Reply to The Origin of Hiragana /wu/ 平仮名のわ行うの字源に対する新たな発見

Disclaimer 記事を読む前の重要事項
This article is a follow-up to my previous article The Origin of Hiragana /wu/; I highly recommend reading that article first to understand my previous theory and its context. However, there will also be a short summary of it in this article.


Introduction 紹介
I find myself in relaxing in a Korean spa. With the peace of the spa, I finally can reflect on my previous research. However, my heart starts racing as I begin to doubt my former conclusion. “Is it too big of an assumption that Katayama would derive his hentaigana from the same kanji?” While I didn’t let it bother me too much, this problem continued to float in my mind. A couple months later, as I was showing my theory to a professor, he was skeptical on the basis that the simplification I proposed had no precedent. After that, I decided it was time to reevaluate my theory and see if I could find any new evidence that either supports or disproves my original theory.


The Question 問題
For a quick summary, in 1873 Junkichi Katayama published Shōgaku Tsuzuriji Hen 小学綴字篇 wherein there is a hiragana gojūon chart with a /wu/ kana. This /wu/ hiragana is boxed in red.



During the Meiji era there were several orthographies which contained a hiragana /wu/; however, none of these contained Katayama’s hiragana /wu/. So naturally this begs the question, does Katayama’s hiragana /wu/ come from a kanji, and if so, which one? Answering this question is more important now than ever as there has been a proposal to encode this character into Unicode, and its origin character is included as a note.


Comments on The Origin of Hiragana /wu/ 『平仮名のわ行うの字源』の論評
In my previous article I claim that the origin character of hiragana /wu/ is 紆 because in the textbook there is also a hentaigana for /wu/ whose origin character is 紆, as seen below.



While there are a few things I would change in that article (such as making the tone less confirmatory), I am still pretty proud of it as my first article. I like to think that the recent influx of discussion on /wu/ kana is partially due to that article. However, I would like to lay out some of the implicit assumptions of that article to better convey its argument. The best candidate for the origin character of hiragana /wu/ would be the one that Katayama explicitly states it is. However, in the absence of that evidence the best candidate should have the following

   1. The character should be mentioned by Katayama indirectly
   2. The simplification of the origin character should have precedent

If no character has both of these than a character with (1) is a better than a candidate with (2). As 紆 was the only character Katayama indirectly mentions it naturally became the character I accepted as the origin. I still think this line of reasoning is the best for this situation, so I will retain it and increase the number of places I search for Katayama to indirectly mention the origin character.


   1. 片山が暗に述べている
   2. 字源の簡略化に前例がある


The Commentaries 附説
The two works I found that would become the basis for this new theory are the Shōgaku Tsuzuriji Hen Fusetsu Upper and Lower Volumes 小学綴字篇附説 巻之上・巻之下, written by Junkichi Katayama in 1873. These are appendixes to Shōgaku Tsuzuriji Hen. In the Upper Volume, Katayama explains his justification for extending kana to fit the gojūon chart. In the Lower Volume, Katayama explains how to teach using Shōgaku Tsuzuriji Hen.


Starting with the Upper Volume, the first thing to catch my eye was a chart with katakana characters next to their origin characters.



I couldn’t believe it; the character I proposed was right there! Now I only needed to find the hiragana chart... But as I was quick to learn, there was no such chart. I searched, and I searched, and I searched, but while the Upper Volume spoke extensively on the /wu/ katakana and its origins, there was nothing about the hiragana. At this point, while I didn’t receive the exact answer I wanted, I felt that this new document overall supported my theory, as it was now confirmed that 紆 was the origin character for the /wu/ katakana... and then I looked at the Lower Volume.


While the Upper Volume was primarily concerned with theory, the Lower Volume was concerned with praxis. The Lower Volume explained how to use the textbook day by day and had very little concerning the new sounds and their characters. But, in this volume, there was one peculiarity that would singlehandedly dismantle my entire theory.


Reference to 汙

In the red box, Katayama is explaining the pronunciation of the /we/ sound. He does so by saying this sound is a contraction of 汙 and 衣, which represent /wu/ and /e/ respectively. Customarily one would represent those sounds by using the origin characters of the consonant and vowel’s kana. This is seen by how every other boxed sound is represented by their katakana origin character. Why is it that Katayama would use 汙 to represent /wu/ over the katakana origin character 紆? To choose a random character of the same sound, over an origin character seems like a very strange decision, especially when the only justification I see for a change in character is just for some variety. I think using the origin character of hiragana /wu/ is much more reasonable, because at least then, Katayama could achieve some variety and still keep with the custom of using origin characters.


Previous Usage of 汙 as Man’yōgana万葉仮名として使われた汙
At this point, we have two candidates that are indirectly mentioned, 紆 and 汙. By the logic I laid out above, the character who has a precedent of simplifying to Katayama’s /wu/ hiragana is the best candidate for the origin character. Starting with 紆, by once again looking at Identifying Japanese Cursive Script by Marcus Sesko, we can clearly see that 紆 does not traditionally simplify to Katayama’s hiragana /wu/.

今までに、片山が暗に述べている候補は「紆」と「汙」だ。上で述べた方法に従うと、片山が選んだわ行うの平仮名の字源への一番の候補は、簡略化の前例がある漢字だ。始めに、マルクス・セス子の『草書の識別』Identifying Japanese Cursive Scriptを見ると、「紆」は片山が選んだわ行うの平仮名に簡略化しないと確認できる。


As for 汙, since it is not in Identifying Japanese Cursive Script, let’s look at its use as a man’yōgana in the Wooden Tablet Database by the Nara National Research Institute for Cultural Properties. In Tablet 1 and Tablet 1806, the character 汙 is being used as a man’yōgana for the /u/ phoneme for personal names.



Clearly, these man’yōgana look identical to Katayama’s /wu/ hiragana. Therefore, only 汙 meets criteria 1 and 2 and thus is the most likely candidate to be the origin character of hiragana /wu/.


And while I knew this for quite some time and think the logic is sound, I still didn’t feel like there was enough evidence for this new theory. So now I would like to go back into the Edo period and try to figure out what led to the weird orthography of Junkichi Katayama.


Various Edo Orthographic Developments 色々な江戸の綴りの発展
Fortunately, the Upper Volume contained several references to Edo period works that discuss kanji readings and kana orthography. Here I will go through each one to establish a timeline.


Let’s go back to 1776, when Motōri Norinaga 本居宣長 wrote Jion Kanazukai 字音仮字用格. This work posits a traditional orthography based on Heian era pronunciation. For example, the character 草 has an on’yomi of さう while the character 走 had an on’yomi of そう. By the Edo period, these on’yomi were both pronounced as そう, but Norinaga’s orthography preserves this classical distinction. While this work wasn’t too influential directly on Katayama, it started getting Edo period scholars to reflect on orthography and ancient pronunciations.


In 1815, Zensai Ōda 太田全斎 wrote Kangōnzu 漢呉音図, wherein he provides the go-on and kan-on readings of kanji juxtaposed with their “original pronunciation” 原音, which is a reconstructed Japano-Chinese pronunciation of that character. How to read these squares is explained well in Katayama’s Lower Volume. Here are the squares for 汙 and 紆 followed by how to read them.


Oda 紆Oda 汙HowToReadYomizu.png

This work also contains a gojūon chart which utilizes 于 for /wu/.



As 于 is used in many subsequent orthographies to represent the phoneme /wu/, I believe that tradition can be traced directly back to this work (though perhaps it can be traced even further).


As for the next two references, they can be seen in the Upper Volume as the creators of the orthography which Katayama is using.



Before discussing Gimon Hōshi, I want to recap what I wrote about Hirokage Shirai 白井寛蔭 in the last article. Shirai is considered the founder of extended kana, as he was the first to designate kana for /yi/, /ye/, and /wu/ in his magnum opus On'in Kanayōrei 音韻仮字用例. I would like to juxtapose his orthography with Katayama’s. Their claimed origin characters are adjacent to the kana. Those that are different are highlighted.


Sound 音素Shirai Hiragana’s 白井の平仮名Katayama’s Hiragana 片山の平仮名Shirai’s Katakana 白井の片仮名Katayama’s Katakana 片山の片仮名
/i/い 伊イ 伊イ 伊
/e/え 衣〈 衣Japanese_Katakana_E.png
/ye/𛀏𛀁ye katakanaエ 廷
/u/𛀍ウ 宥ウ 宥
/wu/う 于wa-wu.png于 宇于 紆

As for the differences between the orthographies, I originally posited that Katayama simply chose more common kana. This is demonstrated by how he slightly modified the katakana for /e/ and /ye/, making them much more similar to the very common katakana エ. As noted in my previous article, several other Meiji orthographies also use the same kana as Katayama. Perhaps these kana were just popular in orthographic traditions? Or maybe these other orthographies were inspired by Katayama’s. But if the creators of those orthographies really were inspired by Katayama, why would none use his hiragana /wu/? As of right now, I don’t have a good answer to explain all of these orthographic variations, but in order to not overanalyze, I will assume that Katayama just opted to use more common kana in his orthography. However, this explanation does not explain why 汙 was chosen. Why would Katayama use a kana so obscure for hiragana /wu/? In Japanese, the kanji variant 汚 is much more prevalent than 汙. For the answer to that, I would like to turn to Gimon Hōshi.


Gimon Hōshi: The Final Piece to This Puzzle 義門法師:このパズルの最終のピース
Gimon Hōshi was a kokugaku scholar who lived from 1786 to 1843 and was also the only other named person whom Katayama accredits for his orthography. Gimon is probably most famous for Namashima 男信 wherein he distinguishes /mu/ む from the moraic nasal /-n/ ん. However, he also extensively studied orthography. As noted in Gimon Hōshi’s National Language Study 東条義門の国語学 by Jirō Kameda 龜田次郎, Gimon made a commentary to Norinaga’s Jion Kanazukai, which Shirai says was very influential to his On'in Kanayōrei. Also, Gimon was in the middle of writing his own treatise on orthography, but it was destroyed in a fire. So that leaves us with another work to look for clues, O/Wo Kyōjūgi 於乎軽重義. This work concerns the difference between /o/ お and /wo/ を, a topic that Norinaga also discussed. It is unknown when it was written, but Gimon died before Shirai was born, so it is impossible that Shirai influenced this work during its creation. And the manuscript that I found actually has comments by Shirai in it, so it must have been very influential to him. The first thing of note is that this manuscript actually references Shirai’s katakana orthography in its chapter titles.


Gimon uses Shirai

I do not know whether this was added by Shirai on subsequent reprints, or if Gimon actually created this orthography first because it is only referenced in the chapter titles, not in the actual contents, but I am going to assume that it was added by Shirai in order to avoid over analysis. As for the content of this work, Gimon discusses the Siddham Script, a South-Asian script which was imported to Japan by Kūkai in 806 CE and is the origin of the gojūon chart. The Siddham script distinguishes between long and short vowels for /a/, /i/, /u/, /e/, and /o/. For example, short /a/ and long /aː/ are contrastive. As seen below Gimon notes that this long /uː/ is actually /wu/.


wagyouu vs agyouu 2wagyouu vs agyouu 1

As seen below, the character that he chooses to represent /wu/ is 汙.


Using 汙 for wu

This makes sense as 汙 was pretty popular for representing the Sanskrit long /uː/ phoneme in Chinese transliterations of the Siddham script, as exemplified by Shu-Fen Chen in Vowel Length in Middle Chinese Based on Buddhist Sanskrit Transliterations.

悉曇の長音「う」の音素を中国の悉曇の翻字で表すために「汙」が一般的なのは理解できる。これは陳淑芬の『仏教悉曇翻字基づいた中古音の長短母音』Vowel Length in Middle Chinese Based on Buddhist Sanskrit Transliterationsで示されている。

sanskirt chart

Gimon also provides an explanation on why he doesn’t use 汚, namely that 汙 is more “closed” than 汚.


Why not 汚2Why not 汚

Moving back to On'in Kanayōrei, Shirai explains why he does not use 汙 or 烏, which is from the same rime, as the basis for his kana. This is because their go-on is う and their kan-on is を, which might cause some confusion.


Why not 烏

So overall, Katayama’s usage of 汙 for /wu/ can be seen a direct reflex of the works of Gimon Hōshi, which ultimately comes from the Chinese tradition of writing the Siddham Script with Chinese characters. I suspect that Katayama made Shirai’s /wu/ hiragana う into his /u/ hiragana because it was more common; then, Katayama chose a kana based on 汙 for his /wu/ hiragana in order to follow the tradition, started by Ōda, of using the 于 radical for /wu/. This would explain why Katayama chose 汙 over Shirai’s /u/ hiragana 𛀍 or another kanji in the same rime such as 烏.


Conclusion 結論
Looking back, I think I was correct to be skeptical of only looking at the shape of Katayama’s kana to deduce its origin character. While this may suffice for Heian era kana, by the Meiji era different orthographic traditions based on different rationales led to some highly improbable conclusions on origin characters. For example, the origin character for ウ is nowadays regarded as 宇 and not 宥 as Katayama states. However, when Katayama expressed /we/ with 汙衣, I knew that my old theory could no longer stand. I’m glad that I could establish a timeline on the usage of 汙 for orthography, and how through Gimon, Katayama would come in contact with it. In addition, the kana shape of hiragana /wu/ matches perfectly to the cursive style for 汙. I think the evidence suggests that the origin of Katayama’s /wu/ hiragana is 汙, and thus, I retract my pervious theory.


However, my research in this area has only just begun. How did Katayama’s orthography impact the other orthographies of the time? What exactly is the contents of Gimon Hōshi’s commentary to Norinaga’s Jion Kanazukai? How do Edo Orthographic studies relate to those of the Heian period and those from China? What little clues am I missing in the manuscripts which I have already read? Fortunately, I intend to keep on researching these orthographies, and I will keep you in the loop if I find anything notable. And who knows, maybe I’ll be seeing you next time in my reply to my reply~






Re: Is it a katakana?


I have sent a response to the email you provided!

> Just asking, is the wu on the picture katakana? if yes, can I please upload it in Wikipedia?