I'm really skeptical of these sorts of memorization systems, not because I think they don't work (I think they do), but because I see it as a brain "hack" that goes against the grain.
What do I mean by this? I think of our memories as a kind of most-frequently-used (MFU) cache. We remember best when we use the information all the time. So when we use this sort of spaced repetition we're trying to artificially "warm up" the cache in our brains. We're not trusting our brains to remember the things that are most important in our lives.
If you're trying to learn a language, you should learn the way children learn their first language: by speaking. Talk to people in that language constantly. If you don't know anyone who speaks the language, what are you learning it for? If you really need to learn the language (for work?) then find somebody who speaks it and arrange to meet them for coffee on a regular basis.
If you're trying to memorize something else, such as theorems in math or standard library functions in a programming language, I'd advise against it. Learn the theorems by applying them and trying to prove stuff. Learn the library functions by doing actual programming.
In the long run, if your knowledge is situated within a context and connected to other things you know (because you learned by doing) then I believe you'll understand it much better. Memorizing by rote, out of context, will help you to recall facts which often don't have any meaning to you.
> What do I mean by this? I think of our memories as a kind of most-frequently-used (MFU) cache. ... We're not trusting our brains to remember the things that are most important in our lives.
Absolutely not. I would never trust my brain to remember the things that are most important in my life. When is my wife's birthday? That happens once a year. My neurobiology is simply unsuited to triggering something at a specific time once a year. Plus our memory is not an MFU cache, not even close. Abandon all computer science metaphors and engage with the actual biology.
> If you're trying to learn a language, you should learn the way children learn their first language: by speaking.
That's like saying that athletes shouldn't do strength training or drills for specific skills. Mind you, I have learned a number of languages for real world use, and SuperMemo is one of my core tools for doing so.
> If you're trying to learn a language, you should learn the way children learn their first language: by speaking. Talk to people in that language constantly.
I agree that this is ideal, but it's also a huge synchronous time sink for the parties involved. When you're learning with flash cards, you can do it on your own time, in a more directed manner with trackable, linear improvement. In less structured environments, you don't get the breadth of vocabulary. In the beginning, you also end up asking "what does _ mean?" over and over again, or scrambling to write it down and look it up later, all while failing to grok the conversation happening around it.
Vocabulary alone is a huge barrier to entry in a language and, in my opinion, worth attempting to memorize in order to tackle a language.
Have you ever learned a language in your 20s? If so, how good are you? How long did it take you?
I personally have tried to learn many languages in my 20s. I’ve became good at figuring out what are the shortcuts to kickstart learning. The best one I found is to learn as many words as possible + simple grammar + useful verbs. If you have the time learn hundreds per day with memory tricks. Once you know enough start trying to have conversations. Grammar is not really important in the beginning but you need to learn conjugations (if there are any) at some point early.
This is a much better approach vs trying to have convos when you know peanuts. If you’re planning to spend some weeks in a foreign country this is what I would advise to you. You are going to learn very little if you get there without having done that
> If you're trying to learn a language, you should learn the way children learn their first language: by speaking.
I respectfully disagree here. After using the methods/strategies in the Fluent Forever book/videos/etc... (of which SRS is but one), while this strategy will work, its ultimately a longer road than necessary.
Kids don't get the chance to learn any other way than by observation and learning what not to do by inference. They're also very willing to make mistakes to tune their understanding. Adults tend not to be, taking the kids learning strategy as an adult is both ignoring the fact that you can take advantage of being an adult that can take in information differently, AND can speak to natives in a more structured way to much more easily learn grammar.
You'll still need to practice with native speakers no doubt, but knowing/learning words based on their frequency in the target language is something that you as an adult can take advantage of that kids realistically have to infer.
Immersion is great, but isn't ultimately a realistic option for most adults.
> but because I see it as a brain "hack" that goes against the grain.
The current, modern, and typical grain if you will: language learning is vocabulary, vocabulary and maybe conversation (maybe).
Although early on in childhood, as we develop "receptive language", we are going against this grain naturally by learning phonetics first then moving onwards to vocabulary through conversation.
I was in the auto parts store and a little boy saw me bring in a battery. He said "that wuks heavy!". I replied "that does LOOK heavy". Such a correction is abnormal from adult to adult. We have more grace depending on age.
For me, an average adult, with a busy schedule, learning the difference between the English T and the Spanish T made some great headway in my language hobby. Moving onward in that direction it has helped me understand spoken language - further aiding my memorization of the actual vocabulary.
My "grain" if you will:
Phonetics - Vocabulary (Anki with imagery) - Conversation (with natives, when I'm able to, it's a hobby)
The current, modern, and typical grain if you will
There is no "modern grain". Our brains haven't changed in any fundamental way over the past 10,000 years of civilization. Humans have been learning language as children the same way, for the entirety of history. The modern method of teaching language goes against the way humans naturally learn languages.
Look, talk to anyone who learned a language in school. None of them are anywhere near as good as a native speaker.
He said "that wuks heavy!"
I wonder how often children who make mistakes like this either have a hearing impairment or have parents who think it's "funny" to imitate their children's mispronunciations. Baby talk is a widespread thing and I think it's harmful to children.
When I talk to kids I treat them like an adult. I try really hard to speak the same way to everyone. If someone doesn't understand, I generally pick it up from their body language and go over what I'm saying in more detail. I ask questions to test their understanding.
> Humans have been learning language as children the same way
But you can't disregard the fact that an adult's brain is different from a child's brain. An adult's brain has a lot of prior knowledge that a child doesn't possess. Concepts such as freedom, money, market are easily grasped by an adult whereas a child has to focus on the meaning rather than the words (the representations).
I think adults can take advantage of their prior knowledge to speed up language acquisition.
Those approaches aren't mutually exclusive. You can easily learn theorems and math concepts using Anki and then apply them in practice.
You can learn words in foreign language using SRS and then use those words in conversations, when reading, watching a movie, etc.
I don't see any value in the purist approach your propose.
Also:
> So when we use this sort of spaced repetition we're trying to artificially "warm up" the cache in our brains.
No idea what this means. It is no more artificial than actively seeking out people to talk to just because you want to learn a new language, or forcing yourself to prove theorems just to better remember a mathematical concept.
I get what you mean and I agree to most of your points. I just want to point out that memorizing something you don’t use often is not completely useless. Some valuable creativity only happens when you connect some obscure and unrelated facts together, or some specific details from another field with what you’re currently working on.
> In the long run, if your knowledge is situated within a context and connected to other things you know (because you learned by doing) then I believe you'll understand it much better. Memorizing by rote, out of context, will help you to recall facts which often don't have any meaning to you.
I feel it's worth mentioning that the creator of SuperMemo agrees with you here. "Learn before you memorize" (Rule 2 in [0]). Learn the information, then commit portions to an SRS tool to keep them more effectively memorized and recallable.
You should be exploiting hacks to make things easier, not purposefully avoiding them to make things more difficult for you out of some misguided attempt at “purism.”
What second languages have you applied this technique to and how many hours of learning did it take for you to approach reading ability, conversational ability, reading/writing fluency, conversational fluency?
I want to learn a language but I don't think I'll be using it with anybody anytime soon. But I'd be happy to understand a newspaper or Wikipedia. What would be the recommended way to learn vocabulary if Anki is not an option?
I've been using Duolingo but it really feels like a glorified Anki
Content! Find video (movies, series), audio (-books, radio, podcasts).
The best way is to get exposed to the language as much as possible (some people prefer humans for this, some people prefer multimedia), and then you can use Anki on the side.
For example, I'm recently watching Fullmetal Alchemist in Japanese. I don't speak Japanese, but I've picked up 50 or so phrases. Sometimes I get a phrase stuck in my head without knowing what it means, or I'll wonder how to say something (that I know I've heard but forgot), and later that day my brain will dig up the answer from the depths (essentially a much longer version of the 'active recall' that Anki uses).
I'm a big fan of the Input Hypothesis (Stephen Krashen) which says the best way to learn a language is "input before output", in other words, immerse yourself in the language for a long period (say, 6-12 months) before trying to produce output (writing or speaking).
In this way, by the time you speak or write, you will know what it sounds like, how the grammar works, and will be able to self-correct, rather than forming bad habits.
We all know people who've lived in a country for decades and their grammar & pronunciation is awful despite constant use. The argument here is that they were forced to produce output (speak) before they were ready (got enough input to develop a workable model of the language), and were not corrected, typically out of politeness.
The key point is that the brain does this modeling on its own, given the input. Explicit learning (eg. Anki) is unnecessary, but it does "cement" knowledge and accelerates things.
If you'd like to read more on this, AJATT (All Japanese All the Time) is a great blog (a bit rambly though) on this method and also goes into spaced repetition quite a bit.
On the other end of the spectrum you have the exact opposite approach, traveling and going out to pubs and speaking the language from day 1: Benny Lewis, Fluent in 3 Months
Start with basic books but try to move up quickly to more challenging material. I read through Lord of the Rings in second grade. It was slow and frustrating at times, but the fact that the book is so engaging helped a lot. I learned so much vocabulary in such a short time this way.
Don't be afraid to keep a hard-copy dictionary of the language you're studying handy.
What do I mean by this? I think of our memories as a kind of most-frequently-used (MFU) cache. We remember best when we use the information all the time. So when we use this sort of spaced repetition we're trying to artificially "warm up" the cache in our brains. We're not trusting our brains to remember the things that are most important in our lives.
If you're trying to learn a language, you should learn the way children learn their first language: by speaking. Talk to people in that language constantly. If you don't know anyone who speaks the language, what are you learning it for? If you really need to learn the language (for work?) then find somebody who speaks it and arrange to meet them for coffee on a regular basis.
If you're trying to memorize something else, such as theorems in math or standard library functions in a programming language, I'd advise against it. Learn the theorems by applying them and trying to prove stuff. Learn the library functions by doing actual programming.
In the long run, if your knowledge is situated within a context and connected to other things you know (because you learned by doing) then I believe you'll understand it much better. Memorizing by rote, out of context, will help you to recall facts which often don't have any meaning to you.