Vocabulary in context: why it beats word lists for retention

Why is a list of isolated words hard to remember?

Your brain doesn't store a word like a dictionary entry. It stores associations. When you learn "casa = house" from a list, you build an artificial link between two labels, with no information about how that word actually behaves in a sentence.

The result: you can translate the word in isolation, but you freeze the moment it shows up in a real sentence, with its usual collocations, its register, its shades of meaning. Spanish "casa" isn't exactly English "house." It can mean a family home, a business ("la casa Ferrari"), or an idiom like "como Pedro por su casa" (acting completely at ease, uninvited). A list captures none of that.

This lack of context has a measurable cost on memory. Without a semantic anchor, without a mental image, without an attached emotion, the word stays a floating piece of information. That's exactly the kind of learning your brain forgets fastest, based on the forgetting curve documented since Hermann Ebbinghaus: without reinforcement, up to 70% of new information disappears within 24 hours.

Word lists also have a misleading frequency problem. They often present words that sound useful in theory, but rarely in the order you'd actually encounter them while reading or speaking. You end up learning "hedgehog" before "despite," even though the second word is a hundred times more common in real text.

What is comprehensible input and why does it change vocabulary learning?

Linguist Stephen Krashen formulated the input hypothesis (i+1) in the 1980s: we acquire a language through exposure to messages slightly above our current level, understood through context, not by memorizing rules or vocabulary in isolation. The "i" stands for your current level, the "+1" is the small dose of novelty that context helps you decode.

This theory draws a line between two processes Krashen considered fundamentally different: acquisition, unconscious and durable, which comes from exposure to meaning, and learning, conscious and fragile, which comes from explicit study of rules or lists. A word list typically feeds the second process. Reading a story that grips you feeds the first.

In a text, every new word arrives surrounded by clues: the sentence's syntax, the familiar words around it, the plot in motion, sometimes an illustration. Your brain guesses the meaning before even checking it, and this act of inference activates different brain regions than rote memorization. That difference explains why children acquire their native language without ever opening a vocabulary list.

Comprehensible input doesn't require understanding 100% of a text. The sweet spot sits around 90 to 95% of words already known. Below that, cognitive load becomes too heavy and you shift from fluent reading into laborious decoding. Our article on comprehensible input and the Lira method breaks down how to calibrate that zone for your level.

How does episodic memory help you retain a word met in context?

Human memory isn't a single system. Researchers distinguish semantic memory, which stores abstract facts ("Paris is the capital of France"), from episodic memory, which stores lived events, with their time, place, and emotional context.

A word learned from a list almost exclusively engages semantic memory. A word met while reading, on the other hand, attaches itself to a whole episode: the character who says the line, the tension of the moment, where you were when you read it. That contextual richness creates more retrieval "hooks." The more anchor points a memory has, the easier it is to retrieve later, even though a single hook is often enough to reactivate the rest.

This mechanism is called contextual encoding, a well-established principle in cognitive psychology since Endel Tulving's work on episodic memory in the 1970s. The context present at the moment of learning becomes part of the memory itself, not just background scenery around it.

Emotion plays a direct role too. A word met in a passage that made you laugh, jump, or feel something benefits from emotional tagging that strengthens long-term consolidation, a phenomenon well documented in the literature on memory and affect. A neutral word list, drilled mechanically, triggers none of these mechanisms.

Does this mean classic flashcards are useless?

No, and it would be a mistake to think so. Flashcards have a real role, just later in the learning process, not at the moment of first meeting a word.

The problem with generic, pre-made flashcards isn't the principle of spaced repetition, which remains one of the most solid tools in memory research. The problem is the word stripped of context. A card reading "efficace = efficient" trains pure recognition, and nothing else: not the collocation, not the register, not the nuance that separates "efficace" from "efficient" in English.

Spaced repetition works better as a consolidation tool for a word you've already met with meaning attached, not as an entry point. Once you've crossed a word in a sentence that meant something to you, seeing it again at increasing intervals reinforces that existing memory trace instead of building an artificial one from scratch.

This is an important nuance: framing "reading in context" against "flashcards" as an either-or is a false choice. The two methods work well together when the order is right: first the encounter in context, then spaced consolidation.

How does Lira apply this principle in practice?

Lira starts from the idea that the best vocabulary to learn is the vocabulary you actually meet while reading a text that interests you, not a generic list built for an average learner who doesn't exist.

When you read an imported book (EPUB, PDF, a web article, or a classic from the Gutenberg library), you simply tap an unknown word to see its contextual translation. It accounts for the full sentence, not just the isolated word, to give you the meaning actually used at that exact point in the text.

That word then goes into spaced repetition using the FSRS algorithm, with one key difference from a generic flashcard app: the original sentence context stays attached to the review card. When the word comes back up for review a few days later, you see the sentence where you first met it, not just its bare translation. Your brain retrieves the whole episode, not an abstract label.

This flips the usual order of generic flashcard apps. Instead of memorizing a list first and hoping to recognize the words later while reading, you meet the word in its natural habitat, while reading a story you already care about, and review consolidates a memory trace that already exists, rather than manufacturing one out of nothing.

Take a concrete example. If you're reading a mystery novel and you meet the word "suspect" in a sentence where a character fiercely denies an accusation, that word attaches to a full scene: the tension of the dialogue, the stakes of the plot, your own curiosity about whether the character is lying. When that word comes back up for review days later with the original sentence displayed, your brain doesn't retrieve an abstract definition. It retrieves the scene, and the translation follows naturally.

This mechanic also explains why some words met just once in a memorable passage stay lodged for years, while dozens of words drilled twenty times on a list fade within weeks. The number of repetitions matters less than the quality of the initial anchor.

FAQ

How many times do you need to meet a word in context before it sticks?

Studies on incidental vocabulary acquisition through reading vary by word and by learner, but broadly agree on a range of 6 to 10 spaced encounters across varied contexts before a word moves into long-term memory. That's exactly what spaced repetition optimizes once the first encounter has happened.

Does context help with grammar too, or just vocabulary?

The same principle applies to grammar. A conjugation rule memorized abstractly sticks less than a structure met repeatedly in real sentences, where your brain infers the pattern without conscious effort.

Should you skip word lists entirely at the start of a language?

Not necessarily. A small list of the most frequent words (articles, pronouns, basic verbs) can give you a useful base to start reading very simple texts. But as soon as you can follow a story, even haltingly, reading in context becomes more rewarding than a longer list.

Does context-based learning work for every language, or only some?

The mechanism itself is language-agnostic. Contextual encoding and comprehensible input describe how human memory and language acquisition work in general, not a quirk of one particular language family. What changes from one language to another is how quickly you reach the 90-95% comprehension zone needed for a text to feel readable rather than exhausting. A language close to your native one, sharing vocabulary roots and sentence structure, gets you there faster than a language with a different script or word order. That's why choosing a text slightly below your frustration threshold matters more than the specific language you're learning.

To learn more about the method, check our FAQ or contact us if you have a specific question about your learning path.