Techniques used to develop Advanced English

There is a number of obstacles that have to be overcome in writing a SuperMemo collection which are not known to lexicographers compiling dictionaries: the knowledge must be represented in simple terms, in minimum words, in agreement with the art of mnemonics, univocally, and with minimum interference between different items. There is no better way to ensure the compliance with the aforementioned principles than to memorize the collection in its entirety. This procedure, though extremely time-consuming, provides an excellent way to weed out ambiguous items, unclear definitions, etc. Only by memorizing the material could the problematic items be redefined, simplified, made more specific or enriched by examples. For instance, the following item seems to be quite correct from the semantic point of view:

Q: to trouble sb with questions or sexual offers

A: molest sb

However, it is strongly interfering with the definition of the word harass, and the learner may often have problems to recognize the difference between the two words. A simple remedy to this problem is to redefine the items in the following way:

Q: to harass sb with questions or sexual offers

A: molest sb

A great help in memorizing vocabulary are vivid examples. For instance, the definition:

Q: a bag like formation of skin

may seem quite enigmatic at first. However, addition of a simple example will turn the item absolutely clear and easy to remember:

Q: a bag like formation of skin (e.g. in kangaroo)

A: pouch

In some cases the need for simplicity might render some items unacceptable from lexicographic point of view. For example, the question of the item:

Q: spider's net

A: cobweb

resembles an entry coming from a cross-word puzzle rather than from a dictionary. However, this form of a question leaves a durable and easily accessible imprint in memory; hence the preference over any other more elaborate definition.

Another difference between SuperMemo collections and dictionaries is that the question should by no means provide other than semantic means of recalling the answer. For example, in items defining names of plants and animals, Advanced English often provides the Latin name to make the definition unambiguous. However, in case of the larch, the Latin name Larix could not be provided as it suggests the correct answer.

As most learners will memorize words in the sequence they are listed in the collection, establishing the order of items appeared to be a major challenge. The simplest solution which comes to mind is to use standard word frequency lists. However, (1) 50% of items in the Advanced English collection are not just words and the frequency data are not easily available, (2) a well-structured SuperMemo collection should give preference to items which can easily be memorized, and should put harder items at the end of the list.

In SuperMemo World, we have developed a simple heuristic algorithm for sorting collections using the following criteria:

The sorting algorithm works quite well in most of cases, and has been used to sort the Advanced English collection as well. Consequently, some of low-frequency and easily remembered items are placed at the beginning of the collection, while more frequent terms, which pose serious problems to memory are listed later. This seemingly controversial solution ensures maximum speed of learning in mastering a given collection, and gives the learner a chance to improve his understanding of SuperMemo and his own memory before he is thrown in at the deep end