Two minutes, one voice: Speaking Part 3
Paired speaking test, B2 First for Schools. Sofia, 15, and Tom, 15 reach Part 3. The interlocutor gives them a spider diagram of five ideas for a school open day - a science show, a food stall, a sports tournament, a music performance, a photography exhibition - and says: 'Talk to each other about which of these would attract the most visitors to your school's open day.'
Sofia names all five ideas in 35 seconds, ranks them, and announces a decision at 0:50. Tom says 'Yeah' and 'I agree' and adds one comment about a football match. At 1:10 Sofia begins listing ideas of her own that are not on the diagram. At 2:00 the interlocutor stops them and asks which two they will choose. Sofia answers both. The interlocutor has to decide whether to intervene, and the assessor has to decide what this performance says about Sofia's Interactive Communication.
1. Should the interlocutor have intervened? Justify your answer with reference to the different roles of interlocutor and assessor.
Read the worked answer
The two roles must be kept apart, and the intervention question belongs to the interlocutor only.
The assessor never intervenes. Marks are awarded on what is heard, not on what could have been heard with help. If the interlocutor rescues the interaction, the assessor is now judging the pair plus the examiner - invalid evidence.
The interlocutor manages the test, not the conversation. The brief is to give instructions, hold time, and ask the follow-up - not to supply ideas or steer content. Joining the discussion would change the task conditions and remove the opportunity both candidates are being marked on.
That said, one narrow move is legitimate. At 1:10, with Tom silent for the best part of a minute, the interlocutor may make a non-content nomination: 'Tom, what do you think?' This adds no ideas; it redistributes the turn. It is permitted when one candidate is dominating, and it should be done once and briefly.
If the pattern repeats across all four parts, the right response is not repeated rescuing but post-test feedback and paired practice, because the underlying problem is turn-taking behaviour, not examiner timing.
The follow-up at 2:00 is the interlocutor's clean instrument: 'Sofia, which one? … Tom, do you agree?' The second half of that question is doing legitimate work.
2. What does this exchange show about Sofia's Interactive Communication, and what would a strong candidate do differently in the same two minutes? Be specific.
Read the worked answer
Sofia is fluent, but fluency is a Language and Discourse criterion, not an interactive one. Interactive Communication rewards initiating and responding appropriately, taking turns, and inviting others in - and she has demonstrated almost none of the second and third.
Diagnostic signs:
- 35 seconds for five ideas means listing, not discussing. Each idea gets named, not developed.
- Announcing a decision at 0:50 is a monologue ending, not a negotiated one.
- Replacing the diagram with her own ideas at 1:10 shows she is treating the task as a long turn she must fill.
- Tom's one contribution never appears in her decision, so there is no evidence of responding.
A strong candidate in the same two minutes takes two or three ideas, not five. For each: a claim, a reason, an example. Then an explicit hand-over: 'A food stall's cheap to run - what do you reckon, would that pull in parents or just the kids?' When the partner speaks, the reply builds on it ('that's true, and if we put it next to the science show…'), and the final decision references something the partner said, with a check: 'Shall we go with the food stall then?'
Provisional band: Interactive Communication around 3 rather than 4–5 - but one part is not enough to band a candidate. The teacher should listen for the same pattern across Parts 2 and 4 before writing the mock grade.
3. Give Tom two concrete phrases he can use to enter the conversation, and explain why they work better than 'I agree'.
Read the worked answer
Phrase 1 - a signposted bid for the floor: 'Can I just say something about the sports tournament?' It announces a turn and names a topic, which makes it easy for Sofia to yield. It is polite enough for an exam and assertive enough to be heard over a fast partner.
Phrase 2 - link, disagree, question: 'That's a good point, but I'm not sure about the music performance - what made you put it first?' This does three jobs in one breath: it acknowledges the partner's utterance, adds a different view with a reason implied, and ends with a question, which forces a response and hands the turn back. That is the whole interactive cycle in a single move.
Why 'I agree' fails. Agreement adds no content and opens nothing. It does not require the partner to respond, so the turn stays with Sofia and Tom is silent again thirty seconds later. Repeated agreement also gives the assessor no evidence of range - no initiating, no negotiating, no disagreeing.
The phrasing is only half the work; Tom needs the behaviour rehearsed. Drill: two-minute pair task where Tom must speak second after every one of Sofia's turns, using one of four moves - build, question, disagree, propose. Rotate the moves so he is not stuck on one. Then swap roles and make Sofia practise yielding: 'What do you think?' after every second idea. Both halves of the pair need the skill, and Sofia's is the harder one to learn.