KANJI + HANZI

Rule: Separate recognition, reading in a word, and writing from memory. Seeing a familiar shape is not the same as being able to retrieve it. Start with 日, 月, 木, 人 and 口. Choose Japanese or Chinese before practicing: readings and sometimes strokes differ.

Worked example: For Japanese, learn 日 in 日本（にほん, Japan）and 日曜日（にちようび, Sunday）rather than assuming one reading works everywhere. For Mandarin, 日 is rì; learn it in a word such as 日期 rìqī, date. Look at the model, say the chosen word, hide the model, write, then compare.

Exercise: Study 木 for 20 seconds, hide it, write it once and use the stroke player to check. What three things should your review record?

Common mistake: Copying a character twenty times while staring at it practices copying. Include a hidden-model attempt.

Rule: The traditional Kangxi system has 214 radical headings. A dictionary radical is a classification choice; a character can contain several visible components but is usually filed under one heading in a given system. Radical forms change by position: 人 often appears as 亻, 水 as 氵, 手 as 扌 and 心 as 忄.

Worked example: In 休, identify 亻 and 木. In 河, identify 氵 and 可. Both have two major components, but 河 is filed under water and 休 under person. The searchable radical table below shows all 214 headings, not 214 independent pronunciation rules.

Exercise: Find 河, filter by radical 85 (water), and name its other major component. Is every part of 河 a separate dictionary radical assignment?

Common mistake: Calling every component “the radical” makes dictionary lookup and structural analysis ambiguous.

Rule: Many characters combine a component associated with meaning and another associated with historical sound. These clues are probabilistic. Sound change, borrowing and differing reading traditions mean the modern pronunciation is not mechanically predictable.

Worked example: Mandarin 清 qīng, 晴 qíng and 情 qíng share 青 qīng. Water 氵, sun 日 and heart 忄 help distinguish meanings; the 青 series helps recall sound, while tone still needs learning. Japanese readings belong to Japanese words and cannot be obtained by reading the Pinyin.

Exercise: For 清 and 情, identify the shared component and the differing components. Can 青 alone tell you the correct Mandarin tone of both?

Common mistake: Treating a modern visual story as a proven historical origin, or treating a phonetic series as a pronunciation algorithm.

Rule: Common tendencies include top before bottom, left before right, and horizontal before crossing vertical. Enclosures often begin before their contents and close afterward. A bend can be part of one continuous stroke. These are learning principles, not a universal program: use the selected regional model for each character.

Worked example: Play 口: its upper/right bent path is one stroke, not two independent strokes. Compare 木 and 国. In 国, inspect the outer enclosure, interior and closing bottom line. Japanese KanjiVG and the Chinese dataset are separate models; a difference need not mean one is universally wrong.

Exercise: Before playing 国, predict which line closes the enclosure. Then step through 口 and count strokes rather than straight segments.

Common mistake: Counting every direction change as a new stroke. The Chinese models here follow PRC conventions, including for many Traditional forms; Taiwan learners should compare the linked MOE model.

Rule: A useful mnemonic links exact visible parts, meaning, and a word you actually want to read. Keep a pronunciation cue separate if it distorts the shape. The stories in this studio are deliberately invented memory aids, not claims about etymology.

Worked example: For 休, picture a person resting beside a tree: 亻 + 木. For 明, picture a sun and moon lighting a room. These stories help recall components; now attach 休む（やすむ）in Japanese or 休息 xiūxi in Mandarin rather than letting the story replace reading practice.

Exercise: Choose one unfamiliar character with a decomposition in the studio. Write a one-sentence scene using its visible parts. Close the model and draw it; name one detail your story failed to preserve.

Common mistake: Making a vivid story that recalls the meaning but not the actual character. Personal notes are stored on this device, not synchronized between phones.

Rule: On readings reflect historical Chinese-derived layers; kun readings connect characters to Japanese words. A character can have several of each. Okurigana are kana written after a kanji, often marking inflected portions of a word. Compound reading is learned lexically, not by one universal “two kanji means on” rule.

Worked example: 山 is やま alone and サン in 富士山（ふじさん）. 生 appears in 学生（がくせい）, 生きる（いきる）and 生まれる（うまれる）. The kana distinguish words and inflection, while the same written kanji has different readings.

Exercise: Explain why knowing 生 as セイ is insufficient to read 生きる. What should be on the back of a useful flashcard?

Common mistake: Reading dictionary on/kun lists as if every listed reading can be substituted in every word. The reference retains some romanized dictionary readings; it is not a full Japanese word dictionary.

Rule: The 2010 Jōyō list contains 2,136 kanji and is a general-use guide, not every character that appears in Japanese. Personal names, place names, specialist writing and historical texts extend beyond it. A list entry is not a mastery certificate: words, readings and contextual use still need practice.

Worked example: Turn on “Jōyō only” in Japanese mode. All 2,136 characters in that list have a stroke model here. Turn it off to inspect the larger 6,432-character Japanese stroke collection. An uncommon lookup can still have a reference entry without a stroke model.

Exercise: If you finish all Jōyō cards, can you guarantee that you can read every surname or technical paper?

Common mistake: Equating the complete Jōyō inventory with complete Japanese, or interpreting missing stroke data as proof that a character is invalid.

Rule: Simplification is a standardized system of mappings. Some recur in components: 門→门, 言→讠, 金→钅 and 糸→纟 in applicable characters. Others replace much of the character: 學→学, 體→体. The transformations apply where the standards prescribe them; they are not permission to shorten any similar-looking form.

Worked example: 語→语 and 說→说 show the left-side 言 component changing; the standalone character 言 remains 言. 銀→银 shows 金→钅 as a component. Learn the complete pair and its words, not merely the shortcut shape.

Exercise: Would the existence of 讠 mean you should write the standalone word/character 言 as 讠? Explain with 語→语.

Common mistake: Applying a component rule to every standalone occurrence or assuming one rule produces the entire simplified inventory. The reference links known variant mappings; context still matters.

Rule: Several Traditional characters can map to one Simplified character. Reversing that mapping therefore needs the word and meaning. Traditional usage also has regional standards. Conversion is not translation and does not change a sentence from Mandarin into Cantonese.

Worked example: 發 and 髮 both map to 发: 發展→发展 (development), 頭髮→头发 (hair). 後 and 后 converge in simplified 后, but 皇后 (empress) remains 皇后 in Traditional while 以後 becomes 以后 in Simplified. Japanese has its own forms: 學→学 and 氣→気, while Chinese uses 气.

Exercise: Convert simplified 头发 and 发展 to the Traditional forms shown above. Why is a one-character global replace unsafe?

Common mistake: Calling Japanese shinjitai “Simplified Chinese,” or treating variant links as a reliable whole-sentence converter.

Rule: Build reading around short comprehensible texts. Select a small number of new words, examine their characters, listen to a suitable model, and return to the whole sentence. Keep recognition and handwritten production as separate goals; typing and reading require different retrieval skills.

Worked example: Choose a menu or a graded passage. Mark five unknown words, not every unfamiliar individual character. For each, record the full written form, contextual meaning, reading and one useful sentence. Review with the sentence hidden, then re-read the original text without opening the dictionary.

Exercise: Create a seven-day plan with five new words on day one, recall checks and a final unseen sentence. What evidence would show more than memorized copying?

Common mistake: Adding hundreds of cards faster than you can review, or using a character frequency list as a substitute for vocabulary and grammar.