Eye-tracking Experiment 1 and 2
Description
This visual-world eye-tracking study examines listeners' strategies and patterns in using and integrating multiple language cues during real-time spoken sentence comprehension of their first language (L1) and second language (L2). Listener listened to English (Experiment 1) and Mandarin (Experiment 2) sentences within a brief discourse context, while their gazes were tracked as they viewed a display of four objects. Each trial began with a statement naming one of the objects (\emph{Here is a cabbage}), after which a second sentence described an action (\emph{Now Susan will shred the cabbage after school}). Participants were instructed to click on the object named in the second sentence. These sentences varied in three binary factors: whether the the target object carried prosodic pitch accent or not; whether the target object was old in the first statement and the target sentence, or new in the second sentence; and whether the preceding verb was semantically appropriate for the target object. Results revealed that both L1 and L2 listeners, in both English and Mandarin, adopted similar cue integration strategies when they were listening to sentences containing a old noun. When sentences contained a new target noun, however, L1 and L2 listeners differed. L1 listeners were sensitive to both semantic and prosodic information and interactively combined them across a variety of contexts. In contrast, L2 listeners showed a reduced ability to integrate different cues, with their performance influenced by the specific prosodic systems of English and Mandarin.
Files
Steps to reproduce
Experiment 1 uses English stimuli, and Experiment 2 uses Mandarin stimuli. The raw data for English and Mandarin listeners is provided in file.txt. Use the Python script script.py to combine the English and Mandarin listeners' data into binned_data.csv. File.rmd is the R script for the GAMMs analysis of the binned data.
Institutions
- University of Glasgow