Cambridge Modern and Medieval Languages interview questions, with model answers

Cambridge does set a written test for Modern and Medieval Languages. It is sat on Thursday 19 November 2026, runs for one hour, and is marked out of 48 — of which only 16 marks are for the quality of your foreign language. This guide rebuilds the interview advice around that paper, the December interview window, and the figures Cambridge itself publishes. Of the packs covering Modern and Medieval Languages, the French pack comes first, £180.

Does Cambridge set a written test for Modern and Medieval Languages?

Yes. Cambridge's course page settles it in one line: "There is an admission assessment at all Colleges for this course. You do not need to register in advance." Read both halves of that sentence, because both matter. Every College uses the assessment, so there is no combination of College and language pairing that quietly avoids it. And because there is no registration step, nothing in the UCAS process and nothing in the My Cambridge Application form will prompt you about it. A candidate who assumes that no news means no test finds out in the second week of November, which is far too late to change how they have been practising.

The assessment is one stage in a sequence, and the sequence is the part that almost every guide on this topic leaves out. Written work comes first. Cambridge asks for "2 pieces of written work", described as "recent examples of writing completed for school, one of which should be in one of the languages you intend to study at University". Note the second half of that instruction carefully: one of the two pieces has to be in a language you are applying to study. Cambridge's generic written-work guidance elsewhere on the site says submitted work must be in English, which genuinely contradicts the course page. For a Modern and Medieval Languages application the course-specific instruction is the operative one, and your College will send you its own upload guidance.

Then comes the written assessment, on Thursday 19 November 2026. Then interviews, in a window Cambridge publishes as "7 December to 18 December 2026", with "Winter pool interviews: around mid to late January 2027" for candidates carried into the pool by their first College. Your assessment script is therefore written, marked and in front of your interviewers before you have said a word to them. Cambridge's specification is careful about how much weight that carries: "Remember that it is only one part of our assessment of you as a candidate, alongside the interviews, schoolwork (if your College has requested it) and everything else we get to know about you from your application form." One part is still a part, and it is the only part where a College watches you produce the language under timed conditions with nobody helping.

The last row is blank on purpose. Cambridge does publish a single decision date for everyone interviewed in the main period — its deadlines page reads "If you were interviewed in the main interview period (December 2026) you will find out the outcome of your application on this date" — but we have not verified the date itself, so we have left the cell empty rather than print a figure you might plan a January around. The same rule governs every number below. If Cambridge or Oxford has not published it, or we have not read it on the page that publishes it, it is not here. That is also why you will not find a total application count for this course anywhere on this page: Cambridge publishes applications per place and it publishes acceptances, and multiplying the two would produce a number Cambridge has never stated.

What is actually in the MML Admissions Assessment?

The paper runs for one hour, and Cambridge splits that hour unevenly between two languages. The published format is "Discursive response in Foreign Language (40 minutes)" followed by "Discursive response in English (20 minutes)" — 40 plus 20, making the one hour the specification refers to when it says "You can expect to take the one hour written assessment for MML at the College where you will be interviewed." Two thirds of your writing time is spent in the foreign language. If your practice does not mirror that ratio, your practice is not preparing you for this paper. It behaves like the other papers we cover under university admissions tests: a fixed date, a published mark scheme, one sitting, no resit.

What makes this assessment unusually coachable is that Cambridge publishes the marking. The mark sheet is headed "TOTAL SCORE (___/48)", and the foreign-language task carries "(Total of 32 marks available)". Inside that task, the published criterion "Comprehension, summary and response (16 marks)" accounts for half of it, which leaves 32 minus 16, or 16 marks, for the quality of the language itself. The English-language section takes the balance of the paper: 48 minus 32, another 16 marks.

Follow that arithmetic through, because it should change what you practise. The quality of your French, German, Spanish, Italian, Portuguese or Russian is worth 16 of the 48 marks on the paper — exactly a third of it. The other 32 marks are for reading a passage properly, summarising what its author actually argues, taking a position on it, and organising the result so a stranger can follow you. Cambridge makes the same point about the English section in a single sentence: "Half the marks for Section B are not for what you say, but how you communicate your observations to us." A candidate with immaculate grammar and nothing to say cannot reach two thirds of this paper. A candidate with sharp ideas and imperfect endings can lose a third of it and still do well.

Where you sit the paper is less settled than when. The University specification describes taking it "at the College where you will be interviewed". Christ's College describes something different in practice for its own applicants: the assessment "will take on Thursday 19 November 2026 at 09:30-11am UK time (or 5-6:30pm if you are in the US/Canada timezone or similar)", sat remotely, with the College handling the admin — "We arrange your assessment automatically if you are selected for interview, so you don't need to register." Two things follow. Read your own College's page and do what it says rather than what a generic guide says. And note the ordering at Christ's: the assessment is set for candidates who have already been selected for interview, so at that College it is not the filter that decides whether you are seen. It is evidence in the room once you are.

The inversion nobody warns you about: you read English and write in French

Every specimen paper Cambridge publishes for this assessment gives the source passage in English. The instruction to candidates reads "Please read the following passage carefully and then answer the questions following it", and then, in capitals in the original, "Your response to Section A MUST be in ONE of the languages that you are applying to study here." That is the exact inverse of the scenario most preparation drills, including the earlier version of this page: an unseen French poem handed over cold, to be picked apart on the spot. Nothing corrects that impression, either, because most published guides to this application do not mention the assessment at all.

The inversion matters because the two tasks train opposite muscles. When the passage is in your target language, comprehension is the hard part and your own vocabulary is scaffolded by the text: you can lift a phrase, mirror a construction, borrow a connective. When the passage is in English and the answer is not, comprehension is nearly free and production is the whole difficulty. There is no borrowing available. Every abstract noun, every subjunctive, every way of saying "this only holds if" has to come out of your own head, under time, with no dictionary. That is why a candidate who has spent a year translating literary extracts can still be caught out. Translation is a comprehension-led skill, and this paper is production-led.

Cambridge rules out the exercises schools most often reach for: "It is not a fill-the-gaps exercise or a verb conjugation test, or a translation. We want to see how well you can express yourself freely in the foreign language." Free expression under time is a trainable thing, but it is trained by writing, not by drilling. The most useful preparation is a short weekly discursive piece in your chosen language on an argument you did not choose, written to the clock and marked against the two published criteria rather than against a general impression of fluency. Read the argument back and ask the marker's two questions separately: did I represent what the author claimed, and did I say something of my own about it?

One strategic decision has to be made before the day, not on it. Cambridge sets out who writes in what: "If you are applying to study two languages which you are studying for A-levels or equivalents, then you can choose to respond in either language (not both). If you're applying to study one language you're already studying and a language you'll be starting from scratch, then you should respond in the language you are already studying." If you have two A-level languages, that is a genuine choice, and it should be made on evidence rather than affection. With 16 marks resting on accuracy, range and complexity of the language itself, the right answer is usually the language in which you write fastest without checking yourself, which is not always the one you like most or the one you plan to specialise in later.

How many interviews will you have, and why the text arrives beforehand

Colleges set their own interview arrangements, and most publish very little. Christ's College publishes the clearest account we found: "Selected candidates are normally interviewed for 35-50 minutes in total. At Christ's, we split the time into two interviews with academics in Modern and Medieval Languages - one for each of the languages you plan to study. You will normally have two interviewers in each interview". Christ's also states that there is no additional general interview on top of those two. Treat 35-50 minutes as that College's published arrangement rather than a University-wide rule. We have not found another College that publishes a duration, and averaging an unpublished one would be inventing a fact.

The detail that should change your preparation most is when you get the text. Christ's says: "Shortly before each interview, you will be sent a short text to read, for discussion during the interview. An interview for a language you have already studied will include a text in that language (and perhaps some questions to think about as you read it); interviewers for a language you wish to learn from scratch will set a text in English". So the discussion is not sprung on you. You will have read the text, and possibly a set of questions with it, before you sit down. That converts the task from improvisation into preparation, and it is why so much interview advice built on the cold-reading assumption misdirects effort.

Do not assume the interview happens on a screen, either. One published College arrangement puts UK-based applicants online "except for the following courses. For these courses interviews will be in person", and Modern and Medieval Languages appears on that list alongside Mathematics, Medicine and Music. If you are travelling in December, the logistics are part of the preparation: a language interview conducted in person has a different rhythm from one conducted over a connection, and the switch between languages between the two interviews is easier to handle in a room than through a headset.

What do they ask? Mostly they ask you to keep talking about the text they sent, in a direction you did not pick. Expect the interviewer to take your reading and push it: to ask what would have to be true for your claim to hold, to offer a rival reading and ask you to weigh it, to move from a sentence to the argument and back to a word. For a language you already study, expect at least part of that conversation in the language. For a language you are starting from scratch, expect an English text and questions about how you think, not about vocabulary you cannot have.

Model answers: what a strong response to the interview text sounds like

Because the text is sent in advance, a model answer is not a script to memorise. It is a shape you can put around any text in the time you have between receiving it and the interview. The shape that works has four moves and survives translation into any of the languages: say what the text argues, in one sentence, without hedging; name the move the author makes to get there; test that move against a case it does not cover; then say what you would need to know to settle it. The last move is the one candidates skip, and it is the one that turns an answer into a conversation the interviewer can extend.

In French, that opening move sounds like this: "Le texte soutient que la langue n'est pas seulement un outil de communication mais une forme d'appartenance. L'auteur passe d'un constat linguistique à une conclusion politique, et c'est ce glissement que je voudrais examiner." Two things are happening. The claim is stated cleanly, and the candidate has already flagged what they intend to interrogate. What follows should be a concrete counter-case, not a general reservation: a bilingual community where the author's equation between language and belonging does not obviously hold, described in a sentence or two, then a return to the author's terms.

In German, the same structure with a different centre of gravity: "Der Text behauptet, dass sich die Bedeutung eines Wortes erst im Gebrauch zeigt. Das überzeugt mich weitgehend, aber der Autor unterscheidet nicht zwischen alltäglichem und fachlichem Gebrauch — und genau dort scheint mir das Argument brüchig." German rewards this kind of answer because the subordinate structures let you hold a concession and an objection in one sentence. If you can build the concession before the objection without losing the verb, you are demonstrating exactly the range and complexity the written paper puts marks against.

In Spanish: "El texto parte de una idea sencilla — que traducir es siempre interpretar — para llegar a una conclusión mucho más fuerte: que no existe la traducción fiel. Estoy de acuerdo con la premisa, pero creo que la conclusión no se sigue necesariamente, y quisiera explicar por qué." Notice what none of these three openings does. None summarises the text at length, none begins with an apology about nerves, and none reaches for a critic's name. The interviewer has the same text in front of them and does not need a retelling; what they need is your reading of it and your willingness to be argued with.

The language quality that supports all this is the same quality the written paper marks. Range means you are not surviving on the handful of connectives everyone uses. Complexity means you can subordinate rather than list. Accuracy means the endings hold when the sentence gets long, which is precisely when they stop holding for most candidates. Because 16 of the 48 marks on the assessment sit there, and because the same range shows the moment you open your mouth in an interview, this is the one part of preparation that pays into both events at once. The rest of your subject preparation — wider reading, a defensible view about a text, an interest you can follow beyond the syllabus — works the same way it does for any Oxbridge subject, but the writing practice above is specific to this paper.

One Discursive Task, Attempted Twice

The passage below is written for this page. It is not from Cambridge, and it is not from the French, German or Spanish pack on the resources hub — it exists so you can see a discursive answer built rather than finished, before you decide whether ten more of these, in your own language, are worth £180. Section A gives you forty minutes to move from an English passage to continuous prose in French, German or Spanish. The shape below holds whichever of the three you choose, and the layers are kept apart the way a pack keeps them: Question, then Prompts, then Suggested answers.

Question

A word is often called untranslatable when no single word in another language covers the same ground. Speakers then treat this as proof that something real is lost between languages. But a dictionary lists single words, not everything a language can say. Given a whole sentence rather than one word, most so-called untranslatable ideas can be rendered, at greater length, in almost any language. If that is true, untranslatability is a fact about dictionaries, not about thought.

In continuous prose, in French, German or Spanish, argue for or against the closing claim. Give yourself the forty minutes Section A allows, and write in the language you would actually choose on the day.

Prompts

Release these to yourself one at a time, and go back to the passage before taking the next.

• Restate the closing claim in one sentence of your own before you agree or disagree with it.
• Think of a word people call untranslatable. Try translating it anyway, in a sentence rather than a word. Did it work?
• If the sentence-length translation succeeds, has anything actually been lost, or only made longer?
• Is there a difference between a concept being expressible in another language and being expressible just as economically?
• Read back what you have written. Does it argue with the passage, or only repeat its move in your own words?

Suggested answers

The attempt that fails. The natural first move is to walk back through the passage in order: first it says untranslatability seems to prove something is lost, then it says dictionaries only list words, then it concludes the problem is really about dictionaries. That is an accurate account of the passage, written in reasonable French, German or Spanish. It is also not an argument, because nothing in it is the candidate's own. Set against the published mark scheme, this attempt earns some of the marks for comprehension — the passage has clearly been understood — and very little for response, because a response has to add a position the passage does not already state. Half a criterion answered is not half the marks; on this paper, summary without response forfeits most of what "comprehension, summary and response" is worth.

The attempt that works. Concede the passage's central move before contesting it, because a flat disagreement against a claim this reasonable reads as reflex rather than thought. The concession: yes, a sentence can usually recover the descriptive content of an untranslatable word — German's Schadenfreude becomes, at length, something like pleasure taken in someone else's misfortune, and a reader who did not know the word before now knows roughly what it names. But descriptive content is not all a word carries. Schadenfreude also carries a register — wry, a little rueful, admitting to something faintly shameful in one breath — and a speed, because a single word lets you name the feeling and move on within the same sentence, where a defining clause forces the sentence to stop and explain itself. Translate the word and you keep the reference; translate it into a sentence and you often lose the tone in the same stroke, because tone lives in economy as much as in meaning.

That is a case the passage does not obviously cover, which is exactly what the prompts were pointing toward. The honest next step is to say what would settle it: whether register and pace count as part of what a word means, or as something separate that translation was never claiming to preserve. A strong answer does not need to resolve that question in the time available. It needs to have found it, stated it precisely, and shown that finding it came from testing the passage's claim against a real case rather than from a general instinct that languages are secretly untranslatable. That is the difference the sixteen marks for language quality and the sixteen for comprehension, summary and response are separately built to reward: one pair of marks for whether the French, German or Spanish holds up under that pressure, the other for whether there was an argument in it at all.

Applying ab initio: the route most guides leave out

Every Cambridge MML student takes two languages, and one of them can be new. The course page puts it plainly: "All our students study 2 languages, one of which can be learnt from completely from the beginning, except French." French is the exception; German, Italian, Portuguese, Russian and Spanish can all be started from scratch. That is a genuine route into the course for a candidate with one strong language and an appetite for a second, and it is missing from almost every interview guide on this topic, including the earlier version of ours.

What you do need is depth in the language you already have. Christ's College sets the benchmark in a form you can measure yourself against: "You do need to have studied at least one of the two languages in which you wish to specialize to A-level or IB Higher Level standard (that is, a standard broadly equivalent to level B2 of the Council of Europe Common European Framework of Reference for Languages)." A CEFR level is a more useful target than a grade, because it describes what you can do rather than what you scored: sustain an argument, handle unfamiliar registers, write at length without collapsing into simple sentences.

The ab initio route changes three practical things. In the written assessment you do not get to choose — Cambridge is explicit that an applicant taking one existing language and one from scratch "should respond in the language you are already studying". In the interview for the new language, the text will be in English, so what is being assessed is how you handle an argument and whether you can be taught, not vocabulary you have had no chance to acquire. And for the sibling Linguistics and Modern Languages course, the assessment calendar splits by route: "Languages from scratch: Tuesday 17 November 2026" and "Languages post A-level: Thursday 19 November 2026". Two different days, which tells you Cambridge assesses the two starting points as genuinely different things.

One combination carries a second assessment. Cambridge states that "Applicants wishing to combine a modern European language with Classical Latin or Classical Greek will have their classical language assessed", and adds that those "applicants will also need to take the MML written assessment". So a Latin-and-Spanish candidate is preparing for a classical language assessment and the MML paper, not one or the other. If that is your combination, build the November calendar around both, and do not let the classical side consume all the preparation time simply because it feels more examinable.

How competitive is Cambridge MML, really?

Two published figures point in opposite directions, and the honest answer sits between them. Cambridge lists Modern and Medieval Languages at 2 applications per place. By the standards of Cambridge subjects that is low, and it is worth saying plainly what a low ratio does and does not mean. It does not mean the course is easy to get into. It means fewer people apply. Modern language entries at A level have been under pressure for years, and a course that requires an A-level language simply has a smaller pool to draw from than one that requires none.

The second figure is where the real bar sits. Cambridge's published minimum is "A level: A*AA" or "IB: 41-42 points, with 776 at Higher Level", but the course page also records what entrants actually achieved: "Most Modern and Medieval Languages students (who had studied A levels and started at Cambridge in 2018, 2019 and 2023) achieved at least A*A*A (71% of entrants)." A*A*A is a grade above the published minimum, and 71% of entrants reached it. So the offer tells you what Cambridge will ask of you, and the entrant profile tells you what the competition is arriving with. Both are true at once, and a candidate planning their A level preparation around the minimum alone is planning around the floor rather than the room.

Note also that individual Colleges can pitch above the University minimum. Christ's, for instance, publishes an IB requirement at the top of the central range rather than across it. Both figures are correct at their own level: the University sets a floor and a College can set its own conditions above it. If you are choosing a College, that is one of the few genuinely comparable things you can read off their pages, along with whether they interview in person and how they run the assessment. The sibling course is a useful comparison too. History and Modern Languages runs at 3 applications per place, higher than MML, which is worth knowing if you are weighing a joint course against the straight languages degree.

Put together, the picture is consistent and not especially comforting or especially alarming. The constraint on this course is not the queue at the door. It is grade profile, an A-level language at something like B2, a written paper in November that most applicants have not prepared for because they did not know it existed, and two interviews conducted partly in a language you are still learning. Every one of those is a thing you can work on, which is more than can be said for an application ratio.

Cambridge or Oxford? The two applications are not comparable

Most comparisons of Oxford and Cambridge Modern Languages are written in terms of style: one is said to be more literary, the other more linguistic, and neither claim survives contact with a College or a tutor who does it differently. There is a much harder difference available, and it is procedural. Oxford states of its Modern Languages course: "You do not need to take a written test as part of an application for this course." Cambridge states of MML: "There is an admission assessment at all Colleges for this course." Oxford also asks for no written work on this course, while Cambridge asks for two pieces. That is not a difference of emphasis. It is two different applications with two different workloads in the same autumn.

The Oxford figures are worth reading slowly. Oxford publishes "Interviewed: 94%", "Successful: 44%" and "Intake: 141" as a three-year average for 2023-25. An interview rate of 94% means the interview is very nearly universal for this course: shortlisting barely happens before December, so the interview is doing almost all the selection work. Cambridge distributes the same work differently, across written work submitted in the autumn, a timed paper in November and interviews in December. Neither is softer. They ask for different things at different moments, and they reward different kinds of preparation.

For a candidate deciding where to apply, the practical reading is this. If your strength is talking — thinking aloud, being pushed, changing your mind in front of someone — the Oxford process gives that a very large share of the decision. If your strength is producing controlled written language under time, Cambridge gives you a paper on which to demonstrate it and publishes exactly how it is marked. And if you are applying to Cambridge, understand what you are signing up for in November: an assessment nobody reminds you about, sat weeks before the interviews you have been imagining, worth 48 marks and taken once.

The year abroad, and a better answer to "why this course"

Somewhere in one of the two interviews, in English or otherwise, you will be asked some version of why you want to do this. Enthusiasm is the standard answer and the weakest one, because every candidate has it and none of it distinguishes anybody. The stronger answer is structural, and Cambridge hands you the material for it on the course page: "You'll spend at least 8 months abroad, during which time you prepare a project that counts as one sixth of your final mark."

Sit with both halves of that sentence. At least 8 months abroad is not a gap year with a reading list; it is a period in which your language is your working environment rather than your subject. And a project worth one sixth of the final mark means that the time abroad is examined, not merely experienced. A candidate who has thought about what they might work on — a body of writing, a regional variety, an archive, a political argument that plays out differently in the other language — is answering "why this course" with the course itself rather than with a feeling about languages.

The same applies to the two-language structure. Every student takes two, and one can be started from the beginning. If you intend to start a language from scratch, the interesting version of the answer is why that pairing: what one language lets you read that the other does not, or what the two together open up that either alone would not. If both are languages you already study, the question becomes what you want to do with the combination rather than which you prefer. None of this needs to be settled. It needs to be thought about, because an interviewer can tell the difference between a candidate who has and one who is deciding at the table.

One correction to a claim this page used to make. It said first-year students all study phonology, syntax and the history of their languages. The course page describes something looser: an introductory first-year paper in which students explore several topics from a list that includes literature, linguistics, history, thought, film and art. Linguistics is one option among several rather than compulsory content. If linguistics is the reason you want the course, say so and be ready to defend it — but do not describe the first year as something it is not, because the person opposite you teaches it.

Preparing in the order the timetable actually happens

Most preparation plans for this course are built around the interview because the interview is what candidates fear. Built that way, they start too late and rehearse the wrong task. Ordered against the published calendar, the plan looks different. Written work first, because two pieces have to be chosen and one has to be in a target language, and the best piece of school writing you have is rarely the one you finished most recently. Then the assessment, which is a writing exam. Then the interviews, which are a reading and talking exam with the text supplied in advance.

For the assessment itself, the useful unit of practice is a timed discursive response written from an English passage. Take an opinion piece in English on something you have no settled view about, give yourself 40 minutes, and write in your chosen language. Then mark it twice, separately, against the two published criteria: once for whether you represented and answered the argument, once for the language. Marking them together produces a vague sense of how it went; marking them apart tells you which 16 marks you are losing. Then do the English section as its own exercise, remembering that half of its marks are for how clearly you communicate rather than for what you notice.

For the interviews, the trainable skill is what you do with a text in the hours between receiving it and discussing it. Read for the argument rather than the subject. Find the sentence where the author moves from evidence to claim. Prepare one counter-case and one question you genuinely cannot answer, because being asked something you have already found interesting is a much better position than being asked something you have never considered. Then practise saying all of it in the language, out loud, because the gap between what a candidate can write and what they can say under mild pressure is usually the gap that costs them.

A last word on what this page does not claim. There is a great deal of confident writing about Cambridge MML interviews online, much of it built on one applicant's memory of one December. Where a College publishes something, we have quoted it and named the College. Where only one College publishes a figure, we have said so rather than promote it into a University rule. Where Cambridge publishes nothing — the total number of applications, the decision date we could not verify, the duration of an interview at a College that does not say — we have left a gap. A gap you can see is more useful than a number you cannot check, particularly in the autumn of an application year when you are deciding what to spend October and November doing.

Every College runs the MML assessment on one November date, and none publishes interview questions. We traced our guide to what Cambridge publishes.

Frequently Asked Questions