0

What is the basis for kanji frequency? For example, 悠{ゆう} has the Frequency of 1921 in the kanji dictionary of Yomichan. What does it mean? Does it mean that in everyday conversation or in newspapers published every day, 悠{ゆう} is being used 1921 times? Also, is the frequency of a kanji the same in all the dictionaries?

2

Your post contains multiple questions. Let me try to answer each of them.

What is the basis for kanji frequency?

As I've understood them, these lists are based on statistical analyses of various corpora -- sets of documents, often newspapers, magazines, websites, and so forth. The basic method is to count the occurrences of each kanji, and then rank them by frequency. Frithany commented with a link to one such frequency list.

For example, 悠{ゆう} has the Frequency of 1921 in the kanji dictionary of Yomichan. What does it mean?

A frequency ranking of 1921 in such a list simply means that that kanji is used less frequently than the preceding 1920 kanji, and more frequently than any kanji coming later in the list.

Does it mean that in everyday conversation or in newspapers published every day, 悠{ゆう}is being used 1921 times?

No. A frequency of 1921 does not mean that that kanji is used 1921 times. See the above comment.

Also, is the frequency of a kanji the same in all the dictionaries?

Frequency lists are quite variable, and are only kinda-sorta useful for getting a ballpark idea of how common each kanji is.

The reason for this variability is that each frequency list is based on a different corpus -- a different set of documents. For instance, frequency lists are often based in part on trends in newspaper usage, and that changes from year to year: not only do writing styles change, but the content of the articles also changes, and even the kanji used to spell people's names change over time as different spellings become popular or fall out of use. (Note: the spelling for a specific person's name probably doesn't change over time. However, there are many different ways to spell the given name "Megumi", for example, and parents will choose different spellings depending on all kinds of factors, leading to different spellings becoming more common or unusual as time passes.)

1

A frequency number is an indicator of how often a particular character/word occurs in a specific pool of words/characters being analyzed, usually a corpus of some kind. The frequency number of any specific item can vary significantly according to which corpus you use as a base. For example, if you used a corpus of newspaper articles to analyze, the word めっちゃ would have a pretty low frequency, since it is informal and unlikely to be used often in the formal style that newspapers use. On the other hand, if you used social media interactions as a corpus to analyze, you would see that めっちゃ would have a much higher frequency as it used far more often in informal language. So a frequency number can vary a lot for the exact same word, depending on what source you are analyzing.

Frequency is generally given a value related to its Frequency-Per-Million-Words (fpmw). Having calculated a fpmw value for each word in a corpus, you can then rank them in order of most frequent (high fpmw value) to least frequent (low fpmw value). In your example, the character listed as 1921 indicates its rank relative to other characters, not the total number of occurrences in the corpus. Specifically, it means that there are 1920 other words that have a higher fpmw value than 悠. Any word with a frequency of 1922 (in your example) or above means that those words have a lower fpmw than 悠.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.