Experimental memory regularity
Ebbinghaus Forgetting Curve
Without retrieval or relearning, accessible memory often falls rapidly after learning and then declines more slowly.
R(t) approximates exp(-t / S)
R is modeled retention, t is delay, and S is memory strength. The exponential is illustrative: real forgetting can be multi-rate, cue-dependent, and nonmonotonic.
A retrieval pulse in the animation raises strength and starts a shallower subsequent decline.
(%)
The animation runs automatically, pauses on the conclusion, and then repeats. The main control changes the scenario rather than scrubbing the timeline.
- CHANGE
- Hours since learning
- WATCH
- retention
- MEANING
- A retrieval pulse in the animation raises strength and starts a shallower subsequent decline.
Forgetting is steepest early, while retrieval can reshape the next curve.
Two traces separate passive delay from spaced retrieval. The display distinguishes recall from Ebbinghaus's original savings measure.
What it actually says
Ebbinghaus memorized nonsense-syllable lists and measured how much work was saved when relearning them after delays. His results established quantitative experimental study of memory.
There is no single universal curve. Meaning, interference, sleep, cues, testing, spacing, and individual knowledge all change retention.
"A useful law compresses a pattern. It does not erase the conditions that make the pattern true."
How the idea developed
The modern form emerged through observation, argument, and later refinement. The timeline separates the first insight from the version now used in textbooks and practice.[1]
Ebbinghaus publishes Ueber das Gedaechtnis.
An English translation broadens access to the work.
Spacing and retrieval research uses richer models and larger samples.
How the pattern works
The relation becomes useful only when its mechanism, measurement process, and operating range are visible.
Memory accessibility changes after encoding.
Later learning competes with earlier material.
Successful recall strengthens future access.
R is modeled retention, t is delay, and S is memory strength. The exponential is illustrative: real forgetting can be multi-rate, cue-dependent, and nonmonotonic.
Where it earns its keep
Applications are strongest when the law changes a decision, measurement, model, or experiment rather than merely providing an analogy.
Schedule spaced retrieval
ApplicationRevisit material after expanding intervals.
Require active recall, not passive exposure.
Measure relearning
ApplicationSavings can reveal retained structure despite failed free recall.
Match tests to operational use.
Where it stops working
The original work was intensive self-experimentation; modern memory science does not support fixed claims such as everyone forgetting a specified percentage after one day.
"One curve fits every subject"
Better: Strength and context alter the trajectory."Forgetting means erasure"
Better: Retrieval failure can occur despite retained information.Sources and further reading
Original publications and serious secondary scholarship are prioritized over summaries.
- Library of Congress - Ueber das GedaechtnisDigitized 1885 primary source.https://www.loc.gov/item/e11000616/
- Classics in Psychology - MemoryEnglish translation.https://psychclassics.yorku.ca/Ebbinghaus/index.htm
- Roediger and Karpicke - Test-Enhanced LearningModern retrieval-practice evidence.https://doi.org/10.1111/j.1467-9280.2006.01693.x