We often assume that when learners make predictable grammar errors, they’re simply transferring habits from their first language. But a recent study from Li Zeng & Fei Gao (2025) suggests that sometimes, second language minds don’t copy—they rebuild.
The researchers re-examined how Chinese learners of English acquire reflexives (words like himself & herself), which work differently across languages. In English, reflexives must refer to someone in the same clause (“Peter blamed himself”), while in Chinese, ziji (“自己”) can also reach across clauses (“John said Peter blamed himself” – where himself could mean John).
The study They tested 98 Chinese learners of English (ages 15–29, proficiency levels from B1–C1) & 12 native English speakers using a story-based truth-value judgement task. Each story was followed by sentences testing four possible reference types:
- Local antecedent (grammatical in English) – the reflexive refers to the subject in the same clause: Peter blamed himself. Used in English.
- Long-distance subject in a finite clause – the reflexive tries to refer to a subject in another (complete) clause: John said Peter blamed himself (meaning John). Not used in English.
- Long-distance subject in a non-finite clause – the reflexive refers to a subject in another (unfinished) clause: John asked Peter to wash himself (meaning John). Not used in English.
- Long-distance object – the reflexive refers to an object, not the subject: John told Peter that Mary praised himself (referring to Peter). Ungrammatical in all languages.
Participants judged whether each sentence “made sense” after reading the story.
The results were striking:
- Learners preferred local antecedents & rejected long-distance ones.
- As proficiency rose, accuracy increased sharply.
- Crucially, none of the learners showed the L1-transfer pattern (Chinese-style long-distance reference).
- Instead, lower-proficiency learners behaved like English-speaking children, initially allowing reflexives to refer almost freely, then gradually narrowing down to local references as proficiency grew.
By high proficiency, 81.5% of learners were fully native-like, while only 14.8% still made inconsistent choices.
So, what does this mean? It challenges decades of L2 research assuming reflexive errors in Chinese learners come from L1 transfer. Instead, the findings support the Lexical Learning Hypothesis (Wexler & Chien, 1985): learners—whether first- or second-language—must discover how a reflexive behaves in each language by learning its specific lexical features, not by copying rules from their first language.
This also echoes broader L2 theories suggesting that Universal Grammar remains accessible to adult learners, guiding them toward target-like systems even without explicit instruction.
Think of it like learning to drive a different kind of car: the instinct to steer remains universal, but you still need to learn where the handbrake is.
Teacher Takeaways?
- Don’t assume L1 interference explains every grammar error. Some patterns reflect universal developmental stages rather than transfer.
- Grammar discovery takes time. Learners may need extensive exposure before rules “click”—especially for abstract forms like reflexives.
- Use comparative reflection. Discuss how English reflexives differ from students’ L1 structures to foster metalinguistic awareness.
This study reminds us that language learning isn’t just a matter of translation—it’s reconstruction. Learners aren’t copying rules from their first language; they’re building new mental grammars from the ground up.
How might similar “universal” patterns reveal themselves in your learners’ development?
Leave a Reply