Herman Ebbinghaus discovered the forgetting curve, by memorizing series of non-sensical syllables like "zod", "byh", "tef", etc., and than seeing how long it took to forget them. The results are interesting, but I'm curious as to how he came up with the number of words to remember (the number of words in a series). Because clearly the more words, the more difficult it would be to remember, which he did mention in his study.

In the study though he mainly used series of 8, 12 or 16 words, and took multiple tests which often came up to be about 104 syllables, but was quite varied. I couldn't find how he came up with those specific numbers of syllables to test, and furthermore it seems that his entire study and curve would be compromised by the amount of words. For instance a 1 word series may never be forgotten while 10,000 may be unmemorizable.

Is his curve a simply an average of testable results? Or is there some other reason for the certain number of words that constituted a series that I didn't quite catch?

  • $\begingroup$ Of course you can forget one single word. I'm sure you have. And of course you can remember 10.000 words. I'm sure you know 10.000 words. $\endgroup$
    – user3116
    Dec 17, 2013 at 9:40
  • 1
    $\begingroup$ Two problems with Ebbinghaus' findings: (1) every test of how much you remember is a new learning situation, which means that you change the forgetting curve by testing it; (2) learning and forgetting of meaningful information shows different forgetting curves, so don't generalize this curve to non-nonsensical words. $\endgroup$
    – user3116
    Dec 17, 2013 at 9:45

1 Answer 1


If you read Ebbinghaus' article carefully, you will note that in Chapter III. The Method of Investigation, Section 13. Establishment of the Most Constant Experimental Conditions Possible, Ebbinghaus gives the following third of seven rules "for the process of memorising":

3 . Since it is practically impossible to speak continuously without variation of accent, the following method was adopted to avoid irregular variations: either three or four syllables were united into a measure, and thus either the 1st, 4th, 7th, or the 1st, 5th, 9th ... syllables were pronounced with a slight accent. Stressing of the voice was otherwise, as far as possible, avoided.

This rule results in 3- or 4-syllable substrings. Of course the whole string of syllables to be memorized must be a multiple of either 3 or 4, if it is to follow this rule.

Your answer is: Ebbinghaus chose 8, 12 or 16 syllables, to keep the experimental conditions constant.

  • $\begingroup$ Hmmm. I did read that, but I think he was only doing that to keep his voice constant and reduce variation in tone from tarnishing the results in anyway. But being a multiple of 3 or 4 is an interesting point and I thank u. That said, the number in a series could still be 3, 6, or 333 for instance, but he still kept his series within a range from 12 - 39 from what I've seen. How could the curve yield any significant results with such a finite range? Would his results imply that in the same time I would take to forget 300 out of 400 words, I would also forget 3 out of 4? Seems doubtful. $\endgroup$
    – Squirrl
    Dec 17, 2013 at 20:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.