Abstract
This paper reports on a study which investigated patterns of collaborative metacognitive activity in senior secondary school classrooms. Although peers working together on mathematical tasks may enjoy the metacognitive benefits of being able to monitor and regulate each other's thinking, collaboration does not guarantee that they will achieve a mathematically productive outcome. Analysis of a videotaped lesson transcript illustrates how metacognitive uncertainty, itself a trigger for collaboration, remained unresolved when students did not have the means of validating their solution.