Forgetting
Also known as the forgetting curve, savings, retention interval
Forgetting is fast, mostly harmless, and the thing every good study schedule is built around.
The claim Newly learned material decays rapidly and then slowly; what is forgotten is usually not destroyed but harder to reach, which is why relearning is fast and well-timed review is cheap.
Strong evidence Ebbinghaus's 1885 curve has been directly replicated with modern methods, and the shape is one of the most stable quantitative results in psychology.
Hermann Ebbinghaus spent the early 1880s memorising thousands of nonsense syllables and measuring how many repetitions it took to relearn them at intervals from twenty minutes to a month. The curve he drew — steep loss in the first day, then a long shallow tail — is the oldest quantitative law in psychology, and it is still roughly right. Murre and Dros replicated it in 2015 using the same savings method, with one subject and seventy hours of work, and got a curve close to the original.
The word 'savings' matters more than the curve. Ebbinghaus did not measure how much could be recalled; he measured how much cheaper relearning was than learning it the first time. Material you cannot produce at all is often still in there, cutting the cost of learning it again. That is the difference between a memory being gone and being unreachable, and almost everything people call forgetting is the second.
This reframes what a review is for. If forgetting were destruction, review would be repair and you would want to review before anything was lost. Because forgetting is mostly loss of access, review is most valuable exactly when access has partly gone — the recall is effortful, the effortful recall is what strengthens, and reviewing too early wastes the opportunity. Every spacing algorithm is an attempt to schedule that moment.
The curve also sets a floor under any honest study plan: whatever you learned today, most of the detail will be gone by tomorrow, and that is normal rather than a sign of a bad memory. People who do not know this conclude they are unusually forgetful and change methods, when what they needed was a second pass on day two.
There is one qualification worth keeping. Ebbinghaus used nonsense syllables precisely so that meaning could not help, and meaningful material — things you understand, connect and care about — decays far more slowly. The curve is the worst case, not the typical one.
What to do differently
- Plan a second pass within about a day of first learning something. That is where the curve is steepest and the review is cheapest.
- Read a blank as 'unreachable', not 'gone'. Relearning is much faster than learning, so material you abandoned is not lost work.
- Do not review while it still feels fresh — you are paying for a review that buys almost nothing.
- Give new material meaning early, by connecting it to something you already know. Meaningful material sits far above the nonsense-syllable curve.
- Measure retention at a delay, not at the end of a session. Everything looks learned at the end of a session.
Limits and trade-offs
- Ebbinghaus's data come from one person memorising meaningless material, and the modern replication used one subject too. The shape is robust; the exact numbers are not a personal forecast.
- The curve says nothing about skills. How quickly you lose the ability to do something is a different literature with much messier answers.
- 'It is still in there somewhere' is easily overread into comfort. Savings shrink with time too, and material left for years costs most of its original price to regain.
- Knowing the curve does not tell you your own schedule. Individual and material differences are large enough that any published interval is a starting point.
What this means for a language
This is why vocabulary you looked up yesterday is gone today and why that is not evidence of a bad memory. It is also the most encouraging finding on this site for anyone returning to a language they studied years ago: relearning genuinely is much faster than learning, and the fluency that comes back in three weeks after a decade away is savings made visible. The practical consequence for a beginner is unglamorous — the second look at a word, the next day, is worth more than the first three looks on the day you met it.
The research
- Murre & Dros (2015), PLOS ONE
A direct replication of Ebbinghaus's 1885 forgetting curve by the savings method; results closely matched the original, with an upward jump around 24 hours.
- Cepeda, Pashler, Vul, Wixted & Rohrer (2006), Psychological Bulletin
Shows the practical consequence: the best gap between reviews depends on how long you need to remember, because the curve does.
Related
Read next
-
839 assessments from 317 experiments. The finding that matters for planning: the best gap between study sessions grows with how long you need to remember the material.
-
A continuously updated survey of the spacing literature with the arithmetic done: when a card is worth making at all, how review cost compounds, where the published intervals come from. Dense, opinionated, heavily cited.
-
A long essay on using spaced repetition to read research papers — memory as a tool for understanding rather than a substitute for it. Also the clearest published account of what daily review actually feels like over years.
-
A cognitive neuroscientist's account, organised around four pillars: attention, active engagement, error feedback and consolidation. The most biologically grounded book on this list, and the most demanding.
-
The reference review for the claim that sleep does not merely protect memories but actively reorganises them. Long and technical; the introduction and conclusions carry the argument on their own.