Abstract This study investigates the role of working memory (WM) in word recognition. By manipulating word frequency and working memory load in a naming task, the researchers analyzed how cognitive resources are allocated during reading. The results indicate that word frequency is significantly influenced by working memory load, with the impact being notably greater for low-frequency words than for high-frequency words. These findings suggest that working memory is more heavily involved in the processing of unfamiliar or rare words, providing important implications for early reading acquisition and educational strategies.
Introduction: The Automaticity Debate
For decades, psycholinguists have debated whether word recognition is a purely "automatic" process—one that occurs rapidly, unconsciously, and without consuming limited cognitive resources. If reading were entirely automatic, performing a secondary mental task should not interfere with a person's ability to identify words. However, contemporary theories suggest a more nuanced reality. While common words may be recognized via a direct visual route, rare or "low-frequency" words likely require a more taxing phonological route, involving the active assembly of sounds. This study aims to test this "dual-route" hypothesis by examining whether taxing working memory interferes with word recognition speed.
Experimental Design and Methodology
The researchers employed a dual-task paradigm involving 60 university students to observe the interaction between memory load and linguistic processing. The experiment focused on three primary variables:
- Working Memory Load: Participants performed a digit-span task. In the "Load" condition, they memorized seven digits; in the "Control" condition, no memory load was imposed.
- Word Frequency: Target words were categorized as High-frequency (common) or Low-frequency (rare) based on the Korean Frequency Database.
- Word Length: The study utilized 48 Korean words, balanced between two-syllable and three-syllable structures.
During the naming task, participants were required to read the target words aloud as quickly as possible while maintaining the digit span in memory. Reaction times (RT) were measured from the stimulus onset to the beginning of the vocal response.
Results: The Interaction of Load and Frequency
Statistical analysis using ANOVA confirmed that working memory load significantly impacts word naming speeds, particularly for less familiar vocabulary. The data revealed a clear interaction between frequency and cognitive load, suggesting that reading and memory draw from the same pool of mental resources.
As shown in the data, the "speed penalty" for low-frequency words increased under load:
- 2-Syllable Words: Under memory load, the gap between recognizing a rare word versus a common word was 100ms. In the control group, this gap was reduced to 74ms.
- 3-Syllable Words: Under load, the gap was 47ms, compared to only 22ms in the control condition.
These results demonstrate that working memory load consistently amplifies the difficulty of processing rare words, indicating that "sounding out" unfamiliar text is highly dependent on available mental bandwidth.
Discussion and Educational Implications
The findings support the Dual-Route Model of reading. Common words appear to be stored as visual snapshots that are retrieved with minimal cognitive effort. In contrast, rare words require phonological assembly—converting letters to sounds—a process that is cognitively expensive and easily disrupted when working memory is occupied.
This research has significant implications for early childhood education. Children, who typically have smaller working memory capacities and are frequently encountering "low-frequency" (new) words, are likely much more sensitive to cognitive overload than adults. Future research should focus on developing training programs that help young learners manage working memory more efficiently to accelerate reading acquisition and literacy development.
Summary updated