A corpus-based exploration of British English impoliteness formulae

(Im)politeness is often said not to be inherent in linguistic forms (e.g., Van der Bom and Mills 2015). If that is true, impoliteness formulae should not exist. This study scrutinises the set of British English conventionalised impoliteness formulae described in Culpeper (2010, 2011). It does so by constructing powerful corpus queries to retrieve those formulae from the spoken component of the British National Corpus 2014. Those queries tap into the words, structures and semantics of the formulae. The success or otherwise of these queries is taken as a proxy for whether those formulae have clearly defined clusters of formal features. Such clusters are not, of course, evidence of impoliteness. Consequently, this study also establishes the degree to which instances of those formulae appear in impoliteness contexts. If they do, that would be evidence that they could become conventionalised for impoliteness. Given the subjective nature of impoliteness, the study uses inter-rater reliability techniques in order to secure robust judgements about whether the contexts of occurrence are really impolite. Five impoliteness formulae variants were established as having clearly defined formal features and strong associations with impoliteness contexts: personalised negative vocatives, one variant of personalised negative assertions, two variants of dismissals, and silencers. Taboo words did not seem to influence the association with impoliteness contexts, though our evidence was not strong. More abstract, more grammatical impoliteness formulae seemed to attract less impoliteness. The study also investigated banter and metalinguistic cases of impoliteness in relation to impoliteness formulae.