Information structure cues for 4-year-olds and adults: tracking eye movements to visually presented anaphoric referents

Abstract

Studies on young children's online comprehension of pronominal reference suggests that children follow similar syntactic, semantic and discourse constraints as adults. However, the observed effects are less stable and appear much later in the eye-movement record than in adults. It is not clear whether this is because children are cued by a different set of factors than adults; or whether children use the same set of constraints, like subjecthood or first mention, but the delay is caused by the developmental stage in which these cues are not yet fully acquired. We added an information structure cue (focus) and asked whether it affects syntactically more- or less-salient discourse referents (subjects/objects) the same way and shows a similar pattern in adults and children or whether it modulates the reliance on syntactic salience in children. Four-year-old German children and adults listened to stories with focused or unfocused syntactically prominent and non-prominent entities, subjects and objects, while we registered their eye movements to visually presented antecedents for ambiguous pronouns. Syntactic and information structural prominence interacted for children: focusing increased the looks to the syntactically salient subject antecedents, but not to the syntactically less salient object antecedents. This suggests that clefting helps children to locate the preferred antecedent. Adults' pronoun resolution in contrast was not modulated by clefting in a clear way. Instead, they showed an overall effect of syntactic prominence. Our study suggests that children and adults are sensitive to the same structural cues in reference resolution but that these constraints may not yet be fully acquired. The process may be enhanced, but not modified, with additional cues such as clefts.

Keywords:

• language comprehension
• visual world paradigm
• eye-tracking
• pronoun resolution
• child language
• language acquisition
• information structure
• focus
• cleft structures

Acknowledgements

This study was carried out while the first author was at the Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands. The authors would like to thank: Laura de Ruiter for her voice, Johanna Tewes and Alexa Weiss for their help in preparing and administering the experiments; Kindergartens Regenbogen, St. Lambertus, St. Willibrord and Zauberfarben in Kleve and St. Barbara in Kranenburg for participation in the study.

Notes

1. According to Dufter (2007), while often used in present-day German, it-clefts in German are much less frequent than their translational counterparts in Romance languages. While there is some recent evidence that adults process the semantic and pragmatic aspects of clefts online (Drenhaus, Zimmermann, & Vasishtsh, 2011), to our knowledge the full semantic interpretation of clefts has not yet been studied in children.

2. Only referents with the grammatical gender masculine were used, because masculine nouns are morphologically marked for (subject and object) case (nominative, accusative) and number (singular, plural) in German. Note, however, that the pronoun er ‘he’ was used ambiguously to refer to either of the animals in the sentences.

3. Previous research has shown that it takes approximately 200 ms to plan and execute a saccade (Matin, Shao, & Boff, 1993). Therefore, in order to allow for this we started the analyses 200 ms after the pronoun onset.

4. Analogous analyses for the adult data indeed showed significantly more looks to subject antecedents overall, but no other effects in all three 600-ms time segments. The models are available from the authors upon request.

5. Further analysis of the 2000–2600 ms time window shows a preference for subject referents (estimate = 0.2056, z=2.050, p=0.0430) but no effect of focus (estimate = 0.0191, z=0.229, p=0.819) showing an overall subject preference for children but one that arises considerably later than in adults.

6. See note 4 above.