It could be said that once you reach a certain threshold the number of words you know isn't as important as the relevance of media to your given level and that you'd be able to use your most preferred method of learning for your motives -- improving listening, reading, writing, or speaking; although it is somewhat difficult to adapt anki to all four, it is possible.
If you've gone through 2/4/6k you may watch an episode or read a novel near your level and simply add words as they come across and a sentence or two. This overemphasis on all encompassing mechanical i+1 corpus and individual sentences tends to overshadow an essential element of study, comprehensible media and not necessarily i+1 words as they don't really exist but words that seem easy to learn because you've seen their parts before and have gripping context. [ie. using Kanji Word Association Tool and parsing for unique words using Mighty morphin and Japanese Text Analysis Tool]
The threshold for simply inputting words and sentences is whenever you feel a word seems important and the context pulls you in, even if just a bit. The primary difference between the effectiveness of vanilla mined words and a pre-made corpus is a sense of novelty because the word came from something you care about. It could also be said that manually selecting from an interesting source makes the learning much easier because it's enjoyable. Find enjoyable content, input simple words, and learn quickly.
But again when you reach that threshold it might be best to manually select interesting media and simply input a word, its definition and a sentence or two; adding a word production model and scrambled gloss if needed or otherwise inputting significant grammar functions into a cloze deletion.
The key to i+1 is not to make learning [x] thousand words easily; but to create a network of knowledge which interrelate through contextual clues and deep word associations. To learn from drab sources with repetitive sentence structures sort of reduces the quality of the networked words inputted into your personal database of terms.
It could be said that even the mechanical process of creating word lists removes the novelty of finding new word associations from the source. Although less efficient, so to speak, at times it's more enjoyable to input words that seem right for the personal corpus and skip over words from sources that won't fit your current level. If too many words are skipped over the content is possibly too difficult and it might be preferable to find easier content or use the methods provided by users of this forum to parse the text.
That's not to say Core 5k or Core 10k isn't a good source; but Core should be a means to an end -- to get you to a base level that lets you learn from original and interesting content.
If you've gone through 2/4/6k you may watch an episode or read a novel near your level and simply add words as they come across and a sentence or two. This overemphasis on all encompassing mechanical i+1 corpus and individual sentences tends to overshadow an essential element of study, comprehensible media and not necessarily i+1 words as they don't really exist but words that seem easy to learn because you've seen their parts before and have gripping context. [ie. using Kanji Word Association Tool and parsing for unique words using Mighty morphin and Japanese Text Analysis Tool]
The threshold for simply inputting words and sentences is whenever you feel a word seems important and the context pulls you in, even if just a bit. The primary difference between the effectiveness of vanilla mined words and a pre-made corpus is a sense of novelty because the word came from something you care about. It could also be said that manually selecting from an interesting source makes the learning much easier because it's enjoyable. Find enjoyable content, input simple words, and learn quickly.
But again when you reach that threshold it might be best to manually select interesting media and simply input a word, its definition and a sentence or two; adding a word production model and scrambled gloss if needed or otherwise inputting significant grammar functions into a cloze deletion.
The key to i+1 is not to make learning [x] thousand words easily; but to create a network of knowledge which interrelate through contextual clues and deep word associations. To learn from drab sources with repetitive sentence structures sort of reduces the quality of the networked words inputted into your personal database of terms.
It could be said that even the mechanical process of creating word lists removes the novelty of finding new word associations from the source. Although less efficient, so to speak, at times it's more enjoyable to input words that seem right for the personal corpus and skip over words from sources that won't fit your current level. If too many words are skipped over the content is possibly too difficult and it might be preferable to find easier content or use the methods provided by users of this forum to parse the text.
That's not to say Core 5k or Core 10k isn't a good source; but Core should be a means to an end -- to get you to a base level that lets you learn from original and interesting content.
Edited: 2016-05-10, 11:01 am