Discussions of folklinguistic accounts of language use are frequently focused on dismissing them because of their limitations. As a result, not a lot is written regarding how such accounts are done and how they "work." This article examines how folklinguistic evaluations are achieved in interaction, particularly through appeals to semiotic registers (Agha 2007). It describes how in explaining their beliefs regarding linguistic variation, speakers frequently produce voicings with varying transparency. These rely on understandings of the social world and bring large collections of linguistic resources into play. They offer rich insights if analytic attention is given to their details because even when evaluating a single variant, whole ways of speaking, and even being, may be utilized. The paper explores in turn how analysis reveals the inseparability of variants, understandings of context and audience, the relationship between linguistic forms and social types, and the performance of social types via the evaluation of semiotic resources. In each section, discussion is grounded in extracts from interviews on Australian English with speakers of this variety of English. Cumulatively they show the primacy of semiotic registers in ethno-metapragmatic accounts. [enregisterment, voicings, identities, discourse analysis, Australian English]