Serendipitous impact of mindless cramming
This text is part of: "I would never send my kids to school" by Piotr Wozniak (2017)
Cramming
Every cloud has a silver lining. Even the worst learning habits developed at school can bring something good. One of the worst habits we take away from school is cramming. In cramming, we learn to pass an exam only to forget the whole material in weeks or months that follow. Hence the phrase: "Cram&Dump!". Cramming is almost useless from the lifetime perspective. Due to the fact that most of learning for exams has a form of cramming, schools are also almost useless when compared with the actual brain potential for true learning. Cramming is encouraged by the system of testing in which a big chunk of material is tested once, and helps the student advance to the next station at the assembly line. Schools teach kids that knowledge is disposable, easily crammed back, and cheap. As a result, schooling conditions the brain to underestimate the value of lasting knowledge and the cost of quality learning.
Paradoxically, my own case of the wasteful procedure of dumb cramming helped me develop what can mathematically be proven as the fastest lifelong learning technique available (at the moment of writing): incremental learning. Some familiarity with SuperMemo would be helpful to understand the paradox described below.
Active learning
I vividly recall Timmy before a test, resting against a school corridor wall, over a book, with his head bobbing up and down like a stressed caged animal in a zoo. This picture is a scary metaphor on the impact of schooling on sensitive minds. This type of behavior might have contributed to Timmy's being bullied. Timmy was reading and re-reading the same passages in a book in hope that some of that reading will stick for the test. Timmy was not focused on learning. He was cramming in the state of utmost anxiety. It was at about that time that I figured out that passive reading is almost useless for passing tests. I needed active recall after reading. I spoke to Timmy about this. He was not receptive. He was just too stressed. That active recall realization later helped develop SuperMemo. In 1982 (aged 20), I started formulating all my vital knowledge as questions and answers (in the form of flashcards). That dramatically improved my learning performance. In the meantime, Timmy did not improve his learning habits. The precocious kid tried twice to pass entrance exams to study medicine. He failed both. I failed to help Timmy. Most of all, Timmy was failed by the education system. By the time of his last exam, he was a nervous wreck. His educational dreams were shattered.
In contrast, I was slowly progressing through. My care-free attitude sheltered me from the impact of stress. By a whisker, I passed my entrance exams to study biology. Had I chosen medicine, like Timmy, I would have failed too. The bar of medicine was just too high. Interestingly, the prospect of digging into corpses in anatomy class was one of the factors that changed my direction. Interestingly, life of Morton Deutsch has been changed in a similar mannerSuccessful cramming
Roots of incremental learning
Origins of SuperMemo
- older layer, dating back to 1985, classical SuperMemo, based on the algorithm that we called computational spaced repetition, makes sure that memories stay with the student for ever
- newer layer, dating back to 2000, based on incremental learning techniques, makes sure that large volumes of information can be processed and fed back to the memory engine of classical SuperMemo
The interesting bit is that the ideas at the core of incremental learning are older than the ideas related to spaced repetition. In my cramming notes used before exams, all pieces of information were marked by two values of a memory attribute: (1) To Cram, or (2) Crammed. The flip of the attribute value would occur at the moment of crossing out the note. I used my cramming techniques in the years 1982-1990 during my university exams, and once, in 1995, before my PhD exams.
In my PhD dissertation, in 1995, I outlined how global hypermedia might be enhanced with tools that employ more processing and memory attributes. For example, a processing attribute might define the priority of a piece of knowledge, while a memory attribute might define how well a piece is remembered.
I knew that those attributes would be vital for implementing fast learning technologies that would satisfy principles of constructionist learning proposed by Seymour Papert of MIT. I used Papert's own term "knowledge machine" to describe a tool that could make super-fast and super-efficient constructionist self-learning possible. Today, this "knowledge machine" is available as SuperMemo for Windows of which versions from before 2013 are free. That lofty concept started from mindless cramming for exams! Schooling can bring some good fruit tooSummary
- schooling schedules inevitably lead to cramming
- cramming is bad for memory, mood, and long-term learning outcomes
- schooling conditions the brain to underestimate the value of lasting knowledge, and the cost of quality learning
- cramming techniques inspired incremental learning, which has nothing to do with cramming
- cramming on paper was important for emergence of incremental reading (see: Handwriting is dead)