Why Kanji Have Multiple Readings

The first major type of reading is *訓読み (kun’yomi), usually called the native Japanese reading. This arises when a Chinese character is matched to an existing Japanese word by meaning. 山 is read yama, 川 is kawa, and 見る is miru*. In these cases, the character is functioning as a graphic symbol for a native Japanese lexical item. This is one reason kun readings often show up in words written with okurigana, such as 食べる, 大きい, or 分かる. The kanji writes the lexical core, while kana write the inflectional ending that belongs to Japanese grammar.

The second major type is *音読み (on’yomi), the Sino-Japanese reading. These are readings derived from Chinese pronunciations as they were borrowed into Japanese. They are not modern Mandarin readings pasted into Japanese; they reflect older layers of Chinese and the historical periods in which borrowing took place. This is why 学校 is gakkō, 文化 is bunka, and 経済 is keizai*. In such words, the characters are not merely standing for native Japanese meanings. They are part of a Chinese-derived lexical stratum that became deeply embedded in Japanese vocabulary.

The official educational materials of Japan’s Agency for Cultural Affairs explain the basic distinction simply: the Japanese reading of 木 as ki is a native Japanese word attached to the character by meaning, while boku and moku are Sino-Japanese readings derived from Chinese pronunciation. Those same materials also point out why one character can have more than one on reading: Chinese pronunciation differed by region and by historical period, so a character could enter Japan more than once or through more than one channel (Agency for Cultural Affairs 2008). That explanation is worth lingering over, because it replaces the idea of random multiplicity with the idea of historical layering.

A kanji can therefore have multiple on’yomi because Japanese did not borrow “Chinese” only once. It borrowed characters and Chinese-derived words over centuries, from different geographic and cultural centers, under different conditions. Traditional labels such as Go-on, Kan-on, and Tō-on describe some of these historical layers. Britannica summarizes the point neatly: Japanese contains many characters with more than one on reading because Chinese words and their pronunciations were borrowed from different parts of China and during different historical periods (Encyclopaedia Britannica 2026a). Multiple readings are thus records of contact history. They preserve traces of different moments in East Asian linguistic exchange.

That explains multiple on readings. But why do some kanji also have multiple kun readings?

Here again the answer is historical, semantic, and lexical rather than chaotic. When a script is borrowed into a language that already has its own vocabulary, there is rarely a perfect one-to-one match between a foreign graph and a native word. A single character may cover a semantic field broad enough to overlap several Japanese words. Or Japanese may distinguish several related actions or states where the imported character tradition groups them together. The character 上, for example, can participate in readings related to “up,” “to rise,” “to raise,” and “above,” because the Japanese language already had a family of native forms for those meanings. The writing system did not erase those distinctions; it housed them under one graph. The famous case of 生 works the same way on a larger scale. Birth, life, rawness, growth, and freshness are linked in the character tradition, but Japanese expresses those ideas with multiple native words and verbal forms. The kanji gathers them; it does not reduce them to one pronunciation.

This is why it is a mistake to think of a kanji as a self-sufficient sound unit. A kanji is better understood as a morphographic anchor: a written sign tied to meaning and lexical history, capable of being realized differently in different words. In that respect, learning kanji one character at a time is useful but insufficient. Serious progress comes when learners start treating readings as properties of words and compounds, not just of isolated characters.

At this point many textbooks introduce a practical rule: on’yomi tends to appear in Sino-Japanese compounds, while kun’yomi tends to appear in standalone native words or in words with okurigana. As a first approximation, that rule is good. It explains why 人口 is jinkō but 人 is hito, why 火山 is kazan but 火 is hi, and why 読書 is dokusho while 読む is yomu. The problem is that some learners treat this as a law when it is really a tendency.

Japanese contains many exceptions and special categories. Some compounds have native whole-word readings that cannot be predicted by reading the characters one by one; these are often called *熟字訓 (jukujikun), with examples such as 今日 kyō or 大人 otona. Personal names bring in 名乗り (nanori) readings that may be perfectly normal in names but unusual elsewhere. There are also cases of 当て字 (ateji)*, where characters were chosen primarily for sound, convention, or layered association rather than transparent semantic composition. On top of that, sound change inside Japanese can alter the shape of a reading once it enters a compound. Rendaku, gemination, and historical phonological change all contribute to the feeling that the written character is less stable than learners would like.

Yet none of this means the system is unusable. It means the system stores history.

There is another reason kanji retain multiple readings instead of being reduced to one official sound each: Japanese gains real advantages from the script. Britannica notes that kanji survive not in spite of their complexity but because they still solve problems that purely phonetic writing does not solve as well. Many homophonous words are visually distinguishable in kanji, and the meaning of an unknown word can often be guessed from the character shapes and their semantic associations (Encyclopaedia Britannica 2026a). In a language with extensive homophony, that is a serious benefit. A reader can distinguish different kō, shō, sei, or kan words far more easily when they are written with different characters than when everything is flattened into kana.

Kanji also preserve the layered structure of Japanese vocabulary. Britannica describes Japanese vocabulary as consisting of four major strata: native vocabulary, Sino-Japanese words, foreign loans, and onomatopoeic expressions (Encyclopaedia Britannica 2026b). The reading system maps onto that stratification. Kun readings connect writing to native Japanese lexicon and grammar. On readings connect it to the Sino-Japanese stratum, which dominates much of abstract, technical, bureaucratic, and academic vocabulary. A one-reading-per-character reform would flatten those distinctions and make large parts of the lexicon harder to recognize historically and semantically.

For practical purposes, the modern state has not left the matter entirely unregulated. Japan’s 2010 Jōyō Kanji Table presents 2,136 characters and specifies standard forms and approved readings for ordinary public life (Agency for Cultural Affairs 2010). That does not mean every reading in the language is contained there, nor does it mean every character neatly has one on reading and one kun reading. It means that the state has stabilized a core written standard while leaving the deeper historical layers of the language intact. The multiplicity remains because the language still needs it.

What, then, should a learner do with all of this?

The most important step is conceptual. Stop asking, “What is the reading of this kanji?” as if every character had one single answer. Ask instead, “How is this kanji read in this word?” That shift sounds small, but it changes everything. Kanji readings are not free-floating labels attached to shapes. They are embedded in vocabulary.

The second step is to learn the historical logic without becoming enslaved to it. It helps to know what on and kun mean. It helps to know that multiple on readings often come from different historical layers of borrowing. It helps to know that native words often surface with kun readings, especially when Japanese inflection is visible. But it does not help to stare at an isolated character and try to memorize every possible reading as an unordered list. That is not how skilled readers know Japanese. Skilled readers know words, patterns, and lexical neighborhoods.

The third step is to accept that a writing system can be efficient without being simple. Japanese did not adopt kanji because they were easy. It kept them because they proved useful. Multiple readings are the visible price of that usefulness. Once a learner sees them not as arbitrary chaos but as the accumulated record of how Japanese made a foreign script its own, the system becomes less irritating and more intelligible. It is still demanding. But it is demanding for historical reasons, not because it lacks structure.

In that sense, the real beginner-friendly answer to the question “Why do kanji have multiple readings?” is not “because Japanese is inconsistent.” It is this: because Japanese kept both its own words and the Chinese-derived words it borrowed, and kanji had to learn to serve both.