Abandoning Anki (at the upper intermediate level)

[davoosh]

I know there have been lots of discussions about anki and other SRS programmes, but I thought I'd share my experience with anki and see what others have to say about their own experience.

 

I have been using anki on and off for a while (several years), and recently decided to abandon it.

 

I am currently around the upper-intermediate level and have recently had a long break from using any Mandarin. When I got back into the routine of studying, I naturally started with anki because it seems an effecient way to keep track of vocab and has worked more or less well for me previously.

 

After using it for a while, I noticed that I would be able to answer questions in anki very quickly and it seemed like the vocab was sticking, but if I saw/heard them in context or a text, I'd still have to pause and strain to get the meaning or my mind would go blank. My anki deck also contained a lot of vocab with example sentences.

 

After continuing like this for a while, words would start to appear a month or two later after I had added new content. I generally read/listen to native materials and add words that I feel are useful. However, these words which were now appearing totally out of context would bore me and make me resent doing any reviews because the number was always increasing and my motivation decreasing.

 

Long story short, I began to find it much more beneficial at this stage just to keep a notebook of words that I can skim through when I'm bored without feeling like I "have" to complete anki reviews. Also keeping any articles I've read and glancing over them once in a while helps reinforce the vocab in a natural context without the tedium of anki. If I notice I've forgotten a word, I just note it down in my vocab book. This way is probably less efficient than that of hardcore anki users, but I feel anki killed my motivation for a while and at this level engaging with native material often is much more beneficial and enjoyable.

 

I didn't know about anki when I first began Mandarin, but I imagine at beginning levels it is probably a lot more effective. I now tend to use anki sporadically if I want to drill a certain vocab list (for example, I recently wanted to learn specific herbs and spices) but I don't keep big decks with huge collections of words.

 

How has everyone else found the use of SRS throughout their studies of Mandarin or any other language?

[Guest realmayo]

I've certainly found that how I use Anki changed once I had 10,000+ cards in it. The main difference is I don't get annoyed when I forget cards. If a card is very leechy, or if I no longer see the point in having it in my deck, I'll cheerfully suspend or delete it. Otherwise it stays. Much of the time it's something I'd usually remember if I see it in the real world, because of context.

 

And in fact I'm now so relaxed about failing cards, often I'll fail ones that I (just) get right but which have long intervals and which I'd like reminding of a few more times.

 

I still think the targeted exposure is useful and I currently can easily spare 30 minutes a day on Anki without sacrificing any other study time.

 

But in terms of adding new vocab I'm now much more selective. I'll "drill" new words in a separate deck for a week or two and then delete most of them, only transferring a few into the main deck.

 

The ones I transfer tend only to be particularly useful, or else they are simple nouns and verbs that carry no nuance.

 

I think some people do get quite emotionally attached to the big deck they've nurtured. There seem to be two ways to break that attachment: either delete the whole thing, or stop caring when you forget cards.

[Silent]

 

After using it for a while, I noticed that I would be able to answer questions in anki very quickly and it seemed like the vocab was sticking, but if I saw/heard them in context or a text, I'd still have to pause and strain to get the meaning or my mind would go blank.

I recognise this. IMHO it's caused by the fact that you 'know' the vocabulary in anki, probably you subconsciously also have an idea when the vocabulary comes up. In the wild however you're not/less able to anticipate the occurrence of vocabulary and you also are uncertain or you know the vocabulary. So a general outline of the character/word is insufficient to be sure it's the same as the character/word you learned. There is also the segmentation issue. In anki you know it's a word, in the wild you're often not 100% sure how to segment the text into words, often it's a bit trial and error to find a segmentation that makes sense.

 

 

Also keeping any articles I've read and glancing over them once in a while helps reinforce the vocab in a natural context without the tedium of anki.

My experience with this kind of approach is that it only works when the text is (very) hard. The first reading gives me the general outline and rereading it I've a framework to help me figure out more details. Rereading a fairly simple text with only a few difficult vocabulary items makes me skim it over, get the general again which makes me feel I know it without really (consciously) reinforcing the unknown vocabulary.

 

To be honest I too start to feel more and more that my anki-deck becomes/is a drag instead of a crutch. For me it's mostly caused due to bad habits develop while getting rid of a big backlog and lacking motivation/discipline. I also feel it keeps me from reading but rationally I know the reading issue has far more to do with discipline then with anki. I'm not yet (mentally) ready to ditch my deck. Nevertheless, despite I strongly pleaded against such strategy in the past the idea of ditching it and starting over becomes more and more attractive.

[Guest realmayo]

Here's something else to consider (I found it very nice): order the cards by difficulty and take out the top 5%. Put them in a new deck and don't touch them for a month or so. Decide what to do with them later. I ended up deleting maybe 75% of them.

 

And stop adding any new words to anki for a month.

 

Again, I think it's part of moving away from the mindset of 'all the words I know are in Anki and I must remember them all' which is fine for, say, HSK1-5/6 if that's how you want to build a vocab base but causes problems later. Because I do agree, once SRS becomes a real drain on your willpower then it's probably no longer an efficient way to study -- that's certainly how it became for me at one point.

[li3wei1]

I was thinking of taking a different approach: abandoning everything BUT Anki - family, job, personal hygiene - but perhaps there's a middle way.

[Shelley]

I took the opposite view and abandoned anki at the very beginning, recognising all the things mentioned by realmayo, silent,and davoosh. Remembering them is different than being able to use them, and recognising them when I hear them.

 

I use Pleco flash cards but not intensively, it is just part of my routine for each new lesson.

[Silent]

 

I was thinking of taking a different approach: abandoning everything BUT Anki

In my perception this is a bad approach. I started my studies with only anki. At the time I felt it worked really well and I still think it was a good approach in the very beginning. However fairly soon (within learning a few hunderd items) you should complement it with actual reading/listening and grammar. Learning only vocabulary makes you a walking dictionary without any real language skills. Just look at the (early) machine translations and you realise that word by word 'translations' are pretty much useless. If you limit yourself to only one thing I'ld say it should be reading or listening.

[edelweis]

I limit my anki time to 15' a day (out of about 1hr of daily study time).

You need to encounter the words in the wild (reading, listening), and ideally actively use them to make sentences, in order to really learn them and how they are used...

[imron]

@davoosh, you may find this thread interesting.

 

In short, there's nothing wrong with abandoning your deck (and in fact this should be your end goal anyway), just make sure you are supplementing it with other learning (reading, watching TV shows etc).

 

 

 

Nevertheless, despite I strongly pleaded against such strategy in the past the idea of ditching it and starting over becomes more and more attractive.

Good to see you slowly coming around

[simc]

The alternative to deleting your deck is to control how you review it.

In Pleco you can add cards to different categories. I usually only review cards for the material I am currently studying.

Another thing you can do is use a frequency list to weed out words that are relatively rare so you prioritize more common words. Pleco can do this automatically if you import a word list to use as filter, see A Frequency List for Flashcard Triage with Pleco.

[Silent]

 

The alternative to deleting your deck is to control how you review it.

In Pleco you can add cards to different categories. I usually only review cards for the material I am currently studying.

What's the difference? You abandon previously studied vocabulary for a new set of vocabulary. Deleting, suspending, excluding.... the effect is the same.

 

 

Another thing you can do is use a frequency list to weed out words that are relatively rare so you prioritize more common words. Pleco can do this automatically if you import a word list to use as filter, see A Frequency List for Flashcard Triage with Pleco.

Sure you have to be critical about what you choose to study. At lower levels it doesn't really matter that much as basically everything is high frequency but general frequency lists and lower hsk levels can provide a decent guideline when in doubt. At higher levels it becomes a personal estimate, vocabulary and frequency lists become pretty useless unless you're able to personalise them. Profession, personal interests and circumstances have a huge impact on the frequency of words you encounter. A general frequency list of newspaper articles sounds very good, but the moment you consistently skip the sports section or the financial section that frequency list has little to do with your actual experience when reading the newspaper.

[simc]

What's the difference? You abandon previously studied vocabulary for a new set of vocabulary. Deleting, suspending, excluding.... the effect is the same.

You get to keep the huge deck that you have sentimental attachment to.    

 

I think you have a point about frequency lists is very true at the higher levels. Use SUBTLEX as I assumed that the words people use in the movies or on TV was a good approximation to what people say in real life. Might be time to revisit that assumption.

 

Does Imorn's program let you build a personalized frequency list based on the corpus of articles saved on your computer? Might be a useful feature.

[imron]

Not yet, but when it's finished, the corpus feature will have that functionality.

[Silent]

 

Does Imorn's program let you build a personalized frequency list based on the corpus of articles saved on your computer? Might be a useful feature.

If you process it in one run it does. Stick it together, mark everything as known and export the known vocabulary (with the frequency selected) I think (but maybe I'm mixing things up with the other program I use) in the past it was also possible to export the unknown vocabulary which imho is more usefull as that's the list you should work on.

 

The best way to go is imho put together what you intend to read analyze it and add the high frequency unknown words to anki. Catch is however names. They tend to appear high frequency, but the characters/words tend to be fairly unimportant. So blindly adding the unknown vocabulary is not recommended. You may use HSK or another frequency list as filter to help throw out the names/less usefull vocabulary.

 

You can do something alike by bunching together some articles about a subject(s) you're interested in in order to get more subject specific frequency lists. To get reliable frequency lists however you need a lot of data from various sources/authors. So I guess the approach of bunching together what you want to read and/or perhaps what you did read is better.

[imron]

Stick it together, mark everything as known and export the known vocabulary

Yes, you can do it if you manually combine all the files. However you don't need to mark everything as known, you can just set the wordlist to 'None' and it will work on all words in the file.

 

in the past it was also possible to export the unknown vocabulary

It's still able to do this currently

[Silent]

 

It's still able to do this currently

Apparently I'm missing something (probably dementia as I feel we discussed this before)

I've two options to export known and none (=all) I don't see an option unknown.

 

Edit: I use version 0.99.9

[querido]

This gets you the unknowns. I also didn't notice it for a while:

 

Wordlist: [Known]

Filter: Exclude words on list

[Silent]

A bit counter intuitive to me but it seems to work correct.

 

thanks！

[imron]

A little bit, but the change facilitates adding other lists in the future.  For example if there is a list HSK6, you can't just change that to say "UnHSK6".  This wording, although slightly confusing initially, will handle multiple different lists in a uniform way.

[MarsBlackman]

This post inspired me to abandon my Anki deck of 6000+ cards. I feel so free. I didn't realize how much time I spent not just reviewing cards, but creating cards. I've just substituted my Anki time with reading. It seems to be a fair trade as I continue to dive into native material (or try to at least).

 

Its been about two weeks and I find my study of new words to be more focused. With my Anki deck, I would passively learn new words, knowing I would put them in my deck later and assume the review process would magically make them stick...which is didn't. 

 

Of course you're going to forget words which is why I think people get attached to their decks. But I blank on English words all the time and I'm a native speaker.