This paper explores how State Space Grids (SSG), a dynamic systems research method, can be used to map teacher-student interactions from moment-to-moment and thereby to incorporate temporal aspects of interaction. Interactions in two secondary school classrooms are described in terms of level of interpersonal control and affiliation, and of variability and recurrent stable states (called "attractors") of teacher-student interaction. Undesirable interactions appeared to feature weaker attractors and higher variability than more desirable interactions. SSGs appeared a promising way to study temporal aspects of classroom interaction, offering variability and attractors as indicators for the quality of classroom interaction. (Contains 3 figures and 1 table.)