Conglomerating items is not a good idea

From SuperMemopedia
Jump to navigation Jump to search

From: L.S.
Country: USA
Sent: Oct 4, 2012
Subject: Conglomerating information in spaced repetition results in slower learning


I tend to disagree with some of your "20 rules claims". For example,

Q: What was decided at the Council of Trent, beginning in 1545, and how long did the Council go on?
A: The basic beliefs of the Catholic Church; 18 years.

Mixing two questions in one is more efficient and also links related pieces of information together that should be linked together. It is more effortful to remember the answers, I agree, and that is a concern. Occasionally I'll rewrite or simplify a question, or break it into two, if it becomes too difficult.

E.g., asking what the Council of Trent was and how long it went on makes sense because they are both very basic questions about the Council of Trent. If I were to separate them, the neurons wouldn't fire together. It would take a little extra effort to discover that the thing that formed the basic beliefs of the Catholic Church lasted for 18 years. Since he'll be learning both facts at about the same time, why not learn them together?

In the same way, contrary to your advice (sorry!), I do still have full sentence and even two-sentence questions--only occasionally. But the point is that sometimes, the thing that needs to be committed to memory is a narrative, a whole sequence of events and not just particular items from the sequence. Maybe I am wrong, but I think a narrative is best remembered by practicing the narrative.


Two separate memories should be separated in SuperMemo due to the fact that they nearly always will require different timing of repetitions. If you can always activate the same mental pathway in thinking about the Council of Trent ("neurons firing together" in the same pattern), your particular item has a good chance of suriving long in the process without a memory lapse. However, once you build a large database of similar items, and you review your sizeable material under the pressure of time, your review will always tend to strip redundant pieces of information. Overtime, your nice item will be reduced to the bare bones of information that will often fail its primary test: applicability in real life. It may happen, that despite zero memory lapses, in 2-3 years, someone will ask you about a Council of Trent in a new context and you will be amazed that you won't be able to reasonably answer the question despite having all the necessary pieces of information included in your item. Two memories of different difficulty might be compared to two different planes of different flying characteristics. The difficult piece (e.g. 18 year duration of the Council) might be compared to a slow flying plane. The easy piece (here the reference to the Catholic Church) might be compared to a modern jet. Review of the conglomerated item might be compared to flying both planes at the same speed. In an extreme case, this might be impossible. The compromise speed might be too high for a slow plane, which might disintegrate beyond a certain speed limit, while the faster plane cannot slow down enough without stalling. In our memory, forgetting is equivalent to forgetting, while stalling is caused by the spacing effect. By doing complex and repeatable reasoning at each repetition, you might act as if handling both planes using remote control. However, this is always difficult and requires lots of focus and deliberation at repetitions. Your brain has natural defenses against such "enforced repetitive reasoning". It is designed to be "intellectually lazy" and thus energetically efficient. Practise shows that incremental reading produces many more items. However, those items are usually much easier to remember. In the end, you spend less time on reviewing 5-10 items than you would spend on an item that would conglomerate information and sufferred repeated memory lapses or very short intervals.

In the course of the evolution, the brain developed strategies for abstracting away from the details and retaining only the most essential, useful and frequently used information. Those strategies are great for survival, but aren't as good in reaching our educational goals. Council of Trent is a typical example of knowledge we wish to have, but that is pretty expensive. This is because, for most people, it does not get reinfornced in run-of-the-mill conversations, TV shows, daily applicability, or at water cooler at work. The situation might differ if you, in particular, read a lot on the subject matter. This might help the memory establish itself in an efficient manner. Incremental reading makes it possible to root such difficult-to-retain knowledge firmly in the context, and still make sure that individual repetitions focus on a very specific and cheap-to-retain memories.

This is how the same paragraph might be processed with incremental reading, and paradoxically cause a significant saving in time in the long run:

Q: The Council of [...], which began in 1545 and lasted for 18 years, made decisions about the basic beliefs of the Catholic Church
A: Trent

Q: The Council of Trent, which began in [...](year) and lasted for 18 years, made decisions about the basic beliefs of the Catholic Church
A: 1545

Q: The Council of Trent, which began in 1545 and lasted for [...] years, made decisions about the basic beliefs of the Catholic Church
A: 18

Q: The Council of Trent, which began in 1545 and lasted for 18 years, made decisions about the basic beliefs of [...]
A: the Catholic Church

Q: The Council of Trent, which began in 1545 and lasted for 18 years, made decisions about [...]
A: (the) beliefs of the Catholic Church

Q: [...], which began in 1545 and lasted for 18 years, made decisions about the basic beliefs of the Catholic Church
A: The Council of Trent

In the end, if you are sure this item works for you, check its performance in the course of the next few years. If you pass the interval of two years without a lapse, you can say that this particular item indeed works for you. In that case, there is no disagreement between you and the 20 rules. It is just that for most people, this item is pretty likely to generate a lapse within two years even if reviewed at correct timing. Depending on the item difficulty, the number of repetitions in the first 2 years might be as low as 3 or well above 20. If your default forgetting index is 10%, this translates to a span from 70% chance of retaining the item to the totally unacceptable 90% chance of forgetting! This last number is little understood and little realized by the users of SuperMemo, and should always make you think a lot about the rules of efficient formulation of knowledge.

For more on the theory of conglomerating information in spaced repetition see Formula 9.4 in the article about building memory stability.