Read and Ask

A pair structure with three steps, the first of which happens only once. First the class forms pairs and agrees who is A and who is B, with one text per pair. Then A reads a passage aloud while B listens without reading along (one minute). Then A asks one question about the passage; B answers in a full sentence, and A checks in the text — right: a compliment; wrong: A points to the place and reads that sentence once more (three quarters of a minute). On the next passage the roles swap. Four passages are the default, so that everyone reads twice and listens twice. Instead of a question you can have the listener retell the passage; you choose that in the preparation.

Because the eye always beats the ear: whoever has the text in front of them reads along and no longer hears what is being said. The structure practises exactly what silent reading does not — holding on to a spoken text and being able to answer a question about it — and that only works if the listener has nothing but their partner's voice. That is why there is one text per pair, in the reader's hands, and why the player never puts the passage itself on the screen: only the pointer to where it is. It is also the reason the reader checks and not the listener — the reader is the only one who has the text.

One passage costs a minute and three quarters with the default times: sixty seconds of reading aloud, forty-five seconds for the question and the answer. Four passages are seven minutes, plus a minute for the pairs and a minute to hand out the texts: about ten minutes. Six passages are a quarter of an hour. Whoever prepares the structure sets the time per step and chooses the number of passages, from one to twelve — and chooses an even number, so that A and B read equally often.

Any text that has to be understood: a section from the textbook, a source, an instruction, a story extract, a text at language level. A passage is what you read aloud calmly in half a minute to a minute — a paragraph or two, five to ten sentences. Cut a longer text into more passages rather than reading longer: after the first minute the listener drops off and the question ends up being about the beginning. Choose the level below the weakest reader in the class, because whoever reads haltingly gives the listener nothing to hold on to. If you get stuck, ask Microsoft Copilot: "Divide the text [paste the text or name the chapter and pages] into 4 to 6 passages that each take about a minute to read aloud, and say for each passage where it starts and ends. One passage per line, without the text itself." Paste the answer into the Passages field — every line becomes a round.

The reader, text in hand. That seems the wrong way round — you would expect the listener to ask what they did not understand — but it is the core of the structure: the reader is the only one who can check whether the answer is right, and making a question about a paragraph is a heavier thinking exercise than reading it aloud. A good question is about the content ("why did the army retreat?"), not about a detail ("which year?"). Model it once the first time, and then walk round and listen to the questions instead of the answers — that is where you see who has understood the text.

Questions if you want to check whether one thing stuck; retelling if you want to hear how much stuck. With questions the reader asks one question and checks the answer in the text — sharp, short, and the reader learns to make a good question. With retelling the listener retells the passage in their own words from beginning to end while the reader follows in the text, does not interrupt, and afterwards adds what was missed with a finger on the sentence. Retelling yields more language and is heavier — count on a silence. For a class doing the structure for the first time: start with questions, switch to retelling later. You choose it per text in the preparation; the steps stay the same.

Three things carry the structure. One: one text per pair, in the reader's hands — check that at the start, because a listener who reads along practises nothing. Two: the reader's pace. Say beforehand that reading calmly is the task, not finishing, and that whoever is done early reads the last sentences once more. Three: the question. Model once what a question about the content sounds like, and walk round during the question step to listen to the questions, not the answers. The player does the rest: the timer, the swap of roles, and the pointer to which passage is up. And the reflection questions at the end — "which question would you ask the class?" — are a good bridge to your own next lesson: write two on the board.

One group of three: one reader and two listeners who answer in turn. At the swap the role moves on — B reads, C and A listen — so in the trio too everyone gets a turn. It is better than pairing a student with you: then you are the listener and cannot see the rest of the class.

Jigsaw divides a text: each member of a group studies their own part, confers with the other experts on that part and then explains it to their own group. That takes three quarters of an hour, groups of four to six and material students can grasp independently. Read and Ask keeps the text whole: everyone reads and hears the same passages, in pairs, in a quarter of an hour. Rule of thumb: Jigsaw when the text is too long for everyone and can be divided; Read and Ask when you want everyone to understand the same text.

Both work in pairs with swapping roles, but the material is different. In Pair Coach one solves a task while the other asks questions and does not prompt; it is about practising. In Read and Ask one reads a text aloud and the other checks what they heard; it is about understanding. Understanding comes before practising: first Read and Ask on the section, then Pair Coach on the exercises that go with it.

Yes, and it is one of the strongest uses. The reader practises pronunciation and reading pace, the listener practises understanding without seeing the text, and the answer in a full sentence is speaking practice — three skills in a round of two minutes. Choose passages of thirty seconds and below the level of the class, allow questions in the home language when comprehension comes first, and use retelling only once questioning runs well. For newcomers it is one of the few structures in which they read aloud without a group listening.

Yes, with one extra rule: the listener does not share their screen and keeps the text closed. Pairs in breakout rooms, the reader with the text, and the player shared in the main room or on everyone's screen through the live link — which only shows which passage is up, never the text. At the swap the reader says where the other should start.

The structure belongs to the family of cooperative pair structures in which two students take a role in turn and check each other — the structures Spencer Kagan described, in which reading in pairs with a check question has a fixed place. Reading in pairs itself was worked out by Keith Topping as paired reading: a reader and a listener who check after every passage what got through. What makes the structure work in the classroom — one text per pair, the reader who asks, the sentence pointed at after a wrong answer — is recorded here the way the player runs it.