No force in the world can convince me to SuperMemo!

Witold Szafraniec, February 2000, Poland

Translation from Polish: Monika Morawska

Disclaimer: This article does not represent the views of SuperMemo World. It presents a frequently encountered set of misconceptions in reference to spaced repetition and learning. However, it is representative of the views of a quite large proportion of people and as such forms a valuable case study on the social acceptance of SuperMemo.

The article was submitted to SuperMemo World in the context of SuperMemo is Useless

The author, Witold Szafraniec, holds a masters degree in agricultural sciences and is an owner of a family farm in central Poland. Computers are Szafraniec's favorite pastime. He is an avid user of spreadsheets and database programs. He is also a fluent DBase programmer. His strong background in literature, humanities, Latin and Greek dates back to his years in secondary school in the 1960s. At that time many schools stressed the importance of classical education. He became familiar with SuperMemo in the early 1990s and has since remained one of the most vocal critics of the method. This article is a true reflection of Szafraniec's beliefs and is not an example of a devil's advocate approach. It excellently illustrates the obstacles SuperMemo will need to overcome for wider social acceptance. Our comments and source footnotes have been placed at the bottom in yellow inserts. and hyperlinked throughout the text via numbers in square brackets

During my recent visit to Piwniczna over Poprad, I had some time to attempt a critical assessment of the SuperMemo method. Unfortunately, I am not sure if I do not have to wage a war against a shadow. Having read the book about SuperMemo (published by SuperMemo World), I have found lots of gibberish that did not really let me truly understand what this sensational method of fast learning is really all about. This makes me wonder if the gibberish wasn't by any chance intentional (to obscure the weaknesses of the technology).[1]

For the sake of this discussion, I will assume that the basic goal of this method is to provide a student with a tool to assist the brain in better or more effective remembering of facts or information. Everything that I write in this text is a consequence of this assumption. 

I will begin with the conclusion. Firstly, I believe that the abilities of the human brain are limited, and no method of this real world can possibly increase the storage capacity of the brain. Secondly, mnemonic techniques (e.g. "cram and plough in daytime relentlessly to see seven hills pile up ...") have been known for ages and I can't see the reason why we should not continue to develop them. Thirdly, the whole idea of SuperMemo calls for bringing up the issue of the so-called erudite trap.

To explain the third point: an honest and hard-working erudite with a wish to reliably and completely explore a problem will at times arrive at an illusion that he has hit upon some original idea, which will enrich the treasure house of human wisdom. However, upon a closer scrutiny, a short search of the historic resources of the mankind will usually reveal that the same idea had much earlier been discovered by someone else. The idea then appears to be more or less an unintentional plagiarism. Ben Akiba used to say: All has occurred before! [2]

In human history, the mankind has come up with at least several breakthrough inventions which superbly improved our access to information. It seem obvious that these milestones were in turn: (1) the development of speech, (2) writing, (3) printed matter, and finally (4) the development of computer technologies. Please note that none of these sensational inventions went in the direction of increasing the storage capacity of the human brain.

SuperMemo method would then be the first invention that has succeeded in aggrandizement of memory. An accomplishment that defies belief! [3]

My remarks that follow are based on a review (by Krzysztof Szymborski) of a book entitled The New Renaissance : Computers and the Next Level of Civilization written by Douglas Robertson. 

Before humans could master the language, information resources accessible to each of our ancestors were limited in principle to the volume capacity of the human brain. That, according to Robertson, would be about 10 million bits. The development of language resulted in this that we did no longer need to search for information in our heads and we could, for example, ask a friend or any other language-capable individual. According to Robertson, this improvement has increased the resources of accessible information more or less a hundred times in an average person's case.

Hand-written texts increased, in Robertson's opinion, the reserves of accessible information by another two orders of magnitude. Dissemination of the printed word in the form of mass-produced books, newspapers, manuals, etc., allowed, according to Robertson, to increase the quantity of publicly accessible information by another million times. The question of how much more information is available to us thanks to the invention of computers boggles the mind. Robertson arrived at a stunning number of 10^25 bits. Most of us (eggheads, managers and plain bread eaters) suffer from the overwhelming information fatigue syndrome. The methods that help us remember more (i.e. create SuperMemory or help us Forget about Forgetting) are precisely in opposition to that what we would expect from the world flooded with information! Our brain uses a signal transmission mechanisms that are a million times slower than those of modern computers. This is still the brain with the processor and memory available at times of Adam and Eve. To absorb new information and learn, we depend on the help of short-term memory with an astonishingly small capacity. This is a biological fact and any attempts at "increasing" this capacity remind an ancient effort of constructing the perpetuum mobile. Collecting and memorizing facts (as in SuperMemo), independent of how fast and efficient (e.g. 40 000 English words in a month), is not yet a sufficient condition of wisdom. Knowledge is rather hidden in relationships between facts. These relationships determine the sense and importance of information. A sensible perception of the reality is not possible without a general hierarchy of value (outlook) that creates an order in the chaos of information flowing in from the outside world. An outlook or hierarchy of values can help us effectively select information and eliminate the excess of facts. Nowadays, with the help of the Internet we have created a sort of a gigantic external human memory. It should be enough to create in our memory some sort of an index to this wealth of information. Instead of remembering something that is written somewhere in generally accessible (public) memory, it is enough to index it in an efficient way in human memory. Creating in memory a database from the wealth of sources is unquestionably foolish and stupid. Our brain will never be a better library of knowledge than, for example, the Internet. Even if the Robertson's or Kurzweil's predictions in reference to the evolution of the human brain come true. We should strive at a situation in which our brain becomes an index to the outside information. An index of relationships between facts and ideas. Once we want to analyze these facts, we will still need to consult the sources. The more we attempt to forget about forgetting, the less memory we have at hand in our heads for "index files". Consequently, we will suffer from more and more information smog as envisaged in David Shenk's Data Smog. Knowledge of literary classics (e.g., biblical scapegoat or La Fontaine's The Fox and the Grapes, etc.), an ordered system of values (i.e. outlook), a conviction that All has occurred before! and the awareness that we are just mere mortals should all contribute superbly to the building of the basis for an efficient and functional index to knowledge!

No force on this planet will ever convince me about the usefulness of SuperMemo. I have tried SuperMemo for some time to learn some vocabulary of English. It might work for vocabulary but ... what's the point of extending the claims into the realm of sciences, literature and arts?

I would like to wish the entire team at SuperMemo World lots of self-criticism and humility in investigating the magic powers of the human mind

SuperMemo World's comments

Unlike the text discussed in SuperMemo is Useless, this article interestingly acknowledges the need to build an index to knowledge in human memory. However, the text fails to observe that the index is a structure that also requires memory. In the light of the omnipresent forgetfulness, the index will equally well need to be preserved with frequent use or with spaced repetition (i.e. SuperMemo)

In simple words, we need SuperMemo to quickly and efficiently build the index to knowledge

In more precise terms, the author's index to knowledge is nothing else than knowledge. Author's terminology seems to emphasize that knowledge stored in human memory should make a microscopic, yet properly selected subset of the entire knowledge available in human accessible resources (e.g. the Internet, Library of Congress, etc.).

Let us consider a simple example. A mere mortal will usually be aware that doughnuts are foods that should better be avoided by people who fear a heart disease. The fact "Doughnuts ain't good for heart" contributes to an average man's knowledge about health. This fact probably does not need SuperMemo. After all, most of us will refresh the knowledge about doughnuts and the heart each time we see a tasty doughnut. The Doughnut Fact contributes also to our index to knowledge. It is enough to jump to AltaVista and type +doughnut +heart to make a good use of this particular entry in our index. The search will help us recover more knowledge about the relationship between doughnuts and the heart

If we want to enhance our ability to think and conclude about doughnuts and the heart, we might try to remember the following:

  1. Fact 1: Doughnuts are high in trans fatty acids
  2. Rule 1: Trans fatty acids in foods tend to lower HDLs (high-density lipoproteins)
  3. Rule 2: Lower blood HDLs are a major risk factor in cardiovascular disorders (e.g. arteriosclerosis)

There is an advantage to knowing the above facts and rules: upon finding out that French fries are high in trans fatty acids, we will be able to use Rule 1 and Rule 2 to derive a new fact: French fries ain't good for heart. The awareness of the above rules will increase our reasoning ability. In the terminology of knowledge engineering, we will be able to derive new facts and rules from the existing set of facts and rules. In plain language, we will know more than we have actually learned. We will be able to conclude more. We will become more intelligent (if intelligence was defined as the inferential ability of the human mind). See also: Roots of creativity and genius

Yet there is a downside to remembering the trans fat rules. They are not as plain as the Doughnut Fact, and they may not effectively be refreshed upon a sight of a doughnut. Consequently, we may simply forget the link between doughnuts, trans fat, and the heart. This is where SuperMemo comes to play a role. Naturally, it will not increase the memory capacity. It will only help you refresh the trans fat rules. It will minimize the number of reviews in lifetime. In other words, it will only help you keep the trans fat rules in your forgetful memory. Using Mr Szafraniec's terminology we could say that SuperMemo makes sure that your index to knowledge remains intact in your mind

As for Kurzweil's predictions ... one of them is that we shall be able to plug in the World Wide Web to our brains by 2019. This still does not ensure our memory of WWW is associative (i.e. we can make instant searches and association). Until we can artificially expand human associative memory ... SuperMemo still has a role to play

Witold Szafraniec's article illustrates how a minor confusion in terminology can result in drawing seriously flawed conclusions. For a concise discussion of the confusion between terms such as data, information, facts, rules and knowledge and how this confusion affects our perception of the learning process, see: It's more than just memorizing by Luis Gustavo Neves da Silva

If the knowledge of literary classics shall be the basis of our knowledge index, we should start from the basic premise and the guiding light:

Let ignorance talk as it will, learning has its value (La Fontaine, The Use of Knowledge. Book viii. Fable 19)



  1. SuperMemo gibberish - the author refers to the description of the SuperMemo algorithm. This description is supposed to help those with some background in mathematics to roughly understand the algorithm. There isn't much to hide as for the general principles on which SuperMemo is based. Please refer to:
    1. How SuperMemo works (principles of SuperMemo in three points)
    2. General principles of SuperMemo (SuperMemo explained in a popular scientific language)
  2. All has occurred before! - while we can find the same themes throughout the ages of human thought and the same mechanisms ruling the history, All has occurred before! cannot be taken in an overly dogmatic manner. Chinese might have invented the print long before Gutenberg. Similarly, Ted Nelson's Xanadu might have preceded World Wide Web by more than a decade. Yet there was no Internet before 1969, no print among Neanderthals and there weren't two Napoleons.
    As for SuperMemo ... it has indeed been invented many times over by a number of individuals. We might have documented over twenty cases of people writing to SuperMemo World: I was just working over something similar when I discovered SuperMemo on the net! There are a few pieces of software that do similar things as SuperMemo and have been written independently; often starting with different premises
  3. Increasing the size of memory - SuperMemo has never claimed to increase the size of human memory. SuperMemo minimizes the number of repetitions needed to retain knowledge in memory. Those familiar with neurosciences will certainly know that recent findings have pointed out clearly that the number of neurons, and the number of neuronal connections in the brain increases with training even in adulthood. In this sense, all forms of learning do increase the size of memory. So does SuperMemo. We do not attempt to quantify such an increase though. It might indeed be negligible. This is irrelevant to the value of SuperMemo


  1. German novelist Karl Ferdinand Gutzkow (1811-78) in his Uriel Acosta included the famous Das Neue ist nur droben! Hier war alles schon einmal da - schon alles dagewesen! (All has occurred before!). This phrase attributed to Ben Akiba was often quoted by some historians to illustrate the cyclical nature of history
  2. Perpetuum mobile - an illusory forever-moving machine. In the past, scores of inventors attempted at constructing one before it was discovered that such machines must infringe upon one of the laws of thermodynamics
  3. "cram and plough in daytime relentlessly to see seven hills pile up ..." is a rough translation of a Polish mnemonic verse for memorizing the pi constant
  4. The New Renaissance : Computers and the Next Level of Civilization by Douglas S. Robertson. Price: $17.50. Hardcover - 224 pages (September 1998). Oxford University Press (Trade); ISBN: 0195121899. Sales Rank: 3,452 - an excellent book with accurate predictions related to the future of technology coming from a mind of an excellent writer with a strong background in mathematics. The book shows how computers should be considered the next great revolution after the language, writing and print
  5. Ray Kurzweil (short bio) is a Boston-based author of The Age of Spiritual Machine ( and The Age of Intelligent Machines for which he won the Association of American Publishers' Award (as the Most Outstanding Computer Science Book of 1990). Kurzweil was awarded the Dickson Prize, Carnegie Mellon's top prize in science (1994). He is also the MIT's Inventor of the Year (1988). In his books, Kurzweil eloquently shows how the evolution will move from biological to the computing platform and will lead to inventing machines that will extend, surpass and ultimately replace the human brain.
    Ray Kurzweil is also a prolific inventor. In the 1970s, he developed an OCR engine and a speech synthesizer. His Kurzweil Reading Machine scanned texts and read them aloud for vision-impaired people. He licensed his OCR technology to Xerox. In the 1980s, he designed the Kurzweil K250 digital synthesizer, which made it possible to sample acoustic sounds and play them via a keyboard. Finally, in the 1990s, his Voice Xpress Plus, was one of the first usable speech recognition PC packages. See: Technology Interview with Ray Kurzweil (!) and Learning in the Age of Knowledge
  6. David Shenk's Data Smog : Surviving the Information Glut discusses the excess of useless information and proliferation of disorganized knowledge that results in phenomena such as: information fatigue syndrome and, as the author calls it, culturally-induced Attention Deficit Disorder. Shenk's book points that information glut can lead to social fragmentation, the breakdown of democracy, the spread of demagoguery, and ... the decline of educational standards. The book also lists the possible solutions. Ironically, we cannot but hesitate if Witold Szafraniec's article does not contribute to Data Smog! Should we publish it or should we rather focus on writing about SuperMemo in plain and precise language. Does it make sense to struggle with articles peppered with references that may obscure the message (as well illustrated in this explanatory section). Or does it serve the purpose of well-rounded education?