The forgetting curve: Hermann Ebbinghaus (1885)
Hermann Ebbinghaus was a psychologist primarily known for his research in the field of memory. His groundbreaking work focused on understanding the processes of learning and forgetting, and he conducted extensive experiments on himself to gather empirical data. He is best known today for the forgetting curve — a graph showing how quickly newly learned information fades from memory over time unless it is deliberately reviewed.
He studied the memorization of nonsense syllables (such as “wid” and “zof”) by repeatedly testing himself after various time periods and recording the results. He plotted these results on a graph, creating what is now known as the “forgetting curve”, which behaves like the exponential curve described by what is called “exponential decay function”.
The forgetting curve shows how learned information slips out of our memories over time – unless we take action to keep it there.
The steepest drop in memory happens quickly after learning, so it’s important to revisit the information you’ve learned sooner rather than later. After that, regular reviews will help to reinforce it. But you can leave longer and longer gaps between these review sessions.
This is known today as “spaced learning” or “spaced repetition”, which is a review technique where you repeat new information that you’ve learned at key moments on the forgetting curve to slow down the rate at which you forget.
Ebbinghaus investigated the rate of forgetting, but not the effect of spaced repetition on the increase in retrievability of memories. Practical applications of spaced repetition were first suggested by Mace in 1932.
The solid lines on the forgetting curve show how memory is boosted by each review session. The dotted lines show the forgetting that would have happened otherwise. Information still fades after each review. But it happens more slowly each time. So, you can leave longer and longer gaps between your review sessions as time goes by.
Ebbinghaus also concludes:
- It’s easier to remember things that have meaning. When Ebbinghaus tried learning meaningless syllables, he found they quickly dropped out of his memory. But, if you’re learning about a subject that really engages and interests you, then you’ll more likely retain that information for longer.
- The way something is presented affects learning. The same set of information can be made more or less memorable, depending on how clearly and powerfully it’s communicated.
- How you feel affects how well you remember. Ebbinghaus showed that physiological factors, like stress and sleep, also have a significant impact on how much information you retain.
In 1885 Ebbinghaus published his book Memory: A Contribution to Experimental Psychology. His work laid the foundation for the scientific study of memory and had a significant influence on subsequent research in psychology and education.
The 1885 experiment: nonsense syllables and the savings method
Ebbinghaus wanted to study memory in a “pure” form, free from the influence of meaning, prior knowledge and personal associations. Since real words always carry associations, he constructed a pool of roughly 2,300 nonsense syllables, each built from a consonant, a vowel and a consonant (such as “wid”, “zof” or “kaf”). From this pool he assembled lists — typically rows of thirteen syllables — and read them aloud to the steady beat of a metronome until he could recite a list twice in a row without error.
To measure forgetting, he invented the savings method. After a fixed interval — from twenty minutes up to a month — he relearned the same list to the same criterion and compared the effort required. The saving is the proportion of learning time spared on the second pass: if the original learning took 20 minutes and relearning took 12, the saving is 40%. This measure has a subtle advantage over simply asking “how much can you recall?” — it also detects memory traces that are too weak for conscious recall but still make relearning faster. For this reason the savings method is still regarded as an unusually pure measure of memory performance.
The discipline of the project is hard to overstate: Ebbinghaus ran the full set of experiments on himself in 1879–1880, then repeated the entire procedure in 1883–1884 to verify his own results, and only then published Über das Gedächtnis (1885). It was the first time learning and forgetting had been measured quantitatively, and it turned memory from a topic of philosophical speculation into an experimental science.
The mathematics of the forgetting curve
Plotting savings against time produced the now-famous shape: a steep initial drop that gradually levels off. Ebbinghaus’ own data points were approximately:
| Time since learning | Savings (memory retained) |
|---|---|
| 20 minutes | 58% |
| 1 hour | 44% |
| 9 hours | 36% |
| 1 day | 34% |
| 2 days | 28% |
| 6 days | 25% |
| 31 days | 21% |
In other words: more than half of the new material was gone within the first hour, about two thirds within a day — but after that the loss slowed dramatically, and what survived the first days remained relatively stable for a month.
Ebbinghaus described his data with a logarithmic formula, b = 100k / ((log t)c + k), where b is the percentage saved, t is time in minutes, and k and c are constants (1.84 and 1.25 in his data). Today the curve is usually approximated by an exponential decay function of the form R = e−t/S, where R is retrievability and S is the stability of the memory — the larger the stability, the slower the forgetting. Memory researchers still debate whether a power function or an exponential fits best, but the qualitative shape — rapid early loss followed by a long, flattening tail — has proven extremely robust across materials, learners and a century of studies.
The 2015 replication: Murre and Dros
Could a single 19th-century self-experiment really be trusted? In 2015 — 130 years after the original publication — Jaap Murre and Joeri Dros of the University of Amsterdam repeated Ebbinghaus’ experiment as faithfully as possible and published the results in PLOS ONE. One subject spent about 70 hours learning and relearning lists of nonsense syllables, with retention intervals of 20 minutes, 1 hour, 9 hours, 1 day, 2 days and 31 days — the same intervals Ebbinghaus used.
The outcome was striking: the replicated curve matched the 1885 data closely at almost every point. The replication also confirmed a detail visible in the original data — the curve is not perfectly smooth, but shows a slight upward bump around the 24-hour mark. Murre and Dros suggest that sleep may be responsible: a night of memory consolidation during sleep partially offsets forgetting. The forgetting curve is thus one of the few findings in psychology that has survived a direct, method-faithful replication across 130 years.
The spacing effect: how reviews flatten the curve
Ebbinghaus himself noticed a second phenomenon in his data: repetitions spread over several days produced better retention than the same number of repetitions crammed into one session. This is the spacing effect, and it has since become one of the best-documented findings in cognitive psychology. A landmark meta-analysis by Cepeda, Pashler, Vul, Wixted and Rohrer (2006) reviewed 839 assessments of distributed practice from 317 experiments and found that spaced learning reliably beats massed learning, and — importantly — that the optimal gap between reviews grows with how long you want to remember: the further away the “exam”, the longer the spacing should be.
The mechanism maps directly onto the forgetting curve. Each successful review does two things: it resets retrievability to nearly 100%, and it increases the stability of the memory, so the next decline is slower than the previous one. After several well-timed reviews, the curve becomes so flat that an item may need to be revisited only after months — this is the flattening shown by the solid lines in the graph above. Scheduling reviews at expanding intervals, just before an item would be forgotten, is the principle behind spaced repetition systems, from paper flashcard boxes to modern algorithms.
What this means for language learning
Vocabulary and grammar are exactly the kind of material the forgetting curve punishes hardest: thousands of small, easily confusable items learned over months. The research points to a few practical conclusions:
- Review early, then stretch the intervals. The steepest loss happens within hours of first contact with a new word, so the first review should come soon — the following ones at progressively longer gaps. This is the core of spaced repetition in language learning.
- Learn with meaning and context. Ebbinghaus deliberately used meaningless syllables and forgot them at maximum speed; he noted that meaningful material is retained far better. Words embedded in sentences and personally relevant examples start higher on the curve and decay more slowly — a principle at the heart of learning with full sentences.
- Do not skip sleep. The 24-hour bump in both the 1885 and 2015 data suggests that consolidation during sleep does part of the reviewing for you.
None of this makes forgetting avoidable — the curve is a description of how human memory works, not a flaw to be fixed. What the century of research since Ebbinghaus shows is that forgetting is predictable, and anything predictable can be scheduled against.
Frequently asked questions
How fast do we forget, according to the forgetting curve?
In Ebbinghaus’ data, savings dropped to about 58% after 20 minutes, 44% after one hour and roughly 34% after a day; after that the decline slowed, levelling off near 21% after a month. The exact numbers apply to nonsense syllables learned by one person — meaningful, well-practised material fades much more slowly — but the shape (fast early loss, then flattening) generalizes.
Has the forgetting curve been scientifically confirmed?
Yes. Murre and Dros replicated Ebbinghaus’ experiment with the original method in 2015 and obtained a closely matching curve, including the slight upward bump around 24 hours attributed to sleep. The broader pattern has also been reproduced with words, sentences, facts and skills across many laboratories.
How can you flatten the forgetting curve?
By reviewing at expanding intervals (the spacing effect), by learning material with meaning and context rather than isolated items, and by sleeping well between sessions. Each well-timed review makes the memory more stable, so the next forgetting cycle runs slower and reviews can be spaced further apart.
Sources
- Ebbinghaus, H. (1885). Über das Gedächtnis. Untersuchungen zur experimentellen Psychologie. English translation: Memory: A Contribution to Experimental Psychology (1913).
- Murre, J. M. J., & Dros, J. (2015). Replication and Analysis of Ebbinghaus’ Forgetting Curve. PLOS ONE, 10(7), e0120644.
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380.
- Murre, J. M. J. (2023). Why Ebbinghaus’ savings method from 1885 is a very ‘pure’ measure of memory performance. Psychonomic Bulletin & Review, 30, 303–307.
- Mace, C. A. (1932). The Psychology of Study. London: Methuen.