Peter Robinson's Cognition Hypothesis (Robinson 2001, 2003, 2005) is a theory of how task design influences second language production and development. Its central claim is that increasing the cognitive demands of pedagogic tasks along certain dimensions does not overload learners, as competing models predict, but instead pushes them toward more accurate and more grammatically complex output and toward deeper processing of input.
The hypothesis frames task design as the principal lever curriculum designers have for sequencing instruction within task-based language teaching. Its core pedagogic recommendation is straightforward: tasks should be sequenced from cognitively simple to cognitively complex along defined dimensions, with linguistic complexity left to follow naturally from the cognitive demands the task imposes.
Robinson argues that complex tasks (those requiring learners to handle more elements, reason about non-present events, or take other perspectives) recruit greater attention to form and meaning simultaneously. The hypothesis predicts that increasing complexity along resource-directing dimensions will:
This cuts directly against the rival Trade-Off Hypothesis of Peter Skehan, which holds that limited attentional resources force trade-offs among complexity, accuracy, and fluency under demanding conditions. The Cognition Hypothesis instead claims that the right kind of complexity expands rather than constrains what learners can attend to.
The most consequential distinction in Robinson's framework, formalised in his Triadic Componential Framework, is between two kinds of complexity:
| Dimension | What it does | Predicted effect on output |
|---|---|---|
| Resource-directing (e.g. number of elements, here-and-now vs there-and-then, reasoning demands) | Forces attention onto specific functional and linguistic resources needed to encode the added complexity | Greater accuracy and grammatical complexity |
| Resource-dispersing (e.g. no planning time, dual-task conditions, no prior knowledge) | Spreads attention thin across competing demands without driving it toward specific forms | Reduced fluency and accuracy across the board |
Pedagogic sequencing should increase resource-directing demands gradually while keeping resource-dispersing demands manageable, then later remove planning time and other supports to push toward real-time use.
Robinson operationalised the Cognition Hypothesis as the SSARC model: tasks are sequenced through three stages.
This gives course designers a principled basis for ordering tasks within a unit or syllabus rather than relying on intuition.
Studies testing the Cognition Hypothesis show a mixed but recognisable pattern. Increasing resource-directing complexity often raises lexical and structural complexity in production, but accuracy gains are inconsistent and sometimes appear instead as losses in fluency. The Skehan vs Robinson debate has driven much of this research and remains unresolved: results often depend on which specific complexity variable is manipulated, the proficiency level of learners, and how output is measured.