Listen, Then Speak
You hear a question, get twenty seconds to think, then speak for up to ninety seconds. Two skills at once — and the listening half is the one people forget to train.
- 20s prep on the clock, Several per full test
- Feeds your Listening and Speaking subscores
- Built for the current July 2025 DET format, including the new Interactive Speaking task
Listen, Then Speak plays a spoken question, gives you about twenty seconds to prepare, and then records your answer for somewhere between thirty and ninety seconds.
It is the most realistic task on the test, because it is what conversation actually is: understand something you only heard once, then produce a coherent response in real time. It is also the task with the most ways to go wrong, since a failure at either end — mishearing the question, or running dry after fifteen seconds — sinks the answer.
The trap is that it looks like a speaking task, so people prepare for it by practising speaking. But if you misheard the question, fluent speech about the wrong subject scores badly. The listening half deserves equal attention.
What you actually see
A play button, a preparation countdown, then a recording indicator. The question is spoken, not written — you never see the text of it.
- A spoken question on an everyday topic — habits, opinions, experiences, preferences.
- The question is audio only. There is no transcript on screen at any point.
- About twenty seconds of preparation time after the audio ends.
- Thirty to ninety seconds of recording. Speaking for the full time is not required, but very short answers give the scorer little to work with.
- Several appearances in a full test, with question complexity adapting to your performance.
Twenty seconds, used well
The question you hear:
"Do you prefer to plan your free time in advance, or decide what to do spontaneously?"
What not to do with the prep time: start composing sentence one word by word. You will get half a sentence written and then read it aloud stiffly before running out of material.
What to do instead — three decisions, roughly six seconds each:
- Position. "Mostly plan, but not rigidly." Committing to a position immediately is what stops you circling.
- One concrete example. "Weekends with friends — someone has to book." A specific instance is worth more than three abstract reasons and is far easier to talk about for thirty seconds.
- An exit line. "So I plan the big things and leave the rest open." Knowing how you will finish means you never trail off, which is the most common way these recordings end.
That structure gives roughly forty-five seconds of speech with no memorisation. The answer sounds spontaneous because it is — only the skeleton was decided in advance, which is exactly what the preparation time is for.
How Listen, Then Speak is scored
This task feeds both Listening and Speaking, which makes it unusually valuable and unusually punishing. Answering a question you misheard damages both.
On the speaking side the assessment covers the familiar dimensions: fluency, how wide and accurate your vocabulary and grammar are under pressure, how intelligible you are, and whether the answer is coherent and actually addresses what was asked. Duolingo has not published the weighting between them.
Relevance is the one that catches people. An articulate, well-organised answer to a question that was not asked is scored as a poor answer, because task fulfilment is part of what is being measured. Our own grading engine treats this the same way, and it is the most common reason a fluent-sounding response scores lower than the speaker expected.
| What the scorer looks at | Why it matters |
|---|---|
| Whether you answered the question asked | Relevance gates the rest. Fluent speech on the wrong topic scores as a failed response. |
| Fluency and pace | Continuous, naturally-paced speech. Long silences and frequent self-correction cost you. |
| Range and accuracy of grammar and vocabulary | Under time pressure, without a script. This is what separates B2 from C1 here. |
| Intelligibility | Whether a listener can follow you without effort. Accent is not the criterion. |
| Coherence and development | Whether the answer has a shape — a position, support, a close — rather than being a list of loosely related sentences. |
The mistakes that cost the most
The failures divide cleanly into listening failures and speaking failures, and they need completely different fixes.
Do this
- Listen for the question word first — do you, why, describe, what would. It determines the entire shape of a correct answer.
- Commit to a position in the first sentence. Answers that hedge for twenty seconds before arriving anywhere sound less competent than they are.
- Use one specific example rather than several general reasons. It is easier to sustain and it demonstrates more language.
- Decide your final sentence during preparation so you can land the answer rather than fade out.
- Keep speaking for a reasonable stretch. A fifteen-second answer gives the scorer very little evidence to score generously on.
Not this
- Do not try to write out your answer during preparation. Twenty seconds is not enough and the result sounds recited.
- Do not answer a question you are not sure you understood. Take the shape you did catch and answer that as directly as possible.
- Do not fill silence with memorised filler phrases. Rehearsed-sounding chunks that do not fit the question are noticeable and read as evasion.
- Do not speak until the recording indicator is live — the opening of your answer is where you sound most fluent, and losing it hurts.
- Do not keep talking past the point where you have run out of things to say. Repetition dilutes an otherwise good answer.
How to actually get better at Listen, Then Speak
Practise the two halves separately before you practise them together. Almost everyone has a clear weaker half and does not know which it is.
Diagnose which half is failing
Record an attempt, then check the question. If you answered something adjacent to what was asked, the problem is listening and no amount of speaking practice will fix it.
Drill the three-decision structure until it is automatic
Position, example, exit line. Do ten questions in a row doing only the twenty seconds of planning, out loud, without speaking the answer at all.
Build a stock of concrete examples, not phrases
Six or seven real experiences you can describe in detail will cover most question topics. Examples adapt to questions; memorised phrases do not.
Practise listening to questions once, at speed
Have someone read questions aloud, or use text-to-speech at natural pace. Reading questions on a page trains the wrong skill entirely.
Record, listen back, and time your silences
Count how many pauses over two seconds you produce. That number, not your vocabulary, is usually what is holding the score down.
How much this one task moves your score
This is one of only two task types that feed two individual subscores at once — Listening and Speaking — and through them Comprehension, Conversation and Production. Three of your eight reported numbers move with it.
That makes it high-leverage and high-risk in equal measure. Sustained improvement here shows up across more of your report than almost anything else you could work on.
It is also slower to improve than the objectively-marked reading tasks, because it requires two skills to be available simultaneously under time pressure. Expect it to lag behind Read Aloud, and use Read Aloud to build the delivery that this task then has to deploy without a script.
Listen, Then Speak: questions people ask
Do I have to speak for the full ninety seconds?
No, and padding an answer with repetition is not rewarded. But very short answers give the scorer little evidence to work with, and thin responses tend to score low simply because there is not enough language to assess. Aim to develop your answer properly rather than to fill time.
Can I see the question in writing?
No. The question is audio only, and that is deliberate — the task is testing whether you can understand spoken English in real time and respond, which is why it feeds both your Listening and Speaking subscores.
What if I did not understand the question?
Answer the part you did catch, as directly as you can. A relevant partial answer scores far better than a fluent answer to a different question, because task relevance is part of what is being assessed.
What should I do with the twenty seconds of preparation?
Decide three things: your position, one concrete example, and your closing sentence. Do not try to compose the answer word for word — twenty seconds is not enough, and a recited answer is audible.
Which subscores does this task affect?
Both Listening and Speaking, and therefore the Comprehension, Conversation and Production integrated subscores. It is one of the highest-leverage tasks on the test for that reason. See how the eight subscores combine.
Is Listen, Then Speak still on the Duolingo English Test?
No. It was retired in July 2025, along with Read Aloud, when the speaking tasks were reorganised. We keep this page as a reference, but the task no longer appears in a live test.
What replaced Listen, Then Speak?
Interactive Speaking, which folds the same core demand — understand a spoken prompt and respond in real time — into a longer, more conversational speaking section. The format changed; the skill it measures did not.
Is it still worth practising now that it is retired?
For the underlying skill, yes: reacting to an audio prompt and speaking under a short preparation window is exactly what Interactive Speaking still asks of you. Just do not rehearse it as a scored item, because it will not appear on your test.
How long do I have to prepare?
About twenty seconds after the question audio finishes. That is enough to decide a position, one example and a closing line — not enough to compose sentences, and attempting to produces an answer that sounds recited.
Can I replay the question?
Treat it as heard once. Building the habit of following spoken questions in real time is more reliable than depending on a replay you may not get, and it is the skill the task is measuring.
What kinds of questions come up?
Everyday topics — habits, preferences, opinions, experiences. No specialist knowledge is required, and no question depends on cultural background you might not have.
How long should I speak for?
Long enough to develop the answer properly, which usually means most of the window rather than all of it. Very short answers give the scorer too little language to assess; padding with repetition dilutes what you did say.
Does it matter if my opinion is unusual?
Not at all. Nobody is assessing what you think — they are assessing the English you use to express it. Choose whichever position you can support fastest with a concrete example.
What if I start answering and realise I misunderstood?
Steer back explicitly rather than continuing. Something as simple as "actually, thinking about it differently" costs a second and recovers the relevance that gates the rest of the score.
Why does this task affect two subscores?
Because you have to understand spoken English and then produce spoken English in the same item, so it provides evidence about both. That makes it one of the highest-leverage tasks on the test — and one where a listening failure damages your Speaking score too.
Where Listen, Then Speak sits in a full test
No task on the DET is scored in isolation. The adaptive engine is building one picture of your English out of every response, so it helps to know how much of the hour this particular task accounts for and which of the others are drawing on the same ability.
The table below is the whole test. Your current page is highlighted.
| Question type | Time | How often | Subscores it feeds |
|---|---|---|---|
| Read & Complete | 3 minutes | 3–6 times | Reading, Writing |
| Read & Select | 5 seconds | 15–18 times | Reading |
| Fill in the Blanks | 20 seconds | 6–9 times | Reading, Writing |
| Listen & Type | 1 minute | 6–9 times | Listening, Writing |
| Speak About the Photo | 90 seconds | at least once | Speaking |
| Interactive Speaking | 35 seconds each | 6–7 questions | Speaking, Listening |
| Write About the Photo | 1 minute | at least twice | Writing |
| Interactive Reading | 7 minutes | 2 sets | Reading |
| Interactive Listening | 6 minutes | 2 sets | Listening, Speaking |
| Speaking Sample | 3 minutes | once | Speaking |
| Writing Sample | 5 minutes | once | Writing |
Two things are worth reading off that table. The first is that the tasks feeding Listening and Speaking are not only this one — improvement transfers, so time spent here shows up elsewhere. The second is that the tasks taking the largest share of the hour are not the ones most people practise most.
Question types that use the same muscles
Speak About the Photo
An image appears and you describe it aloud.
Read the guide → SpeakingSpeaking Sample
An extended spoken response to an open prompt. It is sent to institutions alongside your score.
Read the guide → Listening + SpeakingInteractive Listening
A spoken conversation unfolds in turns. At each turn you choose the best reply, then afterwards answer comprehension questions and write a summary of what was said.
Read the guide →Start a practice test
An hour under real conditions, all eight subscores, and feedback on every written and spoken answer.
Take a test