Embedded NP error in German object relative clause comprehension: A case for a universal developmental pathway

We report the results of an eye-tracking study investigating German children's comprehension of subject relative clauses (SRC) and object relative clauses (ORC) with morphologically unambiguous head and embedded noun phrases (NPs). The experimental paradigm was adopted from Adani and Fritzsche. Children's eye movements were tracked on the visual display while they were listening to an SRC or an ORC. Subsequently, they had to choose the most appropriate visual character on the screen to go with a particular relative clause type in a character selection paradigm. All the head NP and the embedded NPs were of masculine gender. Thus, the relative clause syntax was disambiguated by the ending on the relative pronoun and, subsequently, on the determiner in the embedded NP. We computed fixation probabilities towards the syntactic competitor and the embedded NP character in addition to the proportions of looks towards the target character on the screen. Thematic reversal error remained the dominant error type on children's response accuracy data. The subject advantage was also confirmed on the eye-tracking data, though it was overridden in the post-relative clause time window. However, there was a significant increase in fixation probabilities towards the embedded NP character in the ORC, but not in the SRC condition. While children were less efficient to use the morphological information on the relative pronoun to generate an expectation of a non-canonical ORC structure, they obviously used embedded NP morphology later in the sentence to update their ongoing structural analysis.