This thesis is a corpus based contrastive study concerning the translation of swearing and other taboo words in movie subtitles. The material consists of English and Norwegian subtitles from 15 different movies that were aligned and annotated to form a parallel corpus. The primary aim of the study is to observe and quantify syntactic, semantic (denotative and connotative) and functional discrepancy between taboo words in English movies and their corresponding translations in the Norwegian subtitles and to account for possible linguistic and non-linguistic reasons for the observed discrepancy. In the process of this, inherent differences between English and Norwegian swearing will also be described to some extent. Furthermore, a fair amount of time will be devoted to providing an accurate definition and connotative typology of obscenity – an umbrella term for the type of language subject to analysis in the study. 700 correspondences were categorized and analyzed in the study. The findings were manifold, but in short, syntax, denotative meaning, connotative themes and connotative strength are very often different in the Norwegian subtitles, while function is typically preserved. The observed discrepancies can often be explained as products of various inherent grammatical and semantic differences between the languages or time and space constraints of subtitling. There are strong indications of sanitation, i.e. an apparently deliberate act of choosing conspicuously milder words in the translation. 30% of the English keywords were not translated in the Norwegian subtitles (zero-correspondence).