This distinction, central to the Bryfonski-McKay controversy, separates two fundamentally different ways of using tasks in language teaching.
Tasks are the unit of syllabus design. The curriculum is organised around target tasks derived from a needs analysis, not around linguistic items. Explicit grammar instruction happens reactively (focus on form), when communication breaks down or when learners are developmentally ready.
Key features:
Tasks are added to a traditional syllabus. The curriculum still lists grammar points, vocabulary, and functions to be taught explicitly. Tasks serve as practice activities: opportunities to use the pre-taught language. The syllabus remains synthetic.
Key features:
If a meta-analysis includes both types under the label "TBLT," it conflates two approaches with different theoretical commitments and different instructional realities:
This was a core problem in Bryfonski & McKay (2019): most included studies were task-supported, but the meta-analysis was framed as evidence for "TBLT," allowing proponents of either approach to claim the findings support their position.
Jason Anderson advocates for task-supported approaches. He uses the Bryfonski-McKay findings selectively: dismissing them as evidence for strong TBLT while claiming the same data supports task-supported teaching. The conflation of task-based and task-supported studies in the original meta-analysis is precisely what makes this double move possible.
The 2021 paper and the 2023 follow-up both argue that future meta-analyses must clearly distinguish task-based from task-supported implementations. Without this, the literature will continue to produce aggregate findings that both sides of the debate can cherry-pick.