Calques Natives Never Say

Learning outcomes

You ask a teammate “How is it called, this pattern?” and they answer “Oh, it’s called the adapter pattern.” They understood you. They were not even surprised. And yet no native speaker would have phrased the question that way; they would have said “What’s it called?” Your sentence was grammatical, the words were correct, the meaning was clear, and it was still not English. It was your first language wearing English clothes. This page is about that specific kind of error, the one that hides because nothing in it is technically wrong.

After studying this page, you can:

• Define a calque as a word-for-word translation from your first language that looks grammatical but that native speakers never actually say.
• Recognize the broadly shared calque patterns that show up across learners from many language backgrounds, in fixed expressions, verbs, tense, grammar, and prepositions.
• Replace a calque with the natural English chunk by retrieving a stored unit instead of translating a phrase.
• Run an audit of your own first language to find the calques you personally carry, and tell a true calque from a former calque that English has since adopted.
• Avoid the deeper trap of translating idioms literally, and route any idiom you are unsure of to the right reference.

Before we dive in

Here is the problem that makes this whole topic necessary. You learned English words one by one, and you learned grammar rule by rule, and both of those systems work. So when you want to say something, your mind does the natural thing: it forms the thought in your first language, then swaps each piece for its English equivalent. Most of the time the result is fine, because the two languages happen to line up. But every language stores certain ideas as fixed multi-word units, and those units almost never line up across languages. When you translate one of them piece by piece, you get a sentence that obeys every rule of English grammar and that no English speaker has ever produced.

Think about how you say your age. In English you are a number of years: “I’m thirty.” In many languages you have the years: the literal sentence is “I have thirty years.” A learner from one of those languages, translating faithfully, says “I have thirty years,” and the listener understands instantly, because the words are all real English words in a legal English order. Nothing flags it. But it is not what anyone says. The English for that idea is stored as a fixed chunk, “I’m thirty,” and you cannot reach it by translating; you can only reach it by already knowing it.

This is the trap. Unlike a grammar mistake, a calque does not break any rule you could be taught to follow. Unlike a vocabulary mistake, it uses no wrong words. It is correct at every level except the only one that matters: it is not the unit English actually uses. And because it is understood, no one corrects it, so it can survive for years. The good news, exactly as with the wider problem of why you sound translated, is that calques fall into a small number of repeating patterns. Once you can see the patterns, you can hunt down your own.

What a calque is, and why it slips past you

The word calque comes from the French for “a tracing,” a copy made by laying paper over an original and following its lines. That is exactly what happens here: you lay English over your first language and trace its shape. A calque is a word-for-word translation of a phrase from one language into another that produces a grammatical-looking but unidiomatic result, an expression native speakers do not actually use.

The reason a calque slips past every check you have is worth stating plainly, because it explains why this error is so persistent. A spell-checker sees only real words and stays silent. A grammar checker sees a legal sentence and stays silent. Your listener sees clear meaning and answers it, so they stay silent too. Every alarm you rely on is built to catch wrong words or broken grammar, and a calque has neither. The only thing wrong with it is that it is not the stored English unit, and no tool you own measures that.

To make the shape of the error concrete, look at a spread of common calques next to what is actually wrong with each and the form a native would use. Read the middle column to understand the error, and the right column to fix it.

None of the left-hand sentences would fail a beginner’s grammar test in the abstract, and several of them follow a rule that is perfectly real in the learner’s first language. That is the whole point. The fix in every case is not a rule to apply in the moment but a chunk to have stored in advance. The next section sorts these into the families they come from, because seeing the family is what lets you predict the error before you make it.

The repeating patterns of transfer

Calques are not random. They cluster, because the places where languages disagree are roughly the same places for everyone: the fixed expressions, the empty verbs that carry meaning, the way time is packaged, a few stubborn points of grammar, and the small words like prepositions. We will take these five families in turn. The examples are drawn from errors that recur across speakers of Romance, Germanic, Slavic, and Asian first languages, so treat the family as the lesson, not any single language’s version of it.

1. Fixed expressions translated whole

The first and largest family is the fixed expression: an everyday phrase that your language stores as one unit and that you translate as one unit, getting a result that is legal but foreign. “How is it called?” is the classic case. Many languages ask for a name with their word for “how,” so the learner produces “How is it called?” or “How do you say?” English stores this question differently, as “What’s it called?” or “What do you call it?” There is no rule you broke; you simply reached for the wrong stored question.

The same thing happens with “I have thirty years” for your age, with “make a photo” instead of “take a photo,” and with “make a party” instead of “have a party” or “throw a party.” In each case your language has settled on one verb or one frame, and English has settled on another, and only memory of the English unit gets you there. These are exactly the collocations that another page treats in depth: habitual partnerships that must be stored whole because no logic predicts them.

2. The wrong delexical verb

The second family is a special, high-frequency case of the first: the small set of nearly empty verbs that English uses to carry a noun. “Make a decision,” “take a photo,” “have a party,” “do your homework”: in each, the verb adds almost no meaning, and the choice of which verb is pure convention. Your language has its own conventions, and when they differ, you get “do a decision,” “make a photo,” “make a party.” The verb is the visible part of the error, but the cause is the same as before: a stored unit translated instead of retrieved.

These verbs deserve their own mention because they are the densest source of calques in the whole language. The full treatment of why “make a decision” and not “do a decision,” and how to store the partnerships, lives in collocation, the heart of word choice. For this page, the point is simply that whenever you find yourself attaching “make,” “do,” “take,” “have,” or “give” to a noun by reasoning, you are at high risk of a calque, and you should be retrieving the pair, not building it.

3. Tense and time that your language packages differently

The third family is about time. Languages slice time differently, and two slices cause most of the trouble. The first is duration versus start point. English uses “for” to measure how long something lasts (“for three years”) and “since” to mark when it began (“since 2021”). Many languages use one word for both, so the learner writes “since three years,” which to an English ear says “starting three years,” a start point that is really a duration.

The second slice is the present perfect, the tense English insists on for an action that began in the past and continues now. “I am here since Monday” translates a present-tense construction that is correct in many languages, but English requires “I’ve been here since Monday,” because the staying started in the past and is still true. The diagram below shows how a single thought, “an action that started in the past and is still going,” forces a choice of tense that your first language may not make.

The diagram makes the mechanism visible: the calque is not a careless slip but the predictable result of one language’s time-packaging overriding another’s. When the thought is “started before now and still true,” English wants “have been,” and reaching for the plain present, however natural it feels, traces your first language onto English. The fix is to bind the trigger (“since,” “for,” “so far,” “up to now”) to the tense, so the unit arrives complete.

4. Grammar carried across: countability, comparatives, agreement

The fourth family is grammar itself, carried across whole. Three points recur for almost everyone. The first is countability. English treats “information,” “advice,” and “feedback” as uncountable, like water: no plural, no “an.” Many languages count these nouns, so the learner writes “informations,” “advices,” “feedbacks.” Each is a direct trace of a language where the noun is countable, and each is wrong in English no matter how much information is meant.

The second is the comparative. English forms the comparative and superlative of a short adjective with the endings “-er” and “-est” (“big, bigger, biggest”), and uses “more” and “most” only for longer ones (“more important, most important”). Languages that build every comparison with their word for “more” produce “the most big” and “more easy.” The third is agreement. “People” is already plural, so it takes “are,” not “is”; and after “did” the main verb returns to its base form, so it is “I didn’t go,” never “I didn’t went.” Both “people is” and “I didn’t went” trace an agreement rule that holds at home but not here.

5. Prepositions and small words

The fifth family is the smallest words, which carry the least meaning and cause the most trouble. Prepositions almost never map one-to-one across languages, so translating them is a near-guaranteed calque. “Discuss about it” adds a preposition that English forbids, because “discuss” already includes “about”: you “discuss it.” “Explain me the bug” drops the preposition English requires, because “explain” needs “to” before the person: “explain the bug to me.” “In the Monday morning” picks the wrong preposition entirely, because English uses “on” for days and dated mornings: “on Monday morning.”

Two more small-word calques round out the family. “I am agree” treats the verb “agree” as if it were an adjective needing “to be,” which is how several languages build it; English just says “I agree.” And “open the light” and “close the light” translate languages that open and close their lights, where English turns them on and off. None of these can be reasoned out, because the small words are convention all the way down. They must be stored as part of the chunk: not “explain” plus a guessed preposition, but the whole frame “explain something to someone.”

Mental Model: translate the meaning, retrieve the chunk

Here is the shift that turns this from an endless list of mistakes into a single habit. The wrong model is that speaking English means translating: you form the thought in your first language and convert it, phrase by phrase, into English. Under that model, a calque is just a translation that came out a little off, and the cure would be to translate more carefully. But translating more carefully cannot help, because the calque is a faithful translation. “I have thirty years” is what careful translation produces. The model itself is the disease.

The better model, the one that the idiom principle builds the whole domain on, is that a native does not translate phrases at all. They translate the meaning, then retrieve a ready-made English unit for it. The thought “state my age” does not get converted word by word; it triggers the stored chunk “I’m thirty.” The thought “ask the name of something” triggers “What’s it called?” Fluency is not faster translation; it is the replacement of phrase translation with chunk retrieval. The diagram below contrasts the two routes for one thought.

The diagram is the same fork you met in why you sound translated, now pointed at this one error. Notice what it implies about study. You will never delete calques by getting better at translation, because better translation only sharpens the route that makes them. You delete them by moving more meanings onto the upper route: meeting the English chunk, storing it whole, and retrieving it next time so the phrase translation never even starts. Every calque you fix is one meaning relocated from the lower route to the upper one.

The cure: store chunks and audit your own language

The general cure follows straight from the model: whenever you catch a calque, do not just fix the sentence, store the correct chunk so the calque cannot form again. Bind the whole unit, with its triggers: not “explain” alone but “explain something to someone”; not “since” alone but “for + a length of time” against “since + a start point”; not “decision” alone but “make a decision.” Stored as units, these arrive complete and leave no slot for a translated guess.

But the general cure has a personal half that matters more, because your calques are not everyone’s calques. They are the specific places where your first language disagrees with English, and that set is yours. The most efficient way to improve is therefore to audit your own language directly: list the everyday expressions, verb partnerships, tense habits, and prepositions that your first language builds differently from English, and check each against the natural English chunk. This turns a vague worry into a finite, personal list you can actually finish.

This audit is the same discipline of noticing that the foundations track calls building intuition rather than memorizing rules. You are not trying to learn English grammar again; you are mapping the small, specific border between your language and English where tracing happens, and replacing each traced phrase with the unit English actually keeps.

When a calque has quietly become correct

One honest complication keeps this from being a rule you can apply blindly: language borrows, and a few expressions that began as calques have been adopted into standard English and are now simply correct. English itself is full of calques it took from other languages and naturalized. “Flea market” is a word-for-word tracing of the French marché aux puces; “long time no see” entered English as a loan-translation and is now ordinary; “that goes without saying” traces the French cela va sans dire. These were once foreign tracings and are now native chunks, indistinguishable from any other.

The practical lesson is not to relax, because the borrowed calques are a tiny, already-accepted set, while the calques this page is about are the vast majority that have not been adopted and still sound foreign. You cannot tell which is which by reasoning, only by evidence: a former calque that English has accepted appears constantly in native usage, and a true calque appears almost never. So the test is the same corpus check that settles any collocation question, described in collocation, the heart of word choice: if natives actually say it, it is in; if they do not, it is a calque to delete, however logical it feels.

Common mistakes

The first mistake is the one this whole page treats: translating a multi-word phrase instead of retrieving the English chunk, which manufactures every calque in the table above. The fix is the model shift, translate the meaning and retrieve the unit, plus the personal audit that finds your specific calques.

The second mistake is trusting that “understood” means “correct.” Because every calque is understood, the absence of correction feels like approval. It is not; it is politeness. Treat clear communication as the floor, not the goal, and judge your English against what natives say, not against whether you were understood.

The third mistake, and the most damaging at advanced level, is translating idioms literally. A calque of an ordinary phrase sounds slightly foreign; a calque of an idiom can be unintelligible or accidentally funny, because an idiom’s meaning is not in its words at all. Translating your language’s idiom for “it’s raining hard” or “I’m in a hurry” word by word produces something an English speaker cannot decode. Idioms must never be translated; they must be learned whole, which is the entire subject of idioms that earn their keep. When you are unsure whether a vivid phrase travels, assume it does not, and look up the English idiom instead.

The fourth mistake is over-correcting into stiffness once you become calque-aware. Some learners, having learned that their casual phrasing is foreign, retreat into careful, formal, textbook English to feel safe. That trades one unnatural register for another. The aim is the natural chunk, which is usually the everyday one, not the formal one, as natural alternatives to textbook English shows in detail.

Mastery Questions