Explain a Multilingual Film Without Translating Its World Away

A neighbor buys the family's van after some negotiation.

That synopsis line is not false. It is empty exactly where the scene is interesting. A reader who does not share the film's languages will fill the emptiness with the most familiar arrangement available: one language, spoken by everyone, understood by everyone, a negotiation proceeding the way negotiations proceed in films they have already seen.

The assumption does quieter damage than a missing fact. It rewrites the scene. If all three people in that kitchen understood the same things at the same moments, then the event the scene is built around never happened, and the reader has no way to know they are picturing a different film.

An English-language pitch can carry a multilingual film without pretending the film is in English. It does this by keeping two things explicit and separate: the map of who understands whom inside the story, and the plan for what the intended viewer will be handed at each moment.

The example running through this piece is invented. Language A and Language B are placeholders, not real languages with real speakers. Three people: a parent who speaks A and has no working B; an adult daughter who speaks and understands both; a neighbor who speaks B and has no A. There is a van, and the parent is selling it.

Relationships make a map. A language list makes an inventory.

"The film is in Language A and Language B" tells a reader the contents of the shelf. It does not tell them who can act. The map does.

In this invented household, A is the language of the home and B is the language of everything outside it. That single arrangement generates everything the scene can do. The daughter is the only person who can move in both directions. She is also, therefore, the only person who can decide not to move something across — which is the source of her power in any scene she is in. The parent can speak freely in front of the neighbor without being understood. The neighbor can do the same in front of the parent. Neither of them can be sure what happened on the other side.

That is a map. It answers a set of questions a language list never touches:

• Who is in the room when this language is used?
• Who can follow the utterance, and who can only hear talking?
• What can this person do after the switch that they could not do before it?
• Is anyone relying on a translation they cannot check?

The last question is often the scene. When the daughter carries the neighbor's offer into A, the parent receives it and has to trust the carrier. Nothing in the parent's own knowledge can confirm it.

Switching is an action, not a mood. Not every one of them is architecture, though. In a household where one language belongs to the kitchen and another to the phone, a great deal of switching is simply how the day is shaped, and a film that turns every instance into a turning point will feel like it is announcing itself. A useful test: would the scene still work if this line were spoken in the other language, or in a language everyone in the room shares? If yes, the switch is texture. If no, it is structure, and it belongs in the pitch.

Keep the beat sheet before you write the pitch sentences. Write the exchanges in order, mark the language of each, note who is present and who follows. Most of the contradictions that surface later — a character note that disagrees with a sample scene, a synopsis that accidentally gives someone information they never receive — are caught here, cheaply.

Character understanding and viewer understanding are two systems

They run at the same time and they do not have to agree.

Four mechanisms put meaning in front of a viewer, and they produce different relationships:

Translation for the viewer. Subtitles let the viewer read what a character hears as language. The viewer's access is independent of the scene's social reality; they can be admitted to every exchange in the room.

An interpreter inside the scene. One character carries meaning across. Here the viewer's access and the listener's access travel together, because both depend on the interpreter's choices — including the choice to shorten, soften, or drop something.

Action, object and context. A character who cannot follow the words can still work out what happened from a gesture, a number written down, who stands up, what gets handed over. This channel is real, it is often the cheapest, and it frequently carries more than subtitles can.

Deliberate non-translation. The viewer is left where the excluded character stands, hearing a language they do not have.

These are not interchangeable, and a scene can use different ones at different beats. Say which one you intend, and where. Say it as an intention. Nothing has been produced; there is no subtitle track. The film is subtitled is a claim about a deliverable that does not exist. Intended: A is subtitled into the pitch's language, and B is left unsubtitled is a plan, and it tells a reader exactly what they need to know.

One distinction worth stating outright, because readers trip over it constantly: the language of the pitch and the languages of the film are separate choices. The pitch is written in English because the recipient reads English. It is a description of the film, not a sample of it. A reader who cannot hold those apart will imagine everyone speaking English, and the flattening will have happened before they finish the first page.

Explain the exchange in the pitch's language

A gloss names the language, gives the sense in a few words, and stops. It is not a translation and it is not a transcript. The pitch needs enough to explain why the exchange works, and no more.

Marking the language is not decoration. Compare:

She tells her mother what he said.

She tells her mother what he said, in A.

One word establishes who is in the room and who is outside it, and the second sentence tells the reader why the beat has weight.

Preserve the friction. A mishearing, a hedge that travels and a hard number that doesn't, one person's fluency lagging behind another's, a line that simply never gets said — these are the load-bearing parts. When a pitch smooths them into fluent shared dialogue, the reader receives a scene in which everyone is equally informed, and the disagreement between that picture and the film's actual design will surface later as confusion, or as a note from a producer asking why the scene is there at all.

The worked example, end to end

An invented kitchen, an invented afternoon. The van is outside, and the parent is selling it. Four exchanges:

1. The neighbor, in B, addresses the parent with an offer: two thousand, a thousand now and the rest next month. The parent hears talking and no B.
2. The daughter, in A, carries the offer to the parent: two thousand, half now, half next month.
3. The parent, in A, answers the daughter: "Take it. The clutch is nearly gone. Don't mention it."
4. The daughter, in B, answers the neighbor: "It's a yes. Half now, half next month."

Laid out as comprehension states:

Read the last row slowly. The neighbor accepts the price and leaves believing he has bought a sound van. The parent never learns whether the warning traveled, because the confirmation came back in a language she does not have. The daughter is the only person in the kitchen who is never outside the conversation. And the viewer, if the intended subtitles are what they are meant to be, is the only other party holding both halves of it: that the parent named the clutch, and that the daughter let it go.

The scene's turning point is a sentence the daughter does not say. That is the whole reason the scene is written in two languages rather than one.

Here is a pitch paragraph derived from the table:

THE VAN. A kitchen, afternoon. A neighbor, who has no Language A, offers the parent two thousand for the van — half now, half next month — in Language B. The parent does not follow, and the daughter carries the offer into A. The parent says yes, then adds, still in A, that the clutch is nearly gone and that the daughter should not mention it. The daughter turns back to B, confirms the price and the yes, and says nothing about the clutch. Language A belongs to the household; B belongs to the street. In the intended version, both are subtitled into the pitch's language: the viewer hears the warning and watches it not travel.

And a note that has to agree with it:

Language and viewer note. Parent: speaks A, no B. Daughter: speaks A and B. Neighbor: speaks B, no A. Intended viewer access: both languages subtitled; the viewer finishes the scene level with the daughter and ahead of the neighbor. Intended, not tested; no subtitle track exists.

Now the flattened version of the same scene, which is what a pitch falls into by default:

A neighbor buys the family's van after some negotiation.

Or, with the languages still named but the relationships flattened:

The daughter translates, the neighbor learns the clutch is nearly gone, and the price changes.

The second version keeps the two languages and still loses the scene. Nothing in those four beats produces a moment where the neighbor hears about the clutch. To make that sentence true, you would have to add an exchange that does not exist — and in adding it, you would delete the daughter's choice, which was the only thing happening.

There is a legitimate alternative viewer design worth knowing about. If A is left unsubtitled, the viewer stands roughly where the neighbor stands: listening to a language they do not have while a family talks around them. That is a coherent intention, and it costs something specific. The viewer would no longer receive the clutch line or the parent's answer at the moment they are spoken. The offer and the acceptance stay legible — both are in B, both subtitled — so the sale still reads. What the viewer loses is what the family said around it, and the scene's drama would shift from the daughter's silence to the viewer's own exclusion. Either design can work. Only one of them can be described in the pitch, and the note has to match whichever you chose.

Check what the language description promises

The last job is comparison, and it is mechanical.

Put the synopsis, the character notes, the sample scene, and the stated viewing approach side by side and look for claims that cannot all be true. A character note saying the parent "understands more B than she lets on" sits badly with a sample scene where she cannot follow the opening offer. A viewing note promising the audience a place beside the neighbor contradicts a scene built on the audience hearing the A remark and counting what fails to cross. These are not nuances to be reconciled in a pitch meeting. One of them is wrong, and the reader will find the seam.

If Language A and Language B become actual languages, the glosses need a qualified reader — someone with real competence in the specific thing being checked, which is not the same as someone who happens to be from the right place. Register, idiom, whether a short reply reads as an echo or as a rebuke, whether a joke survives being glossed at all: a literal rendering cannot establish any of that, and neither can birth. One collaborator is also one reader. They can tell you how a line lands for them; they cannot tell you how it lands for every speaker of the language they share.

Two things stay outside this method. Delivery requirements — subtitle duration, character limits, platform specifications — are a separate verification against a separate source, and a placeholder scene makes no claim about any of them. And audience testing has not happened. The intended viewer access described above is a design, argued from the beats, not an observation of anyone watching.

The pitch paragraph and the note beneath it agree about who understands what, and that agreement is the test. A film may hold its audience outside a language on purpose; that is a real design and a defensible one to put in front of a producer. What a pitch can never afford is leaving the reader outside it by accident.

Frequently asked questions

What two systems should an English-language pitch keep separate for a multilingual film?

Keep separate the map of who understands whom inside the story and the plan for what the intended viewer will be handed at each moment. A language list is an inventory; the map answers who follows an utterance, who can act after a language switch, and who relies on a translation they cannot check. The pitch's language and the film's languages are separate choices.

When is a language switch structure rather than texture?

If the scene would still work with the line spoken in the other language, or in a language everyone in the room shares, the switch is texture. If it would not, the switch is structure and belongs in the pitch. Switching is an action, but not every instance is architecture; a film that turns every switch into a turning point may feel like it is announcing itself.

What viewer-access mechanisms does the article distinguish?

Translation for the viewer through subtitles; an interpreter inside the scene; action, object, and context; and deliberate non-translation. They are not interchangeable and can be used at different beats. Say which one is intended and where. 'The film is subtitled' is a claim about a deliverable that does not exist; 'Intended: A is subtitled... B is left unsubtitled' is a plan.

In the invented van scene, how does intended subtitling change what the viewer knows?

With both languages subtitled into the pitch's language, the viewer hears the parent's warning about the clutch and watches the daughter not pass it on. The viewer finishes level with the daughter and ahead of the neighbor, holding both that the parent named the clutch and that the daughter let it go. If Language A is left unsubtitled, the viewer stands roughly where the neighbor stands: the offer and acceptance stay legible because both are in B and subtitled, so the sale still reads, but the viewer loses the clutch line and the parent's answer at the moment they are spoken. The drama then shifts from the daughter's silence to the viewer's exclusion.

What consistency check should the language description pass?

Put the synopsis, character notes, sample scene, and stated viewing approach side by side and look for claims that cannot all be true. A character note saying the parent understands more B than she lets on sits badly with a sample scene where she cannot follow the opening offer. A viewing note promising the audience a place beside the neighbor contradicts a scene built on the audience hearing the A remark. If the placeholders become actual languages, glosses need a qualified reader with real competence in the specific thing being checked; one collaborator is one reader. Delivery requirements are a separate verification, and audience testing has not happened.

Further reading

• How Much Cultural Context Does an Unfamiliar Film-Pitch Reader Need? — Read this when an unfamiliar reader could misread a choice and you need a situated cue that explains the action without turning the film's world into a lecture.
• A Film’s Joke Depends on Its Language. Translate the Words—or Rebuild the Joke? — Read this when a language-dependent joke or ambiguity carries the scene and you must separate explaining it in a pitch from subtitle design or a rebuilt joke.