Comprehension of passive sentences with novel verbs by 25- and 42-month-olds

The acquisition of passive sentence structure has a long history of debate. Early studies using act out methods suggested that children do not understand reversible passives until around 4 years and that preschoolers pass through a stage in which they interpret the first noun phrase (NP) as the agent. However, studies using syntactic priming suggest that three-year-olds have verb-general representations of passive structure. We used forced choice pointing (Study 1) and preferential looking techniques (Study 2) to investigate at which age children show verb-general comprehension of passive structure. Since it is possible that when processing passive sentences children have to overcome a first-NP-as-agent bias, we combined preferential looking with eye-tracking measures in order to understand how children’s comprehension of passive sentences unfolds over time. We adapted a paradigm used by Gertner et al. (2006) whereby participants simultaneously saw two video clips of novel causative events, both involving a boy and a girl, whereby in one clip the boy was the agent and in the other he was the patient. Both Studies 1 and 2 compared 25-month-olds and 41-month-olds in between-subjects sentence structure conditions (Active Transitive vs. Passive). Study 1 found that 3-year-olds pointed above chance for both structures. For Study 2, we examined 800 msec windows that were time-locked to both first and second NP taking into account mean lexical processing speed for each age group. In the first NP region, both age groups showed a bias to map the first NP onto the agent, both for active and passive sentences. Both age groups showed evidence of differentiating the active from the passive sentence structure after the onset of the second NP. Thus, even as early as 25 months, children differentiate different types of two-NP sentences and show signs of an emerging ability to map these incrementally onto semantic roles.