Abstract
 

Humans seem to be especially capable of forming flexible temporal functional systems with others. Mechanisms for imitation, synchronization and alignment are undoubtfully important for this, however even more important is the ability to generate actions that extend and complement the actions of conspecifics. In our work we trace the developmental origins of this capability, i.e. ask how the functional, task-specific structure may "sneak-into" early caretaker-infant interactions? Is a high-level 'theory of mind' a prerequisite? In our analyses of early mother-infant interactions we show that part of the preparation for purposeful co-action may consist in educating child's perception and attention through participation in culturally structured episodes. This participation tunes them to perceive others' actions as affordances for intentional interactions.