9.4 The Official Production Orale Grid & Examiner Expectations

9.4 The Official Production Orale Grid & Examiner Expectations

[!NOTE] You are marked on a public document. Unlike the comprehension papers, which are scored against a closed answer key, the production orale is scored by two accredited examiners filling in a grid that France Éducation international publishes openly. Candidates who have never read that grid spend their preparation on the wrong things: they polish vocabulary lists while three of the five scoring lines reward behaviours — structuring, nuancing, defending — that can be rehearsed directly.

The oral is worth 25 points, exactly like the other three papers, and carries the same 5/25 eliminatory floor. The grid attached to the current format distributes those 25 points across five criteria, each scored on four performance levels.

---

The Five Criteria and the Four Performance Levels

┌────────────────────────────────────────────────────────────────────────┐
│      FEI DELF B2 ORAL PRODUCTION GRID (renovated) — 5 x 5 = 25          │
├────────────────────────────────────────────────────────────────────────┤
│ PRAGMATIC & SOCIOLINGUISTIC COMPETENCE                                 │
│ • Réalisation de la tâche : monologue suivi (5 à 7 minutes)      : /5  │
│ • Réalisation de la tâche : exercice en interaction (10 à 13 min): /5  │
│                                                                        │
│ LINGUISTIC COMPETENCE (assessed across BOTH phases together)           │
│ • Lexique (étendue et maîtrise)                                  : /5  │
│ • Morphosyntaxe                                                  : /5  │
│ • Maîtrise du système phonologique                               : /5  │
├────────────────────────────────────────────────────────────────────────┤
│ PERFORMANCE LEVEL APPLIED TO EVERY CRITERION                           │
│ 0 = non répondu / production insuffisante                              │
│ 1 = en dessous du niveau ciblé (performance reads like B1)             │
│ 3 = au niveau ciblé B2                                                 │
│ 5 = B2+ (solid or superior command)                                    │
├────────────────────────────────────────────────────────────────────────┤
│ TOTAL                                                            : /25 │
└────────────────────────────────────────────────────────────────────────┘

Two structural consequences follow immediately, and both should reshape how you prepare.

First, the debate is worth as much as the exposé. Many candidates rehearse the monologue obsessively and improvise the interaction. The grid gives each task exactly 5 points. Ten silent minutes of poorly defended positions cost precisely as much as a collapsed presentation.

Second, 15 of the 25 points are language points assessed over the whole session. Lexique, Morphosyntaxe and Maîtrise du système phonologique are not scored separately for each phase: examiners judge them across the exposé and the debate. A candidate who delivers a beautifully rehearsed monologue and then reverts to broken syntax during the debate loses on three lines at once, not one.

[!IMPORTANT] The 3 → 5 gap. "Au niveau ciblé B2" is 3 points; "B2+" is 5. The distance between a script that merely satisfies B2 and one that exceeds it is 2 points per criterion — up to 10 points on the paper. That gap is not made of better ideas. It is made of precision, nuance and consistency.

---

Criterion 1 — Réalisation de la tâche : le monologue suivi (5 à 7 minutes)

The examiners are looking for three observable behaviours, inherited from the analytic descriptors that the renovated grid condenses:

The recurring below-level profile named in FEI examiner training is unmistakable: the subject is badly presented, the position taken is too simple, and the talk lacks organisation and clarity. All three are preparation failures, not language failures — which is why the 30-minute preparation protocol in section 8.3 pays for itself twice over.

---

Criterion 2 — Réalisation de la tâche : l'exercice en interaction (10 à 13 minutes)

This criterion also condenses two analytic descriptors:

1. « Peut confirmer et nuancer ses idées et ses opinions, apporter des précisions. » You must be able to re-say an idea differently, add a qualification, and supply the detail the examiner asks for.
2. « Peut réagir aux arguments et déclarations d'autrui pour défendre sa position. » You must engage the objection rather than absorbing it or ignoring it.

The below-level descriptor is equally blunt: the candidate answers the debate questions too simply. Monosyllabic or one-clause answers score 1, however correct the French. The B2+ descriptor is the opposite: the candidate brings new ideas and is convincing.

[!TIP] The three-move answer. Never answer an objection in one clause. Use: (1) acknowledge — « Vous avez raison de souligner que… »; (2) qualify — « … cela dit, il convient de distinguer… »; (3) re-anchor — « … c'est précisément pourquoi je maintiens que… ». This single habit moves the interaction criterion from 3 to 5 more reliably than any vocabulary list.

---

Criteria 3–5 — The Language Lines, Assessed Across the Whole Session

Note where the phonological criterion sits: it is worth 5 points, the same as your entire monologue task. A candidate whose every sentence demands decoding effort loses a fifth of the paper regardless of grammatical accuracy — which is why section 9.3 treats prosody as an examinable skill rather than cosmetic polish.

---

Procedural Facts Worth Knowing Before You Walk In

• Double marking is compulsory. The grid has two examiner name fields; two accredited examinateurs-correcteurs evaluate every candidate at every level and for every competence.
• The grid is attached to your file. After the evaluation, the completed grid is stapled to your DELF B2 script.
• You may read the grid in advance. The published grids carry the line « Le candidat peut prendre connaissance de ce document » — examiners alone may fill it in, but candidates are entitled to study it.
• Legacy grid. Older sample papers still circulate with the previous analytic version, which splits the same 25 points as: monologue suivi 7 points (1.5 + 3 + 2.5), exercice en interaction 6 points (3 + 3), Lexique 4, Morphosyntaxe 5, Maîtrise du système phonologique 3. The behaviours rewarded are identical; only the arithmetic differs.

---

A Self-Assessment Protocol Built on the Grid

Rehearsing without a rubric produces confidence, not points. Record one full mock oral — 5 to 7 minutes of monologue, then 10 to 13 minutes of debate with a partner instructed to object — and score yourself line by line:

1. Monologue /5 — Did I state an explicit problématique and announce a plan in the first 60 seconds? Did every argument carry a concrete example? Did I conclude, or was I cut off?
2. Interaction /5 — Did any answer of mine last under 15 seconds? Did I concede at least once without abandoning my position? Did I introduce at least one idea that was not in my exposé?
3. Lexique /5 — Count repetitions of your three most frequent content words. More than four occurrences of any one of them is a B2, not B2+, profile.
4. Morphosyntaxe /5 — Transcribe 90 seconds. Are the errors scattered or systematic? Did any passive, cleft or subjunctive structure appear?
5. Phonologie /5 — Play the recording to someone who does not know the topic. Did they have to replay anything?

A candidate who runs this protocol three times before the session walks in knowing which of five numbers is lowest — and that is exactly the number the examiners will write down.