to provide an answer to the questions about the nature and purpose of a theory for translation, discusses linguistics as an example for a theory of translation and looks into the theoretical value of machine translation.