How to Remember the Words You Read in a Foreign Language
You look up a word. You understand the sentence. You feel a small flicker of satisfaction, another piece of the language clicked into place. Three pages later, the same word appears again, and you have no idea what it means. You look it up a second time, recognize your own underlining from earlier, and feel a flicker of something less pleasant.
This is the central frustration of learning vocabulary through reading, and almost every reader runs into it. The looking-up is easy. The remembering is hard. You can read for an hour, meet forty unfamiliar words, understand all of them in the moment, and retain perhaps three of them by the next day. The reading felt productive, but the vocabulary leaked out almost as fast as it came in.
The good news is that this is not a personal failing or a sign that you have a bad memory. It is exactly how human memory is built to work, and it has been studied for nearly 140 years. The forgetting is predictable, which means it is also manageable. With a small amount of structure, you can keep most of the words you meet instead of watching them leak away.
This article is about that structure. We will look at why you forget, why a single encounter is almost never enough, why context matters more than word lists, and how to pair spaced repetition with reading without turning a pleasant activity into flashcard drudgery. None of it is magic. All of it is doable inside an ordinary reading habit.
Why You Forget: The Forgetting Curve
In 1885, a German psychologist named Hermann Ebbinghaus did something slightly mad. He memorized long lists of nonsense syllables, meaningless three-letter combinations like “WID” and “ZOF” chosen precisely because they carried no meaning to attach to, and then measured how quickly he forgot them over hours, days, and weeks. He was his own subject, his own experimenter, and the first person to study memory with anything resembling scientific rigor.
What he found is now called the forgetting curve. Memory of newly learned material decays rapidly at first and then more slowly. A large share of what you learn is gone within a day if nothing reinforces it, and the loss continues, though at a gentler rate, in the days that follow. The curve is steep at the start and flattens as time passes.
Ebbinghaus also found the antidote, and it is the foundation of everything in this article. Each time you successfully review something before you have fully forgotten it, the curve resets and flattens. The memory decays more slowly the next time. Review it again, and it decays slower still. His insight was that each repetition increases the optimal interval before the next one is needed, early on you must review within days, but later the same memory can survive months or years between reviews.
A word you meet once while reading is at the very top of the steepest part of that curve. It is a syllable Ebbinghaus memorized at noon and could not recall by dinner. The encounter was real, but a single encounter buys you almost nothing in terms of durable memory. The word will be gone tomorrow unless something brings it back.
Recognition Is Not Recall
There is a second reason the word “vanishes” that is more subtle, and understanding it changes how you should review. When you meet a word again in context and think “yes, I know that one,” what you are experiencing is recognition, the easier of two kinds of memory. When you have to produce the meaning of a word with no hints, that is recall, and it is considerably harder.
The research on this gap is consistent. In studies of vocabulary learned from reading, learners can typically recognize the meaning of substantially more words than they can recall on their own, one body of work found participants recognizing the meaning of around 61% of target items while recalling only about 45%. Recognition knowledge is not only easier to build, it is also retained better over time.
This matters because reading constantly flatters your recognition memory while doing little for recall. Every time a word reappears with the supporting context of a sentence around it, you recognize it, feel that you “know” it, and move on. But strip away the context, ask yourself what the word means cold, and the knowledge often is not there yet. A word you can only recognize is a word you will fail to understand the moment the context is thin, and a word you cannot yet use. Real ownership of vocabulary means recall, and recall takes deliberate effort that passive re-reading does not supply.
Why One Encounter Is Never Enough
If a single exposure does not stick, how many do you need? This is one of the most studied questions in vocabulary research, and the honest answer is: more than you would like, and it varies.
Early work by researchers including Paul Nation, the most influential figure in the field of vocabulary acquisition, suggested that a word needed to be met perhaps five or more times across different texts before it was reliably acquired. Later research pushed that number up. Webb’s work suggested that ten or more encounters were needed to build the fuller range of knowledge a word requires. Nation himself later acknowledged that the real figure for solid acquisition is likely higher than earlier estimates implied. A reasonable working figure, drawn from the spread of this research, is that most words require somewhere in the range of eight to a dozen or more meaningful encounters before they are securely yours, and harder, more abstract words can take more.
The number is not fixed because acquisition is not binary. Knowing a word is not a switch that flips from off to on. The first encounter might teach you the rough meaning. The third might pin down which of two meanings is intended here. The eighth might finally fix the spelling, or the gender, or the preposition it travels with. Each encounter adds a layer, and “knowing” a word is the accumulation of all those layers. Some words are engaging or vivid enough to lodge in fewer exposures; some research finds that highly salient or essential words can be picked up in as few as three. But you cannot count on that.
This single fact reshapes how you should think about reading for vocabulary. The goal of any one reading session is not to learn the new words you meet, that is mostly impossible in one pass. The goal is to log the encounter and make sure the word comes back to you enough times, and at the right moments, to climb that ladder of exposures. Reading provides the encounters. Your job is to ensure there are enough of them.
Why Incidental Reading Alone Is Slow
This is also why reading by itself, without any support, is a slow way to build vocabulary even though it is a wonderful way to consolidate it. Learning that happens as a byproduct of reading for meaning is called incidental learning, and it works, but the gains per session are modest. Meta-analyses of incidental vocabulary learning find that learners pick up a fairly small proportion of the unknown words they encounter in a given text, often under a fifth of them on delayed tests. Deliberate, intentional study of words, actually setting out to memorize them, produces noticeably higher retention rates per word.
The implication is not “stop reading and drill word lists.” We have written at length about why reading is the best way to learn a language, and we stand by it. The implication is that reading and deliberate review are complementary, not interchangeable. Reading gives you a flood of words in rich context, far more than any word list, and tells you which words actually matter by showing you what you keep meeting. Deliberate review takes the handful that matter most and pushes them from recognition into recall faster than re-encountering them by chance ever could. The best results come from doing both, which is the whole argument of this article.
Context Beats Isolated Word Lists
Before we get to the mechanics of review, one principle governs everything: save and review words in context, not as bare items. This is the single most important habit in this article, and it is the one most people get wrong.
A bare word list looks efficient. Madrugada, dawn. Aunque, although. Sin embargo, however. Twenty rows of word and gloss, neat and reviewable. The problem is that this strips away precisely the information that makes a word usable and memorable. A word in isolation is an abstraction. A word in a sentence is an event.
There are several concrete reasons context wins. First, memory is associative. You remember things by their connections to other things, and a word floating alone has few connections, while a word embedded in a sentence is tied to a situation, an image, a moment in a story. Many readers find that a word they cannot recall from a flashcard comes back instantly when they remember the sentence they first met it in, “salia de madrugada,” from a story about a fisherman leaving before dawn. The fisherman drags the word up with him.
Second, most words do not have a single tidy translation. A bare gloss forces a word into one English equivalent and hides the register, the connotation, the typical companions, the situations where it fits and where it would sound wrong. Context shows you the word doing its actual job, not just what it means but how it behaves.
Third, context disambiguates. Enormous numbers of common words have several meanings, and the gloss you write on a flashcard captures only the one you happened to meet. Reviewing the full sentence keeps the meaning anchored to a real use rather than a dictionary’s first entry.
This is why the parallel-reading approach is so well suited to vocabulary building: the translation sits at the level of the whole sentence, so the meaning you absorb is always the contextual meaning, never a decontextualized gloss. We go deeper into the trade-offs of that format in our honest guide to parallel reading. For now the takeaway is simple: when you save a word to review later, save the sentence with it. Never save the word alone.
Spaced Repetition: The Engine of Retention
We now have the two facts that make a method inevitable. Memory decays on a predictable curve, and each well-timed review flattens that curve. The technique that exploits this is spaced repetition, and it is the most reliable tool we have for moving vocabulary into long-term memory.
The idea is exactly what Ebbinghaus implied. Review a word just before you are about to forget it. Each successful review pushes the next review further out. A new word might come back to you after one day, then three days, then a week, then two weeks, then a month, then several months. You are surfing the forgetting curve, catching each word at the moment of maximum efficiency, late enough that the retrieval is effortful and therefore strengthening, but not so late that the word is already gone and you are learning it from scratch.
The first person to apply this systematically to language was the American linguist Paul Pimsleur, who in the 1960s designed an audio course around what he called graduated-interval recall. Pimsleur prompted learners to recall new vocabulary at expanding intervals, the sequence published in his work ran roughly five seconds, twenty-five seconds, two minutes, ten minutes, an hour, five hours, a day, five days, twenty-five days, four months, two years. The exact numbers matter less than the shape: each interval is several times longer than the last. Modern spaced-repetition software automates this, adjusting each word’s schedule based on how easily you recalled it, but the principle is unchanged from Pimsleur and traceable straight back to Ebbinghaus.
Why It Works When Re-Reading Does Not
It is worth being precise about why spaced repetition outperforms simply reading a lot. Two mechanisms are doing the work.
The first is the spacing effect: information reviewed at intervals is retained far better than the same information crammed in one sitting. Ten minutes of review spread across ten days beats a hundred minutes in one evening. Reading naturally provides some spacing, because words recur at random intervals across a book, but the spacing is uncharacteristically uneven, a word might appear three times on one page and then not for ninety pages. Spaced repetition makes the spacing deliberate and optimal.
The second is the testing effect, sometimes called retrieval practice: the act of pulling a memory out of your head, struggling for it, then succeeding, strengthens that memory far more than re-reading the same information. This is the crucial difference between recognition and recall again. Reading lets you recognize. A spaced-repetition review forces you to recall. That effortful retrieval is what converts a word you sort-of-know into a word you own. When you re-read, the answer is always sitting right there in front of you; the muscle never gets worked.
Pairing Spaced Repetition With Reading, Without Killing the Flow
Here is the legitimate objection. Spaced repetition is powerful, but stopping every two minutes to make a flashcard turns reading, the very thing that should be enjoyable and immersive, into a tedious clerical exercise. If review machinery destroys the reading habit, you have traded a long-term asset for a short-term one. Many people have abandoned reading in a foreign language for exactly this reason.
So the real skill is integration: capturing words for later review with as little friction as possible, and reviewing them in a way that does not turn your evening into a chore. Here is how to do that.
Tactic 1: Capture Fast, Review Later, Never Both at Once
Do not try to learn a word the moment you meet it. Trying to memorize mid-sentence breaks your reading flow and rarely works anyway, since one exposure does not stick. Instead, make capturing a word as close to instantaneous as possible, ideally a single tap that saves the word together with its sentence, and keep reading. The reading session stays a reading session. The reviewing happens in a separate, dedicated block, at a separate time. Mixing the two ruins both.
This is far easier when the tool captures the surrounding sentence automatically. If saving a word means opening a separate app, typing the word, typing a translation, and copying a sentence by hand, you will stop doing it within a week. The friction has to be near zero or the habit dies.
Tactic 2: Be Ruthless About Which Words You Mine
You cannot review every unknown word, and you should not try. A dense literary page might contain thirty words you do not know. Saving all thirty floods your review queue, most of them with rare words you will not meet again for months, and the genuinely useful words drown in the noise. Selection is not laziness, it is the whole game.
Two criteria should guide what you save:
- Frequency. Prioritize words that are common in the language and that you keep bumping into. If a word has appeared twice already in what you are reading, it will appear again, and it earns its place in your review queue. Rare, exotic words, a specific bird, an archaic insult, a piece of technical jargon, can usually be understood in the moment and then released. You will likely never need them again, and if you do, you will meet them again. We have written more on the question of how many words actually matter in how many words you need to read in a foreign language.
- Usefulness to you. A word you can already infer from context and will recognize next time does not need a flashcard. A word that is central to the sentence’s meaning, that blocks your comprehension when you do not know it, and that you can imagine using yourself, that word is worth saving. Trust your sense of which words are load-bearing.
A practical rule that many readers settle on: save a word only if you have met it at least twice, or if it is obviously a high-frequency word you will need again. This naturally caps your harvest at a sustainable rate, often around ten to fifteen words per half-hour of reading, and ensures the words in your queue are the ones worth the effort.
Tactic 3: Review in Short, Daily Blocks
Spaced repetition rewards consistency far more than intensity. Ten minutes every day beats seventy minutes once a week, because daily review keeps catching words at the right point on the forgetting curve. A weekly marathon lets too many words slide all the way down before you get to them.
Many readers find the best slot for review is immediately before the next reading session rather than after. A short review block warms up the language in your head, so the first minutes of reading are not wasted shifting mental gears, and it puts the words you have been saving back in front of you while the sentences they came from are still reasonably fresh. Five to ten minutes is plenty. The point is regularity, not volume.
Tactic 4: Keep the Review Itself in Context
When a word comes up for review, do not just stare at the bare word and try to dredge up a translation. Show yourself the sentence, the one you saved it in. Read the sentence, recall what the target word means within it, then check. This keeps the testing effect intact (you are still recalling, not just recognizing) while preserving the contextual associations that make the word memorable. Reviewing the sentence rather than the word alone is the difference between a flashcard you dread and one that actually reflects how language works.
Beyond Review: Re-Reading and Using the Words
Spaced repetition handles the bulk of retention, but two further tactics close the loop and are worth folding in.
Re-Reading
Re-reading a book, or a chapter, that you have already worked through is one of the most underrated vocabulary tactics available. On the second pass, the words you mined and reviewed reappear, now in their original rich context, now mostly known, and each reappearance is another high-quality exposure climbing the ladder toward acquisition. The grammar is familiar, the plot is familiar, so your attention is free to notice the language itself. A book you found hard the first time often reads almost smoothly the second, and that smoothness is the feeling of words consolidating. Re-reading is also gentle and pleasant in a way that flashcards never are, which makes it a habit you can actually sustain.
Productive Use
Recognition becomes recall through retrieval, and recall becomes true ownership through production, actually using the word yourself. You do not need a speaking partner for this. Writing a single sentence of your own using a newly learned word, or even just composing one silently in your head, forces a far deeper kind of retrieval than recognizing the word on a page. The word stops being something you decode and becomes something you can deploy. Even a small amount of this, a sentence or two a day with words you are trying to cement, disproportionately accelerates how quickly they stick. This is also one of the things passive apps tend to neglect, which is part of why tools like Duolingo stall advanced learners: they drill recognition endlessly but rarely push you into genuine production with words you care about.
An Honest Take on Anki vs In-Context Review
No discussion of vocabulary retention is complete without addressing Anki, the free, open-source flashcard program that is, deservedly, beloved by serious language learners. Anki implements spaced repetition rigorously and is enormously powerful. If you are willing to invest in it, it works.
But honesty requires naming its costs. Anki is friction-heavy for readers. To use it well alongside reading, you build each card by hand, the word, the translation, ideally the sentence, sometimes audio and an image. That card-building is real work, and it happens at exactly the wrong moment: while you are reading and want to keep reading. Many people start an Anki habit alongside their reading, maintain it for two weeks, find the card-making tedious, and quietly stop. A spaced-repetition system you do not use beats a perfect one you abandon by exactly zero words.
There is also a subtler problem. Cards built in isolation drift away from context. A standalone Anki deck becomes a list of word-gloss pairs divorced from the sentences that gave the words meaning, and we have already seen why that decontextualization weakens both memorability and usability. You can fight this by always pasting the source sentence onto the card, but doing so by hand, for every word, is precisely the friction that kills the habit.
The alternative is in-context review built into the reading tool itself, where saving a word is a single tap that automatically captures the sentence, and review happens against that sentence, all without leaving the place you do your reading. This loses some of Anki’s deep configurability. For most readers, that is a trade worth making, because the system that survives contact with daily life is the one that actually banks the words. The best tool is not the most powerful one in theory; it is the one whose friction is low enough that you still use it in month three. For a fuller account of how an integrated reading-and-review loop plays out in practice, see how I learned 3,000 words through reading.
Putting It All Together: A Simple Loop
Strip away the theory and you are left with a loop that takes very little time and is genuinely sustainable:
- Read for a set block each day, in real context, for meaning and enjoyment first.
- Mine ten to fifteen words per session, the frequent, useful ones, with a single tap that captures each word together with its sentence. Then keep reading. Do not stop to study.
- Review for five to ten minutes daily, ideally right before your next reading block, recalling each saved word from its sentence so you are training recall, not just recognition. Let a spaced-repetition schedule decide when each word reappears.
- Re-encounter the words naturally as you keep reading and, where you can, re-read material you have already mined, each reappearance is another rung up the ladder of exposures.
- Use a few of the words yourself in a sentence of your own, to push them from recall into ownership.
That is the entire system. It respects how memory actually works, the forgetting curve, the need for many spaced exposures, the gap between recognition and recall, the power of context, without demanding that you become a flashcard hobbyist. It fits inside an ordinary reading habit because most of it is an ordinary reading habit, with a thin layer of structure laid over the top.
How Lingo7 Builds This In
Everything above describes a method you could assemble by hand from separate tools. Lingo7 exists to make the whole loop happen in one place, with the friction removed at every step that usually breaks the habit.
When you meet an unfamiliar word while reading, you save it with a single tap, and it is saved in context, together with the exact sentence you met it in, never as a bare word stripped of meaning. Those saved words go into a built-in spaced-repetition system, so the app surfaces each one at the right point on the forgetting curve and you review against the original sentence rather than an isolated gloss. The capturing and the reviewing both live inside the same app you read in, so there is no separate flashcard program to maintain and nothing to copy by hand.
The reading itself is built for retention from the ground up. Every book offers sentence-level aligned parallel translations, tap any sentence to see its meaning, so the context you absorb is always the contextual meaning of the whole sentence, exactly the kind of input that makes words stick. Synchronized native audio lets you read while listening, adding a second channel of exposure that strengthens memory and connects the written word to its sound. On-demand translation is there for the moments you need it, and Lingo7 supports 90+ languages, on iOS and Android, free to start.
In other words, the structure this article argues for, save in context, review on a schedule, re-encounter in rich input, build recall rather than mere recognition, is the structure Lingo7 is designed around. You bring the reading habit; the app handles the remembering.
The Bottom Line
You forget the words you read for a reason that has nothing to do with your memory being poor. A single encounter sits at the top of the steepest part of the forgetting curve, and most words need eight, ten, a dozen or more meaningful encounters before they are securely yours. Recognition comes easily and flatters you into thinking you know words you cannot actually recall, while recall, the kind of knowing that survives thin context and lets you use a word, takes deliberate, effortful retrieval that passive re-reading never supplies.
The fix is not to read harder or to drill endless lists. It is to add a thin, sustainable layer of structure to reading you already enjoy: save the words that matter together with their sentences, review them briefly and consistently on a spaced schedule so you catch each one just before you forget it, keep meeting them in real context, and occasionally use them yourself. Do that, and the words stop leaking away. They accumulate. And the quiet satisfaction of recognizing a word on page two hundred, and actually knowing it, not just looking it up for the fourth time, becomes the ordinary experience of reading rather than the rare exception.