Discovering impolite constructions in an under-investigated language

Abstract

This article describes a methodology for discovering potentially impolite language constructions in an under-resourced language, with the aim to develop a comprehensive digital database of such constructions. After a brief definitional discussion of (Afrikaans) impolite constructions, we succinctly introduce a multi-stage, multi-phase methodology to inductively find and verify such constructions in a systematic, documented way. The remainder of the article then focusses phase 1, where we apply a bottom-up approach based on qualitative corpus analysis, using incidental/convenience sampling. Specifically, we collect data through personal observations on social media, followed by detailed annotation and analyses of all instances. The seven steps of phase one are spelled out, before discussing detailed results for each of the seven steps. The article concludes with a prospectus regarding the other three phases, which include case studies, a literature review to determine saturation and redundancy, and verification and consolidation of a final candidate list of constructions for inclusion in a digital database.