The SAK input method for looking up kanji

Quick—how would you look up the kanji 毓? Sure, you could cut and paste into your web browser or online dictionary, but what if it was from handwritten materials, or a PDF, or something you saw on a sign or on TV? And good luck with Google Translate’s handwriting recognition.

In days of yore, every student of Japanese worthy of the name had as their constant companion their dog-eared copy of Nelson, the venerable Japanese-English character dictionary (漢和辞典), that we all learned how to find characters in based on their radicals—not at all obvious, in many cases—and stroke count. (For example, to find 毓 in Nelson, you’d have to know that the radical is 母/mother, which is not necessarily self-evident.)

So what’s the equivalent to Nelson in 2022? Enter the SAK input method, where you simply enter the parts of the kanji into your favorite search engine. This turns out to work remarkably, even amazingly, well. In this case we’ll just enter 毎 and 充. Yes, I know the 充 is not exactly right, but it’s close enough. To help the search engine focus on kanji, we also input 漢字. So:

毎　充　漢字

and there’s the answer in all its glory, the very first entry on the results page:

It turns out the meaning of this character has something to do with giving birth to and/or raising a child (the right top part is supposed to be a 子/child turned on its side), but it’s extremely rare in Japanese and so it’s not really worth learning it. In this case it was the name of a person from Taiwan.

嗜

Let’s try another one. How about 嗜? This is easy as pie using the SAK input method. Just enter

口 老 日

and voilà, we get

This one is probably worth learning if you don’t know it. It’s たしなむ/tashinamu, meaning to have a taste for, also found in 嗜好 (shikō/liking).

蔓

Finally, let’s try another one: the 蔓 of 蔓延—you know like 蔓延防止, preventing the spread (of COVID). You knew that, right? This is going to be

草 日 四 又

And just like that up pops a post from chiebukuro:

The kun reading of this character, by the way, is はびこる/habikoru, which you can use to describe all the stuff spreading its tendrils (hence the connection with grass) all over your desk. By the way, instead of 草, you might have better luck with the actual grass radical 艸 if you can figure out how to enter that in your IME.

How does this work?

But how does this actually work anyway? Search engines have no idea of the parts of kanji, do they? No, of course they do not, but there are hundreds or thousands of pieces of user-generated content discussing the characters and their parts, often talking about those parts using stand-alone kanji. And the rest is just the magic of search engines.