The phenomenon of retranslation (the repeated translation of a given work into a given target language) is widespread in practice, and yet its motivations remain relatively underexplored. One very prevalent justification for this repetitive act is encapsulated in the work of Antoine Berman who claims that an initial translation is necessarily 'aveugle et hésitante' (1990: 5), while retranslation alone can ensure 'la « révélation » dřune oeuvre étrangère dans son être propre à la culture réceptrice' (1995: 57). This dynamic from deficient initial translation to accomplished retranslation has been consolidated into the Retranslation Hypothesis, namely that 'later translations tend to be closer to the source text' (Chesterman, 2004: 8, my emphasis). In order to investigate the validity of the hypothesis, this thesis undertakes a case study of the British retranslations of Flaubert's Madame Bovary and Sand's La Mare au diable. A methodology is proposed which allows the key notion of closeness to be measured on both a linguistic and a cultural axis. Given Flaubert's famous insistence on 'le mot juste', Madame Bovary serves as a basis for an examination of linguistic closeness which is guided by narratology and stylistics, and underpinned by Halliday's (2004) Systemic Functional Grammar. On the other hand, Sand's ethnographical concerns facilitate a study of cultural closeness: here, narrativity (Baker, 2006) informs an analysis of how Berrichon cultural identity is mediated through retranslation. In both cases, the thesis draws on paratextual material (Genette, 1987) such as prefaces and advertisements, and on extra-textual material, namely journal articles and reviews, in order to locate specific socio-cultural influences on retranslation, as well as highlighting the type and extent of interactions between the retranslations themselves. Ultimately, this thesis argues that the Retranslation Hypothesis is untenable when confronted with the polymorphous behaviour of retranslation, both within and without the text.