筆順 · semi-cursive compression
Cursive as compressed order
行書 writes neighboring strokes of the regular order as one movement, and sometimes shortens or reshapes a stroke. The join follows that order. Where the order itself changes, the new order has to be learned.
Each pair below is one character. The first plate is the KanjiVG kaisho order. The second plate marks the grouping a published 行書 source numbers for that character. A character with no such numbering is not on the page.
KanjiVG keeps two structures in one file, as Strokes sets out. Nested groups are which component contains which. The -sN order is the kaisho realization. A join shortens that realization: consecutive paths become one movement. It does not change how many component slots the character opens. 草書 is not in the table. The sources read for this page do not number a cursive form stroke by stroke. The 1958 handbook already says that 行書 may use another order, as The rules of stroke order records. On this page that change is the grass radical of 葉.
- 連続
- Neighboring kaisho strokes that the cited source tells the writer to continue as one movement: 連続, 折り返し, or 直接連続. In the table, that run counts as one movement.
- 筆脈、筆意
- The brush’s path from the end of one stroke toward the next. The 教育出版 guide calls 点画の連続 the case in which that path is written through. Sato and others (2007) sometimes score 筆意, or 連続もしくは筆脈, apart from a fully joined stroke. Where that pair is the numbered continuity, this page counts it as one movement and says so on the plate.
- 点画の省略
- A stroke the source says not to write. Only 極 has a numbered omission here. A fold that is still drawn is a change of shape, and the movement count keeps it.
A primary join of k strokes saves k − 1 movements. A dropped stroke saves one. A change of shape, or a change of order, saves none. An alternate grouping that reuses a stroke already counted is stored and is not added. The count is the kaisho length minus those savings. Component slots stay as they are on Strokes.
Each plate draws itself in the kaisho order. Click a plate to replay. On a grouping plate, joins take the slot inks in order: cinnabar, then indigo, then green, then gold. A dropped stroke is gold. A reshaped stroke, or a stroke whose order changes, is cinnabar. Gray digits are that file’s own stroke numbers. With reduced motion the finished character stays still.
口, 国, 心, 之, and 日 and 木 as whole characters are not here: the sources read for this page do not number which of their kaisho strokes join or drop. 日 appears as the enclosed component of 境 and as the split 日 of 由. 木 appears as the left-hand component of 極 and as the lower component of 栄. 糸 appears as the left-hand component of 組. 言 appears as the left-hand component of 語, and only the shape of its third stroke is numbered. Both papers teach an omission in 門構え, and neither numbers which kaisho strokes are dropped. Sato and others (2007) also record a common wrong omission of 門. That shape is not described here.
On this page
Where the source numbers a run of neighboring kaisho strokes and says to continue them, the table counts that run as one movement.
The same paper records 「学の1画目から3画目まで連続」 in 51% of the framed university writing, and calls that continuity unnatural. It is not in the count. Strokes 7–8 are not in the textbook list either.
One line in the same guide’s first-hour plan numbers the changed sweeps as strokes 6 and 7, and tells the writer to make stroke 3 a vertical, stop, and return into a horizontal. In this file stroke 3 is ㇒, stroke 6 is ㇐, and stroke 7 is ㇑. The counted joins are 2–3 and 4–5, and the counted shape change is 8–9, which are the numbers that match the fill-in sentence and this file.
A closed component does not by itself join its strokes. The join is the run the source numbers, including a run that sits inside the frame.
The same checklist scores strokes 12–13 as a separate continuity, the last bar of 日 into the left-falling stroke of 儿. Stroke 12 is already in the counted join, so 12–13 is not added.
Strokes 7–8 are scored as 連続もしくは筆脈. Stroke 8 is already counted, so 7–8 is not added. Strokes 1–2, the left-falling stroke and the horizontal of 禾, are easy to join and common in the fast notes. The paper says that join is in none of the eight textbooks then current, and is rare in classical models. It is not in the count.
A dropped stroke leaves the movement count. A new shape, or a new order, does not. The order change has to be learned; the kaisho order is not what the hand follows there.
The same checklist says the upper dot and the lower dot of 糸偏 are omitted, and that the stroke order of 糸偏 changes. It does not give the new order, or the indexes of those dots. They are not marked.
The same checklist says the 木 of 葉 changes in 行書, without stroke numbers, so that change is not marked. Higuchi 2010’s item for 花 says the grass radical’s order changes, and that strokes 1–2 must not look like the fold of hiragana わ. It does not state the order 2, 3, 1. 花 is not on this page.
The movement count
| Character | Kaisho | Joins | Dropped | Shape or order | Movements | Also scored | Source |
|---|
What this shows
70a0b7ae0c18ceb5cb358274b029cce0234a43bc of github.com/KanjiVG/kanjivg. Plate files in site/assets/kanjivg/ are unmodified. Joins and the other markings are the stroke numbers in Sato, Higuchi, Tsumura, and Honda (2007), 書写書道教育研究 21, pp. 75–80, doi:10.82787/jacse.21.0_75, free to read on J-STAGE, copyright 全国大学書写書道教育学会; Higuchi (2010), the same journal, vol. 24, p. 83–, doi:10.82787/jacse.24.0_83, same copyright (the J-STAGE title string ends in 改普; the article’s headings say 改善); and 教育出版, 『中学書写』令和7年度 内容解説資料, lessons 「大洋」 and 「栄光」, copyright 教育出版. No figure from those publications is reproduced.