Translating Idioms and Jokes for Video

An idiom translated word for word produces a sentence that is grammatically correct and communicates nothing. The reader understands each word and not the meaning, because the meaning was never in the words — it was in the phrase, as a unit, in one language.

This is where translation stops being a mapping between vocabularies and becomes a series of decisions with no correct answer. In video it is harder still, because the picture is fixed: the shot, the timing and anything shown on screen constrain what the target sentence is allowed to be. This guide covers a way to make those decisions consistently, the specific problem of jokes, and what to do about figurative language you have illustrated.

Translate the Function

The question that resolves most cases is not "what does this phrase mean" but "what is it doing here".

Idioms do a small number of jobs. They soften a claim, or intensify it. They signal informality and establish a relationship with the listener. They compress something that would otherwise take a sentence. They mark a shift in the argument.

Identify which job, and the target-language options become clear. A phrase that was softening a claim can be replaced by any softening device in the target, idiomatic or not. One that was compressing an explanation may need the explanation restored, which costs words and may cost time.

Notice how few rows recommend finding a matching idiom. Equivalence is available less often than people expect, and reaching for it produces the characteristic over-translated quality where every sentence contains a phrase that is technically correct and slightly odd.

The Picture Decides What Is Available

Text translation has the whole target language to work with. Video translation has whatever the image permits.

If the speaker says something figurative and the video illustrates it literally — a phrase about being under water accompanied by a shot of water — the image has fixed the metaphor. The target language now needs an expression that works with that picture, which may not exist, and the alternatives are to change the picture or accept an awkward line.

The same applies to gesture. A speaker who counts on their fingers, indicates a size, or points at something has constrained the sentence structure, because the words have to arrive at those moments.

This has a design implication that reaches back before translation. Illustrating a figure of speech literally is a decision that makes localisation harder, and knowing that in advance lets you choose a different image where the video is likely to be localised.

Jokes

Humour divides cleanly into cases that survive translation and cases that do not, and the division is worth knowing before you start negotiating with a line.

Situational humour survives. The joke is in what happens, and what happens is on screen and identical in every language. Translate the words plainly and it works.

Character humour usually survives. The joke is in who says it and how, which the delivery carries. It may need a register adjustment and it does not need re-invention.

Wordplay usually does not. Puns, double meanings, rhymes and sound jokes depend on specific properties of specific words. The honest options are to write a different joke that fits the same moment, or to let the line be plain and accept the video is slightly less funny in this language.

The third case is where most time gets wasted. Hours spent trying to preserve a pun almost always produce something worse than either alternative, and recognising the category early is worth more than persistence.

Where a joke is replaced, it has to fit the same duration and the same picture. That constraint is often what rules out the good replacement and leaves the plain option, which is a reason to decide quickly rather than to explore.

Culturally Specific References

Adjacent to idiom, and more often mishandled: references that assume knowledge the target audience does not have.

A comparison to a sport played in one country, a measurement in units used in one region, a reference to a television programme, a historical event, a brand that exists in one market. Each of these is perfectly clear in the source and blank in translation, and none of them is an idiom in the narrow sense.

The decisions are the same three. Replace it with a local equivalent, which works when one exists and the picture does not contradict it. Explain it briefly, which costs time you may not have. Or remove it and make the point directly, which is usually right.

The reason to prefer removal is that these references are almost always doing decorative work. A sports metaphor illustrating a point about timing is illustrating something the plain statement already contains. Cutting it loses a flourish and keeps the argument, and the target-language version reads as though it was written rather than converted.

Watch particularly for humour that depends on a shared reference, since it fails twice — the joke does not land and the audience knows they missed something, which is worse than a plain sentence.

Register Is Part of the Meaning

A subtler failure than mistranslation: getting the words right and the tone wrong.

Languages differ in how formality is marked and in what counts as neutral. An English instructional register that reads as friendly and professional can arrive as either stiffly formal or inappropriately casual, and neither is a translation error in the narrow sense — every word is correct.

The consequence is that the speaker seems like a different person in each language version, which undermines a series more than an occasional awkward sentence would.

The fix is to decide register explicitly and write it into the brief: how formal, how much distance, whether the audience is addressed directly. A translator given that guidance produces something consistent; one left to infer it from the source text will infer differently each time.

Working Through It

Several of these are much easier when the source script was written with translation in mind — fewer idioms, less wordplay, figurative language kept out of the shots that illustrate it. That preparation is set out in preparing a script so it dubs cleanly.

There is also a timing consequence to watch. Expanding a compressed idiom into plain language makes the line longer, and in a dub the line has a fixed duration — so a translation decision becomes a sync problem, which is the interaction described in why dubbed audio drifts out of sync.

Next Steps

Take a script you are about to localise and mark every idiom in it. Most people find between five and fifteen in a few minutes of video, which is a small enough list to decide deliberately and large enough to cause real damage if left to chance.

Then check which of them the picture illustrates. Those are the ones to resolve first, because they have the fewest options and they are the ones a translator working from text alone will get wrong. How these decisions are tracked across several languages at once is covered in managing subtitles in eight languages, with related material under dubbing and translation.