Abstract
This article presents a study that was initially designed to find evidence of standardisation in translated language, but which revealed a completely different picture from the one expected. Part of the Translational English Corpus and part of the British National Corpus were used to make up a comparable corpus consisting of eight million words of translated and non-translated narrative writing in English. Instances of split infinitives were retrieved from each component (translational and non-translational). The hypothesis was that this kind of structure would be less common in the translational corpora but the total numbers of split infinitives retrieved from the translational and non-translational components were very similar and thus contradicted the initial hypothesis. However, I will show that if we look at them in more detail we can propose new data-driven hypotheses that offer a possible and logical explanation for the results in the light of Baker's theory of translator's style.