An English-only rule sounds rigorous. In practice, a teacher can spend four minutes circling an abstract word while half the class quietly attaches the wrong meaning to it. For Taiwanese learners who share Mandarin, a brief, planned switch can make the next stretch of English more accurate and more productive. The test concerns timing and purpose: those twelve seconds of Mandarin should let the learner understand or say something specific in English.
Use it, then leave it.
The cost of a long explanation
Consider the verb postpone. A teacher could define it with familiar English, give a calendar example, gesture towards Friday, and check comprehension. That sequence supplies valuable input when learners know the surrounding words. At A2, the explanation itself may introduce reschedule, arrangement, and later date. The class now has four uncertain words where it started with one. A quick 「延期」 can secure the core meaning, leaving time to compare postpone the meeting with cancel the meeting.
Lili Tian and Ernesto Macaro tested this issue during listening work with 80 first-year university students learning English in China. The researchers stratified the students by proficiency and randomly assigned them to teacher code-switching or English-only lexical explanations. A further 37 students formed a control group without the same lexical focus. Explicit attention to words beat incidental exposure, and the code-switching condition showed initial evidence of stronger vocabulary learning than English-only information. Proficiency did not clearly favour either treatment.
The result belongs to one university setting, one type of vocabulary instruction, and listening activities. It cannot settle how much Mandarin a Taiwanese secondary class should hear across a semester. It does challenge the idea that maximum teacher English automatically produces maximum learning. Minutes of input matter only when learners can process enough of them.
Translation gives one thin connection
A Mandarin equivalent can establish meaning quickly. It rarely supplies pronunciation, grammar, register, collocation, or the situations where speakers choose the word. Teaching issue as 「問題」 leaves several jobs unfinished: its first vowel, the phrase raise an issue, the softer tone it can carry than problem, and meanings such as an edition of a magazine.
That thin first connection still has value. Choc's article on the mental lexicon and word games explains why usable vocabulary depends on a network of sound, meaning, grammar, and association. A translation can place the first pin. The rest of the lesson has to add lines. Ask learners to choose between issue and problem in two short messages, say raise an issue aloud, and retrieve the phrase after another activity.
I would rather spend eight seconds fixing the core meaning and eight minutes making learners use the word than protect an English-only rule through a long, uncertain definition. The preference has a boundary. If learners can infer the word from a short English example, that inference work deserves time. The teacher can wait, ask for evidence, and confirm the answer in English. Mandarin earns its brief turn when the English explanation itself becomes the obstacle, or when a precise cross-language contrast prevents a predictable error in the next task.
This distinction keeps code-switching in its proper size. Mandarin handles a local blockage. English carries the task.
Teachers need a reason to switch
Ernesto Macaro followed six student teachers across 14 foreign-language lessons. The classes involved English-speaking pupils learning French, so the language pair differs from Taiwan. The teachers used comparatively little first-language speech, and the quantity of teacher switching had little effect on how much first or target language pupils used. Macaro also found tension in the teachers' decisions: their stated beliefs and their moment-to-moment choices did not line up neatly.
That observational study offers no causal verdict about learning. It points to a problem in simple percentage targets. Two teachers might each use Mandarin for 7 percent of a lesson. One spends it translating every instruction. The other uses 11 seconds to contrast borrow and lend, then runs an information-gap task in English. The number hides the teaching decision.
A useful switch should have a named function. It may confirm the meaning of an abstract term, explain a safety instruction, compare a Mandarin and English form that causes predictable errors, or repair a task that nobody has understood. Classroom management with beginners can also justify a short switch when confusion would otherwise consume the activity.
Convenience is a weaker reason. Repeating every English sentence in Mandarin trains learners to wait for the second version. Translating a student's whole contribution can take ownership away from the speaker. Using Mandarin for every grammar explanation prevents learners from meeting the classroom English that could become familiar after five lessons.
Design the return to English
The switch needs an exit. Imagine a teacher preparing 13 adult learners for a supplier call. The target phrase is We're running behind schedule. The teacher gives one natural example, checks the phrase, and sees blank faces. A six-second Mandarin gloss, 「進度落後」, settles the message. The next move happens in English: learners decide which of four project updates fit the phrase, tell a partner why, then call back the phrase from memory after 17 minutes.
The gloss saved time because the lesson spent the saved time on use.
There are boundaries. In a pronunciation task, Mandarin cannot replace listening closely to English sound. During a fluency turn, a teacher who interrupts with translations may break the very retrieval process the task was built to exercise. In a multilingual class, switching to the majority language can exclude learners who do not share it. Advanced students working on inference may need to tolerate ambiguity rather than receive an instant equivalent.
Teachers can audit their own choices with three notes after class: the word or instruction that triggered the switch, its length, and the English action that followed. If the third column stays empty, the Mandarin probably occupied time without opening a path. If learners used running behind schedule correctly at minute 17, those six seconds earned their place.