This article deals with the phenomenon of junction, a cohesive device for signalling interclausal or inter-sentential relations, and its translation. The predominant finding of recent French-English contrastive studies on the topic of junction has been that, whereas there is a trend to junction-less juxtaposition in French, explicit co-ordination is preferred in English. Doubts concerning the universal validity of such a norm constitute the motivation for this study, which aims to consider: (1) the status of translator behaviour as evidence of norms of language behaviour; (2) the status of contrastive linguistics within translation studies. Examples of translations of writings by Albert Camus are then discussed in an attempt to show that translators’ decisions are sensitive to a number of contextual factors including genre, discourse and text type. The conclusions lead the author to suggest some limitations on the use of quantitative studies within translation studies, including those based on analysis of machine-readable corpora.