Two learners receive the same restaurant menu and the instruction “Discuss it for five minutes.” They can finish after agreeing that the pasta looks good. Give each learner a different page, with prices missing from one and allergy details missing from the other, and silence becomes costly because group speaking earns its class time when each learner holds information the others need.
Here, the intended reader is a teacher planning online lessons for 13 B1 adults in Taiwan.
More breakout rooms will increase private speaking time. Task design decides what learners have to do with that time.
A gap gives talk a purpose
An information-gap task distributes facts unevenly. Learner A may have a train timetable with platform numbers. Learner B has delays and transfer rules. They share a goal, such as finding the last route from Banqiao to Taichung, and neither sheet is sufficient alone. This arrangement creates ordinary conversational work. A learner asks for a missing detail, checks whether thirteen or thirty was said, requests a repetition, and reformulates an unclear direction. Those moves are negotiation of meaning. The task makes them relevant because a wrong interpretation changes the route, price, or final answer.
Because neither learner can complete the route alone, every request and confirmation carries a consequence that a free discussion prompt rarely supplies: the pair either reaches the correct platform before departure or misses the train.
Catherine Doughty and Teresa Pica compared classroom patterns in their study of required and optional information exchange, finding more conversational modification in tasks that required learners to exchange information. Pair and group patterns produced more modification than teacher-fronted work when the exchange requirement was present.
Simply placing chairs in a circle did little.
Doughty and Pica examined interaction produced during tasks. Durable language development lay outside their measures. The strongest classroom claim concerns opportunity: required exchange makes learners more likely to signal trouble and work on it together.
Prepared input cannot react
A teacher can simplify directions before class by shortening sentences and repeating vocabulary. Interactive speech has another property: it changes in response to the learner's exact misunderstanding. Teresa Pica, Richard Young, and Catherine Doughty tested this with 16 non-native speakers of English, who followed directions under two conditions: input simplified in advance with extra quantity and redundancy, or unmodified directions with access to interaction with a native speaker. Comprehension was best supported when content was repeated and rephrased during interaction. Lower syntactic complexity in the prepared version had no clear effect by itself.
Sixteen participants cannot settle how a Taiwanese online class should spend a semester.
The study also focused on understanding task directions rather than free conversation. Its design captures a useful feature of live exchange: a partner can respond to the particular word, reference, or step that failed.
That response might be a confirmation check such as “Platform thirteen?” It might be a clarification request: “Which train do you mean?” The repair can stay inside the task because both speakers still need the timetable answer.
Task type matters more than room type
Susan Gass, Alison Mackey, and Lauren Ross-Feldman compared task interaction among 74 university learners of Spanish, who worked in pairs on three communicative activities in classrooms and laboratories. Interaction patterns differed little between the two settings, while the type of task accounted for more of the observed differences. That finding is useful for online teaching. A Microsoft Teams breakout room does not have a fixed pedagogical effect. The same room can hold a lively information exchange, two people reading identical answers, or 91 seconds of polite silence.
Technology supplies privacy and channels. The materials distribute knowledge and responsibility.
When each screen contains a different deadline, price, and restriction, the breakout room becomes a place where learners must make their own language clear enough for another person to act on it.
Alison Mackey's study of 34 adult ESL learners went further by testing development of English question forms. Learners took part in different kinds of task-based interaction across four experimental groups and a control group. Results linked interaction with grammatical development and gave particular weight to active participation and the learner's developmental level. With 34 learners split across five groups, the target was narrow. The safest interpretation stays close to English question forms and active participation. A learner who asks, checks, and repairs has a different experience from a learner watching a stronger partner finish the sheet.
Build pressure without hiding the lesson
Imagine a supplier-call task. Learner A sees the factory's available dates: September 17, 19, and 24. Learner B sees the client's deadlines and a note that the 19th is a public holiday. Both must choose a delivery date and prepare a 40-second update for a manager.
Four design choices keep the gap productive:
- Each person owns at least two facts needed for the decision.
- The shared product records one answer, so ignoring a partner has a visible cost.
- Language support sits on both sheets, including Could you confirm…? and Did you say…?
- One detail is deliberately confusable, such as fourteen and forty, while the task stays solvable.
Teachers should listen for the point of friction. If six pairs struggle to distinguish dates, the class can pause for a 52-second pronunciation repair. If one pair misunderstands the public-holiday constraint, the teacher can ask them to reread their own cards. Feedback follows evidence from the task.
This is one reason Choc argues that a smaller class works only when the lesson changes. Fewer learners create room for diagnosis. A designed gap gives the teacher something worth diagnosing.
Know when to remove the gap
Information gaps can become puzzles that reward speed more than language. A beginner who lacks the phrases to request help may let a partner take over. Learners can also read their cards aloud mechanically, completing the grid without attending to meaning. Sensitive personal facts should never become compulsory classroom currency. I would use a gap when the lesson needs clarification, confirmation, or precise transfer of facts. I would remove it when secrecy adds paperwork without changing the language. For the next pair task, print two versions. Delete the platform number from A, delete the delay from B, and give both learners the same last train to catch.