Memory
Why you forget vocabulary
You learned thirty words on Monday. By Friday you have five. This is not a failure of discipline, and the fix is smaller than you would expect.
Everyone who has studied a language knows the feeling. A session goes well, the words make sense, you close the app satisfied. A week later someone says one of them out loud and you have nothing.
The instinct is to blame your memory, or your age, or the app. None of those is usually the problem. Forgetting is simply what happens to anything you meet once, and it happens on a schedule that is well understood.
The curve everything falls down
In the 1880s Hermann Ebbinghaus spent months memorising nonsense syllables and testing himself at intervals, producing the first map of how memory fades. The shape he found is the important part: a steep drop in the first day, then a long, slow flattening.
Modern replications have found the same shape, so the curve is not a Victorian curiosity. It is worth being careful with the exact percentages, though — Ebbinghaus tested meaningless syllables on himself, and real words met in real sentences fade more slowly than that. The steepness is overstated by the famous version of the graph; the pattern is not.
The practical consequence: a word you meet once is not learned, however well you understood it at the time. Understanding a word and retaining it are separate events.
Why cramming fails and spacing works
The counter to forgetting is the spacing effect, one of the most reliably reproduced findings in memory research. Take a fixed amount of study time and split it across several sessions, and you retain more than if you spend the identical time in one sitting.
This is worth sitting with, because it is not an argument for working harder. It is the same effort, rearranged. Three ten-minute encounters spread across a week beat one thirty-minute session, and it is not close.
Cramming feels more productive because everything stays fresh while you are doing it. That fluency is the trap: it measures how available a word is right now, not whether it will still be there next month. The mild difficulty of half-remembering something is what makes the memory durable, so the harder-feeling method is usually the better one.
Recognising a word is not knowing it
There is a second distinction that catches people out. Being shown a word and understanding it is recognition. Producing it when you need it, without prompting, is recall. They feel similar when you are studying and behave completely differently when you are speaking.
Being asked to retrieve something — and even struggling to — does more for retention than being shown it again. This is why flashcards work better than rereading a list, and why rereading feels so much more pleasant: it never makes you fail at anything.
Most people build a large recognition vocabulary and a small production one, then conclude they are bad at speaking. They are not. They have practised one skill and not the other, a gap we went into in how many words you need to speak a language.
What each method actually gives you
Once you separate spacing from retrieval, it becomes obvious why some study habits work better than others, and why the pleasant ones tend to be the weak ones.
| Method | Spacing | Retrieval | What it really does |
|---|---|---|---|
| Rereading a word list | No | No | Feels productive because the words look familiar. Familiarity is not memory. |
| Cramming before a trip | No | Some | Works for a few days, then falls off sharply. Fine for a holiday, useless for a language. |
| Flashcards, all at once | No | Yes | Real recall practice, but packed into one session so the spacing does no work. |
| Spaced flashcards | Yes | Yes | The strongest combination, and the most effort to keep up with daily. |
| Reading and listening | Yes | No | Repeated encounters in context, but the word is always shown to you. |
| Widgets and ambient review | Yes | No | Solves frequency without effort. Does not ask you to produce anything. |
Nothing in the bottom half is a substitute for the row above it. They solve different halves of the problem.
How often a word needs to come back
The honest answer is more often than feels necessary, and nobody can give you an exact number. Estimates for how many encounters it takes to pick up a word from context vary a great deal, and commonly land somewhere around a dozen or more, depending on the word and how much the sentence around it gives away.
Spaced repetition software schedules those encounters formally, stretching the gap each time you get one right. That is the rigorous approach and it works, provided you keep up with the reviews — which is where most people come unstuck, because the queue does not care how your week is going.
The version that survives a bad month
There is a trade-off worth naming plainly. The most effective method is spaced retrieval practice, done daily. The most effective method you will still be doing in March is whichever one asks least of you.
Since forgetting is driven by gaps in contact, a weaker method applied consistently often beats a stronger one applied in bursts. Not because the science says intensity is bad, but because a review you skip has an effectiveness of zero.
That is the gap Glancealot is built for. It puts words on your lock and home screen so they come back on their own, in the seconds you spend checking the time — the spacing half of the problem, solved without a session to start or a queue to clear. It does not ask you to recall anything, so it is a complement to flashcards and speaking practice rather than a replacement. There are packs for French, Spanish, Japanese and others, or you can load in whatever your main course is teaching you.
If you are trying to work out where the repetition fits into an ordinary week, we wrote about that in learning a language on your commute.
Frequently asked questions
Forgetting is the default for anything met once. Memory for new material drops steeply in the first day or two and then levels off, so a word learned on Monday and never seen again is largely gone by the weekend. It is not a sign of a bad memory; it is what happens without repeated contact.
A description of how memory for new material fades over time, first mapped by Hermann Ebbinghaus in the 1880s using nonsense syllables. The steep early drop and gradual levelling off have held up in modern replications, though real vocabulary learned in context fades more slowly than the nonsense syllables he tested.
The spacing effect is one of the most consistently reproduced findings in memory research: the same total study time produces better long-term retention when spread across separate sessions than when packed into one. That holds across ages, subjects, and study methods.
Estimates vary widely and depend on the word, the context, and whether you are being tested or just reading. Figures in the range of a dozen or more meaningful encounters are commonly cited for picking up a word from context alone. The practical point is that once is never enough, and the encounters need to be spread out.
Both matter, but they are different skills built in different ways. Recognition comes from exposure, and recall comes from being asked to produce a word without seeing it first. Recall is harder and does more for speaking, which is why exposure alone leaves people able to follow a conversation but not join one.
Passive exposure builds recognition well and is far easier to sustain than daily study, which matters because consistency is the thing most people fail at. It will not build active recall on its own, so it works best as the repetition layer alongside something that asks you to produce the words.
