Kanji

A small number of kanji are pictographs, simplified drawings of things: 山 (mountain), 川 (river), 木 (tree), 日 (sun). Others combine ideas: two trees 林 make a grove, three 森 a forest. By far the largest group, however, are phono-semantic compounds, in which one part hints at the meaning and another at the Chinese pronunciation.

The components used to classify characters in dictionaries are called radicals (bushu). The water radical 氵 appears in 海 (sea), 池 (pond) and 泳 (swim); the person radical 亻 in 休 (rest) and 体 (body). Learning the common radicals is the most efficient first step.

Most kanji have at least two kinds of reading. The on'yomi, or 'sound reading', is an adaptation of the Chinese pronunciation at the time the character was borrowed; some characters even have several on'yomi from different periods. The kun'yomi, or 'meaning reading', is a native Japanese word attached to the character.

As a rule of thumb, kanji compounds use on'yomi and single kanji with okurigana use kun'yomi. 食 is read shoku in 食事 (shokuji, meal) but ta in 食べる (taberu, to eat). There are many exceptions, and the reading of a word is best learnt with the word itself.

Large dictionaries list tens of thousands of characters, but most are rare. The reference for everyday literacy is the jōyō kanji list, set by cabinet notice in 2010 with 2,136 characters. Primary school covers 1,026 of them; compulsory education finishes the rest.

For learners, the Japanese-Language Proficiency Test gives a rough guide, although it has published no official kanji lists since its revision in 2010. Commonly cited estimates, based on the old pre-2010 lists, put N5 at around 100 characters and N1 at around 2,000.

Learners who succeed with kanji usually combine three things: knowledge of components, so that new characters can be broken down; vocabulary, so that each character is anchored in real words; and spaced repetition, so that reviews happen just before forgetting. Writing by hand helps many people remember shapes, even if most writing later happens on a keyboard.

Reading graded texts with furigana, and later ordinary texts with a dictionary app, connects characters to context. Learning isolated characters with lists of every reading is widely considered inefficient.

Japan simplified many characters after 1946, independently of the simplification in the People's Republic of China from the 1950s, so Japanese forms often differ from both simplified and traditional Chinese: Japanese 学 and 国 match simplified Chinese, but 売 and 気 are Japanese forms. Readers who know Chinese characters have a head start with meaning, but must still learn Japanese readings.