When we demonstrate children (or adults) a sequence of visibly causally irrelevant and relevant actions to reach a goal, they tend to imitate both, the irrelevant and the relevant actions - they 'overimitate'. Strikingly, apes don't do so. It is much debated what the underlying motivations for this human-specific apparently inefficient behavior could be. Do we overimitate because we draw erroneous causal conclusions from action observations or is it rather a strategy to affiliate with another person? Do we simply try to follow a social convention or a rule? In this talk we present a set of behavioral experiments, in which we contrasted explanatory models focusing on erroneous causal reasoning or social motivations; and present a new integrative explanatory model for overimitation. [mehr]