Different Methods for Using SRS Sentence Flashcards

[Yadang]

This post is in an attempt to find out different ways one might use a flashcard system with sentences.

 

I’ve searched the forums a bit, and have found references to a few different methods, but nothing very specific or comprehensive.

 

Firstly, I think the first thing that may be said is that how one uses sentence flashcards will depend on what one’s purpose is in using such flashcards. Is it character recognition, active character recall (for just one word in the sentence, or the whole sentence), passive understanding of the sentence and its structure, active recollection of the sentence (perhaps so that one has it so ingrained in their head, it can serve as a foundation for similar sentences), or perhaps just used for either shadowing or chorusing, to get more native-like pronunciation.

 

What should also be considered is how much information one has. Is it just a translation, pinyin and characters - or do you also have the audio, or a movie screenshot or even movie clip?

 

 

Character Recognition

 

If this is the goal, than you could have the whole sentence on the front in characters (if you wanted to recognize the whole thing), or the whole sentence in pinyin, except for the certain vocabulary words you wanted to learn to recognize, which would be in characters. On the back, you might have the same sentence, the audio of the sentence (if you have it) and maybe an English translation of the sentence? However some people dislike translations... How would you solve this problem? Could you just not have a translation, and go for a general understanding (and if you felt like you didn’t understand it well enough, you could look up the constituent words of the sentence)? The problem with just looking up the constituent words is that it’s not always enough to know the grammar structure and all of the words... You could use a Chinese to Chinese dictionary to not run into the English problems, but this might pose the same problems of knowing all the words but not the sentence.

 

 

Active Character Recall

 

If character recall is the goal, one could have pinyin on the front, or, if they don’t mind the whole translation thing, English on the front, and the sentence in characters and pinyin on the back. If they were just testing some of the words, instead of the whole sentence, they could do the sentence in characters on the front with cloze deletion of the words they want to actively recall. On the back would be the complete sentence. If this was the case, they might not even need any English translations of the word itself, because context would be enough for them to start getting a feel for the word, but the same problems still arise of how to know what the sentence means as a whole.

 

 

Passive Understanding of the Sentence and its Structure

 

Not really sure how this would work... Maybe you’d have the characters on the front; on the back the characters with pinyin. If you understood the general gist of the sentence and get the pronunciation right, you'd move on.

 

 

Active Recollection of the Sentence

 

This could be only recalling it by saying/thinking it, or writing it down in characters. I think the main strength of this would be that you could start to internalize sentences and sentence structures, and so similar sentences would come more and more naturally with each sentence you internalized. The problem is how the test would be set up for this. If you tested by having English on the front, there again is the problem of translation. Pinyin on the front wouldn’t really allow you to internalize the sentence - only to write the characters of the sentence. Audio I think has the same problem - it turns it into passive studying instead of active recall.

 

But what if you were using a movie as a source? Would it be plausible, say, using subs2srs, to have just a screenshot on the front, and everything else (with or without an English translation) on the back? I feel like this gives the least amount of information, and demands the most.

 

But is it plausible to recall many many sentences in a movie from just seeing screenshots of where that sentence took place?

 

 

For Shadowing or Chorusing

 

This I think could be done a few different ways, just as long as on the back you have some audio which you either shadow or chorus. This approach could also be incorporated into the total active recollection method above, where after the person recalls the sentence from just a snapshot, he uses audio on the back and shadows/choruses it.

 

All of this is highly speculative by me, because I haven’t done any sentence based flashcard testing. I’m wondering if I’ve missed any different methods or goals, or if any of these methods could be improved upon.

 

 

As for my personal goals, I want to internalize whole sentences, to create a foundation or core of sentences with different grammatical patterns that I can use as a template to base sentences I make up in daily speech off of. I want to basically achieve what OneEye did in this post, where he describes using the movie 那些年 and striping the audio away form the movie, then chorusing and internalizing many of the sentences, leading to much better understanding of native speech and much closer to native-sounding pronunciation (he does not I don't think use subs2srs however).

 

I think for this goal, I should see if I could use a screenshot of the movie for the front, and then do active recall of the sentence which would be in characters and pinyin on the back, with the audio, with which I would chorus. Any words that I didn’t understand, I would look up, and maybe make separate flashcards for them. As for English - I think I would have the English information, but I wouldn’t make it part of the card (in Anki terms, it would be on the "note" but not on the "card"). Therefore, if I really didn’t get a sentence, I could go to the browser and see what the English subtitle translation was. However, I have watched the movie once in Chinese with Chinese subtitles and once in Chinese with English subtitles. I think this is good, because while I have a pretty good idea of what to expect them to say, I didn’t memorize any of the English, because I only watched it once in English. So, I hope I won’t have to refer to the english subtitles very much...

 

Do you think this is the best way to accomplish the goal of internalizing natural sentences for the use of building a foundation of sentences with a variety of different sentence structures upon which to base similar sentences? Or could it be done more efficiently in a different way?

 

Thanks for all of your ideas and input!

 

[JenniferW]

I have one ANKI deck which I set up for reading practice - so I handle strings of characters rather than reading character by character.  I take sentences from the current lesson in my coursebook, ones which have new vocab in, or are good examples of the grammar for that lesson.  Other times I take example sentences from my grammar book, sometimes from jukuu.com. 

 

So, my 'flashcard' presents me with a sentence in characters.  Sometimes, I read it aloud, sometimes I read it silently.  Then I look at the 'reverse', which gives me the sentence in pinyin.  If I need it, it also gives me a translation - but I only give myself that for a minority of sentences.  If there's a problem over a grammar or vocab item, I also add in a note about it, to remind myself.  So, while I basically get reading practice, I also get reminders and practice of other things. 

 

But the reading's the priority for me.  I was reading at the stage of going character by character, and this helped me move up a level, to processing chunks of a sentence at a time, and to start skimming and scanning.

[tysond]

I use sentences and paragraphs all the time.  My purpose is to learn the pronunciation, writing and meaning of words/characters, in context.

I use MCD technique for this -- Massive Context Deletions (or, in the case of a single sentence, Micro Context Deletions).

 

At the moment I do them bilingual style.  So on the front of the Anki card, I have:

• Hanzi sentence/passage with one character clozed.
• English translation (doesn't have to be exact in any way, it's just context, you are not learning to translate, you are learning to fill in the gaps of a Chinese sentence using English to give a hint)
• Image for context if available (sometimes I use screenshots via subs2srs, sometimes I use images from the internet, sometimes I use the English scanned page of a comic).

On the back of the card:

• Hanzi sentence/passage with one character un-clozed.
• English translation (doesn't have to be exact in any way, it's just context)
• Pinyin
• Audio if available (from subs2srs, recorded native speaker, Chinesepod mp3, or Text-To-Speech sometimes... it's not great)
• Heisig keywords for every character (from an Excel formula I wrote that looks them all up and concatenates them into a single line)
• Definitions for anything HSK4 and above (copy pasted from MDBG Chinese Reader app on Windows)
• Any notes on usage I feel are useful to remember, characters not in Heisig, etc.

Then I rate myself on how well I remember the character - writing + pronunciation in Anki.  Sometimes I shadow the whole sentence as well for good measure, especially if it's a bit longer. 

 

A typical sentence might have between 4 - 10 clozes, which come up as different cards, so I get plenty of repetition of the sentences.

I have around 600 base cards active, generating 2000 cloze cards right now.  I choose clozes based on what is not obvious about the sentences - it could be a new vocab item, but equally it could be a 来 or 下 or something really simple which co-locates... like 留下来.  Or the 掉 in 关掉闹钟.

 

What is useful about this?

• Learn grammar and word usage in context - you begin to get a feel for what goes where without really thinking.  Sometimes my sentences finish themselves without me thinking because they fall into a pattern that I've done many times before.
• Read lots of sentences quicker and quicker - many times I don't even look at the English portion so it's all Chinese reading.
• Learn all the small but important words and how they are used 又， 再， 跟， 对， 了， 的/地/得， etc.  
• Learn to write everything you know
• The audio starts to repeat in your head after a while so you just "know" the pronunciation without even thinking of the pinyin
• You can drill into very complex material ... e.g. I have learned the first paragraph of Obama's wikipedia entry despite it being full of political and ethnic jargon.
• You can go completely monolingual after a while if you include enough context or C-C definitions

However, it has some drawbacks:

• Slower to make cards (subs2srs is a good kickstart.  I also paid someone to make cards for me)
• Slower to do reps than pure recognition of hanzi -> english (time spend on this could be spend on having a huge vocabulary of nouns, verbs, etc which would help listening enormously).  
• My insistence on writing also slows it down (however I could just accept a correct pronunciation, wrong character to go much faster)
• Few pre-made resources to draw upon, unlike vocabulary lists

As part of your overall learning strategy I think it's valuable and I'm planning to keep using it.

[abcdefg]

But the reading's the priority for me.  I was reading at the stage of going character by character, and this helped me move up a level, to processing chunks of a sentence at a time, and to start skimming and scanning.

That looks like it might be useful for me, JenniferW, since I'm at a similar stage. Thanks. 

 

Read lots of sentences quicker and quicker

The "Tysond method" looks promising too.

[Yadang]

Thanks Jennifer and tysond, both of these approaches look really good - I'm looking forward to trying them out! Also - thanks for giving detailed explanations!

[roddy]

Had a good read of this topic, as I was trying to figure out the best way to use a deck of Spanish sentences + audio - one of the Tatoeba decks, I think there are Chinese equivalents around. 

 

It's set up as cloze cards, which I wasn't too keen on initially. I'd rather be doing what @Yadang described as "Active Recollection of the Sentence", rather than recalling one word with the heavy translation and context hints. So I spent hours (ok, it felt like hours, and also had to reposition by difficulty) figuring out how to make new English: Spanish+Audio cards. Which I managed to do, but the process didn't quite work. Is "There you are" as in "I've been looking for you" or as in "Here is the thing you wanted", etc? You could probably remember how it's meant to be translated after a few showings, but I'm not sure that's what I'm trying to remember. It would be possible to go through the deck and make edits so the English sentences are closer to the Spanish, or add notes where necessary to point in the right direction.

 

I think I'm going to give the cloze cards another shot, but make a point of shadowing the Spanish audio for the structure / pronunciation benefits. Will see how that goes. Other ideas welcome. I like @tysond's method of having multiple cloze cards per sentence, but not sure I want to spend the time on that.