As the title says. I stumbled on the Reddit post today and became curious if that's true? Here's a quote from that post:

Kanji don't actually have a meaning on their own, they get meaning from the words that they spell. If a kanji has multiple 'meanings' that's because it's used in multiple words.

Generally when learning the kanji via vocabulary (rather than learning vocabulary for its own sake) I would go with the most common word/meaning that I can find, with a slight bias towards picking kun-reading words since those tend to involve only the character I'm learning and not confuse the issue by adding other kanji. (A number of people have the exact opposite tendency, wanting the most common ON reading to prepare for reading kanji compound words, but I figure that will come quickly enough when learning those words).

  • 2
    How do you make new words with kanji if that's the case?
    – aguijonazo
    Commented Dec 5, 2023 at 16:45
  • 2
    Another astounding example of Reddit wasting spectacular amounts of time on nonsense.
    – Fattie
    Commented Dec 6, 2023 at 22:18
  • @aguijonazo Phonetically? There's no reason someone creating a neologism in Japanese can’t come up with a sequence of mōra and then choose characters based on those. In fact, that’s essentially how gairaigo work, though they use katakana instead of kanji. Commented Dec 7, 2023 at 3:13
  • @AustinHemmelgarn - You are not suggesting all recent coinages are phonetic, are you?
    – aguijonazo
    Commented Dec 7, 2023 at 3:32
  • @aguijonazo No, I’m pointing out that a majority of gairraigo are phonetically adapted in their written form (that’s the whole reason they usually use katakana in the written form), and that that general methodology would be possible with kanji as well (which your original comment seemed to imply was not the case). The fact that this is generally not done with kanji except for some cases of ateji where phono-semantic matching is possible is simply a result of kanji having semantic meaning. Commented Dec 7, 2023 at 16:11

5 Answers 5


Just to add to @A.Ellett's fine answer post, I must emphasize that the Reddit poster is off base.

Copying what I just replied to them a moment ago:

Kanji don't actually have a meaning on their own

Where on earth did you get this idea?

Kanji absolutely have a meaning on their own. It might be obscure or abstract, and difficult to define, but the meaning is there.

Kanji are not like Latin-alphabet letters, which are purely abstract signifiers for sound. Kanji are instead graphical units of meaning, broadly speaking.

Kanji originated in Chinese as the written form of Chinese words. Would you say that "words don't actually have meaning on their own"? Some words are obscure or abstract, and difficult to define, but the meaning is there. This is precisely how and why Chinese characters can be used meaningfully across cultures -- because the characters themselves are assigned meanings, and meanings are independent of the spoken language.

(Granted, as Chinese, kanji are not independent of the spoken language -- the character components themselves are broadly morphophonemic. That said, in the context of Chinese characters as used in non-Chinese languages, the morphophonemic nature is less prevalent, sometimes even obscured entirely.)

In terms of the sound-vs.-meaning dichotomy, as A.Ellett pointed out with the 寿司 (sushi) example, written Japanese makes liberal use of kanji purely for their phonetic values. In fact, this is where hiragana and katakana come from -- using the entire kanji (for hiragana) or just a representative part of the kanji (for katakana) to represent the sound of that kanji (as imported centuries ago into Old Japanese from Middle Chinese).

That said, even when using kanji phonetically, writers will choose characters that have meanings that fit their intention. For 寿司 (sushi), the characters are used phonetically, but still have an underlying meaning component that still lends a nuance to this spelling of the word -- literally 寿 ("long life") and 司 ("to carry out, to work towards"), alluding to the idea that eating sushi will bring you a long life -- favorable marketing, at least. 😄 By contrast, we could write the word sushi using the kanji spelling 酢死, literally "vinegar death" -- but that is a terrible connotation, and no one would use this spelling unless intentionally trying to paint that sort of macabre picture for the reader.

I will not go into learning approaches, as that is highly dependent on the individual themselves and how best they absorb information. I will simply reiterate that kanji absolutely do have meaning, and that meaning is still a factor even when a given kanji character is used for its phonetic value.

  • 4
    I might never again think of sushi the same. :-)
    – A.Ellett
    Commented Dec 5, 2023 at 19:18
  • 2
    @A.Ellett — 🤣 Oh dear, I hope I haven't ruined things for you! Commented Dec 5, 2023 at 20:25
  • Has the original Reddit OP changed your perspective after you posted this SE answer? You seem to partially agree with him: " I think I understand much better where you're coming from" Commented Dec 8, 2023 at 18:57
  • 1
    Is it kind of like saying currency is only valuable because enough people believe in its value to someone who wants advice on how to earn money?
    – aguijonazo
    Commented Dec 8, 2023 at 23:42
  • 1
    @EiríkrÚtlendi: he possibly wanted to express the distinction between an ideogram and a logogram. A kanji is not just a sign---idea, but a sign---phonetic component---idea. Commented Dec 9, 2023 at 0:25

The claim that kanji don't have meaning on their own seems outlandish. It sounds like the opinion of someone who doesn't have much familiarity or facility with how printed Japanese is used.

Look at any kanji dictionary (published for Japanese even) and you will find an entry for each kanji listed with its meaning, and perhaps some details on when a particular pronunciation was borrowed from Chinese, and other details of historical interest.

Kanji (and their underlying meanings) are particularly helpful when encountering them in a compound you haven't seen before. But usually, context, the meaning of both kanji, and a little thought will help you figure out what the new word most likely means. Often enough, you can even guess the reading of the particular compound (it's never 100 percent guaranteed, but you can often get surprisingly close).

There are words such as 寿司 (sushi) that bare little resemblance to the meaning of the characters, only the phonetic value of the kanji are used. Other words, like 煙草 (tabako), play on the literal meaning of the kanji, but the pronunciation is completely different (I think for 煙草 it is far more common just to write タバコ instead of using kanji, but that doesn't diminish the value of the example).

If you're trying to learn kanji, then I would suggest start reading a lot. You can find books geared toward different levels of fluency. My personal preference was to use books written for middle school students. I could build a lot of vocabulary, furigana were often used (easing the familiarity of how to read the characters), and the grammar was generally not particularly challenging. And yet, middle school level reading was sufficiently advanced that I could pick stories (novels) of interest or books on hobbies or even textbooks for history, science, etc that were of interest to me.

Learning kanji one at a time with their various readings and meanings has its use but it's a very unproductive path forward to learning a lot of vocabulary and kanji-recognition if you're not also simultaneously pairing your study with lots of reading.


Good question. Kanji characters are basically Chinese characters (but slightly different sometimes). They do have meaning on their own. Often the meaning is embedded in parts (or radicals or components) of the pictogram itself. In the same way that an English word when combined with another word gives a different meaning, a Kanji character when combined with another Kanji character gives a different meaning. It would be enlightening to learn about the radicals or basic building blocks of each Kanji or Chinese character. A good app is PlecoDict which handles Japanese characters too.


Knowing the etymology of the English word "welcome", would you say that the "wel" part doesn't have meaning?

Each kanji ordinarily is a content morpheme. It has meaning even if okurigana, or another kanji, might be required to make a coherent unit. As such, they absolutely have meaning. It's only natural to expect such meaning - they are ideographs, generally imported from Chinese.

There are, of course, kanji where the actual "artwork" is misleading - because some part of it is intended to indicate pronunciation, except that it's some classical Chinese pronunciation that then got warped into an on-reading.

And there are, I suppose, some cases where they don't form a free morpheme by themselves. For example, the 好 of [好]{す}き can be treated as a prefix for some compound words, pronounced as こう, but isn't a "word" (lexeme) by itself.

And yes, there will be some words that have idiosyncratic gikun, jukujikun or ateji readings. This is a natural language, after all. Similarly in English, a carpet has nothing to do with either cars or pets, but "car" and "pet" are still distinct and meaningful concepts.

But they are still fundamentally ideographs that correspond to morphemes that have meaning.

If kanji are listed as having multiple "meanings", those are glosses for a core meaning that has become extended through metaphor - in the same way that every language develops. It just happens that different languages (language groups, really, since Japanese borrows a lot of "results" from Chinese) get to do so separately.

For example, for 分 as a kanji, Jisho lists:

part, minute of time, segment, share, degree, one's lot, duty, understand, know, rate, 1%, chances, shaku/100

But there is really only one meaning: splitting (the act, or the result thereof).

"part", "segment", "share" mean the same thing, in the sense that they're intended here.

"minute (of time)", "degree" (either in the temperature sense, or in the sense of "extent out of 100%"), "1%", "shaku/100" (shaku referring to a traditional unit of length) are examples of something having been split up - units that you get by splitting a larger unit.

"rate" and "chances" are more practical applications: a rate is the part of something that will happen in a set amount of time. The chance of an event is the part of the time that it will happen, if you repeat the experiment.

"one's lot" is straightforwardly the part of a whole that one deserves or is owed.

A "duty" is the part of a task for which the individual is responsible. Duties are assigned in contexts like the military, where some larger goal must be attained by a collective effort of people doing possibly different things.

分 as in the verb 分かる is somewhere along the lines of "know" or "understand", accounting for those glosses. The meaning doesn't map directly to English, because the boundaries of words in natural language are fuzzy - for English speakers, there isn't an objective, clear, unambiguous border between knowing vs. understanding something. Anyway, this meaning derives from metaphor: by splitting a complex problem into parts, we can come to understand and study it more easily (and thus know the underlying concepts), a piece at a time. Personally, I prefer the gloss parse, which retains a similar metaphor in English (splitting e.g. text into pieces and determining its structure).

If you look at compound words that include it in a J-E dictionary, you can similarly see how the 分 kanji imparts the sense of "splitting" to more complex ideas, for example:

  • [分数]{ぶんすう} - fraction (in the mathematical sense); formed by splitting (whole) numbers

  • [分娩]{ぶんべん} - childbirth; the child is split from its mother

  • [分量]{ぶんりょう}する - to measure; a specific amount is split off from the supply

Though we think of these words as wildly unrelated in English, there is a commonality.

However, that doesn't mean that drilling yourself on individual kanji meanings will be the most productive thing. And really, when people make statements like that Redditor did, we all know that it's really the lead-in for advocating for one study method or another. ;)

My suggestion is:

  • For individual kanji, learn one "core" meaning - if you find a complex dictionary entry like above, try to come up with the "story" behind it, and choose a single representative gloss.

    • If it's a word by itself, or is used to write a one or more common verb or i-adjectives, learn those words and their readings (which will normally be a common kun-reading plus okurigana). For example, you might want to learn that the 好 of [好]{す}き is also used for [好]{この}む; you probably won't care that it's occasionally used as a variant of [良]{い}い/[良]{よ}い, since that's very commonly just written in hiragana anyway.

    • If you can't come up with any of those, learn the most common on-reading.

  • For words that are just multiple (usually two) kanji without okurigana, study the pronunciation and meaning of the word. You'll come to recognize that the same kanji are appearing in multiple words, and you'll already have a gloss from the other study, and you'll pick up the necessary on-reading automatically.

    • Then, in the cases that don't use on-readings consistently - even if it's something as obscure as the surname [小鳥遊]{たかなし} (my text input doesn't even offer it as a suggestion) - well, you already learned the pronunciation, so no problem, right?
  • If you do the flashcard deck thing: make flashcards with words first. Use a separate deck for studying individual kanji (even if that kanji could be a word by itself); in that deck, only use meaning (your personal gloss), not pronunciation.

The above approach is absolutely not tested! I personally took a very long and winding journey, and have not-great results to show for it. But it fairly closely resembles what I've been doing in the last couple of years, when I started taking things much more seriously and made much more progress as a result - and it's what I think I would have told myself to do, with the benefit of hindsight.


The opinion that each kanji don't have a meaning sounds strange and is unacceptable to most people (including me), but it also seems to me to point out some important fact.

First, each kanji can have more than one meanings. Sometimes they shares or once shared the same "original" meaning, but not always. In a most extreme case, a kanji has multiple meanings simply because multiple different kanji happened to converge into the same appearance (and came to be treated as one kanji). Kanji dictionaries try their best to list up many different meanings to cover as many words as possible, but in parctice, even native speaker sometimes do not know what each kanji means in certain word (no matter how often or how appropriately they use that word!)

Second, kanji is not always a word by itself. When multiple kanjis are combined into a word, the meaning of entire word is not always a simple combination of that of each kanji nor something otherwise predictable (especailly where metaphorical or idiomatic change in meaning occured).

Theoretically kanjis have meanings, but practically not all of them are explicitly recognized even when actually used. Therefore, to know meanings of kanjis is not necessary nor sufficient to understand words. However, in learning the language, it is a powerful clue to efficiently guess or memorize the meaning of word.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .