This study examined the comprehension of actional and psychological verbs in both their active and passive (short and long) forms by 99 Serbian-speaking children. The children, whose age ranged between 3 years, 6 months (3;6) and 7 years, 6 months (7;6), were divided into three groups: 3;6–5 (M = 4.3), 5;1–6;1 (M = 5.6), and 6;2–7;6 (M = 7.0). All groups reached ceiling performance on actives of both actional and psychological verbs. They showed a good performance on passives of actional verbs, both short and long, but an exceptionally low performance on passives of psychological verbs: only the oldest group, 7-year-olds, reached 80% correct performance on psychological passives. There were no differences in the children's performance on short versus long passives of either type of the verb. These results are in line with studies reporting delayed passive comprehension in children younger than 5 in a number of languages, suggesting a delay in the comprehension of this structure cross-linguistically. The discrepancy observed in the comprehension of psychological passives, as opposed to actional actives, fits neatly into Borer and Wexler's account, which relates children's difficulties in this area to their underlying knowledge of argument chains created by movement of sentential elements.