Abstract
With reference to two recent doctoral research projects on ELF, the present article examines the characterisation of language attitudes as either stable or variable evaluative phenomena, and provides a detailed account of methodological practices that may be favoured from each ontological position. The durability of language attitudes is more specifically conceptualised as a stable (but not enduring) construct directed to a linguistic phenomenon in one thesis, and as variable and emergent forms of evaluative social practice around a language-related issue in the other. With these two different approaches in conversation, the authors consider the extent to which stability and variability of language attitudes may be two sides of the same coin, and question whether it is safe to assume a priori the inferability of stable language attitudes from the observation of evaluative practice. This article evidences the need for ELF researchers working in this area to contemplate what and how it is being researched in the name of language attitudes while having awareness of possible alternatives in any given study.