Skip to main navigation Skip to search Skip to main content

A corpus-based exploration of impoliteness formulae: methodological challenges

Activity: Talk or presentation › Science to science

Description

Through corpus-based methods, this paper explores different impoliteness structures in the Spoken British National Corpus 2014 (Love et al., 2017), a 11.5-million-word corpus of everyday spoken British English conversation which has been tagged with a range of social metadata. Previously, most corpus approaches to impoliteness have focussed on single words, such as swear words or taboo words (e.g., Jucker and Landert, 2023).

The present paper builds upon previous work by van Dorst, Gillings and Culpeper (2024), who examined different conventionalised politeness formulae and their variation according to speakers’ social background based on politeness types (solidarity, tentativeness and deference) and different levels of formality (formal or informal). The structures examined in this present paper are a subset of Culpeper’s (2010) conventionalised impoliteness formulae. Our selection is driven by the extent to which the structure is searchable via corpus queries, the extent to which it occurs frequently enough to pattern, and the extent to which genuine cases of impoliteness can be found. This reduces the number of structures to just a few, including, for example, the YOU+NP structure, recently examined in Van Olmen, Andersson and Culpeper (2023). The conventionalised impoliteness formulae from Culpeper (2010) are all converted to their linguistic structure, through the use of CQP syntax, part-of-speech tags and semantic tags, to find out how conventionalised for impoliteness their base form are.

Through this iterative process, three main methodological challenges arise. The first challenge comes from the conventionalised impoliteness formulae that do not follow a set linguistic pattern. The lack of such a pattern means that it is ill-fitted for a corpus search. The second and third methodological challenges pertain themselves to precision and recall. The precision challenge occurs when a linguistic structure is too generic and has a high likelihood to be used outside of impolite contexts. The recall challenge results from the data. Corpora as a whole, and specifically the Spoken BNC2014, does not contain much impoliteness. The reason for this is that the types of people who are happy to be recorded for corpus compilation together are likely to get along with each other, and thus we would expect impoliteness rates to be lower. Therefore, for the impoliteness formulae that do form a precise corpus query, only very few cases occur in the data.

Despite these challenges, there a number of conventionalised impoliteness formulae that lend themselves well to corpus research and open up a new approach to impoliteness studies.
Period17 Jul 2024
Event titleIVACS 2024
Event typeConference
LocationCambridge, United KingdomShow on map
Degree of RecognitionInternational