Spaced repetition works. That is not in dispute, and anyone who has used it to get through a first thousand words knows it works. The question worth asking is not whether flashcards are effective but what they are effective at, because the answer changes as you progress.
What a flashcard is good at
Getting a large number of discrete facts into memory on a schedule. If the task is "know that 犬 means dog", a card is close to optimal — it isolates the item, tests recall directly, and spaces the reviews so you spend time only on what is fading.
For the first stretch of a language this is exactly the task. You need a large stock of frequent words fast, most of them concrete, most with a clean one-to-one gloss. Nothing beats spaced repetition for that.
What a flashcard cannot carry
The problem arrives when words stop having clean glosses.
Knowing German warten means "to wait" is incomplete: what you need is that it takes auf plus the accusative. Knowing Spanish realizar means "to realise" is actively wrong; it means to carry out. Knowing a Japanese word without knowing whether it is plain, polite or humble means you can produce it in the wrong register.
These are properties of a word in use, and a two-sided card has nowhere to put them. You can cram an example sentence onto the back, at which point you are no longer really using a flashcard — you are reading, one sentence at a time, inefficiently.
The signal that you have hit the limit. You recognise a word in your deck instantly and still cannot use it in a sentence. That gap means the card has taught you everything it can, and the remaining knowledge — collocation, register, which of three synonyms fits — only comes from meeting the word in context.
The arithmetic behind the switch
There is a numerical reason the turning point lands where it does. The most frequent thousand words cover roughly eighty per cent of everyday text; three thousand gets you to about ninety-five per cent.
Below that coverage, reading real material is exhausting — an unknown word every few words breaks the thread continuously, so you cannot learn from context because there is no context, just gaps. Deliberate memorisation is the only route in.
Above it, reading becomes possible, and reading teaches things cards cannot. So the switch is not a matter of taste: it is the point at which a second mechanism becomes available and starts outperforming the first.
What replaces them
- Volume reading with instant lookup. The mechanism is repeated meaningful encounter rather than scheduled recall. It requires that looking a word up costs nothing — the moment lookup has friction, you start guessing, and guessing is how half-learned words fossilise.
- Material you actually care about. After the frequency lists are exhausted, the most useful vocabulary is whatever appears in the things you read anyway. Interest is not a motivational nicety here; it determines which words you meet often enough to acquire.
- Sentence context over isolated words. If you keep any deck at all, keep sentences rather than words. What you are learning by this stage is behaviour, and behaviour needs a sentence to be visible in.
- Production practice. Recognition arrives long before use. Writing and speaking is what converts a word you know into a word you have.
A reasonable overall shape
Flashcards for the first two thousand words, aggressively and without guilt. Then taper: keep them for genuinely stubborn items and for anything you need on a deadline, and move the bulk of your time to reading and listening with a dictionary within reach.
Learners who never start using spaced repetition take much longer to get going. Learners who never stop plateau with a large recognition vocabulary they cannot deploy. The tool is right for one stage and wrong for the next, which is unremarkable — it is just worth noticing when the stage has changed.
Look words up as you read
Reading with instant lookup is what replaces flashcards. 45 offline dictionary pairs, no connection needed, no data cost per lookup.
Open the dictionary appFrequently asked questions
Are flashcards actually effective for language learning?
Very, for a specific job: getting a large number of discrete items into memory quickly. Spaced repetition is well-supported for exactly that. What flashcards do not teach is usage — which preposition follows a verb, how formal a word is, which of several near-synonyms fits a sentence.
How many words should I learn with flashcards before switching?
Somewhere around the first two thousand is a common turning point. That is roughly the vocabulary where general frequency lists stop delivering words you meet often, and where you can begin reading real material with enough coverage to learn from context.
What replaces flashcards?
Reading and listening to material you care about, with frictionless lookup. The mechanism shifts from deliberate memorisation to repeated meaningful encounter, which teaches usage and register alongside meaning — the parts a card cannot carry.