Abstract

To what extent can swearwords be seen as carrying impolite meanings? Is their use generally – or perhaps even by definition – linked to impoliteness? These are important questions if swearwords are to be used in a corpus-based assessment of im/politeness levels in specific contexts. In a first-wave approach to politeness, linguistic expressions were taken to have relatively fixed politeness values. In a second-wave approach, im/politeness values were taken to be locally and discursively negotiated. A third-wave approach provides a rapprochement between the first and second. It assumes some default values that can be re-negotiated in specific contexts. But how can modifiable default values be used in a corpus-based investigation?

Corpus-based investigations generally rely on (relatively) fixed values in order to provide a basis for large-scale generalisations across entire corpora. They also tend to require time-consuming and laborious manual inspections of the retrieved hits in order to eliminated deviating instances. To solve this problem, Lutzky & Kehoe (2016, 2017) proposed a method of using contextual information in order to automatically sort through lists of swearwords and apology expressions, such as sorry, pardon or excuse, respectively.

In this presentation, I introduce a further development of this approach. Lutzky & Kehoe rely on collocates of their search terms, i.e. expressions that regularly occur in close vicinity to the search term. I propose a more global approach – a topic approach – which relies on the keywords that characterise those texts in which a specific search term tends to occur. Keywords are expressions that appear in a specific text with a frequency that is much higher than their frequency in the entire corpus. They are a good indication of the topic structure of the text in which they occur, and they provide not only a useful tool to disambiguate the search terms (as in Lutzky & Kehoe’s approach), but they also offer nuanced information on the im/politeness level of the search terms themselves, and they afford an empirical basis for the evaluation of the severity of certain swearwords. As a result, they help to uncover additional swearwords with similar distribution and usage patterns, and they turn swearwords into useful diagnostics in contrastive or diachronic investigations of varying im/politeness levels in specific contexts.

References

Lutzky, Ursula, and Andrew Kehoe. (2016) Your blog is (the) shit: A corpus linguistic approach to the identification of swearing in computer mediated communication. International Journal of Corpus Linguistics 21.1, 165-191.

Lutzky, Ursula, and Andrew Kehoe. (2017) “I apologise for my poor blogging”: Searching for apologies in the Birmingham Blog Corpus. Corpus Pragmatics 1, 37-56.