Pancyprian French common mistakes
Pancyprian French · Understand the traps and how to avoid them.
During the listening section, the candidate hears a word that resembles a familiar French word and writes down its meaning immediately, then stops following the rest of the recording, so the answer to the item is built on one sound rather than the full utterance (for example hearing "poisson" and answering about fish when the recording said "boisson").
The recordings are played a limited number of times and spoken at near-natural speed, so candidates who translate word by word run out of processing time. False friends and near-homophones (poisson/boisson, verre/vers/vert, mer/mère/maire) are deliberately common in distractors, and the candidate's attention collapses after the first recognised item.
What to do instead
Train with a two-pass routine: on the first listening, note only the situation (who speaks, where, about what, positive or negative) and never write full answers; on the second pass, answer the items. Practise dictation of minimal pairs (poisson/boisson, dessus/dessous, sur/sûr) and, after each recording, write one sentence summarising the speaker's intention before checking options.
The candidate answers a listening item from the words that appear in the question rather than from what the speaker actually says, choosing the option whose vocabulary is echoed in the recording even when the speaker's meaning is the opposite (for example the speaker says the museum is closed on Mondays but the option says it is open).
Exam items routinely include a lexical trap: the option reuses the recording's words while reversing the logic (negation, comparative, tense). Candidates under time pressure match surface forms instead of processing negation and connectors such as "sauf", "au lieu de", "cependant", "ne...que".
What to do instead
Before listening, underline the key word of each question and predict what kind of information is needed (a time, a price, a reason, an opinion). While listening, mark the exact words that carry the answer and check for a negation or restriction around them. Build a personal list of trap markers (ne...que, sauf, plutôt, en revanche, pas du tout) and rehearse them with short audio clips.
In reading comprehension, the candidate reads the whole text at normal speed, then answers from memory of the general topic, producing answers that are plausible in the real world but not stated or implied in the document (for example inferring that a scheme is free because it sounds like a public service).
The texts are short but information-dense (notices, advertisements, letters, press extracts) and distractors are written to sound reasonable. Candidates who do not locate the exact supporting sentence fall back on general knowledge, and the exam penalises answers that go beyond the text.
What to do instead
Use a locate-and-justify method: read the questions first, then scan for the paragraph containing the relevant keyword, underline the exact sentence, and write its line number next to your chosen option. Reject any option you cannot tie to a specific phrase in the text. Practise with short authentic documents (small ads, instructions, horoscopes, letters) and time yourself at roughly one minute per item.
The candidate treats every unknown word in a reading text as a blocker and stops to guess its meaning, losing time and often misreading the sentence; alternatively the candidate skips the connectors ("donc", "pourtant", "en effet", "or") that carry the argument, so questions about the writer's attitude or the purpose of the text are answered wrongly.
School habits reward understanding every word, but exam texts contain deliberately unfamiliar vocabulary that is not needed to answer. Meanwhile the items that test inference depend on logical connectors and on register (irony, reservation), which candidates have rarely been trained to notice.
What to do instead
Practise guessing from context: cover the unknown word, read the sentence, and decide whether it is positive, negative or neutral before checking. Separately drill connectors by classifying them into cause, consequence, opposition and concession, and for each practice text write one line on the author's aim (informer, convaincre, avertir, vendre).
In the grammar and vocabulary section, the candidate answers by ear, choosing the form that "sounds right", which leads to errors in agreement ("les personnes que j'ai rencontrées" written as "rencontré"), in mood after conjunctions ("bien qu'il est" instead of "soit"), and in pronoun choice ("lui" for "elle", "y" for "lui").
Spoken French neutralises many written distinctions: past-participle agreement, the imperfect versus passé composé contrast, the subjunctive after "il faut que", and the difference between direct and indirect object pronouns are often inaudible or blurred in speech. Candidates who rely on intuition reproduce spoken patterns in a written test.
What to do instead
Work rule by rule with a written checklist: for each verb, identify the auxiliary, the subject and any preceding direct object before writing the participle; for each conjunction, memorise whether it triggers the subjunctive (bien que, pour que, avant que) or the indicative (parce que, puisque, après que). Do short targeted drills of ten items per rule and keep an error log with the rule name, not just the corrected sentence.
The candidate writes the written production in one continuous draft with no plan, produces a text that answers only part of the task (for example describing a place when the task asked to describe it and invite a friend), and mixes registers, using "tu" and "vous" in the same letter or opening with "Salut" in a formal message.
Time pressure pushes candidates to start writing immediately, and the task instructions usually contain two or three components plus a format constraint (letter, article, message, blog post) that are easy to miss. Register confusion comes from not deciding the relationship with the addressee before writing.
What to do instead
Spend the first five minutes planning: underline each instruction verb, list one bullet per required content point, and state the format, the addressee and the register at the top of the plan. Write with a clear opening and closing formula suited to the register, check off each bullet after writing, and reserve the last five minutes for a grammar pass focused on agreements, tenses and accents.
The candidate links sentences only with "et", "mais" and "parce que", writes very short sentences with repeated basic vocabulary, and produces a text that is grammatically survivable but scores low because it shows no range of structures or connectives.
Candidates prioritise avoiding errors over demonstrating range, and they have not automated a small set of reusable complex structures. Under time pressure the brain falls back on the most frequent, simplest patterns, so the script reads like a beginner text even when accuracy is acceptable.
What to do instead
Memorise and rehearse a bank of ready-made complex structures ("ce qui me plaît, c'est...", "si j'avais le choix, je...", "il est vrai que... cependant...", "non seulement... mais encore...") and force at least four of them into every practice text. Build a connective ladder (d'abord, ensuite, en outre, en revanche, par conséquent, enfin) and use it as a paragraph skeleton in every essay.
In the oral production, the candidate prepares full written sentences and recites them, so the delivery is flat, hesitations cluster at the points where memory fails, and the candidate cannot respond when the examiner asks a follow-up question that departs from the script.
Writing before speaking is the default school strategy, but reading aloud produces unnatural rhythm and blocks interaction; the candidate has practised a monologue, not a conversation, and has no prepared repair strategies for when a question is not understood.
What to do instead
Prepare keyword notes only, never full sentences, and rehearse aloud with a timer so the monologue fits the allotted time. Record yourself, listen for pauses and pronunciation problems, and re-record. Practise interaction with a partner using follow-up questions, and memorise repair phrases ("Pardon, pourriez-vous répéter?", "Vous voulez dire que...?", "Je n'ai pas bien compris, mais je pense que...") so a misunderstanding does not end the exchange.
The candidate pronounces written French as it is spelled, sounding final consonants that are silent ("ils parlent" with a pronounced t, "beaucoup" with a pronounced p), ignoring liaison and the difference between "é" and "è", and speaks so softly that the examiner cannot assess fluency.
French spelling is historically conservative, and candidates learn vocabulary visually; without systematic oral practice the written form dominates pronunciation. Anxiety about mistakes also lowers volume, which further harms intelligibility and the impression of fluency.
What to do instead
Drill pronunciation from audio rather than from spelling: shadow short recordings, mark silent final consonants and obligatory liaisons on a transcript, and practise minimal pairs (parlé/parlais, dessus/dessous, tous/tout). In practice sessions, speak at full volume and record from across the room to check audibility.
The candidate runs out of time or space because of an unbalanced allocation: spending too long on one reading text or on the planning of the written task, leaving the last exercise or the conclusion unfinished, or writing far over the required word count and losing accuracy in the rush.
The exam is divided into timed sections with fixed word limits, and candidates rarely rehearse under real time constraints; perfectionism on early items consumes the budget, and the final items, often the easiest, are answered in panic or not at all.
What to do instead
Rehearse full papers under strict timing at least three times before the exam, with a visible clock and a target time per section. Adopt a rule to move on after a set number of minutes on any single item, mark doubtful answers and return only if time remains, and practise written tasks at the required length so you know what the word count looks like on the page.