Abstract: This study utilized a lexical decision task to investigate the impact of varying time intervals (500 ms, 750 ms, 1000 ms) on lexical processing in advanced Chinese L2 learners, with particular attention to the interaction between time interval and word frequency. The results indicated that a 750 ms interval optimised performance, facilitating both higher accuracy and rapid responses, especially for high-frequency words. In contrast, the 500 ms interval, while reducing reaction time, posed greater challenges for mid-and low-frequency words as well as non-words, without significantly improving accuracy. These findings contribute theoretical insights for L2 vocabulary instruction and assessment, and offer practical implications for the design of more effective learning tools.
Introduction
In the field of cognitive linguistics, the "word frequency effect" is a cornerstone principle: common words are processed more efficiently than rare ones. While well-documented in native speakers, the dynamics of this effect in second-language (L2) learners are more complex, influenced by cognitive load and proficiency levels. This paper investigates how advanced L2 Chinese learners navigate word recognition under varying temporal constraints, specifically manipulating the Interstimulus Interval (ISI)—the pause between stimuli—to identify how time pressure interacts with word frequency to facilitate or hinder lexical access.
Methodology
The study employed a standard lexical decision task (LDT) involving 18 advanced L2 Chinese learners (HSK 5 to HSK 6 proficiency). Participants were required to determine if a presented character string was a legitimate word or a non-word.
Experimental Design
- Temporal Manipulation: The ISI was set at three levels: 500ms, 750ms, and 1000ms.
- Stimuli: Vocabulary was sourced from the Modern Chinese Frequency Dictionary, categorized into high-frequency and mid-to-low-frequency sets, with visual complexity (stroke count) strictly controlled.
- Statistical Modeling: Reaction times (RT) were analyzed using a Mixed-Effects Linear Model, while accuracy was assessed via a Generalized Linear Model (GLM) with a binomial distribution.
Results
The statistical analysis revealed a distinct "Goldilocks" effect at the 750ms interval, where learners achieved an optimal balance between speed and precision.
Reaction Time (Speed)
Both the 500ms and 750ms intervals significantly reduced reaction times compared to the 1000ms baseline. Specifically, the 750ms interval reduced RT by approximately 200.4ms ($P=0.001$), while the 500ms interval reduced it by 161.7ms ($P=0.007$). However, word rarity imposed a heavy penalty; low-to-mid-frequency words increased RT by ~263ms ($P=0.000$).
Accuracy
While the 500ms interval increased speed, it did not yield a statistically significant improvement in accuracy ($P=0.083$). In contrast, the 750ms interval produced the most significant accuracy gains (coefficient +2.11, $P=0.006$). Accuracy dropped sharply when participants encountered low-to-mid-frequency words (coefficient -1.52, $P=0.000$) or non-words (coefficient -2.23, $P=0.000$).
Discussion
The findings suggest that 750ms provides a sufficient cognitive window for L2 learners to process vocabulary without losing the momentum of residual activation from previous stimuli. At 500ms, the cognitive load becomes overwhelming for less familiar words, causing a bottleneck in lexical access. Conversely, at 1000ms, the immediate priming effect fades, leading to a plateau in processing speed.
These results provide empirical support for the Logogen Model of word recognition. High-frequency words, having lower activation thresholds, require less cognitive effort and can thus be recognized even under the high-pressure 500ms constraint. Low-frequency words require more time to reach the necessary activation threshold, making the 750ms interval critical for successful retrieval.
Conclusion and Practical Implications
This research highlights the importance of temporal dynamics in L2 processing. For educational technology, these findings suggest that language learning software (such as spaced-repetition systems) could be optimized by dynamically adjusting display intervals based on word frequency. By providing a "Goldilocks" interval of approximately 750ms for challenging vocabulary, tools can maximize both the speed and accuracy of L2 learners, leading to more efficient acquisition.
Summary updated