“I disagree” is grammatically complete and socially unfinished. The same two words can sound efficient in a planning meeting, severe in a conversation with a client, or oddly formal between close friends choosing lunch. Disagreement lessons should organise language around relationships and consequences, then let learners practise the turns that follow.
One phrase list cannot do that work.
Taiwanese learners already make social choices
Avoidance, delay, reasons, partial agreement, and direct contradiction are choices about interaction as well as vocabulary. A Taiwanese employee may remain silent when a manager proposes an impossible deadline, then explain the problem privately to a colleague. English ability is only part of that decision. Status, distance, and the cost of opposing the proposal also matter.
Miao-tzu Chen compared disagreement by 60 Mandarin speakers, 60 English speakers, 30 lower-proficiency Taiwanese EFL learners, and 30 higher-proficiency learners, using written and spoken-response instruments rather than recordings of naturally occurring meetings. Chinese speakers avoided disagreement more often, while English speakers used direct disagreement more frequently and often softened it with varied positive remarks. Learner proficiency influenced available language, yet cultural and situational choices remained visible.
A separate Taiwanese study used four scenarios varying institutional power to compare learners with American English speakers. Role-play questionnaires can elicit comparable responses, though they cannot reproduce the timing, facial expression, or repair found in a live argument. Together, these studies justify teaching disagreement as a situated act rather than assigning one expression to a CEFR level.
They do not establish an American norm that every Taiwanese speaker should copy. Directness varies within every group, and English often serves communication among people with different first languages. The instructional aim is a larger set of choices and clearer control over their likely effects.
Build contrasts into the task
Start with one proposition across two relationships. A team leader tells a close colleague, “We can finish the migration tonight.” Later, a new employee hears the same claim from a division director. Learners rank possible openings and state what each one risks:
- “I don’t think tonight is realistic. The backup alone takes three hours.”
- “Can I check one assumption? The last backup took three hours.”
- “No, that won’t work.”
- silence followed by a private message
The task has four options because disagreement offers more than a polite/direct binary. Learners can discuss how the factual reason, the question form, and the timing change the turn. They should also be allowed to prefer one response. Pragmatic competence includes choosing a voice, not imitating a single supposedly native script.
Details give the language pressure. “I disagree because it is difficult” asks for little. A migration starts at 18:40, customer traffic falls at 23:15, and rollback needs 47 minutes. Now the speaker must decide whether to challenge the schedule, one assumption, or the acceptable risk. That is a better reason to use English.
Teach the next turn
Many materials stop immediately after the softened objection. Real disagreement creates a response: a request for evidence, a defence, a concession, or a sharper challenge. Learners need to remain in the exchange.
Practise four-turn sequences in which speaker A proposes the plan, speaker B marks a concern and gives one concrete reason, A asks for clarification or pushes back, and B restates the concern, narrows it, or offers an alternative. Rotate the roles and change the relationship. The useful language includes “The part I’m worried about is…,” “Are you saying that…?,” and “I could support that if…”. These forms manage the emerging problem instead of decorating an isolated sentence with politeness.
Instruction can change performance, although proficiency sets limits. In an online collaborative project, Maíz-Arévalo analysed 25,347 words produced before and after explicit instruction. B1 participants improved their expression of disagreement. Lower-level learners remained further from the target pragmatic and linguistic performance. The participants were an international group of master’s students, so the result is evidence for explicit attention, not a direct forecast for a Taiwanese evening class.
Choc’s article on repair phrases for stalled conversations fits after this sequence. Disagreement can produce exactly the misunderstanding those phrases are meant to repair.
Assess the decision as well as the sentence
A grammar correction catches “I am not agree.” It misses whether the turn fits the relationship or gives the other speaker a usable next move. Feedback can separate four dimensions: intelligibility, stance strength, supporting reason, and response to the following turn. A learner may produce accurate language and still choose a force that surprises the listener.
Do not grade every indirect answer as superior. During an urgent safety check, “Stop. The pressure reading is 187” may be the responsible choice. Between close colleagues, direct disagreement can display trust. A long preface may sound evasive when the group needs a decision in two minutes. Context changes the evaluation.
Self-assessment becomes more precise when learners name that context: “I softened my opening because I was speaking to a new client, then I gave the delivery date as evidence.” That sentence can be checked against the recording. “I was polite” cannot.
Change one variable
End the lesson with a transfer task outside the original script. Give pairs a Taipei office scenario: a client wants 26 slides added before a 09:30 presentation, or a colleague wants to hold an outdoor event despite a typhoon warning. Change one variable on the second attempt, such as status or urgency. Listen for the moment when the disagreement has to change with it.