This paper investigates whether normalization and shining-through effects in translations lead to a hybridization of target texts and whether this has an impact on the target language. The case studies presented in the paper deal with the quantification of these phenomena on the basis of a monolingual comparable corpus of translated and non-translated text and their explanation using a parallel corpus. By including diachronic texts, the relation between translation properties and language change can be investigated. Finally, psycholinguistic studies reveal how translated texts are received and accepted by the target audience. The product- and process-oriented methods used for these different analyses are chosen to complement each other in order to pave the way for empirically modeling the nature of translation.