Principle 01
Faithful, Fluent, Refined: Literal vs Free Is a Dial
Xin, Da, Ya — A Dial, Not a Dogma
Translation · Trade-off
One-line principle + a master's words
Translation is not a choice between "literal" and "free." It is setting a ratio, for this job, among fidelity (xin), fluency (da), and refinement (ya). Yan Fu named the difficulty over a century ago.
"Translation involves three difficulties: fidelity, fluency, and elegance. To be faithful is hard enough; yet fidelity without fluency is no translation at all — so fluency must be sought too."
— Yan Fu, preface to his translation of Huxley's Evolution and Ethics (1898)
In short: fidelity alone is already brutally hard; but faithful-yet-unreadable is as good as untranslated. The three pull against each other — set the ratio by purpose.
Why it works
Word-for-word keeps fidelity but often loses fluency; free translation gains fluency but may shed the original's flavor. Lu Xun once argued for "faithful even if not smooth," because he wanted to import new syntax and ideas through hard translation. But for everyday expression the ratio flips by context: technical docs turn the dial toward fidelity (better stiff than ambiguous), marketing copy turns it toward refinement (rewrite freely, but land the punch). Decide which end this job needs, then choose how close to hug the literal.
Literal / FaithfulFree / Refined
hug the words · keep structure · kill ambiguitychase the effect · sound native · move the reader
Technical docs sit left, ads and literature sit right — same sentence, purpose sets the dial
Revision in action
It's raining cats and dogs. → literal Chinese "the sky drops cats and dogs" (absurdly faithful, reader baffled)
大雨倾盆 / 雨下得跟泼水似的 — the latter is both fluent and keeps the colloquial wink; a better ratio
胸有成竹 → "have bamboo in one's chest" (keeps the image, but no one understands)
"to have it all worked out in advance" (clear; for a literary reader, add a note: "lit. bamboo already in the chest" to keep some refinement)
When to use + common mistakes
- ✓ Technical terms (lean faithful), slogans and brand lines (lean refined), bilingual onboarding docs
- ✓ Self-check: first ask "does this job want zero ambiguity, or maximum impact?", then set how literal to go
- ✗ Blind literalism to flaunt "loyalty" — produces translationese no one can read
- ✗ Lazy free-rendering — losing the original's precision and personality along the way
This week's exercise + question
Take an English maxim you love and write three Chinese versions: ultra-literal, ultra-free, and your chosen best ratio — with a sentence on where each fits. Question: when fidelity and fluency cannot both be had, what tells you which to sacrifice? Is it the same instinct you use when writing original prose?
Principle 02
The Untranslatable: When There Is No Match, Pick One Facet
The Untranslatable — Saying Almost the Same Thing
Translation · Culture
One-line principle + a master's words
Some concepts have no equivalent in the other language. Admit that, and translation rises from "find the match" to "re-express at the highest level" — what Qian Zhongshu called huajing, the realm of transformation.
"The highest standard of literary translation is 'transformation.' To turn a work from one language into another so that it betrays no trace of stiffness from the difference in habits, yet fully preserves the original flavor — that is to enter the realm of transformation."
— Qian Zhongshu, "Lin Shu's Translations" (1964)
Compare: Umberto Eco summed translation up as "saying almost the same thing" (dire quasi la stessa cosa) — everything hangs on that "almost."
Why it works
"江湖" rendered as "rivers and lakes" means nothing, because depending on context it is the martial-arts world, society's grey underbelly, or a drifting life — one word carrying a whole culture. Conversely, "serendipity" and "privacy" once had no ready word in Chinese. Faced with the untranslatable, forcing a fake equivalent is the worst option; the right move is to see which facet this job needs, choose a word for it, and add a line of explanation if needed. For a leader communicating across cultures, that is exactly the self-awareness required when introducing a native framework to a global team: don't pretend a perfect equivalent exists.
Revision in action
Rendering 内卷 as "involution" (the academic source word; a lay audience is lost)
Pick a facet by context: "rat race" / "zero-sum overcompetition," then add "everyone works harder for the same slice" (facet + explanation)
"He has a lot of privacy concerns." → literal Chinese leaves the concept floating and stiff
他很在意私人信息不被随意收集和公开。(Don't chase one word; spell the concept out — that is the "transformation")
When to use + common mistakes
- ✓ Culture-loaded words, introducing native terms outward, explaining Chinese concepts to a global team
- ✓ Introducing a foreign concept in your writing: first admit it's untranslatable, then "transform" it in a sentence
- ✗ Worshiping "one word for one word" — culture-bound terms least tolerate a 1:1 swap
- ✗ Dropping the source word with no gloss (e.g. "involution," "guanxi") — building a wall for outsiders
This week's exercise + question
Pick a Chinese word you feel "English lacks" (e.g. 缘分, 孝, 面子) and write 80 words of English that make a foreign colleague understand it — no coinage, only explanation and examples. Question: when a concept is "untranslatable," is the language missing a word, or do the two cultures simply not care about the same thing? What does that mean for your cross-cultural expression?
Principle 03
The Translator's Invisibility: Read Like the Original, or Show the Foreign Grain?
The Translator's Invisibility — Domesticate or Foreignize
Translation · Stance
One-line principle + a master's words
Readers prefer a translation that "doesn't read like a translation" — fluent, transparent, as if the original author wrote it. Venuti warns: this invisibility is a choice, not a neutral default.
"A translation is judged acceptable... when it reads fluently, when the absence of any linguistic or stylistic peculiarities makes it seem transparent, giving the appearance that it... is not in fact a translation, but the 'original.'"
— Lawrence Venuti, The Translator's Invisibility (1995)
In other words: a translation passes as "good" largely because it reads smoothly and shows no rough edges of language or style — transparent, as if it were not a translation but the original itself.
Why it works
Venuti named two roads: domestication (move toward the reader, smooth away the foreign flavor) and foreignization (move toward the original, keep the strangeness). Domestication reads easily but tends to swallow the original's personality; foreignization stays true but raises the reading cost. This is not only about translating books — when you "translate" a Western management framework to your team, do you package it to feel like native wisdom (domesticate), or keep its original context and tension (foreignize)? Both roads are legitimate; the error is defaulting to domestication unconsciously, flattening the meaning without noticing.
Domestication
move toward the reader
fluent, familiar, zero friction
cost: grinds off the original's edges
Foreignization
move toward the original
keeps strangeness and source context
cost: the reader works harder
Before choosing: do I honor the reader's comfort, or the original's true face?
Revision in action
如鱼得水 → "like a fish in water" (foreignizing — keeps the image, slightly odd to English ears)
Domesticated: "like a duck to water" (idiomatic, frictionless, but the Chinese "fish" image is gone — choose by reader and purpose)
Introducing "disagree and commit" to a team by importing the English phrase raw — no one really grasps the culture behind it
Domesticated: "argue it out in the room, then carry it out together" — landed in native idiom, but mind that you don't lose the "voice your dissent first" layer of the principle.
When to use + common mistakes
- ✓ Localizing product copy, importing outside methodologies, explaining cross-cultural concepts to a team
- ✓ Audit your translation: am I helping the reader, or quietly rewriting the original?
- ✗ Unconscious over-domestication — reads smoothly, but the meaning has been swapped out
- ✗ Stiff foreignization for "fidelity" — if the reader can't get through it, fidelity is wasted
This week's exercise + question
Take a foreign proverb and write one fully domesticated version (like a native saying) and one deliberately foreignized version (keeping the exotic image); note who each suits. Question: when you "translate" a foreign idea to your team so it sounds like native wisdom, is that skillful communication, or a quiet distortion?
Principle 04
CN–EN Traps: Your Mother Tongue's Grammar Leaks In
CN–EN Traps — Your First Language Leaks In
Translation · Practice
One-line principle + a master's words
The most insidious translation disease is carrying the source language's syntax into the target: Chinese picks up "Westernized" clutter, English carries a Chinese accent. Yu Guangzhong sounded the alarm decades ago.
"English was not learned well, yet Chinese was learned badly... formulaic translationese is polluting the ecology of the Chinese language."
— Yu Guangzhong, "The Normal and the Pathological in Chinese" (1987)
In short: translators force English structures (passives, abstract nouns, long pre-modifiers) into Chinese; it looks "accurate" but corrupts the language's native economy.
Why it works
Chinese favors verbs, short clauses, and few passives; English favors nouns, long sentences, and subordinate clauses. Translate without actively "re-boning" the sentence and both sides go wrong. The classic Chinese symptoms are empty helper-verbs (進行/作出/加以), overused 被 and 們, and adjectives piled into long modifiers; the classic English symptoms are missing articles, the "although...but..." pairing, and rendering a Chinese topic-comment sentence into loose English. There is only one cure: after translating, reread with the target language's native ear and delete the source language's shadow.
Revision in action
我们对这个问题进行了深入的讨论，并作出了相应的决定。(Westernized clutter; verbs hollowed out into nouns)
我们深入讨论了这个问题，并作了决定。(verbs restored to primacy — clean)
Although the deadline is tight, but we can make it. (Chinese 虽然...但是 pairing; English keeps only one)
Although the deadline is tight, we can make it. / The deadline is tight, but we can make it.
When to use + common mistakes
- ✓ Translating English source into a Chinese report, writing email in your second language, auditing the tone of bilingual docs
- ✓ A "re-boning" read before the final draft: judge only whether the target language flows, ignore the source
- ✗ Believing "hugging the source" equals "accurate" — you've imported the disease along with it
- ✗ Chinese piled with 性/化/被/們/endless 的; English dropping articles and running loose
This week's exercise + question
Dig up a recent bilingual email; mark every 進行/作出/加以/被……所 and every redundant 的, and fix them one by one; then read the English aloud to catch missing articles and "although...but..." Chinese tells. Question: AI translation rarely makes these grammar errors anymore — so what judgment in "re-boning" remains irreplaceably human?
Reflection
Going Deeper
Push the edges
AI translation is already fluent — is studying the craft still worth it?
More than ever, just shifted in focus. AI excels at fluency — smooth, grammatically clean; but the fine trade-offs of fidelity (which facet of a word to keep) and the style judgments of refinement (solemn or playful here?) still need someone who understands the context, the reader, and the original's intent. Your value is no longer "able to translate" but able to direct and calibrate: deciding which way to turn the dial this time, and reclaiming the meaning when AI translates "too domesticated." The deeper you know the craft, the better you command the tool.
Isn't "the realm of transformation" just an excuse for over-free, self-indulgent translation?
No — the line between them is whether the original's full intent is preserved. Transformation demands "fully preserving the original flavor," which is harder, not looser: leave no stiff trace, yet lose not a drop of meaning. Over-free translation sacrifices accuracy for smoothness and feels good about losing things. The test is simple: translate your "transformation" back to the source — would the original author nod, or say "that's not what I meant"?
How does the skill of translation transfer to writing original prose?
Translation is "dancing in chains" — the meaning is locked, so you can only maneuver in expression, which is the purest diction training there is. Frequent translators grow acutely sensitive to "ten ways to say the same thing," and choose words more precisely when writing originally. Deeper still: writing itself is a kind of translation — rendering a vague notion in your head (source language) into clear words (target language), facing the same fidelity trade-offs and the same "untranslatable" snags.
For the truly bilingual, should you aim for "seamless switching," or keep a distinct persona in each language?
The masters keep the latter. Languages are not interchangeable containers; Chinese economy and reserve and English precision and directness naturally summon different expressive personae. Forcing the two selves to be "identical" usually flattens both. A better goal: be idiomatic and forceful in each language, letting the Chinese you and the English you each have their strengths — which is exactly the counter-lesson of this whole Day 33: don't let one language colonize the other.
For global teams, should you standardize on English, or let multiple languages coexist?
It depends on what you optimize for. Standard English lowers collaboration friction but systematically disadvantages non-natives and loses precise expression native to other languages. The pragmatic middle is "English as the collaboration base, key concepts kept in the original + explained": write important decisions in clear English, but when you hit an untranslatable native concept, don't force it — name the source word and "transform" it open. Let English be a bridge, not a steamroller of all difference.