1. How Ebbinghaus Measured Forgetting with Nonsense Syllables
By 1880, the German psychologist Hermann Ebbinghaus had completed experiments to measure forgetting [1]. He used lists of nonsense syllables—letters that do not form words. He chose these because real words and poems trigger associations and meaning, which would make the measurement of raw memory unstable [2].
Ebbinghaus measured forgetting using the "savings method." He compared the time it took to learn the list the first time with the time it took to relearn it later. The percentage of time saved is the "savings percentage." If the savings are 100%, the memory is fully intact. If it is 0%, nothing remains [1]. Using this method, Ebbinghaus spent a total of 70 hours relearning the lists at intervals of 20 minutes, 1 hour, 9 hours, 1 day, 2 days, and 31 days [1]. This raises the question: how much easier is it to relearn the list as more time passes?
2. How Quickly We Forget: Fast at First, Then Slower
Commonly cited values show that savings are about 58% after 20 minutes, 44% after 1 hour, 36% after 9 hours, 34% after 1 day, 28% after 2 days, and 21% after 31 days [2]. These numbers come from a secondary source summarizing the original data table. After 20 minutes, about 42% of the original learning effort (calculated as 100 minus 58) has already returned. The decline then becomes gradual. Even after 31 days, a savings rate of about 21% remains, meaning relearning is still easier than starting from scratch. Forgetting does not go to zero [2].
In 2015, Murre and Dros repeated the experiment using the same method. They reported that the results were similar to the original data. However, they noted that the curve might not be perfectly smooth and could show a slight upward turn at the 24-hour mark [1]. This experiment relied on data from one author and was not verified across many people. Does this shape stay the same no matter what you do?
3. Why We Forget Less After Sleeping Than After Staying Awake
In 1924, Jenkins and Dallenbach asked two subjects to memorize nonsense syllables and recall them up to 8 hours later. They found that subjects forgot less when they slept during the interval compared to when they stayed awake [3]. This result was interpreted as being due to interference from new information, rather than old memories simply fading with time [3]. However, with only two subjects, we cannot say that "sleeping prevents forgetting" for everyone.
In 2025, a study from the University of Tsukuba showed that in mice, an average of 2.4 "newborn neurons" in the hippocampus were active during REM sleep, just as they were during learning. When this activity was stopped, memories became harder to fix [4]. This research does not prove that the same thing happens in humans. So, how can we learn better while we are awake?
4. Why Testing Yourself Works Better Than Re-reading
In a 2006 experiment by Roediger and Karpicke, college students learned passages of text. One group took three recall tests, while another group re-read the text multiple times. One week later, the testing group remembered 61%, while the re-reading group remembered 40% [5]. The testing group had read the text an average of 3.4 times, while the re-reading group read it 14.2 times. That is about 4.2 times more reading, yet the testing group remembered about 1.5 times more (61 divided by 40) [5].
The re-reading group felt more confident that they had learned the material [5]. When tested just 5 minutes later, the re-reading group performed better. The difference appeared only after time had passed. These results were for college students reading text, not for children or nonsense syllables. What happens if we space out our practice?
5. How Spaced Learning Helped Mice Keep Memories
In a 2011 study from RIKEN, mice learned an eyeblink response. When the learning was spaced out with breaks, the mice retained nearly 100% of the memory after 24 hours. When they learned continuously without breaks, retention dropped to about 50% [6]. The researchers suggest that when short-term memories become long-term, the memory trace is stored through changes at different synapses in the brain [6]. This study used motor learning in mice and cannot be directly applied to human studying. So far, we have looked at how to keep memories. Why does the brain have a mechanism for forgetting in the first place?
6. How the Brain Actively Wears Down Memories
In 2017, Davis and Zhong summarized that the brain has biological mechanisms to wear down memory traces. In fruit fly studies, "forgetting cells" release dopamine, which changes the structure of memory-making cells through a molecule called Rac1 [7]. Three types of forgetting are known: new nerve cells growing, interference from new information, and the brain constantly wearing down memories [7]. This research is on fly olfactory memory, so the same pathways may not work in humans.
A press release from Scripps Research explains that this mechanism creates room for new information [10]. Richards and Frankland argue that the purpose of memory is not to keep old information perfectly, but to keep useful information to help with judgment [8][9]. By discarding details, memories become easier to use in new situations. This is a hypothesis in a review article, not proof that "forgetting makes you smarter."
7. How to Test Your Own Forgetting at Home
You can try this at home. Write down 5 words from your house on a piece of paper. Look at them for 1 minute. Close the paper and write them down on a separate sheet without looking. Do this again 1 hour later, before bed, and the next morning. List how many you remembered each time. This uses different material than Ebbinghaus’s syllables, but you can see if the shape "drops quickly then stays non-zero" applies to you.
You can also compare two methods. Use the same 5 words. On one day, re-read them 10 times. On another day, practice recalling them 3 times without looking. Swap the order of the methods and see which method lets you write down more words after 1 week. This small test mirrors the question in Roediger’s research [5]. However, one trial cannot prove a difference. Your health that day or interference from earlier memories can change the results. If you want to read the original studies, look for Murre and Dros (2015) or Roediger and Karpicke (2006). The figures and tables may be a good place to start when reading these papers in English.