Linguistic typologists (e.g. Ramat 1987, Seiler 1995) tend to assume that there is a close connection between translatability and language universals (research) but this assumption has recently been called into question by Comrie (1986, 1989, 1998), who claims that such a connection should be ignored in the practical enterprise of doing linguistics or will have a debilitating effect on the progress of linguistics as a discipline. This paper will demonstrate that translation is not only possible in principle but also has much relevance to language universals (research). First, to argue that translation is, in principle, not possible is to deny the basic relation between cognitive-conceptual content and language, mediated by translation. Moreover, the question of translatability is no longer looked upon as something either possible or impossible but translation is recognized as a process in which optimal solutions to translation are sought. Comrie’s argument for the impossibility of translation is based on formal equivalence, a concept which is highly debatable, particularly within Translation Studies. Second, translation has much relevance to language universals (research), because universal constraints (for example, on relative clause formation) are formulated or defined in terms of grammatical context, which will only come to light through translation. This does not come as a surprise because the mediation between language and cognition is also done through translation. The role of translation in language universals (research) is, pace Comrie (1986, 1989, 1998), much more than a mere heuristic. Translation is crucial to language universals (research), as recognized by Ramat (1987) and Seiler (1995).