The method
You already know the French. The lab makes it arrive on time.
A drill does not teach a structure. It takes a structure a lesson introduced and turns it over, out loud, under a clock, until you produce it without translating. That is a different job from a lesson, and no lesson app does it.
The four phases
Hear, answer, hear, repeat.
- 01
Écoutez
Hear the question
Tu as fait la valise ?
- 02
À vous
Answer out loud in the pause
(you) Oui, je l’ai faite hier.
- 03
Écoutez la réponse
Hear the correct answer
Oui, je l’ai faite hier.
- 04
Répétez
Say it again, matching the voice
(you) Oui, je l’ai faite hier.
Each exercise starts with a model: one exchange, written on the worksheet, that shows the transformation. Then eight to ten variations you have never seen. The pause is sized to the answer: about twice the time it takes the voice to say it, never less than three seconds.
The slow pace lengthens every pause by forty percent. Use it for the first play of a lab and drop it on the second. On a third play the pauses shrink below the first pace, which pushes you to answer faster than comfortable. That is deliberate.
Try one · Modèle 3, the pronoun y
Modèle 1 :
Vous allez à la gare ?
Répondez oui :Oui, j'y vais.
Listen to the model, then answer each variation out loud during the pause.
What it means
"Automatise what you already know"
At B1 you know that the past participle agrees with a preceding direct object. You could explain it. Asked Tu as fait la valise ? in a real conversation, you still say je l'ai fait, because the rule lives in the part of you that reads and the sentence is made by the part of you that speaks. Automatisation is moving the rule from the first part to the second.
It happens through repetition with variation: the same operation applied to many different sentences, with the correct form heard immediately after each attempt. A single drill will not do it. Ten exposures over three weeks, spaced, will. The scheduler in the app keeps count so you do not have to.
Honest limits. A lab will not teach you a structure from zero; you need to have met it. It will not replace conversation; it prepares you for one. It will not fix a pronunciation habit you cannot hear; it will give you many chances to hear it. And if you are below A2, the labs are too fast for you for now.
The evidence
Three findings the lab is built on.
Output matters
Understanding a language does not produce the ability to speak it. Learners have to be pushed to produce, and producing reveals the gaps that listening hides. This is Swain's output hypothesis (1985), confirmed in many classroom studies since.
Retrieval under pressure beats re-reading
Recalling something from memory strengthens it far more than seeing it again. Roediger and Karpicke (2006) showed this for facts; the same effect holds for sentence forms. The pause is a retrieval test with a clock.
Spacing works, and the intervals matter
Cepeda and colleagues (2006, 2008) reviewed decades of results: practice spread over days beats the same practice in one sitting, and the best gap grows as the material settles. Structures come back at one, three, seven and twenty-one days.
References: Swain, "Communicative competence: some roles of comprehensible input and comprehensible output" (1985); Roediger & Karpicke, "Test-enhanced learning" (Psychological Science, 2006); Cepeda et al., "Distributed practice in verbal recall tasks" (Psychological Bulletin, 2006).
The daily programme
Twenty minutes, in this order.
| Block | Minutes | What you do | Why it is there |
|---|---|---|---|
| Listen | 2 | The day's dialogue once, no text. | Ears before mouth; sets the structures in context. |
| Lab | 15 | Six exercises and the dialogue, out loud. | The output practice. The only block that cannot be skipped. |
| Revision | 3 | Six to eight items from structures that are due. | Spacing. Confusable pairs are interleaved on purpose. |
| Shadow | 1 | The dialogue again; speak over the voices. | Rhythm, liaison, speed. |
Compared with
What the others do, and what they leave out.
| Timed spoken production | Transformation drills | Structure scheduler | No speech recognition | Screen-off audio | |
|---|---|---|---|---|---|
| Parlessa | Yes | Yes | Yes (beta) | Yes | Yes |
| Pimsleur | Yes | Translation prompts, fixed script | No | Optional | Yes |
| Glossika | Repeat only | No | Sentence-level | Yes | Yes |
| Kwiziq | No (typed) | Yes, silent | Yes | n/a | No |
| AI conversation apps | Untimed | No | No | No: the mic is the product | No |
Each of these is good at its own job. Pimsleur for a first month, Kwiziq for knowing which rule you keep breaking, a conversation app for the courage to improvise. Parlessa sits between the lesson and the conversation.
Who it is not for
Four honest cases.
- Complete beginners. The labs assume you have met the passé composé and the object pronouns. The A1 course comes after launch.
- Learners who want a score. The lab never grades you; you grade yourself against the recording.
- People who want conversation practice. A lab prepares you for one; it is not one.
- Anyone who cannot practise out loud. Whispering works; silence does not.
Do a whole lab. Then decide.
Laboratoire 1, Le départ: 7 exercises, about fifteen minutes, free.