PLEASE REMOVE THE GLYPH FOR "の" (U+306E)

[akira7095]

I am using the Mullvad Browser, and this browser comes bundled with STIX Two Math.

Please remove the "の" (U+306E) glyph from STIX Two Math. Because of this glyph, when reading Japanese text written in Hiragana, only the character "no" is displayed in an unnatural form. This is cultural appropriation.

[apoorvpotnis]

It could probably be removed if it's not really required in mathematical notation (decision of the maintainers). But there's no cultural appropriation here at all. If the の character appears in Japanese text, and not as a mathematical symbol, on a webpage, then the Stix Two font shouldn't be used at all for rendering it in the first place. This looks like a bug in the webpage configuration or the browser. Also do note that sometimes, non-Latin or non-Greek glyphs/characters are used in math notation in English. See https://meta.mathoverflow.net/q/5856/516900 and #269 (comment).

[davidcarlisle]

You haven't shown the CSS styling being used, I would expect whatever font is chosen for Japanese text to be used here in preference to the math font?

I don't speak for stipub but the inclusion here isn't just a quirk of the stix font, it was listed in the AMS column of the original stipub submission of math characters to Unicode

https://www.ams.org/STIX/bnb/stix-tbl.ascii-2006-10-20

and (not unrelated) is listed as an isolated character sometimes used in a math context by Unicode TR25

https://www.unicode.org/reports/tr25/

which says

Additional Characters. In addition to this basic set, mathematical notation also uses the bold
upper- and lowercase digamma (U+1D7CA and U+1D7CB), and the four Hebrew-derived
characters (U+2135..U+2138), for example in ℵ0 for the first transfinite cardinal. Occasional
uses of other alphabetic and numeric characters are known. Examples include U+0428 Ш
CYRILLIC CAPITAL LETTER SHA, U+306E の HIRAGANA LETTER NO, the ideograph U+4E2D 中 and
Eastern Arabic-Indic digits (U+06F0..U+06F9). However, unlike the characters in the
mathematical alphabets, these characters are only used in a single, basic form. See also the Arabic
Mathematical Alphabets discussed later in this section.

So it may be that this character should be declassified for math use, but if so that should be done more generally not just in one font.

[tiroj]

This glyph was inherited unchanged from the original STIX fonts. As @davidcarlisle notes, there are a few characters from various writing systems that may be used in math, and historically mathematicians and scientific publishers have borrowed symbols from other alphabets or syllabaries. This is in the nature of a productive notation system hungry for symbols for new concepts.

That said, I would love to know why this Hiragana character was included in the STIX set and whether it actually has any use in math. I suspect it was among a set of characters compiled from existing piublications from STI Pub member publishers, but I don’t think those were documented in detail.

If STI Pub tell me this character can be removed, I will be happy to do so, but I would also be pleased to update the glyph with one that is less ‘unnatural’. I have Japanese type design colleagues whom I can ask for advice and models.

[RuixiZhang42]

Haruhiko Okumura (奥村晴彦), a prolific figure in the Japanese TeX community, twitted the inclusion of の in the STIX font about nine years ago. He described this decision: 「の」の謎 (the mystery of “no”).

[…] only the character "no" is displayed in an unnatural form. This is cultural appropriation.

But there's no cultural appropriation here at all.

I would also be pleased to update the glyph with one that is less ‘unnatural’. I have Japanese type design colleagues whom I can ask for advice and models.

The STIX の is very reminiscent of the “chop suey” font style, which is considered by many as cultural appropriation. Linus Boman had a YouTube video on this topic, and Raven Mo did a talk about it last month at ATypI Stanford. So I can totally see where @akira7095 is coming from.

This looks like a bug in the webpage configuration or the browser.

@apoorvpotnis Yes, but the blame shouldn’t entirely fall back to font fallback configuration. When Discord first introduced gg sans a few years ago, that new font broke Japanese (and Chinese). Celestial Phineas (Yehang Yin) and Toshi Omagari might also have twitted about this. Discord ended up removing the two katakanas ツ and ノ, along with the full-width right parenthesis (U+FF09), from the font. It turns out that partial coverage of an entirely different script is a bad idea—who would’ve thought that! Joking… just ask any Greek how they feel about their typography “ruined” by math symbols.

Unicode TR25 […] says

[…] Occasional uses of other alphabetic and numeric characters are known. Examples include U+0428 Ш CYRILLIC CAPITAL LETTER SHA, U+306E の HIRAGANA LETTER NO, the ideograph U+4E2D 中 and Eastern Arabic-Indic digits (U+06F0..U+06F9). […]

The Unicode’s claimed “examples” are unusual. I’m not aware of a single use of の or 中 as a symbol in math formula inside Japanese or Chinese text, respectively. Sure, there are instances of kanji and kana appearing in formulae, but they appear as text (similar to $\frac{\text{Height of the tree}}{\text{Length of the shadow}}$), not as individual symbol.

The closest examples of ideographs used as symbols (i.e., doing algebra), are found in Chinese math books dated pre-Qing dynasty. For instance, 𝑎²+𝑏² would be 甲^二丄乙^二. However, so many of these ideographs are not of common use today, and some of them aren’t even encoded, though Monu Math tries to support such archaic way of writing math in Classical Chinese (文言).

[apoorvpotnis]

I am pretty sure that whoever intended to use the の character in math intended to use it without any racist intentions. It is very much possible that it was from a Japanese person who thought that the character was suited for whatever mathematical purpose it was intended. It turns out that the Hiragana "No" character の was included in the Stix symbols list because it appeared in an AFII* list (source: discussion with David Carlisle and Barbara Beeton on TeX.SX chat). Its original intended use, the person(s) who used it and added it to the AFII lists, and the original publication(s) is unknown.

Symbols, words, sounds, concepts, and ideas have always traveled between cultures and languages throughout history, as long as human societies have communicated with one another. There is absolutely nothing wrong with this. I consider this to be a rather good thing.

It seems that the の notation has never become widely adopted or gained mainstream popularity (with my limited knowledge of mathematics), especially when compared to something like the use of the Aleph symbol (ℵ) to represent cardinalities. The appropriate action would be to fix the bugs in the browser or the webpage code, so that the の character appears in the appropriate font and not Stix Two Math. And John Hudson is willing to change the glyph style to a more natural one. So, I see no issue here.

The Unicode’s claimed “examples” are unusual. I’m not aware of a single use of の or 中 as a symbol in math formula inside Japanese or Chinese text, respectively.

But there's no claim of using の or 中 as mathematical symbols in a Japanese or Chinese text. The symbols were probably used as math symbols along with English (Latin script) text.

*The Wikipedia article is in French. I could not find an English version. One can use translation tools.

[tiroj]

If we keep the character in the STIX Two Math font — for backwards compatibility if no other reason — I could replace the glyph with one from an open source Japanese font. There are plenty of options on Google Fonts:
https://fonts.google.com/?preview.text=%E3%81%AE&script=Jpan

I am assuming something in a Mincho style.

[RuixiZhang42]

The appropriate action would be to fix the bugs in the browser or the webpage code, so that the の character appears in the appropriate font and not Stix Two Math.

Agreed. A well-ordered CSS font fallback chain and a language-tagged HTML plain text is the way. But people expect things to work in their default states ¯_(ツ)_/¯. That’s why I’m citing the Discord gg sans example.

But there's no claim of using の or 中 as mathematical symbols in a Japanese or Chinese text.

That’s not what I meant. My perspective offers absence of evidence in Japanese and Chinese text, and now yours (which is very illuminating) also suggests absence of evidence in Latin-script-based text. The Unicode’s claimed “examples” are indeed unusual, in either perspective. Like you said, there’s so much unknown about the origin of の in the AFII list. There’s no known actual use of の (or 中) in math publications either, which I’d be delighted to be proven wrong. All we have is speculation, no use evidence.

[apoorvpotnis]

Coindicentally, I encountered the a similar issue today. I saw that Devanagari letters from NewCMMath were being used in preference to other Devanagari fonts (including NewCM Devanagari surprisingly) for the "Google offered in:" language options on Falkon on CachyOS. I tried to debug using ChatGPT. The conclusion was that "QtWebEngine’s glyph fallback is ignoring the family sequence added to the Arial pattern. The Arial → Noto Sans Devanagari rule is valid in Fontconfig, but it cannot solve this Falkon problem." We couldn't find a solution at the browser level. I had all the NewCM fonts installed in my /usr/local/share/fonts directory. As a workaround, removing the fonts solved the issue. LuaTeX can access the fonts from it's own directories. Perhaps something similar is going on with the Mullvad browser.