Core 2.3k Anki Deck - Version 3 (Core2.3k) - Core2k/6k with improvements!

Core 2.3k Anki Deck - Version 3 (Core2.3k) - Core2k/6k with improvements! #

High Quality Cards for the Most Common ~1,970 Words from Visual Novels, Light Novels, Newspapers, Magazines, Blogs, Internet Forums and more #

How To Use (IMPORTANT) #

Install and setup Anki following these settings:

Unless you are actively reading Japanese media appropriate for your level, there is virtually no benefit to doing a core deck. So be sure to skim through a grammar guide and start reading something easy like Yotsubato, Takagi-san, Pokemon, or anything else that interests you.

When reviewing cards only pass a card if you can remember the reading of the word. You should also try to recall the general meaning of the word but don’t grade yourself harshly on it, as meaning is better acquired from reading native material (seeing the word used in many different contexts) and not from seeing dictionary definitions.

Do not worry about the example sentences, just learn the words. The sentences are there to help you learn the word but if you can’t read them then please ignore them. The Japanese vocabulary Core 2k anki deck contains images that are in this deck, cards without images were not in Core 2k.


What is this? #

If you’re just starting to learn Japanese it may be a little overwhelming to jump straight into reading native content. Learning a list of common words can ease you into reading more enjoyably. This is an Anki deck based on iKnow’s Core 6k deck with the following improvements:

  • All duplicates have been either removed (if they were not needed) or reworked to not be ambiguous using the hint field.
  • Words manually identified as redundant given other words in the deck have been removed.
  • Katakana loan words have been removed.
  • Card order is now sorted by average frequency of vn, narou, and bccwj (all frequencies are normalized in the same way as my new freq dicts)
  • Removed any word whose freq is greater than or equal to 89 in ANY of the freq dicts (if a particular freq dict doesn’t have it then that’s ok, as long as the ones that do say under 89)
  • If for some reason vn and narou don’t recognize the word then use mecab to parse it into component parts, treat least frequent part as the frequency of the whole (also if still not found in all freq dicts by this point then remove it)
  • Every kana word comes with a context snippet on the hint field to make it easier. The reason this is done for kana words and not for kanji words is because kanji carry meaning (which you should become familiar with) within themselves. Kana on the other hand don’t. The result is that reviewing kanji words context free is both viable and recommended (no reliance on specific contexts, faster reviews) while kana words are better learned using a context as assistance.
  • The card styling CSS has been tidied up and the default fonts now actually support Japanese. The previous deck caused a lot of clueless people to learn Chinese variants as the default font was ‘arial’ which has no Japanese defined for it causing your system to default to something else (i.e. often Chinese)
  • Images are already included meaning you don’t have to separately download them and put them inside your media folder.

The result is a deck made from the highest quality cards, in the most efficient order possible, and as much extra fluff/non-core words removed as possible. Also, despite keeping the same name for historical reasons, the deck is actually only ~1,970 cards in total.


Example core card