Mechanisms of instantiation: towards an account of how language becomes text

Resumen

The paper interrogates the theoretical distinction between instantiation and stratification in Systemic Functional Linguistics by treating ‘text as instance’ as a probabilistic configuration – one that is related to, yet distinct from, ‘text as language in context’. While standard SFL accounts view instantiation primarily as the manifestation of system potential, we complement this view by posing that instantiation constitutes a discrete domain of constraint in which systemic resources are probabilistically redistributed. In order to do so a specialised corpus in Brazilian Portuguese was compiled across three genre families: Evaluating (academic research articles), Informing (educational booklets), and Engaging (oral narratives). The analysis reveals three instantiation mechanisms: redistribution, whereby register constraints reframe baseline system probabilities; associative keyness, which locates co-selection networks; and system dynamics, showing that instantiation unfolds as a conditional probabilistic patterning rather than a static index. The findings show that text typology is driven by localised patterns of systemic co-selection rather than mere frequency, and offer a replicable methodology for tracking how systemic constraint scales up into text types. The paper argues that instantiation and stratification must be maintained as complementary but non-reducible dimensions.