DALF · C2 · CHECKED 2026-09-24

DALF C2

Permanent advanced CEFR diploma

01 · VERIFIED FORMAT

Sections and timing

30 MINUTES

Compréhension et production orales

See the official format for task detail.

Preparation: 60 minutes after two listens
210 MINUTES

Compréhension et production écrites

See the official format for task detail.

02 · ESSENTIAL FACTS

Before you register

Official data · checked 2026-09-24

This diploma is valid for life.

Pass rule: 50/100; minimum 10/50 per component.

03 · BUILD EACH SKILL

Prepare for DALF C2 by transforming complex sources into precise communication

Independent Freniva practice — not an official exam paper

Independent DALF C2 preparation for the integrated oral and written components: faithful reporting, personal development, debate, source synthesis, genre control and precise reformulation.

LISTENING

Build a reusable map across the two hearings

Understand the sound document deeply enough to report its content, expose its problem and support a personal development without distorting the speaker.

  1. During the first hearing, map the situation, speakers, central problem, main claim, stages of reasoning and conclusion. Use short labels rather than attempted transcription.
  2. During the second hearing, attach precise evidence, examples, qualifications, tone and useful formulations to the map; separate what was stated from what you infer.
  3. Use the sixty-minute preparation after the two hearings to turn the same evidence into three distinct jobs: an accurate report, an independent development and material for debate.

L’intervenante ne condamne pas la technologie ; elle conteste l’idée qu’un outil puisse, à lui seul, corriger une inégalité institutionnelle.

The speaker does not condemn technology; she disputes the idea that a tool alone can correct an institutional inequality.

A C2 report preserves the speaker’s distinction instead of reducing the position to 'for' or 'against'.
SPEAKING

Move from report to position to debate without blending them

Deliver the three oral parts coherently in a thirty-minute examination: content report, personal development from the problem, and debate with the jury.

  1. In the report, reconstruct the document faithfully and signal its organisation. Do not insert your opinion or reproduce a chronological transcript.
  2. In the personal development, reformulate the problem, state a defensible position and develop it with distinctions, examples, consequences and a justified limit.
  3. In debate, answer the jury’s exact premise first. Concede where warranted, expose an assumption, refine your scope and reconnect the answer to your position instead of repeating the prepared monologue.

Votre objection serait décisive si l’efficacité était notre seul critère. Or la mesure engage aussi l’égalité d’accès ; je maintiens donc le principe, tout en limitant son application.

Your objection would be decisive if efficiency were our only criterion. But the measure also concerns equal access; I therefore maintain the principle while limiting its application.

The answer identifies the objection’s assumption, adds a competing criterion and qualifies the proposal.
READING

Turn a 2,000-word dossier into an evidence system

Read across multiple documents, identify relationships and select only the source material needed for the required written genre.

  1. Identify each document’s genre, author position, purpose, evidence and limits, then build a cross-document table of agreements, tensions, causes, consequences and proposals.
  2. Group evidence by the problem you will address rather than by document order. Track source ownership so a striking idea is not silently presented as your own.
  3. Use the permitted monolingual dictionary selectively for decisive meaning or register questions; do not let dictionary searches replace structural reading.

Les données du rapport confirment l’évolution décrite dans l’entretien, mais elles n’établissent pas le lien de causalité que l’éditorial tient pour acquis.

The report data confirm the trend described in the interview, but they do not establish the causal link that the editorial takes for granted.

The sentence compares the evidentiary force of three source elements instead of merely listing them.
WRITING

Choose a genre, a reader and a line of argument

Produce a structured article, editorial, report or speech from the dossier while integrating sources, controlling voice and fulfilling the genre’s real purpose.

  1. Define who is writing, for whom, why and with what expected effect. Let those choices control title, opening, register, organisation and conclusion.
  2. Formulate one governing problem and a precise line of argument. Integrate dossier evidence through reformulation, comparison and attribution rather than stitched quotation.
  3. Revise on four levels: task and genre, argument and evidence, cohesion and reference, then grammar and lexical precision. Check the exact instructions on the paper rather than relying on a memorised template.

Ces résultats invitent moins à abandonner le dispositif qu’à redéfinir les conditions dans lesquelles il peut remplir sa promesse.

These results argue less for abandoning the scheme than for redefining the conditions under which it can fulfil its promise.

The formulation interprets evidence, rejects a false binary and prepares a qualified recommendation.
04 · SEE A COMPLETE ORIGINAL MODEL

A complete response that develops every instruction

Original DALF C2-style editorial model: A dossier examines whether cities should use predictive digital systems to allocate public services. Write an editorial for an informed general readership. Use the issue and evidence relationships without pretending this is an official paper.

Prédire pour mieux servir : la formule résume l’ambition des systèmes numériques auxquels certaines villes confient désormais une partie de leurs décisions. En croisant demandes d’aide, données de transport ou signalements, ces outils promettent d’identifier un besoin avant qu’il ne devienne une crise. Leur utilité n’est pas imaginaire. Une équipe limitée peut repérer plus vite un quartier mal desservi, planifier une intervention ou tester plusieurs scénarios. Encore faut-il distinguer une aide à la décision d’une décision rendue incontestable par son apparence mathématique. Les données municipales ne décrivent jamais la population de façon neutre. Elles enregistrent plus facilement les besoins des habitants qui savent signaler un problème, possèdent une connexion et font confiance à l’administration. Un quartier peu visible peut donc recevoir moins de ressources précisément parce que ses difficultés ont été moins documentées. Le système risque alors de transformer une absence de données en absence de besoin. Refuser tout calcul automatisé serait pourtant une réponse trop simple. Les décisions humaines sont elles aussi traversées par l’habitude, l’urgence et des biais rarement mesurés. La vraie question concerne donc les conditions de contestation. Une ville responsable devrait publier les critères utilisés, vérifier régulièrement les écarts entre quartiers, conserver une voie de recours accessible et donner aux agents le pouvoir de contredire une recommandation motivée par des données incomplètes. Elle devrait surtout évaluer le résultat concret : le service atteint-il mieux les personnes concernées, ou produit-il seulement une distribution plus facile à administrer ? Un outil prédictif mérite sa place lorsqu’il rend l’action publique plus attentive et plus révisable. Il devient dangereux lorsqu’il remplace la justification politique par un score que personne ne peut expliquer. La modernité d’une ville ne se mesure pas à la quantité de calcul qu’elle automatise, mais à sa capacité de rendre chaque décision compréhensible, discutable et corrigible.

Predict in order to serve better: the phrase sums up the ambition of digital systems to which some cities now entrust part of their decision-making. By combining requests for help, transport data or reports, these tools promise to identify a need before it becomes a crisis. Their usefulness is real. A limited team can spot an underserved neighbourhood sooner, plan an intervention or test several scenarios. But decision support must be distinguished from a decision made indisputable by its mathematical appearance. Municipal data never describe the population neutrally. They record more easily the needs of residents who know how to report a problem, have a connection and trust the administration. A less visible neighbourhood may therefore receive fewer resources precisely because its difficulties were less documented. The system can then transform an absence of data into an absence of need. Rejecting all automated calculation would nevertheless be too simple. Human decisions are also shaped by habit, urgency and rarely measured biases. The real question is therefore the conditions for challenge. A responsible city should publish the criteria used, regularly check disparities between neighbourhoods, preserve an accessible appeal route and let staff contradict recommendations based on incomplete data. Above all, it should evaluate the concrete result: does the service reach the people concerned more effectively, or does it merely produce a distribution that is easier to administer? A predictive tool deserves a place when it makes public action more attentive and revisable. It becomes dangerous when it replaces political justification with a score nobody can explain. A city’s modernity is measured not by how much calculation it automates, but by its ability to make every decision understandable, debatable and correctable.

  • The opening establishes the issue and intended readership without a mechanical plan announcement.
  • The argument accepts real utility before identifying the evidentiary problem.
  • A false choice between algorithms and human judgement is rejected.
  • The recommendation turns transparency, appeal and outcome evaluation into concrete conditions.
  • This is original independent Freniva material, not an official response or score guarantee.
05 · TRY

Read for proof, then make the decision

Original source-relation practice: measuring public success

Une administration peut atteindre son objectif chiffré tout en manquant son objectif réel. Réduire le délai moyen de réponse paraît positif ; si les dossiers complexes sont simplement fermés plus vite, l’indicateur récompense pourtant une dégradation du service. Une mesure utile doit donc être accompagnée d’un contre-indicateur capable de révéler ce qu’elle encourage à négliger.

An administration can meet its numerical target while missing its real objective. Reducing average response time appears positive; yet if complex cases are simply closed faster, the indicator rewards a deterioration in service. A useful measure must therefore be paired with a counter-indicator capable of revealing what it encourages people to neglect.

1What is the text’s central distinction?
2What does pourtant introduce?
3Which oral report remains faithful?
4Which debate move is strongest?
5Which written plan fits an editorial?
WRITE

From a 2,000-word practice dossier, write one complete article, editorial, report or speech. Define the reader and purpose, integrate and attribute evidence, distinguish correlation from causation, answer a serious objection and end with a genre-appropriate outcome.

SPEAK

After two hearings, use sixty minutes to prepare a faithful report, a separate personal development and debate notes. Deliver all three stages, then answer four objections by conceding, distinguishing, reformulating or narrowing where justified.

06 · CORRECT

Repair the reason, not only the sentence

Before: Le document dit que les chiffres sont mauvais.

After: Le document montre qu’un indicateur isolé peut récompenser un résultat contraire à l’objectif du service.

The revision restores the precise conditional criticism.

Before: Malgré que le système soit rapide, il reste opaque.

After: Bien que le système soit rapide, il reste opaque.

Use bien que plus subjunctive in standard formal French; malgré takes a noun phrase.

Before: Les données permettent de mieux décider et que les erreurs diminuent.

After: Les données permettent de mieux décider et peuvent réduire certaines erreurs.

Coordinate compatible structures and avoid claiming an unsupported automatic effect.

Before: Tous les auteurs sont d’accord, donc mon avis est correct.

After: Les documents convergent sur le diagnostic, mais leurs critères de solution restent différents.

Source convergence is not proof of the candidate’s opinion; compare the actual relationship.

Before: Je ne réponds pas à votre objection parce que mon exposé est terminé.

After: Votre objection porte sur la fiabilité des données ; je maintiens le principe à condition qu’un recours humain reste possible.

The debate requires a direct answer and a justified qualification.

07 · REVIEW

Retest the exam decisions

  • Can I map the document’s structure, stance, qualifications and evidence across two hearings?
  • Can I report the source faithfully before introducing my own position?
  • Can I develop the source problem independently rather than repeating the recording?
  • Can I compare dossier evidence by relationship and evidentiary strength?
  • Does my writing fulfil a real genre, reader and communicative purpose?
  • Can I answer an objection directly and refine my scope without abandoning coherence?
  • Do I use the monolingual dictionary selectively rather than as a substitute for reasoning?
  • Have I checked current FEI material and my test centre’s instructions before registering?
08 · PRIMARY SOURCES

Check the authority

CHECKED 2026-09-24

Droits et devoirs des candidats au DELF-DALF

France Éducation international

Open official source
INDEPENDENT PREPARATION

Test one skill, then study the error.

Freniva practice is original learning material, not an official exam simulation.

Start practice