People willingly accept personal costs to sanction norm violations even if they are not personally affected by the wrongdoing and even if their sanctioning yields no immediate benefits—a behavior known as third-party punishment. A notable body of literature suggests that this behavior is primarily driven by retribution (i.e., evening out the harm caused), rather than by the utilitarian motives of special prevention (i.e., preventing recidivism), or general prevention (i.e., preventing imitation). This has led to the conclusion that laypeople are “retributivists” in general. More recent evidence, however, raises doubts about the ubiquity of retributivism, showing that punishment is driven by multiple motives. The present research adds to this debate by investigating the motives underlying punishment in children around age 10. Specifically, we investigate children’s (N = 238) punishment motives in an economic game paradigm, isolating punishment motives by experimentally manipulating the extent to which the offender and a bystander learn about the punishment. This offers the possibility to examine whether (and to what extent) children engage in punishment even when it is devoid of any preventive effects. Results show that children’s punishment is motivated by retributive, special preventive, and general preventive purposes. These results point to a clear need for further theory specification on the motivational basis of punishment in humans and provide practical implications for the treatment of child misbehavior.