Incremental reading

Using incremental reading with an Anki add-on: extracting a portion out of an article and creating a cloze deletion

Incremental reading is a method for learning and retaining information from reading that might otherwise be forgotten. It is particularly targeted to people who are trying to learn for life a large amount of information, particularly if that information comes from various sources.

"Incremental reading" means "reading in portions". Instead of a linear reading of articles one at a time, the method works by keeping a large reading list of electronic articles or books (often dozens or hundreds of them) and reading parts of several articles in each session. Articles in the reading list are prioritized by the user.

In the course of reading, key points of articles are broken up into flashcards, which are then learned and reviewed over an extended period of time with the help of a spaced repetition algorithm.

Incremental reading is based on psychological principles of long-term memory storage and retrieval, in particular the spacing effect and the testing effect.

The method itself is often credited to the Polish software developer Piotr Wozniak. He implemented the first version of incremental reading in 1999 in SuperMemo 99, providing the essential tools of the method: a prioritized reading list, and the possibility to extract portions of articles and to create cloze deletions.

Until recently, Wozniak's SuperMemo was the only implementation of incremental reading. Anki now has an implementation available as an add-on.

There is also incremental reading support for the text editors Emacs and Yi.

Method

When reading an electronic article, the user extracts the most important parts (similar to underlining or highlighting a paper article) and creates flashcards. Flashcards are information presented in a question-answer format (making active recall possible). Cloze deletions are often used in incremental reading, as they are easy to create out of text. Both extracts and flashcards are scheduled independently from the original article.

With time and reviews, articles are supposed to be gradually converted into extracts, and extracts into flashcards. Hence, incremental reading is a method of breaking down information from electronic articles into sets of flashcards.

Contrary to extracts, flashcards are reviewed with active recall. This means that extracts such as "George Washington was the first U.S. President" must ultimately be converted into questions such as "Who was the first U.S. President?", "Who was George Washington?", etc, or cloze deletions such as "... was the first U.S. President", "George Washington was...", etc.

This flashcard creation process is semi-automated – the reader chooses which material to learn and edits the precise wording of the questions, while the software assists in prioritizing articles and making the flashcards, and does the scheduling: it calculates the ideal time for the reader to review each chunk, according to the rules of spaced repetition. This means that all processed pieces of information are presented at increasing intervals.

Individual articles are read in portions proportional to the attention span, which depends on the user, their mood, the article, etc. This allows for a substantial gain in attention.

Without the use of spaced repetition, the reader would quickly get lost in the glut of information when studying dozens of subjects at the same time. However, spaced repetition makes it possible to retain traces of the processed material in memory.

External links

• Incremental reading (supermemo.org)
• FAQ : Incremental reading (supermemo.com)
• Incremental learning (super-memory.com)
• Kevin Purdy, Use Incremental Reading to Memorize Large Batches of Data (lifehacker.com)