You know the feeling. Someone asks you a question in English. The answer forms in your head in your own language first. Then you translate it. Then you check the translation. By the time you are ready to speak, the moment feels stiff. The sentence you finally say sounds nothing like the way you think.
This is called mental translation, and it is one of the most common reasons it takes so long to sound fluent. The good news is that it is not a fixed trait. It is a habit, and habits change with practice.
Why translation slows you down
In the 1980s, the psycholinguist Willem Levelt mapped out what the brain does when a person speaks. His model, described in a standard psychology of language textbook, involves four stages: forming the idea you want to express, turning that idea into words and sentence structure, sending those words to your mouth, and monitoring what comes out.
Mental translation adds extra steps before any of this happens. Your idea forms in your first language. You find words for it in your first language. Then you look for the equivalent words in English. Then you run the full formulation process again in English.
Each extra step costs time. In a conversation where responses are expected within two or three seconds, those added steps are what create the lag. They are also what creates the self-consciousness: you can feel the delay, so you get nervous, which makes the next retrieval slower still.
Research on bilingual word access shows how real this shift is. A 2021 study in Frontiers in Psychology used ERP brain measures with Chinese speakers of English and found that advanced learners’ brains activated the sounds of English translations automatically, even while they were reading words in Chinese. Their English connections had become strong enough to fire on their own, with no deliberate translation step. Less advanced learners did not show the effect. In plain terms: direct links to English words are not a talent some people are born with. They are built with use.
The habit trains itself
The shift from translation to direct retrieval does not happen because you decide to “think in English.” It happens because English words get used often enough that they build their own direct connections to the concepts they represent.
Every time you retrieve an English word by going through your first language, that indirect route gets a little more practiced. Every time you use an English word directly, without stopping at the first-language equivalent, the direct route strengthens instead. Repetition determines which path becomes the default.
Three drills accelerate the process.
Name what you see, in English, as soon as you see it. Not translated. The word chair, the word traffic, the word cloudy. If you do not know the word, look it up right then, say it once, and move on. Over time the English label is what surfaces first.
Narrate small actions out loud. “I’m opening the fridge. There is no milk. I’ll have tea instead.” This forces sentence-level construction in real time, not just single words. The habit of forming complete thoughts in English builds the same direct connections, at the level of ideas rather than isolated vocabulary.
Build single sentences about what you are thinking right now. Not paragraphs. One sentence, immediately, about the present moment. “This queue is longer than I expected.” “I’m not sure what I want to eat.” The point is not eloquence. The point is retrieval under zero preparation time.
Why you still need a conversation
These drills move you toward direct retrieval. They share one limit: you choose what to say. In a real conversation, someone else sets the topic, asks the question, and expects an answer on their schedule.
The brain’s job under those conditions is harder. It has to retrieve English words directly while it also processes what the other person said and decides what to say next. Research using pupil-dilation to measure cognitive effort found that even proficient bilingual speakers show measurably more effort during word retrieval than native speakers of the same language. The same research found the gap shrinks with second language experience: the more you use the language, the less each word costs. Conversation is where that experience piles up fastest, because someone else keeps choosing words you were not prepared to find.
Drills are preparation. Conversation is the training itself. Each real exchange in English, where someone asks something you did not expect and you have to find words for it right now, is a repetition that makes direct retrieval a little faster.
What to do when you go blank
The blank moment, when you reach for English and your first-language word surfaces instead, is not a failure. It is a sign that a particular word or idea is still coming through the long route. The fix is not to freeze. It is to use the English version you can find, even if it is simpler, and keep going.
“I mean…” followed by a more basic sentence is better than silence. The simpler sentence is still an English sentence produced under pressure. That counts as a repetition.
Consistent speaking practice over time rewires the default. The day you notice you thought of something in English before you thought of it in your own language is not a milestone you can set out for. It arrives quietly, after enough repetitions.
When that starts happening, you will also notice fewer pauses and fewer filler words. Slower word retrieval is one of the main reasons speakers say “um” more than they want to, and direct retrieval reduces it at the source.
