Children rapidly acquire their first language in the early stages of their life. Not surprisingly, not all aspects of language are present from the beginning. For that reason, psycholinguistic research has focused on how children process language in comparison to adults. However, unlike research with adult participants which has identified a bi-directional relationship between spoken utterances and visual attention, research on children was often conducted offline (i.e., using offline measures only). With this thesis, we tried to extend existing findings on real-time child language comprehension which suggest pronounced adult-children differences, especially in the pragmatic domain.

In six eye-tracking studies (grouped into 3 Experiments), we investigated the influence of case-marking, prosody, and non-linguistic visual cues (depicted actions/a wiggling target character) on real-time thematic role assignment in children and directly compared the children’s behaviour to that of adults (Exp 1, Exp 3, but Exp 2 adults only). In addition to monitoring eye-movements, we recorded accuracies for active and passive voice post-sentence comprehension questions. Participants listened to unambiguously case-marked subject-verb-object (SVO) and object-verb-subject (OVS) sentences (Exp 1, Exp 2a), unambiguously and ambiguously case-marked OVS sentences (Exp 2b), or ambiguously case-marked OVS sentences (Exp 3). Sentences were assigned an SVO- or OVS-biasing prosodic contour (SVO: L*+H accent on the NP1 and H* accent on the verb; OVS: L+H* accent on the NP1; as proposed by Weber, Grice, & Crocker, 2006). As a baseline, we added a neutral prosodic contour (Exp 1), or directly contrasted the SVO- with the OVS-biasing prosodic contour (Exp 2). In Exp 3, prosody was kept constant across conditions (OVS-biasing prosodic contour). Visual scenes contained three clipart animal characters, two of which were depicted as performing identical actions (Exp 1, Exp 2). Thus, the depicted actions did not disambiguate role relations early in the sentence. In Exp 3, we manipulated the target animal role filler: i.e., the target was either depicted as a) it was (no-cue baseline), b) performing an action (action-cue), c) wiggling up and down (wiggle-cue), or performing an action and wiggling (action plus wiggle cue). Visual cues were limited to the lifetime of the verb.

The results (Exp 1, Exp 2) revealed no clear effects of prosody on real-time thematic role assignment in children and adults but we observed slightly more target inspections in a) the biasing prosody conditions (Exp 1: Adults and children) and b) in ambiguous OVS sentences with an SVO-biasing prosody (Exp 2b: Adults). However, adults rapidly used case-marking to anticipate the target role filler as early as the verb (the patient in SVO sentences and the agent in OVS sentences) even though actions were depicted ambiguously (i.e., they did not disambiguate role relations). By contrast, children did not use case-marking for early target anticipation. In both sentence structures, they anticipated the patient (vs. agent). Responses to post-sentences questions corroborated our eye-movement result: Adults’ responses were high independent of word order whereas children’s response accuracies were only high for SVO sentences but below chance level for OVS sentences.

Action and/plus wiggle cues elicited clear effects on target anticipation during verb and adverb in adults and children (peaking during the verb in adults and the adverb in children). However, adults’ offline responses were marginally higher in the action plus wiggle cue conditions (vs. action-cue, wiggle-cue, no-cue baseline) whereas children’s responses were higher for active (vs. passive) voice questions. Within the subset of active-only questions, the wiggle-cue elicited marginally more correct responses than in the other conditions.

Taken together, differences between children and adults emerged during thematic role assignment in unambiguously case-marked SVO and OVS sentences (online and offline) but target anticipation was not clearly influenced by prosody in either of these two age groups. Visual cues boosted attentional responses but elicited different accuracy results in children versus adults.