The forgetting curve and vocabulary: why words slip away
Oct 6, 2026 · 12 min read
A word you looked up this morning can feel gone by the evening. The gap between meeting a word and keeping it is the territory of the forgetting curve, a long-standing description of how memory for new material fades when nothing is done with it. The curve does not tell you what you personally will remember tomorrow; it describes a general shape, and that shape is useful for planning.
This page covers what the curve describes, why it matters for a word list, what makes the drop steeper or shallower, and how spacing, retrieval and a short review calendar fit together.
What the forgetting curve describes
Hermann Ebbinghaus, a German psychologist, ran memory experiments on himself in the 1880s. He learned lists of nonsense syllables and then re-tested himself after set delays: twenty minutes, an hour, nine hours, a day, two days, six days, thirty-one days. What he recorded was a steep early fall followed by a much slower tail. Most of the loss appeared within the first day, and the remaining trace decayed gently after that.
The finding was published in 1885 and is still the reference point. Later work with real words tends to produce the same general shape at different rates: material that is not revisited fades fastest right after it is first met.
Why this matters for a vocabulary list
Reading a word once is an encounter, not a durable trace. That is why a list reviewed carefully on Sunday can feel unfamiliar by Wednesday, and why learners report that they know a word but cannot produce it in a sentence.
The words that disappear first are usually the ones met only once, in a list, without an example. Words met repeatedly in reading, or used in speech, sit on a flatter part of the curve.
What makes the drop steeper or shallower
Four factors come up again and again in the research literature.
- The material. Concrete words with a clear image hold better than abstract ones, and a word close to something you already know starts from a higher point.
- Whether it was understood. A word learned inside a sentence you understood holds better than a translation pair, and depth of processing matters more than repetition count.
- Attention and context. Words met while tired, distracted or in a rush are weakly encoded, and weak encoding shows up as a fast early drop.
- Sleep and downtime. Consolidation research links sleep after learning to how new material is stabilised, which is one reason a late-night cramming session can look worse than it feels.
Spacing and retrieval, not more minutes
Two practices take the curve into account. Spaced repetition places reviews at widening intervals, so a second look arrives as the first is starting to fade rather than while it is still fresh. Active recall changes what happens during the review: instead of reading the word again, you try to produce the meaning, the spelling or a sentence of your own.
A word retrieved with effort and then checked is more useful than the same word read three times, which is the idea behind desirable difficulties. The aim is not a longer session but a better-placed one.
A seven-day review calendar
A simple pattern for a set of new words, with each round kept short.
| Day | What to do | Time |
|---|---|---|
| Day 1 | Meet the new words in context; make a first note for each. | 15–20 min |
| Day 2 | Cover the meanings and test yourself on all of them. | 5–8 min |
| Day 4 | Test again; mark the ones that failed and read their examples. | 5–8 min |
| Day 7 | Write one sentence per word that still feels weak. | 5–8 min |
| Week 2 | Fold the survivors into a mixed set with other topics. | 5 min |
Three short rounds are easier to keep up than one long evening, and the spacing is what does the work. Mixing topics at the end is interleaving practice.
Choosing which words to review
- Test before you look. A one-minute self-test on a set of twenty words tells you where to spend the next ten minutes.
- Errors first. Words you missed are the ones the curve has already started to bend; words you produced easily can wait a week.
- Group the confusables. Pairs that look or sound alike are best reviewed together, in one comparison, rather than as separate entries.
- Demote recognition-only words. If a word only matters for reading, a quick check is enough; production practice is better spent on words you need to write or say.
Keeping the record light is part of the method; the ideas in using word lists effectively are built around this problem.
Common misuses
- Treating the curve as a formula. It describes a pattern, not a prediction of your Tuesday. Rates differ between people, materials and days.
- Copying words out repeatedly. Writing a word five times keeps the answer in front of you, which is the opposite of retrieval.
- Saving everything for one long session. A single block of review flattens the spacing that gives the method its point.
- Rereading instead of testing. The page cannot tell you what is in your memory; only producing the word can.
Where the games on this site fit
The arcade sets here work better as review rounds than as study in themselves. Memory-match is recognition practice: match a word to its meaning, quickly and without notes. Speed-recall pushes the same material into timed production. Whack-a-word asks you to pick the right word from a moving set, which adds distraction.
Each round is voluntary practice and short by design. Use one after a review session, not instead of one.
Four steps to try this week
- Pick ten words you have met but cannot yet produce, and write one sentence for each.
- Leave them a day, then cover the meanings and test yourself.
- Mark the misses, reread those examples, and test again on day four.
- On day seven, write a sentence per weak word and mix the set with another topic.
Six questions to ask yourself
- Could I use this word in a sentence without looking?
- When did I last test myself on it, rather than read it?
- Have I met it in real text or in conversation, not only in a list?
- Do I confuse it with a similar-looking word?
- Do I need it for reading only, or for writing and speaking?
- Is my review spaced, or does it all land in one evening?
Frequently asked questions
Who was Ebbinghaus and what did he measure?
Ebbinghaus was a German psychologist who tested his own memory in the 1880s, learning lists of nonsense syllables and re-testing himself after set delays. The pattern he recorded, a steep early drop and a slower tail, is now called the forgetting curve.
Does the forgetting curve predict exactly what I will forget?
No. It describes a general pattern, not a fixed timetable. The speed of the drop varies with the material, how well it was understood, how tired you were and how much sleep you had, so treat it as a rough shape rather than a schedule.
What is spaced repetition?
Spaced repetition means reviewing an item at widening intervals, so the second look arrives as the first is fading. A common pattern for new words is a quick self-test on day 1, then day 2, day 4 and day 7.
Why is retrieval better than rereading?
Producing a word from memory shows you what is actually available, while rereading keeps the answer on the page and can hide gaps. A short self-test shows which words need another round.
How long should a review session be?
Five to eight minutes is enough for a quick pass over a small set. Short rounds done often are easier to keep up than one long session, and they leave the rest of your study time for reading.
This page describes a pattern of practice, not a promise of results. How much of a word stays with you depends on how often you meet it, test yourself on it and use it yourself.