In linguistic research, opinions vary on how language variation functions and how it is structured. One vision emphasises that free variation does not exist, that language variation is structured along parameters such as speech setting or the social profile of the language user, making it possible to distinguish linguistic systems such as standard languages or sociolects. In another vision however, it is argued that there is less structure in language variation than traditionally assumed, stressing the agency each language user has to creatively combine language variants while constructing a unique identity. According to this view, linguistic systems are mere theoretical constructs not observable in everyday language use.

In this project, the ontological status of linguistic systems is studied empirically from a usage-based perspective. Statistical analyses are performed to verify whether certain groups of language features covary systematically in the light of varying social and situational parameters. I compare two areas – Suriname and Flanders – to test the validity of theoretical models characterising language communities on the basis of the presence or absence of systematic covariance (also named coherence). The project is innovative in that it proposes a systematic empirical method to test coherence – taking both production and perception into account – and as such allows optimising existing theoretical models on coherence in language variation.