Ok, so i've been doing the setence method for a couple of weeks now and I'm about up to 500 (Not including 3/4 of tae kim). All is going very well. My ability to read is rapidly increasing alongside my vocab and comprehension is coming along nicely. However... seeing sentences a few times only isn't really enough to internalize the patterns you're seeing (which I'm guessing is why like 10,000 approx is needed?). So I'm trailing a new addition to my reviewing to see if I can help this along.

A thought occurred to me that If I hear something in English, I can repeat it no problem. Even expand and elaborate on it if I wish. However, In Japanese, the only things I can repeat after hearing are things I know very very well. Things that I could probably output. For example, I'd have no problem telling someone in a very rude tone of voice that I was going to kill them. Haha.

Anyway, I was doing reviews and comprehending things fine when I decided to close my eyes and without looking at the sentence, after having read it once, repeat it aloud correctly. I noticed myself making one or two mistakes. I found this interesting and trialed it out with a few more cards and again, a mistake here and there and some cards completely correct.

I would assume doing this for every sentence would seriously help to internalize the given sentence structure you are emulating. It's like output with a safety net. You attempt to output something you understand and if you get it wrong, you know instantly, the answer is right in front of you. You try again and this time get it right, thus strengthening the neural pathway for the output of the current sentence.

Some may just call this "parroting" or other such things but given I can get wrong what i've JUST read, it goes to show I'm not nearly ready to output it on the fly! I would rather like to think of it as "Output Training Wheels"

I do this as well, for sentences I know I have badly internalized. It makes sure you're simply not reading the sentence kana per kana or kanji per kanji, but really analyze and parse the sentence. There's no way you can "re-state" a sentence you haven't understood decently.

This is especially useful for sentences where grammar is trained.

Definitely a good thing to implement into reviews. I've done it a few times without really thinking, but actively repeating what was read without looking for the whole review is something I think I'll try, especially since you've had good results. Thanks for letting us know about it

Do you do this only with text sentences, or also with audio ones? It occurs to me that doing audio ones as well would train a different bit of your brain as well.

Im not doing audio cards yet cos I dont have the matierials. More like the bandwidth. I notice when I read it back after I check if I'm correct things start to get a lot smoother and more natural sounding. Makes reviewing take longer though. I'm definately going to keep it up.

Any other suggestions?

mezbup Wrote:Im not doing audio cards yet cos I dont have the matierials. More like the bandwidth. I notice when I read it back after I check if I'm correct things start to get a lot smoother and more natural sounding. Makes reviewing take longer though. I'm definately going to keep it up.

Any other suggestions?

Have you checked out the iKnow Kanji Odyssey lists made by the incredibly helpful members of this site? You can get a few hundred sentences with audio into Anki for a couple of MB and I'm sure that is managable even on dial up. http://forum.koohii.com/showthread.php?tid=2419

Thanks for the tip, since I started doing audio sentences I have been making sure I can say an entire sentence and understand all the words from the audio before passing it. However, I hadent thought of doing it for written sentences. I think I will give it a go, it sounds like a good idea that is likely to embed things better into memory without too much extra effort.

Something else you might want to experiment with that I do out of circumstance/habit is repeat sentences from cards as/after I grade them as pass or fail, so that as audio is playing from the previous card or as I'm calling it to mind, I'm also reading or listening to the next card. (The way it usually works is that I'm listening to the previous card while reading the next card, or visualizing the text of the previous card while listening to the next, and variations and combinations of them all.) Good practice for comprehension skills, because part of getting better at a language is learning to process information more efficiently so you can multitask, as it were, and skim through information w/o worrying about 'interference' distracting you. (See 'verbal interference' and 'articulatory rehearsal' studies.) It also keeps me from 'restudying' cards too much rather than promptly failing them when I make mistakes, as I'm quite OCD and tend to make judicious use of ctrl-z to go back and double-check, so I'm always in these transitory periods of overlap.

Lastly, if you're doing this for overall output practice instead of compartmentalizing and focusing on particles or grammar and relying on dictation/muscle memory/etc to reinforce nuances, probably best to focus on choice phrases that resonate instead of every card, because that could be just as exhausting and time consuming as memorizing every sentence (and increase temptation to actually memorize entire sentences).

It could be fun to try taking choice mature cards and remixing them into bits of output, then parlay those into microblog entries. That's something I've been meaning to do, but now I'm distracted with my 'imaginary conversations' methods for output practice, plus Nukemarine's recent comments have spun me into thoughts on the immersion environment. Anyway, rambling again.