Introduction
Learning a second language (L2) involves more than just memorizing a new vocabulary; it requires learning how a different culture categorizes the world. Lexical categories—the boundaries that define where one word ends and another begins—rarely overlap perfectly between languages. For example, what an English speaker distinguishes as a "mug," "cup," or "glass" might be grouped under a single term in another language. This study investigates how Chinese-English bilinguals navigate these shifting boundaries and what factors predict "native-like" naming patterns in their second language.
Methods and Datasets
The researchers conducted a picture-naming study involving 63 Chinese-English bilinguals across two distinct environments: a non-immersed group in Beijing, China, and an immersed group at Penn State University, USA.
Participants were tasked with naming 67 photographs of common household serving dishes, such as cups, bowls, and plates, in both Chinese and English. To establish a "ground truth" for native-likeness, naming data was also collected from functionally monolingual Chinese speakers in Beijing and English speakers in Pennsylvania.
The analysis employed two statistical approaches:
- Subject-wise regression: To determine how individual language history (e.g., age of exposure, years of study, immersion) predicted overall native-likeness.
- Item-wise regression: To predict the probability of a participant providing the "dominant" native name for a specific object, accounting for linguistic community norms (the level of agreement among native speakers).
Key Insights and Results
The study revealed several critical factors influencing L2 lexical categorization:
- The Classroom Penalty: Surprisingly, more years of formal English study outside of an immersion environment were associated with lower native-likeness. Traditional classroom settings often emphasize one-to-one translations, which may reinforce first-language (L1) category boundaries and make them harder to unlearn.
- Immersion and Code-Switching: Length of residence in an L2 environment was the strongest predictor of native-like categorization. Immersion provides the diverse, often inconsistent input necessary to realize that L1 rules do not apply. Furthermore, frequent code-switching among immersed learners correlated with higher native-likeness, suggesting that active language management helps refine category boundaries.
- L1 Entrenchment: Objects that native Chinese speakers named with high consistency were more difficult for learners to name correctly in English. This suggests that highly "entrenched" L1 categories are more resistant to being overridden by L2 norms.
- Cross-Language Convergence: Bilinguals' naming patterns in English and Chinese were more similar to each other (correlations of 0.92–0.95) than the naming patterns of monolingual speakers of the two languages (correlation of 0.64), indicating a blending of the two linguistic systems.
Discussion
The findings highlight that mastering a second language requires exposure to the "messiness" of native naming—the complex, many-to-many mappings found in real-world usage—rather than the simplified translations found in textbooks.
While this study focused on concrete objects (household dishes), future research is needed to determine if these patterns hold for abstract concepts or other domains like clothing and furniture. Additionally, the "classroom penalty" observed here might be specific to traditional instructional styles; exploring more immersive or diverse teaching methods could provide pathways to mitigate this effect. Ultimately, native-likeness in L2 is a dynamic outcome of both individual history and the specific linguistic norms of the communities in which a learner operates.
Summary updated