A Film’s Joke Depends on Its Language. Translate the Words—or Rebuild the Joke?
A Film’s Joke Depends on Its Language. Translate the Words—or Rebuild the Joke?
A pitch recipient who does not speak the film's language needs to understand the joke's mechanism. That does not automatically mean the film needs a different joke. First decide whether you are explaining an existing scene, designing subtitles for it, or proposing a target-language version of the scene itself. These are different jobs, with different permissions to change things.
An explanatory translation can succeed by making the intended turn clear. A playable adaptation has to work alongside speech, image, performance, and time. The first can tell the reader what an audience is meant to discover; the second must decide when and how that audience discovers it.
Do not judge both by whether their sentences mean approximately the same thing. A translation may preserve the information and remove the surprise. A rewrite may preserve a surprise while turning a restrained character into a performer of jokes. Either loss can matter more than the surface similarity of the words.
Identify what changes in the viewer's understanding
Here is an original, deliberately small English rehearsal. It is not a borrowed film scene, an observed audience response, or a validated bilingual solution. English lets us inspect the mechanism before asking a collaborator to solve a different language's problem.
We begin close on a worker's face. The room and the object being discussed are outside the frame.
Customer, offscreen: Can you save my sole?
Worker: Depends. Leather or rubber?
A boot lands on the counter.
Spoken aloud, sole can sound like soul. The proposed first line leaves room for a personal plea; the reply redirects it toward material, and the boot confirms the ordinary repair transaction. Written on the page, “sole” has already supplied information that the proposed spoken line withholds.
That difference is the first translation problem, even before another language enters. The pitch's printed spelling is not the audience's initial listening experience.
Now identify who misunderstands. In this version, neither customer nor worker has to be confused. They know they are discussing footwear. The intended misdirection belongs to a viewer who cannot yet see the setting. If the actor playing the worker looks puzzled by a spiritual request, the scene changes: a character now makes the mistake rather than the framing briefly inviting the audience to make it.
Neither construction is automatically better. But they are not the same comic event. If the film depends on the worker being practical and unflappable, introducing their confusion may damage characterization while seeming to rescue the pun.
The mechanism note can therefore be exact:
The first line permits two interpretations in sound. The worker treats it as an ordinary repair question. Their material-specific reply begins the reveal; the boot confirms it. The viewer's restricted view creates the ambiguity. The worker is not misunderstanding the customer.
This note states more than “a joke about shoes.” It identifies an ambiguity, a knowledge difference, and the moment of reinterpretation. Those are the things an adaptation would have to preserve or consciously replace.
The same analysis works when the language feature is not a pun. A very formal expression might puncture an intimate moment. A familiar institution's name might let a joke land without explanation. A final verb might revise the apparent meaning of a sentence. Name the actual mechanism in the scene instead of assigning every problem to “cultural differences.”
Give the deck an explanation, not a translation audition
Suppose the film remains in its source language. The recipient needs to assess its tone and dramatic choices, not personally experience every joke at full performance speed during a read-through.
For the shoe scene, show the brief exchange and add the mechanism note. A translated gloss for a non-English-speaking reader would distinguish the two possible meanings, then explain that the spoken word does not disclose its spelling. It should not claim that the reader has just heard an equivalent joke.
There is no need to hide the explanation to protect a surprise the recipient cannot otherwise perceive. The pitch can reveal how the scene works. That is different from putting the same explanation in a subtitle for a future viewer.
Keep the note close enough to the example to be useful, but outside the dialogue. An annotation reading “sole/soul homophone; practical reply reframes the plea” belongs to the pitch's explanation, not necessarily on the cinema screen. Separating those channels prevents explanatory copy from being mistaken for a proposed subtitle file.
This route also protects the scope of the conversation. A request for a translated pitch is not automatically a commission to rewrite the screenplay, dub the film, or replace its cultural references. Before solving a larger problem, identify whether that larger problem is actually open.
A useful deck sentence might be:
The scene is intended to play in English. The annotated exchange explains its sound-based misdirection for this pitch; it does not specify the film's eventual subtitles in other languages.
That leaves a real adaptation question visible without pretending the film's language is itself a defect.
A documented substitution has a purpose, not universal permission
Darcy Paquet described a specific choice in his English subtitles for Parasite: a reference to Seoul National University was changed to Oxford. In an Arirang News interview, he said Bong Joon-ho proposed a quickly recognizable institution because the line was a joke and immediate understanding mattered.[^1]
That is a translator's account of one decision. It is not a rule that local institutions should routinely be replaced with English-speaking ones. The relevant question is which function the name served in that particular line and what the substitution was intended to preserve.
In a separate interview, Paquet described working through subtitle revisions with the director and other participants. He also explained the invented word “ramdon” in relation to a later image of food packages. His account connects wording to information supplied elsewhere in the scene, rather than treating the isolated term as the whole problem.[^2]
These published accounts give us reasons expressed by the translator. They do not provide a frame-accurate comparison of source performance and final subtitles here, and they do not validate a new solution for the shoe scene. The transferable practice is to make the choice specific: what information must arrive, through which channel, and how quickly?
A substitute may carry the immediate function but alter a later reference. A line that works on its own may conflict with an image. A name may establish not just prestige but where a character studied, what they aspire to, or what another person assumes about them. Keep those consequences attached to the decision.
Route one: retain the performance and explain the limitation
For the scene as written, preservation may be the appropriate first route. Keep the spoken scene and its visual reveal. Give the pitch recipient the gloss and the mechanism, while acknowledging that the final subtitle solution is not yet settled.
This is especially coherent when the source performance matters more than producing the same verbal surprise in every language. The worker's matter-of-fact reply, the withheld setting, and the boot's arrival can remain central even if a later subtitle has to reveal “shoe” sooner.
What this route preserves is clear: source wording, vocal performance, character behavior, and the existing shot design. Its loss is also clear: someone relying on a translation that distinguishes sole from soul immediately may not experience the same ambiguity.
Do not conceal that loss behind “the humor is universal.” Explain which part of the scene can still be followed and which part depends on sound. The pitch recipient can then evaluate the actual proposed film rather than a claim that no translation problem exists.
Preservation is a decision about this film, not a declaration that translation is impossible. A qualified translator may find a viable construction you have not considered. Leaving that work open is different from announcing a solution before one exists.
Route two: adapt the subtitles without contradicting the scene
Subtitles add a visible information track. Their wording, segmentation, and arrival can alter the reveal even when the source audio stays untouched.
Take a deliberately literal English diagnostic paraphrase of the first line:
Can you repair the bottom of my shoe?
This is not a target-language translation. It shows a particular loss: the object is named before the worker answers. The scene can still make sense, but the intended double reading has gone.
Now try:
Can you save it?
This preserves uncertainty about the object while removing the specific personal meaning of “soul.” The viewer is waiting for a referent, not necessarily hearing a plea about an inner life. That may support the visual reveal, but it does so through a different mechanism. It also depends on whether the target language can make the pronoun natural without information the frame withholds.
A third diagnostic line is:
Can you save me?
This moves closer to the personal plea, but changes the grammatical object from a thing to a person. It can make the customer sound knowingly theatrical or distressed. If the performance is entirely casual, the translated line may create a mismatch. A clever ambiguity is not a complete solution when it changes the person speaking.
The comparison is not a contest between three finished translations. It identifies distinct costs a collaborator can assess in the actual language and scene: early disclosure, weakened personal meaning, or changed characterization.
Timing introduces another decision. Suppose the opening line occupies a proposed first beat, the reply a second beat, and the boot a third. A subtitle that combines the opening question with “Leather or rubber?” may expose the material clue while the actor is still delivering the ambiguous plea. Splitting the lines can preserve the order of information, but only if each cue remains readable within the actual performance.
Netflix's own subtitle timing guidance explicitly warns against revealing punchlines early and requires the file to be watched back. Those are instructions for Netflix deliveries, not a complete specification for every pitch clip; the relevant principle here is that wording and timing must be assessed together.[^3]
Do not solve the problem by withholding a subtitle until after the boot arrives if that leaves the viewer without a readable version of the question. Nor should you assume that a cue boundary automatically controls the exact instant a person understands a phrase. The proposed timing is something to inspect with the scene, not a promise about every reader's speed.
Set side by side, the three variants order the information differently: sole names the object before the reply, save it withholds it, and save me replaces the object with the speaker. That comparison is a text analysis, not a tested subtitle track.
Route three: rebuild the target-language scene
A reconstruction has more freedom because it changes the scene, not just the written translation. That freedom needs an explicit brief. Is the film being remade? Is a pickup possible? Can the props, shot order, or dialogue change? A subtitle commission usually cannot assume those options.
Here is an English reconstruction to expose the trade-off:
Customer, offscreen: Is it beyond saving?
Worker: Not until you've seen the price.
A boot lands on the counter.
The scene no longer depends on a homophone. It shifts from a potentially grave question to a joke about the repair price. The worker now makes a dry remark. In the original rehearsal, the worker could remain entirely literal.
We have preserved the hidden object and a turn toward ordinary commerce. We have lost the precise double meaning and changed who produces the humor. Calling this an “equivalent translation” would hide the most consequential edit.
A different reconstruction might put the misunderstanding into the customer's behavior: they arrive treating a battered shoe like a patient, while the worker refuses the melodrama. That would create actions and perhaps a recurring character trait not present in the original. It might be useful for a new film, but it is not a repair to a subtitle.
Check later scenes before accepting either route. Does the phrase recur? Does another character imitate the worker's literalness? Does the film build a pattern in which the framing deceives us while the characters remain practical? A local rewrite can solve the immediate line and weaken that larger pattern.
Also inspect social direction. Is the humor aimed at the repairer's craft, the customer's attachment, the size of the bill, or the audience's initial assumption? Moving the laugh from one relationship to another is a creative change, not just a linguistic adjustment.
A reconstruction brief should therefore state what can change as clearly as what must survive. “Keep it funny” delegates the whole argument without defining its boundaries.
Give collaborators the scene, not a puzzle detached from it
The translator needs the surrounding action, character relationship, relevant later references, and current cut or performance material. A line without its shot can conceal the very condition that makes it work.
For our example, the important constraints are that the room initially remains unidentified, the worker understands the transaction, and the material-specific reply begins the reveal before the boot confirms it. A proposed subtitle should be evaluated against those constraints, not only back-translated into tidy English.
Ask the collaborator to describe what their version makes the character seem to know and intend. Ask where the key information appears, which meaning is lost, and whether the register belongs to this speaker. A fluent explanation of the words is useful evidence about wording; it is not evidence that the audiovisual beat has been tested.
Keep the writer's proposed effect, the translator's interpretation, and an observed response separate. A collaborator may judge a phrase idiomatic while recommending a different pause. A viewer may understand the setup and not find the turn amusing. Neither observation should be quietly upgraded into “the joke works internationally.”
When reviewing a performance, preserve the variant that was actually tested. Record its text, cue placement, cut version, and who assessed it. Otherwise, a later wording change can inherit approval that belonged to a different version. Approval of a pitch explanation is not approval of the subtitle timing.
For this article's rehearsal, no target-language performance, qualified bilingual review, or audience trial has occurred. The published translator accounts are contextual evidence, not collaborators' approval of these invented lines. The mechanism comparison is useful now; an actual bilingual solution remains work to be done with authorized material and the relevant expertise.
State the decision the pitch can honestly make
For the original shoe scene, the defensible pitch-stage route is to retain the English performance and explain the ambiguity. It preserves the proposed film while making the language-dependent effect intelligible to a recipient. It does not settle future distribution subtitles.
The brief can finish like this:
Preserve the worker's practical response and the viewer-only ambiguity created by the restricted view. The reply should begin the material reveal; the boot should confirm it. This pitch includes an explanatory gloss, not a production subtitle solution. For other-language subtitles, test whether the ambiguity can survive without making either character misunderstand the transaction. Where it cannot, state the loss. A rewritten target-language scene requires a separate creative decision, because changing the mechanism may change the worker.
The reader now knows what the scene says, what it is intended to do, and which decision has not been made. That is more useful than a smooth translation that quietly promises a different film.
Sources
[^1]: Arirang News, Lee Min-sun, “Interview with English subtitle translator for 'Parasite'”, published interview and accompanying text, paragraph beginning “There's a scene in the film…” Inspected June 18, 2026. The accompanying text was inspected; this is not a claim of a timed audiovisual review of Parasite or of the interview video.
[^2]: Lee Hana, “Comedies Are a Fun Challenge”, KOREA magazine / KOCIS, September 2019, questions about joint revision, “ramdon,” and language choices. Inspected June 18, 2026. First-person translator account; statements about critics or universal audience response in the introduction are not adopted.
[^3]: Netflix Partner Help Center, “Timed Text Style Guide: Subtitle Timing Guidelines”, Introduction and “Timing rules.” Inspected June 18, 2026. Used for avoiding early punchline disclosure and full watchback, not numerical settings, certification, or a claim this example was delivered to Netflix.
Frequently asked questions
What are the different jobs a language-dependent joke might require?
The article separates three jobs: explaining an existing scene, designing subtitles for it, and proposing a target-language version of the scene itself. They carry different permission to change things. An explanatory translation can make the intended turn clear; a playable adaptation must work alongside speech, image, performance, and time.
In the shoe-repair rehearsal, who is meant to be confused by sole sounding like soul?
Neither customer nor worker has to be confused. They know they are discussing footwear. The intended misdirection belongs to a viewer who cannot yet see the setting. If the worker looks puzzled by a spiritual request, a character now makes the mistake rather than the framing briefly inviting the audience to make it.
What should a mechanism note state instead of just calling it a joke about shoes?
It should identify an ambiguity, a knowledge difference, and the moment of reinterpretation. The sample note says the first line permits two interpretations in sound, the worker treats it as an ordinary repair question, the material-specific reply begins the reveal, the boot confirms it, the viewer's restricted view creates the ambiguity, and the worker is not misunderstanding the customer.
What does the Parasite subtitle example support and not support?
It is one translator's account of a specific choice: Darcy Paquet said a Seoul National University reference was changed to Oxford, and Bong Joon-ho proposed a quickly recognizable institution because the line was a joke and immediate understanding mattered. It is not a rule that local institutions should routinely be replaced, and it does not provide a frame-accurate comparison or validate a new solution for the shoe scene.
How should subtitle timing and wording be evaluated?
Together. The article says a subtitle that combines the opening question with the reply may expose the material clue while the actor is still delivering the ambiguous plea. Netflix's guidance warns against revealing punchlines early and requires watchback, but those are delivery instructions, not a complete specification for every pitch clip.