CPE Speaking Part 2: How to Work With Your Partner and the Pictures

The short answer

CPE Speaking Part 2 is a four-minute conversation with your partner, built around a set of pictures. It looks like the easiest part of the test, and that is the trap. Strong candidates use the pictures as shared material: they point to them, refer to them, connect them to each other and to what their partner says, and work towards a decision together. Weaker candidates glance at the pictures once, then talk in parallel, and the examiner hears two people with good English who never quite have a conversation.

What Part 2 actually asks you to do

Part 2 lasts about four minutes for a pair. The examiner gives you a set of pictures on a theme. First, you talk together for about one minute about one or two of the pictures, usually reacting to them or speculating about them. Then the examiner gives you a longer task, about three minutes, in which you discuss all the pictures and reach a decision, for example choosing which would work best for a particular purpose (Cambridge English teacher guide for C2 Proficiency Speaking).

Cambridge describes what candidates should do as sustaining an interaction, exchanging ideas, expressing and justifying opinions, agreeing and disagreeing, suggesting, speculating, evaluating and reaching a decision through negotiation. Notice how many of those depend on another person.

Talking past the pictures

The pattern I see most often in CPE Part 2 is a lack of visual interaction. The candidates look at the pictures for a second, then look at the table, the examiner or into the distance. Neither of them points to a picture or says “this one here”. Each candidate gives a well-formed opinion about the theme in general, and the pictures, which are the whole point of the task, quietly disappear.

At C2 this matters, because the language is usually not the problem. Most CPE candidates can speculate and evaluate with real sophistication. What gives them away is that the conversation isn’t anchored to anything. Without the pictures, the discussion becomes abstract and repetitive, and without looking at each other, it becomes two monologues. It is the same habit I describe in CAE Speaking Part 3, only at a higher level, where examiners expect it to have disappeared.

What the examiner is listening for

At the top band for Interactive Communication, the examiner is listening for a candidate who interacts with ease, links their contributions to those of other speakers, widens the scope of the discussion and negotiates towards an outcome. At C2, they expect this to happen naturally, not as a technique.

In terms of the Four Pillars, Part 2 tests Flexibility (responding to an idea you didn’t expect and developing it) and Natural Expression (sounding like two people genuinely deciding something together). The pictures are your best tool for both.

Use the pictures as shared ground

The pictures are not a prompt you read once. They are the thing on the table between you, and the most natural way to build a conversation is around something you can both see.

Point and refer

Physically point, or name the picture clearly: “This one, with the crowded market…” It tells your partner exactly what you mean and keeps you both anchored.

Connect pictures to each other

“What links these two is…”, “This one is almost the opposite of that.” Comparison across pictures widens the scope without changing the subject.

Look at your partner, not the examiner

Turn slightly towards them. Watch their reaction as you speak. The examiner is listening, not taking part.

Invite a response about a specific picture

“Which of these would you rule out?” is easier to answer, and more natural, than “What do you think?”

The three minutes: reaching a decision together

The longer task always ends in a decision, and many pairs either rush to it or never reach it. A simple rhythm works well.

Explore first, for about two minutes. Go through a few pictures together, building on each other’s ideas. Don’t decide too early.

Then narrow down: “I think we can rule these two out, can’t we?”

Then negotiate. Weigh the remaining options, disagree if you genuinely do, and reach a decision you can both justify: “We see it slightly differently, but if we had to choose…”

It doesn’t matter what you decide. It matters how you get there.

Before and after: the same task, rebuilt

The task: pictures of different ways of spending free time. “Talk together about how these activities might affect people’s wellbeing, then decide which one would be most beneficial for someone with a stressful job.”

Before

A: “In general I think free time is very important for people’s mental health, because modern life is extremely demanding and people need to switch off.”

B: “Yes, I would agree. Nowadays many people suffer from stress, and leisure activities can help them to relax and recharge their batteries.”

Both turns are fluent and accurate. Neither mentions a single picture, and neither responds to the other. It could be any task.

After

A: “If we look at this one, the people hiking, I’d say it’s probably the most obviously restorative, because you’re outdoors and away from screens. Would you agree?”

B: “Up to a point, but I’m not sure it suits everyone. For someone who’s exhausted, this one here, reading at home, might actually be more realistic. A long hike could feel like another demand.”

A: “That’s a fair point. So maybe it’s less about the activity and more about how much energy it asks of you?”

The language level barely changes. What changes is that every turn is anchored to a picture, picks up the previous turn, and moves the idea forward. By the third turn, the pair has found a genuinely interesting angle together, which is what widening the scope sounds like.

How to practise it

You need pictures and, ideally, a partner. Any four or five photos on a theme will do.

With a partner, set a four-minute timer. Every turn must name or point to a picture. Afterwards, count how many turns responded directly to the one before.

On your own, speak for one minute about two pictures, then pause and imagine your partner disagreed. Answer that disagreement out loud.

For the decision, practise the rule-out step until it’s automatic. It stops the rushed final ten seconds.

In my sessions, I take the role of your partner as well as the examiner. I respond the way a real partner might, including disagreeing or taking the conversation somewhere unexpected, so you practise genuine interaction rather than a rehearsed exchange.

The bigger picture

Talking past the pictures is one of the patterns I see most often in CPE candidates, and it is almost never a language problem. It comes from preparing for the test alone.

If you’d like to find out which patterns are affecting your speaking, the Initial Assessment is a 60-minute diagnostic session followed by a written Speaking Performance Profile. For the rest of the test, read CPE Speaking Part 3: How to Structure Your Two-Minute Long Turn and the CPE speaking preparation guide, or find out more about how I coach CPE candidates.

Frequently asked questions