Japanese and Kanji: Why a Non-Sinitic Language Still Writes with Chinese Characters

Japanese was spoken long before Chinese writing entered the archipelago. What changed in the first millennium CE was not the existence of the language but the technologies available for recording it. Contact with Chinese civilization, much of it mediated through the Korean peninsula, brought both literacy and the prestige of Classical Chinese. The earliest writing used in Japan was therefore not “Japanese writing” in a mature sense but Chinese writing employed for diplomatic, religious, and administrative purposes. The earliest substantial Japanese texts that survive date from the eighth century, but the process of contact and adaptation began earlier. (Britannica 2026f; Britannica 2026b; Seeley 1991)

That chronology matters because it shows that kanji did not create Japanese. They created, or at least enabled, a new relation between Japanese speech and writing. At first the easiest solution was simply to write in Classical Chinese, as educated elites elsewhere in East Asia did. But Japanese was not Chinese. If one wanted to write Japanese as Japanese, Chinese characters had to be repurposed. This is the background to man’yogana, the early system that used Chinese characters phonetically to represent Japanese syllables. From the simplification and abbreviation of man’yogana emerged the kana syllabaries: hiragana and katakana. Japanese writing thus grew not by rejecting kanji, but by splitting their functions and domesticating them. (Seeley 1991; Frellesvig 2010)

Modern written Japanese is a mixed system in which different scripts specialize in different tasks. Kanji typically write lexical roots: nouns, the core stems of many verbs and adjectives, and large numbers of compounds. Hiragana usually writes inflectional endings, particles, native function words, and many grammatical elements. Katakana is commonly used for foreign loans, scientific names, emphasis, onomatopoeia in certain conventions, and various stylistic purposes. This division is not mathematically rigid, but it is robust enough to define the orthography. A verb such as 食べる combines a kanji stem with hiragana inflection. A compound such as 学校 is written wholly in kanji. A loanword such as コンピューター is normally written in katakana. The system works because each script signals a different kind of linguistic information. (Britannica 2026e; Britannica 2026f; Lunde 2009)

This distribution makes Japanese writing unusually efficient for experienced readers. Script alternation helps mark morphological boundaries and lexical type. Function words are visually distinct from content-heavy roots; inflection is distinguished from lexical base; foreign material is typographically separated from older lexical strata. What looks to beginners like needless complexity is, for skilled readers, part of how the sentence is parsed. One can write Japanese entirely in kana, but kana-only prose is often slower to process and visually monotonous because so much lexical differentiation disappears. The continuing use of kanji is therefore not merely traditionalism. It reflects the way the mixed script encodes different layers of the language. (Lunde 2009; Shibatani 1990)

The most famous difficulty of kanji is that a single character can have several readings. This is not an arbitrary defect. It follows from the history of borrowing and adaptation.

The first major distinction is between on’yomi and kun’yomi. On’yomi are Chinese-derived readings: the approximate Japanese renderings of pronunciations associated with Chinese characters at the times and from the regions in which those characters entered Japan. Kun’yomi are native Japanese readings: Japanese words that were matched to characters by meaning. If a character meant “mountain,” for example, it could be used to write the native Japanese word yama; that is a kun reading. The same character could also participate in a Sino-Japanese compound with a Chinese-derived reading such as san; that is an on reading. The character 山 therefore does not have “one true pronunciation” in Japanese. It participates in multiple lexical histories. (Britannica 2026e; Britannica 2026i; Britannica 2026j)

Even the on’yomi are not one thing. Chinese words entered Japan in different periods and through different channels, which is why scholars distinguish reading strata such as Go-on, Kan-on, and To-on. These correspond, roughly, to different times, places, and prestige models of Chinese pronunciation. A character may therefore have more than one Chinese-derived reading because it was borrowed or regularized more than once. Japanese phonological change then acted on those readings, and later conventions sometimes preserved several of them in different lexical environments. The result is a layered system rather than a chaotic one. (Seeley 1991; Frellesvig 2010)

The kun’yomi side is equally historical. A single character might be mapped to more than one native Japanese word if its semantic range overlapped several local lexemes. Some readings become tied to particular derived forms or compounds. Some compounds are read as whole words with pronunciations not recoverable character by character; these are often called jukujikun. Other words use characters more for convention or phonetic approximation than for transparent semantic composition; these are cases of ateji. Personal names introduce yet another layer, often called nanori, in which character readings may be traditional, archaic, or locally conventional. The practical lesson is that readings belong to words and contexts, not to characters in the abstract. Asking “How is this kanji pronounced?” is often incomplete; the more precise question is “How is this word read here?” (Lunde 2009; Seeley 1991)

The most common answer is “homophones,” and that answer is real but insufficient. Japanese has a relatively constrained syllabic structure and a very large Sino-Japanese vocabulary, which creates many homophonous or near-homophonous sequences in speech. Writing those sequences in kana alone can make lexical identification harder. Kanji help disambiguate. A reader can distinguish 同音異義語 visually in a way that kana alone may obscure. (Britannica 2026e; Shibatani 1990)

But homophone resolution is only part of the story. Kanji also provide morphological economy and lexical chunking. They compress information. They let readers identify roots rapidly. They preserve ties to etymology and to the broader sinographic lexicon of East Asia. They also sustain stylistic range. A Japanese sentence written in mixed script looks and feels different from the same sentence flattened into kana. The difference is not just decorative. It affects readability, register, and genre expectations. Over centuries, readers, teachers, editors, and institutions built norms around those effects. By the time modern reform movements considered large-scale kanji abolition, the mixed system had become deeply integrated into ordinary literacy. (Seeley 1991; Lunde 2009)

That is why technological modernization did not make kanji disappear. Typewriters, word processors, and smartphones changed how Japanese is produced, but they did not eliminate the orthographic advantages readers perceive in mixed script. In fact, digital input systems often reinforce kanji use. A user can type phonetically and select the intended kanji from conversion options, which lowers the burden of production while preserving the visual and lexical benefits of character-based reading. What has changed more noticeably is handwriting ability: many people can still recognize and select kanji more easily than they can produce them from memory by hand. Digitality weakened one argument against kanji — that they are cumbersome to write — without weakening the main arguments for keeping them in print and on screen. (Lunde 2009)

The answer depends on what “used” means. There is no single number that captures all actual character use in literature, history, religion, law, personal names, and specialized fields. For general public literacy, however, modern Japan relies on an official benchmark: the Joyo Kanji list. The current list contains 2,136 characters and is treated by the Japanese government as a guide to ordinary written usage in law, public documents, newspapers, magazines, and broadcasting, not as an exhaustive inventory of every character that may ever appear. Of those, 1,026 are taught in elementary school as the Kyōiku Kanji; the remainder are typically completed during secondary education. (Agency for Cultural Affairs 2010; Britannica 2026e)

That still does not exhaust actual practice. Personal names may use additional characters from a separate name-use list, and specialized fields routinely go beyond the general-use inventory. Historical scholarship, Buddhism, legal archives, classical literature, place names, and personal names all preserve characters outside the ordinary benchmark. Modern digital standards, especially Unicode, include tens of thousands of CJK characters. So the realistic answer is hierarchical: around two thousand characters form the state-endorsed general-use core of modern literacy, while the total reservoir available to specialists and in proper names is much larger. (Agency for Cultural Affairs 2010; Lunde 2009)

The history of this benchmark is itself revealing. Postwar reforms drastically narrowed and regularized kanji usage, first through the 1946 Toyo Kanji list and later through the Joyo Kanji reforms of 1981 and the 2010 revision that brought the total to 2,136. Reformers never fully abolished kanji; instead, they standardized a manageable core. This is typical of Japanese script policy. The state has repeatedly tried to tame complexity without abandoning the mixed system that readers actually use. (Agency for Cultural Affairs 2010)

An important part of the Japanese-kanji relationship is that the characters did not remain simply Chinese after arriving in Japan. Japanese not only assigned native readings and developed kana from sinographic material; it also created its own character practices. There are kokuji, characters coined in Japan, such as 働 and 峠. There are compounds whose meanings and distributions are strongly shaped by Japanese usage. There are modern technical and philosophical terms formed from Chinese morphemes in Japan and then re-borrowed into Chinese and Korean. Kanji in Japanese are therefore both inherited and transformed. (Lunde 2009; Seeley 1991)

This domestication is why the claim that kanji are “foreign” can be misleading. Historically, yes, they originated in China. Structurally, however, they have become part of Japanese. A character in Japanese can participate in a native word, a Sino-Japanese compound, an idiosyncratic name reading, or a modern coinage that later travels abroad. No single description covers all these functions. Kanji are not external to Japanese any more than Latin and Greek elements are external to English scientific vocabulary. Borrowed material can become constitutive. (Frellesvig 2010; Lunde 2009)

None of this means kanji are easy. They impose heavy educational costs. They create barriers for children, second-language learners, and sometimes native speakers themselves. Public discussion in Japan often returns to the problem of forgotten handwriting, the difficulty of name readings, and the burdens placed on schooling. These concerns are real. But reform has usually stopped at pruning and standardization because the perceived payoff remains high: kanji support compactness, disambiguation, historical continuity, and orthographic nuance. The Japanese script is not a fossil because it continues to solve problems that a kana-only system would solve less well for most readers under current norms. (Agency for Cultural Affairs 2010; Lunde 2009)

The deeper point is that writing systems persist not because they are abstractly optimal, but because communities build literacies around them. Kanji survive in Japan because they are woven into education, publishing, bureaucracy, and the mental habits of readers. Their continued use is neither irrational conservatism nor inevitable necessity. It is an historically evolved settlement between a morphosyllabic script, two kana syllabaries, and the linguistic structure of Japanese.

The relationship between Japanese and kanji is thus best understood as a long-term accommodation between language and script. Japanese is not Chinese, and kanji do not turn Japanese into Chinese. But Chinese characters gave Japanese a way to build a literate culture of great lexical density, then remained useful after kana emerged because the language itself made a mixed system advantageous. Multiple pronunciations are the trace of layered borrowings and native semantic assignment. Continued heavy use reflects not only habit, but real functional benefits in parsing, disambiguation, and stylistic differentiation. The number of kanji “used in Japan” is therefore not one simple number, but a structured set of circles: an official core for ordinary literacy, broader layers for names and specialists, and a much larger historical archive behind them. That complexity is not incidental. It is the whole point of kanji in Japanese.