This paper uses the interactional spoken data contained in the Scottish Corpus of Texts & Speech (SCOTS) to investigate "friendly" language and shows how its principles can inform a model of language for learners of English as a second or foreign language. Pragmatic markers used in local speech varieties are in danger of being neglected in an educational environment that privileges a common core lingua franca. This paper makes the case for raising awareness of local speech varieties in English as a lingua franca (ELF) and EFL curricula and demonstrates how corpora such as SCOTS can be exploited in the classroom. The SCOTS corpus contains several features useful to educators: it is freely available online and may be browsed and searched using integrated tools, including a concordance facility and a map showing the geographic location of authors/participants. The spoken language drawn on here is available as video/audio recordings with orthographic transcriptions and extensive demographic and textual metadata.