Fifteen minutes, four parts, two candidates, under exam conditions. Record each part in one take, without stopping to think and without reading the notes first — the notes are the debrief, for afterwards. Then listen back against the model, read what the examiners were listening for, and record the part once more.
Under exam conditions Speaking is timed as the real test is: record each part in one take, without stopping to think, then listen back against the model. There is nothing to hand in — the recording is the paper.
The interlocutor asks each of you a few questions in turn. Answers are short — two or three sentences — but never one word. This part is not assessed heavily; it is where the examiners tune their ear to your voice, so speak up and do not warm up slowly.
Question 1, at the top of the band — about 25 seconds.
“In Porto, in a flat above a bakery, which I'd have found hard to believe at twenty. I assumed I'd be in Lisbon by now, doing something with a title, and instead I've stayed somewhere smaller and stopped minding. What I hadn't predicted was that the staying would be the decision — I never chose Porto, I just kept not leaving it.”
Why this scores. Both halves of the question are answered, in that order. The expectation is set up and reversed in one sentence, which gives the answer a shape. The last sentence does something a B2 answer never does — it notices something about itself — and the structures are doing that work: I'd have found, What I hadn't predicted was. Three sentences. It does not go on.
You are each given three photographs and asked to talk about two of them for a minute. There are two questions printed above the pictures and you must answer both — this is not a description task. When your partner has finished, you are asked a short question about their pictures and you have about thirty seconds. The photographs are described in words here.
Candidate A — three photographs of people working outdoors.
1 A fisherman on a harbour wall at dawn, mending a net across his knees; the boats behind him are still tied up.
2 A woman in a high-visibility jacket on a building site, holding a tablet, looking up at a crane; the ground is mud.
3 A market trader under a striped awning in heavy rain, arranging vegetables; there are no customers in the frame.
Candidate B — three photographs of people learning something difficult.
1 A driving lesson: a teenager at the wheel, both hands rigid, an instructor beside him with one hand near the brake.
2 An adult swimming lesson: a woman in her sixties holding the side of the pool, a young instructor in the water beside her, both laughing.
3 A pottery class: a man at a wheel, clay collapsing sideways, a teacher standing behind him with her arms folded, not intervening.
Candidate B's long turn, pictures 1 and 3, at the top of the band — about 60 seconds.
“I'll take the driving lesson and the pottery, because the teachers are doing opposite things. Driving is hard to learn for an obvious reason — a mistake has consequences — and you can see that in his hands; he's gripping the wheel as though it might leave. The instructor's hand is near the brake, and I'd say that's the whole relationship: the teacher is there to make the mistake survivable, and the learner presumably knows it, which is probably the only reason he's able to drive at all.
The pottery is difficult for a different reason — nothing happens if you get it wrong except that it collapses, which it's doing. And the teacher is standing with her arms folded, which looks unkind and I suspect isn't. She may well have decided that the collapse is the lesson. So the teacher matters in both, but in the car she matters by intervening and at the wheel she matters by refusing to — and if I had to say which is harder to be taught by, it's the second.”
Why this scores. It answers both questions, and it takes the second one seriously — it judges each teacher and finds them different. It speculates in every sentence — I'd say, presumably, I suspect, may well have — which is what the task asks for. It picks one detail per picture (the hands, the folded arms) and builds the turn on them. And it closes with a judgement, so the examiner hears an ending rather than a stop.
You are given a written prompt with a question in the middle and five options around it. You talk together for about two minutes about the options; then the interlocutor asks you to decide something, and you have one minute to reach agreement. It is a conversation, not two speeches — and the decision minute is assessed on whether you actually negotiate.
How important is it for a country to preserve these things?
· regional accents
· traditional crafts
· historic buildings
· local dishes
· folk songs and dances
First (2 minutes): talk to each other about how important it is for a country to preserve each of these.
Then (1 minute): decide which one of them is most at risk of being lost.
Forty seconds from the middle of a strong collaborative turn.
A: “I keep wanting to say accents, and then I stop, because I'm not sure you can preserve an accent — you can record it, but the moment somebody's taught to speak it, it's a performance. Does that seem right to you, or am I being too strict about the word?”
B: “No, I think that's exactly the difficulty, and it's true of the songs as well — a folk song kept alive by a committee isn't quite a folk song. Whereas a building either stands or it doesn't. I'd have thought that's why buildings get the money: they're the only thing on the list you can actually save.”
A: “Go on — so is that a reason to prioritise them, or a reason to worry about the others more?”
B: “The second, I think, having said the first. The buildings will look after themselves because they're easy to argue for. It's the crafts I'd be frightened for — one generation not learning it and it's gone, and nobody notices until it has.”
A: “That's persuaded me. Shall we take crafts as most at risk and argue about whether the songs are second?”
Why this scores. Every turn is short and every turn acts on the previous one — extends it, questions it, changes position because of it. Both candidates hedge and then commit (I'd have thought · The second, I think, having said the first), and B openly reverses, which is worth more than defending a first opinion. The last line moves the pair towards the decision without cutting the discussion off.
The interlocutor asks broader questions arising from Part 3. This is the longest part, it is the most abstract, and it is where C1 is genuinely tested: you are being asked to hold a position, qualify it, and defend it against a counter-argument — sometimes your partner's, sometimes the examiner's.
Question 2, at the top of the band — about 45 seconds.
“Let it change, and I'd accept what that costs, which is that some of it will change into something its own grandparents wouldn't recognise. If you fix a tradition in place, you've turned it into a museum piece — it's preserved in the way a specimen is preserved, which is to say it's dead and well lit.
But I'd want to be honest that changing isn't free either. A festival that adapts to whatever the current audience wants can end up as an advertisement for itself, and at that point nobody can say what it was for. So I'd let it change on one condition, which is that the people changing it are the people it belongs to. A tradition altered by its own practitioners is alive; a tradition altered by a tourist board is a product. The first can become something new and still be itself. The second can only become something else.”
Why this scores. It answers in the first three words and never wavers. It states the cost of its own view before the examiner can, and then narrows the claim with a condition instead of abandoning it. The language is doing work — a cleft, a conditional, a pair of parallel sentences at the end that make the distinction audible. And it ends on a distinction rather than a summary, which is the difference between an opinion and a position.
The one habit to take from this. At C1 you are not rewarded for balance. You are rewarded for a position that has been through the counter-argument and come out the other side. Say what you think, say what it costs, say why you'd pay it.
Play each part back once with the model in front of you, and listen for one thing only at a time. Listening for everything at once tells you nothing.
Pass 1 — did I answer the question that was asked? Pass 2 — did I develop, or did I stop? A C1 answer gives the point and then the reason, every time. Pass 3 — did I commit? Part 4 is marked on whether you can hold and defend a position, not on whether you can list both sides. This is the last mock: if one pass fails on one part, that part is the week's work, and nothing else is.