Best Flashcard Setting to Actually Memorize Vocabulary

[Caleb Williams]

I read somewhere that SRS isn’t the best method to memorize and accumulate vocabulary. I just want to find out how to I should set my test setting to help myself accumulate vocabulary. I’ve been studying for 7 years.

[imron]

The best way to accumulate vocabulary is to read/listen/otherwise consume content appropriate for your level.

 

SRS can then be used as a secondary tool to supplement this learning.

 

If you only use an SRS to accumulate vocabulary you will find that when you try to read something, many of the words you supposedly know have only been learnt to a superficial level that lets you get them correct in flashcard reviews but still not understand them in real life content.

 

For what it's worth, here is my Pleco setup, and here is where I talk about using reading and memorizing vocab.

 

 

[wibr]

I use the standard Pleco SRS algorithm. Previously I followed the approach to just dump everything into SRS and let the algorithm take care of it, however at more than 10k entries the reviews start to pile up and you review many words that are not very relevant. Now I use a more targeted approach and learn only words that I am about to see in  a book or tv series, using precompiled word lists. Afterwards I remove those words from my flashcards until they come up again in another list. That way all the words I study are immediately relevant and I can see how they are used in the real world. It also makes it easier to limit the amount of reviews I have to do each day.

[suMMit]

22 hours ago, imron said:

The best way to accumulate vocabulary is to read/listen/otherwise consume content appropriate for your level.

I have to say I agree with this. I think you get "true" spaced repetition this way. You also get full sentences with collocations, context, grammar... A few months ago, I decided to try not using any kind of vocab lists or flashcards as an experiment. I feel like I've learned/remember just as many words, so I don't plan on going back to flashcard drudgery.

[jannesan]

I use Anki to study pre-study vocabulary of books that I am reading and words that I encounter when speaking with teachers/language partners.

I have the feeling that the creation of the flashcards is already a good start to memorize as I pick example sentences for each usage of the word to include on the flashcard.

 

As some words that I have learned for a book may become less relevant for me later, I am planning now to be more strict with suspending these words if I forget them again, to limit the amount of reviews every day.

[realmayo]

I think you should memorise vocabulary before using an SRS. Once you've learned the words, then use an SRS to stop you forgetting them.

 

I don't think SRS is good for learning because it's too shallow - it's just: "look at word, do I remember it?". To really learn it requires more brainpower than that. Or alternatively, coming across a word in a book or situation that is fairly memorable itself.

 

Maybe SRS is okay for learning English to Chinese, because you've got to think hard about the word first.

And I think it's fine for learning how to write characters if you've got some kind of system in your mind to help you recall them (because you'll be actively using your brain as you use that system).

 

But for Chinese to English vocabulary learning - I'd advise another method, before putting it into a normal SRS programme.

(And that other method could actually use a heavily modified SRS ... but that's another story.)

[wibr]

@realmayo Interesting, that's almost the opposite of what I would recommend. Typical for language learning advice.

 

I think SRS is best suited for the initial learning, to bring words from unknown into long term memory. Then, when the SRS intervals stretch into weeks and months, the model of the algorithm becomes less and less precise and you will get natural repititions in the real world that the algorithm doesn't know anything about. So that's why I stopped using SRS for long term maintenance. Depending on the word, just SRS is not enough to really learn it but at least you have some understanding of the word already in long term memory after the initial SRS phase.

[somethingfunny]

It's easy to conflate two aspects of learning: recall and understanding.  SRS will help me remember Henry VIII was king in the 1500s and had 6 wives, but it won't help me tell you how that caused a split in the church.

 

Recall is improved with spaced repetition.  Hopefully, this would be done mainly through natural language use - reading, writing, and conversing in Chinese as part of your daily life.  However, for most learners this isn't possible, so we artificially supplement it with SRS.

 

Understanding is improved with explanation and recall in different contexts.  The more I read and talk about the life of Henry VIII, the more I understand why he had 6 wives, and why this was important.  Hopefully, this would also be done mainly through natural language use - reading, writing and conversing in Chinese about things that interest you as part of your daily life.  However, for most learners this isn't possible, so we artificially supplement it with something called 'interleaved practice' where textbooks and teachers expose us to the same content in different contexts.  A good Chinese textbook will still be using as much of the vocabulary from chapter 1 in chapter 20 as possible.

 

Focussing on recall before understanding works.  I could memorise a list of facts about Henry VIII's life in isolation and then go and read a biography.  The facts I'd learned would then be fleshed out by the contextual information provided by the biography.  However, learning facts out of context is a bit demoralising, and it also limits the schema forming capacities of the brain (the facts are emphasised, rather than the connections between them).

 

Personally, I'd rather take something in context first, with explanation, and then derive factual information for retrieval practice based on that.  In other words, if you've really got your heart set on learning the life of Henry VIII (where learning is defined as the alteration of long term memory) then the best strategy is probably go read a biography, inputting factual aspects into an SRS system as you do so, using that system for a few months, and then reading another biography.  Then repeat.

 

It doesn't really make sense to say things like "learn before SRS" or "memorise before SRS", because "learn" implies retention in long-term memory.  Continued retrieval of information is a part of the learning process.

 

In terms of Chinese, this means that there isn't much point in SRS if you never actually use the language.  But if you do have the opportunity for language use, then SRS is a powerful tool for making that language use as efficient as possible in terms of learning.  This is because your time using the language won't be spent on costly recall, but on contextual understanding and schema formation.

 

tl;dr - what Imron said.

[realmayo]

1 hour ago, wibr said:

Interesting, that's almost the opposite of what I would recommend. Typical for language learning advice

 

  Very typical! But maybe not so far apart: I suppose what I meant was: spend some effort on a brand new word, rather than just pressing "Again" the very first time you see it.

 

I found saying a new word aloud, looking up to see what kind of context it might appear in, and then testing myself on it a few times within the space of a few minutes to work best for me. I'd also write it out: Chinese - pinyin - English.

 

(Then I would in fact use a special SRS deck which would show me it once per day until I got it right three days in a row, after which it would "graduate" into the main SRS deck which used the normal spaced repetition algorithm. But now I'd do those first few tests in a notebook, not on screen, because I prefer pen and paper. In an ideal world I'd also be seeing and using that word over the first couple of days too.)

 

As far as I can see, the alternative - where a new word popping up in an SRS deck is the first time you encounter it - would be:

   Day 1: see the word for the first time. Get it wrong. See it again in 10 mins. Get it right.  (10 seconds total)

   Day 2: see the word for the third time. Get it right.  (5 seconds)

   Day 5: see the word for the fourth time. Get it right.  (5 seconds)

   Day 11: see the word for the fifth time ...

 

For me that's triage, not learning! It's sorting the easy words out from the ones you will have to Fail on Day 11.

 

But @wibr I think we're probably only disagreeing about "initial": I'm focussed on the very very initial. For me, the first few times you encounter a new word of moderate complexity ideally would not, I think, be in an SRS system.

 

 

[carlo]

I never had much success with SRS. The best method for me personally has been to (1) mark up key words in text or audio/ video files, (2) read through/ look at that list of marked-up words frequently over a period of time, and (3) reading/ watching original content repeatedly, absorbing the words in context. And, much more infrequently, (4) record myself repeating those words and phrases as accurately as possible.

 

I believe that doing this over time with content I enjoy has been the greatest contributor to my language learning success. I’ve been doing this with German, Swiss German, Mandarin, Cantonese, Japanese for many years now (and even with lyrics in languages I don’t speak at all but wanted to memorize effectively in order to do a cover of the song).

 

One efficient way to do that would be making my own subtitles/ wordlists for audio/video using Aegisub (open source subtitle editor).