Most people try to learn vocabulary by looking at it repeatedly. That is the slowest method there is. Two changes will roughly halve the time it takes for a word to stick.
Stop reviewing by rereading and start reviewing by retrieving — force yourself to produce the answer before you check it. Then space those retrievals out over increasing gaps instead of cramming them together. Everything else on this page is detail hanging off those two points.
Look at a word list five times and something reliable happens: it starts to feel familiar. Familiarity is a genuinely pleasant sensation and it is almost completely uncorrelated with whether you will be able to recall the word tomorrow.
This is the trap at the centre of vocabulary learning. The method that feels most productive while you are doing it — smooth, easy, no failures — is the one that produces the weakest memories. The method that feels worst, sitting there struggling to dredge up a word and sometimes failing, is the one that works. If your study session feels comfortable, it is probably not doing much.
Every review should begin with you trying to produce something from memory, with the answer hidden. Cover the translation. Say the word out loud before you turn the card. Write it before you look.
The act of failing to recall and then seeing the answer is worth more than several passive readings. The effort is the mechanism, not a side effect of it, which is why making your reviews easier makes them less useful.
Twenty repetitions in one evening produce a memory that is gone by the weekend. The same twenty spread over a month produce one that lasts for years. The gap is the active ingredient: each time you nearly forget something and then retrieve it anyway, the memory is rebuilt stronger and decays more slowly.
This is what spaced repetition software automates. It is not magic and you do not need an app to benefit — you need the schedule.
| Review | Gap after the last one |
|---|---|
| 1st | Same day, a few hours later |
| 2nd | Next day |
| 3rd | 3 days |
| 4th | 1 week |
| 5th | 3 weeks |
| 6th | 2 months, then let it go |
Six correct retrievals on that schedule and a word is effectively yours. Note how few they are. The work is not heavy; it is just spread out, and spreading it out is the part people skip.
A card with a word on one side and a translation on the other is the weakest card you can build, because it teaches you a link between two labels rather than a piece of language. Four cheap upgrades:
Put the word inside a short sentence you would actually say. Words learned in isolation come back in isolation, and you cannot speak in isolated words.
If you learn a word silently you learn a spelling, not a word. Say it out loud on every review, and use audio where you can — otherwise you will not recognise it when somebody says it to you.
For concrete nouns, an image beats a translation. It skips your first language entirely, which is where the delay in speaking comes from.
Three meanings on one card means you never learn any of them cleanly. Split them, and add the second meaning only when the first is solid.
Seeing Wasser and knowing it means water is recognition. Wanting to say water and producing Wasser is production. Production is several times harder and it is the one that decides whether you can speak.
You do not need both directions for every word. A sensible split: recognition only for words you merely need to understand — which is most of them — and both directions for the two or three hundred words you use constantly. Testing everything in both directions doubles your workload for a small return.
The uncomfortable rule of thumb: if you finish a study session feeling pleased with how smoothly it went, make it harder. Difficulty during practice is what buys you recall later.
Somewhere between five and fifteen successful retrievals, depending on how similar the word is to one you already know and how vivid its context was. Passive exposures do not count for much — the number that matters is retrievals, not sightings.
No. They automate the schedule, which is genuinely useful once you are tracking hundreds of words, but paper cards sorted into boxes by interval work on exactly the same principle. The schedule is the method; the app is a convenience.
Yes, from the very first day, in languages that have them. Attaching gender later is far harder than learning it attached, because you have to overwrite a memory rather than form one. Learn the article as part of the word, not as a fact about it.
Usually not. Learning ten fruits together means confusing them for weeks, because similar items learned together interfere with each other. Mixed batches are less satisfying to build and noticeably easier to retain.
Excellent for the small number of words that refuse to stick — build a vivid, slightly absurd link between the sound and the meaning. Do not build them for everything, though. Constructing a mnemonic costs more time than the word is usually worth, so save them for repeat offenders.
The rest of the beginner questions, answered the same way.