Zhenghua Zhang Qingfang Zhang ^*

• Renmin University of China, Beijing, China

We employ visual world eye-tracking to investigate whether processing of focus and accentuation follows linear incrementality or hierarchical incrementality. Participants were asked to complete a scenario description task where they were prompted to use a predetermined sentence structure to accurately convey the scenario, thereby spontaneously accentuate the corresponding entity. We manipulated the positions of focus with accentuation (initial vs. medial) by changing the scenarios. The initial and medial positions correspond to the first and second nouns in sentences like "N1 is above N2, not N3". Our findings revealed that speech latencies were significantly shorter in the sentences with initial focus accentuation than those with medial focus accentuation. Furthermore, eye-tracking data demonstrated that speakers quickly displayed a preference for fixating on initial information after scenarios onset. Crucially, the time-course analysis revealed that the onset of the initial focus accentuation effect (around 460 ms) preceded that of the medial focus accentuation effect (around 920 ms). These results support that focus and accentuation processing during speech production prior to articulation follows linear incrementality rather than hierarchical incrementality.