Ideal spacing before fullwidth opening punctuations at the beginning of sentences or lines

[kidayasuo]

There was an extensive discussions on the email list regarding the amount of space before U+300C LEFT CORNER BRACKET (and fullwidth opening punctuation in general) at the beginning of a sentence or a line (started with 行頭役物 Koji Ishii (Monday, 16 October 2023)).

This issue captures the current conclusion. Further discussions should use this thread.

The conclusions are:

• The logical default spacing is ツメ, i.e. removing preceding ½ em space before opening punctuations at the beginning of text such as beginning of sentences, folded lines and lists. This applies to both indented and blank line sentence styles.
• Nonetheless there are different house styles and preferences and they should also be honoured. One piece of caution is that these house styles assume indented sentence style and therefore they do not necessarily applicable to blank line sentence style, which is a default on the web.
• The discussion of what should be the default on the web is a different issue because of compatibility requirements.

[kidayasuo]

Below is summary by Bin-sensei (@KobayashiToshi) analysing various styles under the indented sentence style. The translation to English (by chatGPT) is attached after the original text.

行頭の括弧については，これまで何回か私の考えを記してきましたが，部分的な説明でしたので，まとめておきます．

出てくる箇所は，以下の2つ
　段落の先頭行（全角下ガリ）とする場合　以下，“段落先頭”という
　段落の2行目以下とする場合　以下，“折返し先頭”という

考えられる組合せとしては以下の7つ，なお，括弧の字幅は二分として説明
　a　段落先頭：全角二分下ガリ　折返し先頭：二分下ガリ
　b　段落先頭：全角二分下ガリ　折返し先頭：ベタ（以下，天付きという）
　c　段落先頭：全角下ガリ　折返し先頭：二分下ガリ
　d　段落先頭：全角下ガリ　折返し先頭：天付き
　e　段落先頭：二分下ガリ　折返し先頭：天付き
　f　段落先頭：二分下ガリ　折返し先頭：二分下ガリ
　g　段落先頭：天付き　折返し先頭：天付き

実際に多く見掛けるのは，a, d, eの3つです．書籍ではaはあまり見掛けませんが，Webではよく見掛けます．それ以外は，cやbもあり得ますが，整合性がとれない組み合わせですので，除外します．また，段落先頭を全角下ガリとする方針を前提とします．

まず，前提条件から

1　縦組でいえば横の並び，横組でいえば縦の並びは，どの程度問題にすべきか

行頭では，気になるが，2字目以下では，あまり気にしている人はいない可能性が高い．行頭での問題は，以下の2例です．小見出しと直後の段落先頭の括弧，段落先頭の括弧と折返し先頭の括弧です．

　　「原稿編集」の問題　←（行頭の全角下ガリの小見出し）
　　「原稿編集」は，著者から入手した……
　「原稿」をどのようにして行えばよい……
　した作業を行うには，著者の意向をま……

2　行の調整処理は，どの程度考慮しないといけないか

組版エンジンの負担にはなるが，それは大きな問題ではないと考える．字間の乱れは，行長や調整量にもよるが， 行頭の括弧に伴う調整量は二分です，これだけを考えると，あまり気にならないという可能性が高い．ですから，aがよいということは，この際は考えない．

次に，“折返し先頭”から，

最近の書籍・雑誌等の行末に配置する句読点や括弧の後ろを空けないでベタ組とする例が増えています．これは行末の並びが気になり，句読点や括弧の後ろの空きを削除した方がよいという考えからきていると思われます．一般に，縦組でいえば横の並び，横組でいえば縦の並びは，行頭の方が行末よりは揃えたいという意識が強い，と思われる．先頭は気になるが，末尾はなんとなく見てしまう，という傾向はあると思います．ですので，この行末が気になるのであれば，行頭はさらに気になる．

であれば，折返し先頭は天付きがよいとなる可能性が高い．

段落先頭では，どうか

まず段落先頭が二分下ガリという形式について

1）これは段落先頭を全角下ガリとする方針が前提ですが，この全角下ガリという原則のルール違反です．これはいかなる理由があるのか．それは納得できることなのか．いまのところの説明としては，会話文がある場合に下がりすぎになるのを避けたいという考えです．これは一般の場合には適用できないことでしょう．

2）行頭では，気になる例の小見出しですが，括弧のある小見出しは全角下ガリとすることが多いので，これと行頭はそろわない．前にもお送りした画像ですが，添付いたします（行頭の括弧例.JPG）．このケースは数は多いとはいえないが，私は，けっこう見ています．

3）段落の先頭を全角下ガリとするのは，段落の区別を付けるためです．段落先頭の二分下ガリは，その区別としては，ややあいまになる可能性がある．特に問題となるケースは以下の2つの場合です．
　①段落の先頭に括弧のある段落の直前の末尾行が，行の末尾近くまで配置されている場合．これは，それなりにある．（私は，この経験は何回かしている．）
　②本を読む場合は，私は勝手に“段落読み”と読んでいるのですが，主に段落の先頭にある文だけを読んでいく方法をとることがあります．段落の構成を考えたテキストでは，これで結構意味を理解していくことが可能です．この方法は，私だけでなく，2冊ほど，この方法があるよと書いた本を読んだことがあるので，数は多くないが，こうした読者がいると予想されます．この場合に段落の先頭の括弧の二分下ガリはあいまいになる．

次に，段落先頭が全角二分下ガリという形式について

1）句読点や括弧が重なった場合，あるいは全角スペースと括弧が重なった場合は，そのアキを調整します．となれば，段落先頭を全角下ガリとする場合，全角二分下ガリはルール違反になります．

2）段落先頭を二分下ガリとする方針が採用されている例が多いという現状を考慮すると，この段落先頭が全角二分下ガリは，下ガリすぎと思われる．

3）括弧のある小見出しの全角下ガリと段落先頭が全角二分下ガリでは，行頭はそろわない．（段落先頭が全角二分下ガリとする書籍でも，小見出しの行頭は“5行頭の括弧例.JPG”と同じ形式にする例が多い）．

4）段落先頭が全角二分下ガリは，折返し先頭は二分下ガリとなります．段落先頭を二分下ガリとするeの方式が多いことから，この折返し先頭の二分下ガリを段落先頭と誤解される恐れがある．

以上です．ですので，aとeという方式は問題が多いということです．

---

Additional comments

もちろんaやeという配置方法の選択はあると思います．それを否定しようとするわけではありません．例えば，会話の多い場合は，eになるでしょうし，句読点や括弧が連続した場合の処理ができないアプリケーションも，けっこう多く，こうしたアプリケーションではaという選択になるでしょう．また，文字の字幅を全角として，それを単純に並べていくという，ある種の簡便な処理にしたい場合にも，aの選択は考えられます．

ただ，基本はdであろう，ということです．

行頭を下げない方針の場合はどうなりますか？

前に送ったメールの考えられる組合せの種類でいえば，理屈からいえばgでしょう．下げないということですから．行頭に括弧がない場合，段落先頭と折返し先頭も天付き（差異はない），括弧がついても，段落先頭と折返し先頭ともに天付きでよく，差異を付ける必要はないでしょう．

---

I have previously noted my thoughts several times about the opening brackets at the start of lines, but as those were partial explanations, I will summarize them here.

The instances where these occur are as follows:

1. At the start of a paragraph (fullwidth hanging punctuation) - hereinafter referred to as 'paragraph start.'
2. At the start of the second and subsequent lines of a paragraph - hereinafter referred to as 'wraparound start.'

The possible combinations are the following seven, and for these explanations, the width of the brackets is considered to be half-width:
a. Paragraph start: fullwidth and a half-hanging, wraparound start: half-hanging
b. Paragraph start: fullwidth and a half-hanging, wraparound start: solid (hereinafter referred to as 'top aligned')
c. Paragraph start: fullwidth hanging, wraparound start: half-hanging
d. Paragraph start: fullwidth hanging, wraparound start: top aligned
e. Paragraph start: half-hanging, wraparound start: top aligned
f. Paragraph start: half-hanging, wraparound start: half-hanging
g. Paragraph start: top aligned, wraparound start: top aligned

In practice, the three most commonly seen are a, d, and e. While a is not often seen in books, it is common on the Web (kida's comment: it naturally happens when the engine lacks ability to control space). Others like c and b are possible, but they are inconsistent combinations and are therefore excluded. Furthermore, we will base our discussion on the policy of making the paragraph start fullwidth hanging.

First, let's consider the following premise:

1. In vertical typesetting, this would be the horizontal arrangement, and in horizontal typesetting, the vertical arrangement - how much should this be a concern?

At the start of the line, it's noticeable, but beyond the second character, people don't seem to pay much attention. The issues at the line start are as follows: brackets at the small heading and immediately following paragraph start, and brackets at the paragraph start and wraparound start.

The issue of 'manuscript editing' ← (small heading with fullwidth hanging at line start)
'Manuscript editing' involves obtaining materials from authors...
How to proceed with 'manuscripts'...
To perform these tasks, it is necessary to consider the author's intentions...

2. How much should we consider line adjustment processing?

While it does burden the typesetting engine, I don't see it as a major problem anymore. The disorder between characters depends on the line length and adjustment amount, but the adjustment amount due to the bracket at the line start is half-width. Therefore, it seems not too concerning, so we won't consider option a for now.

Next, considering the 'wraparound start':

Recently, there has been an increase in the practice of solid setting (not leaving space) after punctuation marks and brackets at the end of lines in books and magazines. This seems to stem from the concern about the arrangement at the line end, suggesting it's better to remove the space after punctuation marks and brackets. Generally, in vertical typesetting, this would be the horizontal arrangement, and in horizontal typesetting, the vertical arrangement - there's a stronger desire to align these at the start of the line than at the end. It seems there's a tendency to be concerned about the start but somewhat overlook the end. Therefore, if the end of the line is a concern, the start of the line should be even more so.

Hence, the likelihood is high that top alignment is better for the wraparound start.

Then, what about the paragraph start?

First, regarding the paragraph start being half-hanging:

1. This assumes the policy of making the paragraph start fullwidth hanging, but it is a violation of this basic rule. What could be the reason for this, and is it acceptable? As far as current explanations go, it's to avoid excessive dropping in cases with dialogue. This might not be applicable in general cases.

2. For the line start, the concerning example is the small heading with brackets, which is often made fullwidth hanging, so it doesn't align with the line start. I have sent an image before, but I will attach it again (example of brackets at line start.JPG). This case might not be frequent, but I have seen it quite a bit.

3. Making the paragraph start fullwidth hanging is to distinguish paragraphs. However, the half-hanging at the paragraph start might be somewhat ambiguous as a distinction. The particularly problematic cases are as follows:
   a) When the last line before a paragraph with opening brackets at the start is almost fully extended. This is fairly common. (I have experienced this several times.)
   b) When reading a book, I sometimes adopt what I call 'paragraph reading' - focusing only on the sentences at the start of paragraphs. In well-structured texts, this allows for a good understanding of the content. I am not

the only one who does this; I have read a couple of books mentioning this method. So, while not numerous, there are readers who adopt this approach. In such cases, the half-hanging of the opening brackets at the paragraph start becomes ambiguous.

Next, considering the paragraph start being fullwidth and a half-hanging:

1. In cases where punctuation marks or brackets overlap, or when a fullwidth space and bracket overlap, we adjust the space. Therefore, if we adopt fullwidth hanging at the paragraph start, fullwidth and a half-hanging would be a rule violation.

2. Considering the current situation where many examples adopt the policy of half-hanging at the paragraph start, this fullwidth and a half-hanging seems excessive.

3. The fullwidth hanging of brackets in small headings and the fullwidth and a half-hanging at the paragraph start do not align. (Even books that adopt fullwidth and a half-hanging at the paragraph start often follow the same format as in 'example of brackets at 5 lines start.JPG' for the heading.)

4. If the paragraph start is fullwidth and a half-hanging, the wraparound start becomes half-hanging. Considering the prevalence of option e, where the paragraph start is half-hanging, there's a risk of confusing the wraparound start half-hanging with the paragraph start.

In conclusion, this means there are many problems with the approaches a and e.

---

Additional comments

Of course, I believe there is a place for layout methods like a and e, and I am not trying to negate that. For instance, in cases with a lot of dialogue, e might be appropriate, and there are quite a few applications that cannot handle consecutive punctuation marks or brackets, which would lead to the choice of a. Also, when adopting a simplistic approach of arranging characters with full-width measurements, option a becomes a viable choice.

However, the basic approach should be d.

What about policies that do not involve indenting the line start? (i.e. blank line style)

Speaking in terms of the combinations I mentioned in the previous email, logically, it would be g. Since it's about not indenting, in cases without brackets at the line start, both the paragraph start and wraparound start would be top aligned (no difference). Even with brackets, both the paragraph start and wraparound start should be top aligned, and there is no need to create a difference.

---