Linguaskill Speaking Parts 3 and 4: Satisfying a Brief

Our earlier post on this module showed that the audience changes between Part 1 and Part 2 without the test announcing it. Parts 3 and 4 continue that pattern, and they add something the first half never asks for.

Part 3 hands you somebody else's requirements and marks you on meeting them.

Then Part 4 takes the person away again. SpeakShark is free if you want to practise talking to a brief rather than at a topic, three AI conversation sessions a day with no card.

In this guide: the brief · the message format · how to answer Part 3 · Part 4 · the conclusion · the audience arc · what is not published · practice · method · FAQ

Key takeaways

Part 3 is a recommendation task. You see a friend's or colleague's requirements in a text message format, read supplementary information, then recommend something that meets them.
It is the only task in the module with two inputs to reconcile, and the only one with external criteria for being right.
Part 4 is argument presentation and evaluation. A statement, arguments both for and against, then a conclusion with reasons.
The conclusion is named in the task description, so an answer that argues well and stops is incomplete.
Cambridge publishes durations for Part 1 and Part 2.1 only. Parts 2.2, 3 and 4 have no published per part timing.
The audience appears in the middle of the module and disappears again by Part 4.

Part 3 is the only task with a brief

Read Cambridge's descriptions of the five sections next to each other and one of them is doing something the others are not.

Section	What you are given	What decides a good answer
Part 1 Presentation	A topic and three bullet points	Covering the bullets for the minute
Part 2.1 Summary	An audio extract with four key points	Compressing four points into forty seconds
Part 2.2 Discussion	Three follow up questions	Your opinions, connected to the summary
Part 3 Recommendation	Someone's requirements plus supplementary information	Whether your recommendation meets their requirements
Part 4 Argument	A statement	Both sides, then a conclusion with reasons

Only Part 3 sets a target that belongs to somebody else. In every other section the material is yours to shape. Here the friend or colleague has already said what they need, and a recommendation that ignores it is off task no matter how well it is delivered.

This is the trap for confident speakers. Fluency carries you through Part 1 and Part 2.2 comfortably. In Part 3 fluency without accuracy to the brief sounds impressive and answers the wrong question.

The format the requirements arrive in

That Part 3 synthesises two inputs is covered in our format hub. The detail worth adding is the shape the first input takes.

Cambridge says the requirements arrive in a text message format. Not a brief, not a specification. A message.

Requirements written as a message are casual and compressed, and they routinely carry more than one condition inside a single sentence. Someone asking for somewhere cheap near the station that is open late has given you three conditions in eleven words.

So Part 3 opens with a reading comprehension problem wearing everyday clothes, inside a speaking test, against a clock you cannot see because Cambridge publishes no duration for this section. Misreading the message is a quieter way to fail than running out of language.

How to answer Part 3 without drifting

Three habits, in order of how much they protect the mark.

Count the requirements before you speak. A message with two conditions and a preference has three things to satisfy. Knowing the number stops you satisfying one and improvising the rest.
Name each requirement as you meet it. Saying that something suits them because it fits the condition they gave makes your alignment audible instead of implied.
Choose, then justify. A recommendation is a decision. Listing options without landing on one leaves the task unfinished.

The same discipline pays in every task that gives you a stated brief, which is why it transfers to OPIc, to the TOEIC speaking test and to CELPIP speaking tasks, all of which reward answering the question asked over speaking well in general.

If your English is comfortable but your answers wander, this is the habit that moves you. Practising against a brief is what the free plan is for: three conversation sessions a day with no card, and the pricing page sets out what the paid tier adds.

Part 4 a statement and nobody to say it to

Part 4 gives you a statement. You consider arguments for and against it, then reach a conclusion with reasons supporting your final argument.

There is no listener in the description. No colleague, no friend, no examiner asking anything. After two sections built around a specific person, the audience simply goes away.

That changes how the answer should sound. Part 2.2 is opinion delivered to somebody who asked. Part 4 is an argument built for its own sake, closer to a short spoken essay than to a conversation. Candidates who carry the chatty register from 2.2 into 4 often produce something that sounds friendly and argues thinly.

Cambridge also does not say the statement will be one you agree with. Considering arguments both for and against is the instruction, so having a personal view is not the same as having done the task.

The conclusion is part of the task

This is the smallest sentence in the description and the easiest mark to leave behind.

Cambridge writes that candidates should reach a conclusion, providing reasons to support their final argument. Not summarise. Not weigh up. Reach one, and support it.

So an answer that presents three arguments for, three against, and then stops has done most of the work and skipped the part that was named explicitly. It is the spoken equivalent of an essay with no final paragraph.

Two practical points follow. Leave time for it, because a conclusion that arrives after the recording stops does not exist. And make it a decision rather than a hedge: saying that both sides have merit is a description of the arguments, not a conclusion drawn from them.

The habit of landing an argument is the same one assessed across Cambridge speaking exams, which our post on speaking criteria by level sets out, and on what examiners write about interactive communication.

The audience arc across the module

Put the five sections in order and the person you are speaking to appears and vanishes.

Section	Who is listening
Part 1 Presentation	Nobody is named
Part 2.1 Summary	A colleague or classmate who was absent
Part 2.2 Discussion	The same person, now asking
Part 3 Recommendation	A friend or colleague with requirements
Part 4 Argument	Nobody is named

The module starts alone, spends three sections with a specific person, and ends alone. Our post on Parts 1 and 2 covers the first half of that arc in detail.

Practically, that means your register should move twice, and a candidate who sounds identical in all five sections is misreading two of them. It also means the two ends of the test, Part 1 and Part 4, share a problem that the middle does not: nothing comes back at you, so nothing rescues a stalled sentence.

Tests where a human sits opposite you work differently, which is worth knowing if you have prepared for Trinity GESE or an IELTS style interview. Recorded formats such as Versant and Aptis share Linguaskill's silence.

What Cambridge does not publish

Two gaps, stated rather than filled.

Per part timings for Parts 2.2, 3 and 4. The exam formats and task types page gives one minute for Part 1 and forty seconds for Part 2.1, and no duration for the other three sections. Sixteen minutes covers the whole module including listening to prompts, reading on screen and preparation, so the total does not divide cleanly into speech.

Add up what is published and only one minute forty seconds of the sixteen has a stated task duration. Anyone giving you exact seconds for Part 3 is inferring, and we are not going to join them.

How hybrid marking divides. Cambridge states that human examiners and auto marking technology assess the module together and does not publish which criteria each contributes to. Our post on the two examiner roles in Cambridge speaking exams covers the face to face equivalent, which is a different arrangement.

How to practise Parts 3 and 4

Invent a brief, then satisfy it. Write three requirements on a card, read them once, and recommend something aloud that meets all three while naming each.
Argue a statement you disagree with. Both sides, then a conclusion, out loud, in one take.
Practise the last ten seconds. End on a decision, not a fade. This is the most improvable part of Part 4.
Alternate register deliberately. One task to a person, one task to nobody, back to back, so the switch stops feeling strange.

Drills two and four need something that answers back or they stay theoretical. That is what SpeakShark is built for: unscripted conversation with native accent AI teachers, plus instant phoneme level feedback on the sounds that keep tripping you. For the honest picture of what tools like ours do and do not do, we keep an open comparison of speaking apps and a plain how it works page.

Related Cambridge and recorded test reading: Oxford Test of English, LanguageCert academic versus general, MET speaking task timings, the pronunciation myth in Cambridge scores, and China's CET oral test.

How we researched this guide

Every description of Part 3 and Part 4 is quoted or closely paraphrased from Cambridge English's own exam formats and task types page for Linguaskill, opened in a browser with the Speaking tab expanded rather than summarised from a preparation site. The module length, the five sections, the unmarked warm up and the hybrid marking statement come from the same page. Where Cambridge gives a number we use theirs, and where they publish none we say so above instead of inferring one.

We are a speaking improvement tool. We are not an exam preparation provider, and we are not affiliated with Cambridge English or any exam board. For test format, booking and official practice material, go to the exam body directly. Use SpeakShark to make your spoken English stronger, and use official material to learn the test.

Site: speakshark.com
Support: speaksharksupport@gmail.com
Free tier: 3 AI conversation sessions per day, no card

Sources

FAQ

What happens in Linguaskill Speaking Part 3?

Cambridge describes it as a recommendation task. You see a friend's or colleague's requirements in a text message format, you read supplementary information that is given to you, and you then make a recommendation that meets those requirements. It is the only task in the module with two separate inputs to reconcile, one stating what the person wants and one giving you the material to choose from, and it is the only task where somebody else's stated criteria decide whether your answer is on target.

What happens in Linguaskill Speaking Part 4?

Cambridge describes it as argument presentation and evaluation. You are given a statement and you have to consider arguments both for and against it, then reach a conclusion and provide reasons that support your final argument. Two things follow from that wording. Presenting only one side is an incomplete answer even if it is fluent, and stopping after the arguments without landing a conclusion leaves out something the task explicitly asks for.

How long are Linguaskill Speaking Parts 3 and 4?

Cambridge does not publish per part timings for Part 2.2, Part 3 or Part 4 on its exam formats and task types page. What it publishes is that the whole module is sixteen minutes across five sections, that Part 1 is one minute, and that Part 2.1 gives forty seconds. We are not going to infer the rest, because the sixteen minutes also covers listening to prompts, reading material on screen and preparation time, not just speech.

Is Part 3 about being fluent or about being accurate to the brief?

Both are marked, but only one of them is the task. The requirements are stated in the prompt, so a recommendation that ignores them is answering a question nobody asked, however smooth the delivery. In practice the highest value habit is to name the requirements back as you satisfy them, which keeps you on target and makes it audible that you have understood the brief rather than simply talked around the topic.

Who are you speaking to in Parts 3 and 4?

In Part 3 there is a named person, a friend or colleague whose requirements you are meeting. In Part 4 there is nobody at all, only a statement to argue about. The module opens the same way, with a presentation that names no listener, so the audience appears in the middle of the test and disappears again at the end. Candidates who keep the same register throughout are answering two of the five sections in the wrong voice.

How is Linguaskill Speaking marked?

Cambridge states that Linguaskill Speaking and Linguaskill Business Speaking are assessed by human examiners together with auto marking technology, which it calls hybrid marking. No person is in the room while you speak, because the module is taken at a computer with a microphone and headphones and your responses are recorded, but a human examiner is involved in the score. Cambridge does not publish how the two contributions divide.