Better than Spaced Repetition: 3 Points Against the Forgetting Curve


Spaced-repetition flashcard systems address a fairly plausible idea: that we forget things over time, and that by reminding us of these things at a regular schedule, we’re more likely to remember them over the long term.

They also take seriously the idea of overlearning — that drilling something too much actually makes it harder to remember in the future.

Both of these concepts were articulated by psychologist Hermann Ebbinghaus, who worked on problems related to memory in the late 19th century. He saw forgetting as a curve where most was lost in the first 20 minutes. After this, we continue to forget, but at a rate that gradually slows down over time.

Most of his research he did on himself, but when it comes to the memorization of pure facts, there doesn’t seem to be much reason to take issue with the model or the narrative behind it.

But if we think of learning in broader terms than just rote memorization of facts, how useful is this model?

Here are three reasons why we might want to look at other models when designing digital learning tools.

1. It’s Based on Memorizing Something That Doesn’t Need to be Memorized

When Ebbinghaus developed his theory, he tested himself with nonsense syllables like “ZOF” and “WID.”

One can see the logic behind this: a study on pure forgetting should start facts that are at once isolated from one another and challenging to remember.

And yet, it is a bit too abstract. When have humans ever inhabited an environment where forgetting was pure, where the facts around us were truly isolated from each other?

Ebbinghaus also described how mnemonic representation could improve memory.

In other words, memory is boosted when facts are tied to something else.

Probably the simplest process for doing this is to reimagine the concept as a visual representation and/or connect it with an existing network of thoughts or memories.

This is all well and good, but in this case, reshaping the mental landscape to learn nonsense is like cutting down trees to build a vacant house.

It belongs to a category of meta activities like speaking backwards or searching for anagrams — ones that, while perhaps useful as a form of mental training, have little in common with the ordinary processes of thinking and learning.

2. It Focusses on Material That Doesn’t Improve Over Time

Even if the material were something meaningful — let’s say Cantonese vocabulary words, for example — the process imagined by the forgetting curve requires us to imagined some kind of fixed text.

This might be, for example, a list, or a series of facts that are collected under a single organizing structure, such as “SAT Vocabulary.”

Concerning this, there’s a relevent passage in the Phaedrus, where Socrates describes the discussion between two Egyptian divinities: Theuth, who is boasting about the invention of writing, and Thamus, who critisizes this invention for undermining readers’ and writers’ ability to remember.

The usual understanding of this passage is that Thamus sees writing as a kind of cheat sheet; he fears that people will rely on external aids rather than on themselves. In this way, modern commentators have sometimes compared it to Google.

But the subsequent passage in the Phaedrus elaborates on the shortcomings of writing. In particular, it criticizes the way it doesn’t evolve, comparing it to a person, whom, if asked a question, will always give a single, unchanging answer.

In contrast, Socrates goes on to extol the virtues of conversation and dialectic, which allow for ideas to be challenged, tested, and create the basis for a shared knowledge to grow over time.

Writing, by contrast, never changes, never challenges, in the same way as a knowledgeable interlocutor.

In the scenario under the forgetting curve, though, the learner is simply going through the process of internalizing a single, unvarying perspective, one which doesn’t evolve over time. It describes the acquisition of a set of facts, not the development of knowledge, which comes about by challenging facts and then developing connections between them.

3. Knowledge Ultimately Rests on Thinking, Which is Better Served by Conversation

The acquisition of pure facts can be a practical and necessary part of learning, but it’s usually not its ultimate objective.

The use and development of knowledge instead requires thinking.

As thinking is refined over time. Thoughts, which structure concepts and things we need to remember, become sleeker, more powerful. The usage of ideas becomes less about practicing their recall than the development of their substratum.

When it comes to language learning, you’ll often hear arguments about language ultimately being social, about it being about communication.

This is certainly a valid perspective. But I think one could make the argument that communication-based practice is not valuable in and of itself (even though the ultimate goal may be to speak), but because it serves as an effective tool to develop thinking.

Conversation gives us the opportunity to think, to test, to develop thoughts, and get feedback about them. Socrates believed that conversation between intelligent people could lead to the best of insights.

Digital Thinking

But there may be other ways to stimulate, to prod, to get learners to think. This includes the use of digital tests and tools.

In the world of traditional language learning drills, for example, certain substitution drills can force learners to make connections between one form of knowledge and another.

These sometimes take the form, for example, of a list of several nouns, and then a sample sentence is given that students need to practice plugging the noun into.

In contrast to simple memory drills, they get learners to think about the language in broader terms — and create a sentences that they’ve possibly never seen or heard before.

More sophisticated forms of these substitution exercises exist – in particular the conversation drills in Japanese: the Spoken Language by Elizabeth Harz Jorden. These incorporate a piece of dialogue that the learner must respond to, using a phrase created by substitution as a piece of a conversation.

The pieces of conversation also are highly idiomatic, often resisting direct translation. They help learners adopt the mindset of native speakers.

It must be said, however, that these drills can be relatively dull. Learners usually find them unrewarding in the short term and become tired with them over time.

But I’d say they’re still in a primitive state.

A future set of digital tools, based on something like these substitution and response drills, could help learners not only memorize facts, but place them in a context that shapes their thinking.

If the technology were adaptable, possibly social, it might let us get past the anxiety of remembering, and focus on shaping our thoughts.

Then, we could focus not on beating the forgetting curve, but on making our thinking effective and true.