Abstract
In Mandarin Chinese, wh-phrases remain in situ and do not undergo overt movement, allowing subject and object wh-argument questions to be compared without the confounding effects of wh-movement. Such questions, in which a wh-phrase functions as a predicate-licensed subject or object, provide an important window into how children use syntactic position to identify grammatical roles. For example, Shéi zài zhuī tiān’é? “Who is chasing the swan?” is a subject wh-question, whereas Tiān’é zài zhuī shéi? “Who is the swan chasing?” is an object wh-question. Previous findings have been inconsistent, reporting a subject advantage, an object advantage, or no reliable asymmetry. This thesis addresses this discrepancy by examining when subject-object asymmetry emerges in Mandarin-speaking children’s acquisition of wh-argument questions, how its direction and strength vary, and under what conditions it is attenuated.These inconsistencies suggest that asymmetry cannot be explained by argument position or structural complexity alone, but reflects dynamic multi-cue integration during sentence processing. Building on the Competition Model and the vulnerability of interface processing, this thesis proposes an Interface-Weighted Cue Integration (I-WCI) account. It argues that children integrate cues at the syntax-semantics-discourse interface - argument position, wh-type, animacy, discourse accessibility, and task demands - and that the relative weights and configurations of these cues shape processing load, thereby determining the direction and strength of asymmetry. The account predicts that asymmetry varies with cue configuration rather than forming a fixed subject or object advantage; that comprehension and production may follow partly dissociable developmental trajectories; and that children with autism spectrum disorder (ASD) may retain basic cue sensitivity while showing greater difficulty under high integration load.
To test these predictions, the thesis adopts a staged design moving from naturalistic corpus evidence to controlled cue manipulation, real-time processing, and group comparisons. Study 1 analyzes longitudinal and cross-sectional spontaneous-speech data from 15 typically developing (TD) children and 7 children with ASD to examine the everyday distribution of wh-argument questions. Study 2 tests 84 TD preschoolers using comprehension and elicited-production tasks manipulating argument position, wh-type, and animacy, establishing a TD baseline for how cue configurations shape asymmetry. Study 3 combines visual-world eye-tracking with elicited production to compare 120 TD preschoolers, 60 children with ASD, and 70 adults, examining real-time cue integration, group differences, and the roles of vocabulary, nonverbal IQ, and executive function. Data are analyzed using mixed-effects models, with eye movements further examined through time-course analyses.
Across the three studies, Mandarin-speaking children showed no uniform subject-object asymmetry. Object wh-questions showed the most consistent advantage when wh-type, animacy, and task demands jointly supported interpretation; under cue conflict or high integration load, this advantage weakened or reversed. Thus, difficulty lay not in argument position per se, but in coordinating argument-role assignment, referential identification, discourse-pragmatic constraints, and cognitive control. The I-WCI account explains both the task-dependent behavior of which-NP questions and the persistent comprehension-production gap. Eye-tracking revealed online differences not captured by final accuracy: adults showed the clearest and most stable target-looking patterns, especially in lexically specified object conditions. Children with ASD broadly followed a TD-like cue-sensitive profile, but showed less stable target preference under high integration load, particularly in later or post-sentence windows, along with lower target-like production.
The thesis makes four main contributions. Theoretically, it reframes subject-object asymmetry in Mandarin wh-argument questions as dynamic multi-cue integration, thereby unifying previously inconsistent findings. Developmentally, it shows that children gradually calibrate the weights of wh-type, animacy, discourse, and task cues rather than acquiring fixed subject- or object-question templates. Methodologically, it integrates corpus data, offline comprehension and production tasks, visual-world eye-tracking, and individual-difference measures to capture language ability across usage, behavior, real-time processing, and cognitive resources. Clinically, findings from children with ASD suggest reduced cue-integration stability under high processing demands rather than qualitative differences in core grammatical representations. Overall, subject-object asymmetry should be interpreted in relation to cue configuration, task demands, and processing load, not as a standalone developmental or clinical marker.
| Date of Award | 23 Jul 2026 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Chun Yu KIT (Supervisor) & Shenai HU (External Supervisor) |
Keywords
- Mandarin wh-argument questions
- subject-object asymmetry
- Interface- Weighted Cue Integration
- sentence processing
- autism spectrum disorder
Cite this
- Standard