Who else goes both ways with SRS?

[feihong]

I should clarify something: I use my English to Chinese cards mostly to practice my writing. Therefore I have zero problems with synonyms, because the pinyin is also on the front of the card. But I don't practice writing because I like to write characters by hand, I do it because it helps me remember the characters.

[icebear]

Therefore I have zero problems with synonyms, because the pinyin is also on the front of the card. But I don't practice writing because I like to write characters by hand, I do it because it helps me remember the characters.

Yea, that reminds me to clarify also - going English to characters I allow myself to listen to the sound of the word and then write it - I grade myself only based on how well I remember the strokes for that direction.

[realmayo]

I would love to hear (more) experiences of people who stopped doing production cards and instead went recognition-only.

I stopped doing production somewhere between the 5000 and 10000 mark (I think): it was just taking too long.

I've since restarted production simply for individual characters, because although I don't expect to ever have to write anything by hand, I feel that being able to do so gives me more confidence when it comes to recognition -- this is the case for me but not for most people I think.

Currently, then, it's only the single-character deck which is production, the vocab one is all recognition.

But, I'm thinking to restart some production in the vocab deck, though certainly not for all the vocab in the deck.

Even if you're not doing any SRS you will have some vocab which you can produce in conversation and so on, and some vocab which you can only recognise. It seems to me that there should at this moment be a chunk of vocab which is not too easy and not too difficult which it would be worthwhile me SRS-ing on a production basis. If I was speaking Chinese every day there might be no need for the artificiality of SRS, but I don't normally speak Chinese these days.

[Ludens]

Thanks for the input everyone! It confirmed my inclination to change my srs habits. I think I will go with a model similar to that of Realmayo, so now the next step will be figuring out how to convert my vocab production cards into single character production cards without losing scheduling data.

They (flashcards) are a means to an end, and even when I tend not to be extremely harsh with them they seem to be doing a reasonable job at that. Still, I think rounding that corner towards 'natural SRS' is important and hope I'm ready for it in the not too distant future. It sounds like you may be nearing that point also (5000 items - is this 5000 words or 2500 words?).

That's almost 5000 words (which translates to 10.000 cards now), using just over 2600 characters. I think it would be possible to keep up with my current model for another 5000 or so words, but it would take so much time that I hardly have time to study reading, listening and speaking (which is the main problem now already).

[icebear]

That's almost 5000 words, using just over 2600 characters. I think it would be possible to keep up with my current model for another 5000 or so words, but it would take so much time that I hardly have time to study reading, listening and speaking (which is the main problem now already).

Exactly! At 4000 words/2000 characters I'm able to manage all my SRSing in around 30 minutes a day, which is reasonable in my opinion. I don't plan on significantly increasing that amount of time, however; I'd rather invest more time in reading, TV or conversation. Don't lose sight of the forest for the trees...

[imron]

I hardly have time to study reading, listening and speaking (which is the main problem now already).

If you read enough material, and read regularly, you will cover the vast majority of words in your flashcard deck in probably at least the same amount of time or earlier than they were scheduled for revision - especially for reasonably well known cards. This means that as long as you are doing regular reading, you can probably discard most of the cards in your deck without any real loss of knowledge.

At 5,000 words and 2,500 characters you should definitely be looking to branch into native material.

Personally I wouldn't remove the production cards. That's where I find the real learning occurs (or rather the learning that makes me really grasp the word). If it's taking too long I would say the problem is more too many cards and not enough supplementary activity that allows you to reduce the number of cards without any negative impact.

Don't lose sight of the forest for the trees...

This.

[Ludens]

Imron, I get that you are promoting the deletion of cards instead of building an ever growing deck. You might have a point there, but I'm not ready for that yet psychologically :-)

Out of curiosity, how do you decide a card is ready for removal? Especially with production cards, which you seem to use, it would be hard to know when you can do without the artificial repetition of srs for more obscure words (which you don't 'produce´ regularly in daily life).

Deleting my production (vocab) cards will hopefully free up a lot of my time, so that more can be spent on actual production (and consumption) of the language. If I feel the need, I can then always use your method of temporary production cards alongside my recognition deck.

[imron]

I get that you are promoting the deletion of cards instead of building an ever growing deck

It's not so much that I promote the deletion of cards, but rather I want to promote regular reading and usage of Chinese as having a stronger focus than SRS reps. Regularly deleting cards is just a means to accomplish that end.

Out of curiosity, how do you decide a card is ready for removal?

That's easy, I simply delete everything once revisions start to regularly get over 1/2 an hour an day

The logic is as follows: If I know the card well, deleting it will have no negative effect. If I don't know the card well but the card is useful for me, it will show up again soon enough in my reading and I can decide to add it back to my current list if need be. If I don't know the card well, but the card isn't useful for me, then there's no point me spending my time learning it and it can be safely deleted. 'Useful for me' is defined as likely to be encountered during my usage of Chinese.

For production cards, I do them because I find them a useful way to ensure the word is firmly entrenched in my mind, not necessarily because I'm going to actually produce them myself. More often than not I will only encounter them in a recognition context during my usage of Chinese. I've found that learning them as production makes this recognition faster, because I know the word more thoroughly.

People always worry about obscure cards, and if they delete them, they may never see them again and they'll be losing that knowledge unless they keep it using SRS and so on. To me that just indicates someone who doesn't do enough reading, because something I've found through regular reading is that even obscure words aren't really all that obscure. For example, I try to read about 1/2 an hour a day. For me this currently equates to about one book a month. The average book in China contains about 250,000-300,000 characters, so a book a month equates to over 3 million characters a year. Plug whatever frequency statistics you like into that figure, and even really infrequent words/characters are going to show up a couple of times, and I'm always surprised at how words I thought I'd never see again turn up in completely different books written in completely different styles by completely different authors.

For example, have a think about how many characters out of the 2,600 you know are ones you would consider obscure. Plugging in the Junda frequency data for imaginative texts, at 3,000,000 characters a year, the 2,600th most frequent character is likely to show up 63 times, or about 5 times a month. That's almost certainly more than you would see just doing SRS reps for that character, and that repetition helps keep it fresh in your memory.

Even if your reading speed is much slower, say 100 cpm, if you read regularly (1/2 an hour a day, everyday) it should still be possible to cover over 1 million characters total in a year. (Note: that's 1/2 an hour spent reading, so time taken to look up words should be outside this figure). At that rate you'll still see the 2,600th most frequent character 21 times year, or just under twice a month. Your reading speed will also certainly increase over that time if you are doing it regularly.

At 5,000 words and 2,600 characters, you might still find novels require a lot of dictionary lookups, so comics, and newspaper articles might be a better place to start, although whatever content you choose you'll still require a fair amount of dictionary lookups in the beginning to account for content/genre specific vocab so don't let that put you off. Simpler novels like 《活着》might also be suitable.

The other thing to make sure of, is to make sure you limit the amount of new words you are adding per day to a reasonable amount (5-10 works well for me), even if you encounter many more new words than that during your reading. This helps keep the burden of learning new words low, and as long as you are reading enough, you'll encounter the useful ones you leave out on one day soon enough on another day.

[realmayo]

I started this thread a long time ago but now that I've got some time to spend studying Chinese again, I've started flashcarding again too. But unlike previous decks, for this one, each item generates three cards - I now go three ways - testing me on:

 

1. Chinese characters -> Pinyin + English

2. Pinyin -> English

3. English -> Pinyin

 

(All three parts are shown in each answer).

 

I added numbers 2 and 3 because I worry that if I just test myself on Chinese characters, I'm often not remembering a two-character word, but am instead deciphering two characters in order to (re)construct the two-character word. And I want to focus on word recognition, not word reconstruction.
 

Other benefits for numbers 2 and 3 that I feel exist but may not:

 

- Forcing myself to produce the Chinese from the English seems harder but also stickier (people discussed "production" a lot earlier in this topic).

 

- More fluent reading. Do we read a bit with our ears? If I see 职位 but do not *instantly* recall the meaning, it seems that the next step my brain takes is to sound out the 职 and then the 位. And with zhíwèi suddenly in my ears, the meaning immediately comes with it. Is that faster recall than the alternative of isolating the meaning of 职 as "profession-ish" and then 位 as "place-ish"? It feels that it's faster. And it might not be an either/or, but instead, perhaps I'm doing both at the same time, sound and meaning? In which case I'm likely to recall more quickly.

 

- Better automation of tones. If I see 职 in a brand new context, I'll know it's a 2nd tone because I'll know it's the first part of zhíwèi. Perhaps I'll remember the meaning of 职 better too - it's the "job" part of "position" in zhíwèi.

 

- Modern Chinese is mainly a bisyllabic language and focusing on bisyllabic words as bisyllabic words feels more robust: zhíwèi isn't a word with a rising tone character and then a falling tone character, instead, it's an "rising then falling" word.

 

- During the initial rote-memorisation phase, I'm attacking the same word from three different angles, often only minutes apart, which has got to help with the learning.

 

 

Downside: more work! It's harder and less pleasant to produce the Chinese word from the English equivalent. Flashcard decks can look intimidatingly large very quickly. Need to put in some clunky hints (e.g. "not chouchu but the other one"). Same for homonyms when testing from pinyin.

 

But: the goal isn't to get good at doing flashcards, the goal is to remember words. And because I automatically suspend indefintely any card I get wrong twice (outside of an extending 'learning' period), I figure that the time spent trying to learn the #2 or #3 side of the card will have helped me me with the #1 side, even if for some words their #2 or #3 cards prove too tricksy to keep getting right.

 

 

[imron]

5 hours ago, realmayo said:

Better automation of tones. If I see 职 in a brand new context, I'll know it's a 2nd tone because I'll know it's the first part of zhíwèi. Perhaps I'll remember the meaning of 职 better too - it's the "job" part of "position" in zhíwèi.

I try to make sure that I only ever create flash cards of 2 or more characters. When I come across a single character that I don’t know I search the dictionary for a 2 character word that means the same thing (or a close approximation of the main meaning in the context where I encountered it).

 

This is partly for the reason you mentioned here, and also because it significantly reduces homonyms. 

[Sreeni]

On 4/5/2021 at 10:48 PM, realmayo said:

each item generates three cards - I now go three ways - testing me on:

 

1. Chinese characters -> Pinyin + English

2. Pinyin -> English

3. English -> Pinyin

 

Are you using Anki? how are you generating 3 cards? I am using Anki free app.

 

My Anki one side just 职责, other side as below: If i can recollect, swap to second side and go to next card faster,

If i can not recollect spend time to connect it again. some times it works for 1 time and some times even 5.....

 

and if any character starting with 职 i come across, i will update the same card:

 

Anki card hidden side:

职责 zhízé duty, responsibility, obligation

 

 

职 zhí office, duty; (orig.) “one who hears and remembers; to manage.”

只 zhǐ only, just, merely, (orig.) to cry

责 zé duty, responsibility; (orig.) to tax someone

 

On 4/5/2021 at 10:48 PM, realmayo said:

And because I automatically suspend indefintely any card I get wrong twice (outside of an extending 'learning' period), I figure that the time spent trying to learn the #2 or #3 side of the card will have helped me me with the #1 side, even if for some words their #2 or #3 cards prove too tricksy to keep getting right.

 

automatically suspend, if twice wrong? then how can you learn that word?

[realmayo]

6 hours ago, Sreeni said:

how are you generating 3 cards?

 

Anki (on a PC): Tools -> Manage Note Types -> (select note type) -> Cards... -> Options (the top one) -> Add Card Type

 

6 hours ago, Sreeni said:

then how can you learn that word?

 

Anki automatically decides if cards are "learning", "young" (the next phase), and then "mature". If a card is still in the "learning" phase, getting it wrong is not a problem - it doesn't count as a "lapse". So I change the settings to give myself a longer "learning" period, during which I hope to learn the card.

 

Only after it moves out of this phase and into the "young" or "mature" phase, does getting it wrong count as a lapse. And after two such lapses, the card is suspended.

 

Edit:

@Sreeni I started learning Chinese by talking to people all the time, and with a bit of pinyin. I didn't start learning characters until I could already have quite long conversations in (not very good) Chinese. From some of your posts, it seems like you're focussing almost all your energy on written words/characters first. Maybe I'm wrong, but that seems like a big mistake. You need to be able to use the words that you're learning: if you learn the meaning of a word but your Chinese isn't good enough to use that word in several natural sentences, then I think you've wasted your time learning that word, you should have used that time to learn more grammar etc. Perhaps you're doing that as well, of course - I don't know much about how you're studying. But please don't think that by memorising words and characters you'll be able to understand Chinese. Very sorry if I've misunderstood your approach. Either way, what I wrote in the earlier post in this thread is unlikely to apply to beginners so I don't think it will be a useful approach for you.

[Jan Finster]

2 hours ago, realmayo said:

From some of your posts, it seems like you're focussing almost all your energy on written words/characters first. Maybe I'm wrong, but that seems like a big mistake. You need to be able to use the words that you're learning: if you learn the meaning of a word but your Chinese isn't good enough to use that word in several natural sentences, then I think you've wasted your time learning that word, you should have used that time to learn more grammar etc. Perhaps you're doing that as well, of course - I don't know much about how you're studying. But please don't think that by memorising words and characters you'll be able to understand Chinese. Very sorry if I've misunderstood your approach. Either way, what I wrote in the earlier post in this thread is unlikely to apply to beginners so I don't think it will be a useful approach for you.

This is super debatable (to say the least). Quite a lot of language expert recommend focusing on input first or at least foremost. Also, when you look at some older posts, some of the better known names on this forum started learning 3000 Hanzi before anything else...

[realmayo]

22 minutes ago, Jan Finster said:

This is super debatable

Good point - I should have added "or understand", so we get: "but your Chinese isn't good enough to use or understand that word in several natural sentences".

Then, I think, the statement is unlikely to be debatable.

[Sreeni]

6 hours ago, realmayo said:

From some of your posts, it seems like you're focussing almost all your energy on written words/characters first. Maybe I'm wrong, but that seems like a big mistake. You need to be able to use the words that you're learning: if you learn the meaning of a word but your Chinese isn't good enough to use that word in several natural sentences, then I think you've wasted your time learning that word, you should have used that time to learn more grammar etc.

 

@realmayo my source of the vocabulary is text book and not from random lists and I am re-reading the textbooks as well, so taking care of using them. even checking in PLC dictionary on how they are using these words in different contexts/sentences.  but working on retaining all the meanings..My Anki shows 78%. trying to reach 98% words retention.

[realmayo]

5 hours ago, Jan Finster said:

some of the better known names on this forum started learning 3000 Hanzi before anything else

trying to think who?

[Sreeni]

13 hours ago, realmayo said:

Anki (on a PC): Tools -> Manage Note Types -> (select note type) -> Cards... -> Options (the top one) -> Add Card Type

 

I never used Anki on desktop. I had created my cards in Anki mobile and reviewing. I want to merge 2 sets of cards, can I do it on my mobile itself?

 

In mobile, if it is not possible, If I login to PC, can I merge both sets to one ? 

[imron]

Use paid Pleco. It won’t be as flexible as anki, but it’ll get you 90 of what you need and you’ll never have spend any time managing and massaging your decks (it can also generate this sort of reverse card stuff automatically)

[imron]

Use paid Pleco. It won’t be as flexible as anki, but it’ll get you 90 of what you need and you’ll never have spend any time managing and massaging your decks (it can also generate this sort of reverse card stuff automatically)

[Sreeni]

On 4/16/2021 at 8:31 AM, imron said:

It won’t be as flexible as anki,

 

get used to Anki. I tried on Android phone to create a new words list, creation is simple with just + and sound is added automatic. as you mentioned it might add reverse type. next week chinese semester exams. So let me focus more on remembering technique and get ~100% in Anki this weekend (not planning to add more for this semester except very few). Thanks for valuable advise