New evidence for the development of hiragana?

« previous post | next post »

An article by Tomoyoshi Kubo in The Asahi Shimbun, "Poem on 9th-century wood could provide missing link between kanji, hiragana" (11/27/15), may provide evidence for the development of hiragana (cursive syllabary) from Chinese characters.

…The entire verse of famed tanka poem “Naniwazu” was inscribed in ink on Japanese cypress in an intermediary syllabary between manyogana, one of the earliest Japanese writing systems dating back to the fifth century, and hiragana, the Kyoto City Archaeological Research Institute said on Nov. 26….

The kanji were originally semantic but were read phonetically to suit the Japanese language.

The characters were later simplified and turned into hiragana, but the process of that transformation remains a mystery.

The writing on the latest discovery is neither manyogana nor hiragana, but something in between, the institute said. It is also the first finding of the entire poem written in the intermediary system….

Comments from Bob Ramsey:

I have to say that I found this article on a “missing link” strange and unprofessional. Sure, it was written by a reporter for a completely lay audience, but I’m having trouble figuring out just what it is that apparently has professional Japanese philologists excited. First of all, there’s no such thing as an “intermediary writing system between manyogana and hiragana”. Really, there’s no qualitative difference in any case between manyogana and (hira)gana; everything functional about kana was already present in the styles of manyogana used in at least three books of the Manyoshu, books where Chinese characters were used only as phonograms. As you well know, (hira)gana was nothing more than a further scriptification of the cǎoshū 草書 ("cursive") forms of Chinese characters being used as phonograms, and throughout the Heian (794-1185) period calligraphic practice varied almost at the whim of the writer; manyogana never switched clearly to (hira)gana at any one point.

What I remember from my textbooks is that Japanese philologists rather arbitrarily decided to draw a line along the continuum from manyogana to (hira)gana and call a particular 10th-century version of the Akihagi-jo 秋萩帖 the “first” (hira)gana document. But if you take a look at the text, you’ll see that there’s nothing really different about it from lots and lots of other calligraphic works being written around that same time in the Heian. Sure, the first character ān / an 安 is written in a cǎoshū 草書 ("cursive")  form that looks pretty much like a あ, but you don’t have to go very far down the line of characters to see that most of the characters are still just cǎoshū 草書 ("cursive") .

Boy. And then I see “The kanji were originally semantic but were read phonetically to suit the Japanese language.” How unprofessional can the writeup get?

The Japanese truly did something important by simply starting to use Chinese characters purely as phonograms. Of course, that step had already been taken, in part, elsewhere, including in China (and Korea). But using symbols solely as phonograms without any semantograms mixed in is something that was certainly important for the development of Japanese writing. That practice, however, was already in place in some circles in the Nara (710-794) period. It just became more common in the Heian.

What all of this boils down to is that stages in the evolution of writing systems typically do not occur at a certain moment in time, but rather that they occur over a period of time that may be quite long.  As such, the boundaries between the stages will almost inevitably be blurred.

[h.t. Petya Andreeva]


  1. satkomuni said,

    November 29, 2015 @ 10:13 am

    So, it's. . . not really new evidence for the development of hiragana at all. Right. "False alarm, misleading title, back to work, everyone; sorry to've wasted your time." : ?

  2. Dan Lufkin said,

    November 29, 2015 @ 12:21 pm

    It has to be a general rule that any radical change in communication has to come into force gradually. Think of the guy who bought the first commercial fax machine.

  3. said,

    November 29, 2015 @ 1:01 pm

    Both the cursive visual forms and the phonogram technique are already present in China, and the Japanese versions are direct descendants of them via Korea, down to the same characters being used as phonograms (Bentley has an article on the topic). I think the Japanese innovation was the graphical specialization : for me, modern-style "kana" began when they started writing mixed texts using cursive to signal phonograms, and full forms for morphograms.

  4. Victor Mair said,

    November 29, 2015 @ 2:01 pm


    Thanks for reminding me to replace the question mark at the end of the title. I had it in the draft, but somehow it got lost while posting.

  5. Ross Bender said,

    November 29, 2015 @ 7:46 pm

    Whatever the merits of the Asahi Shinbun article, the Naniwazu poem cited has a very important place in the development both of Japanese poetry and its writing system. Joshua Frydman gave a paper on this poem in a 2012 panel on Nara Japan at the AAS conference in Philadelphia. His Yale dissertation (2014) is titled Uta mokkan: a history of early Japanese poetry through inscription .

    I am not sure what the earliest attested version is, nor the script in which it was written. Below is a modern iteration employing kanji and hiragana.

    Naniwa-zu ni/Sakuya kono haana/Fuyu-gomori/Ima o haru-be to/Sakuya kono hana
    難波津に 咲くやこの花 冬ごもり いまを春べと 咲くやこの花

    Naniwa Bay, now the flower blooms, but for winter. Here comes spring, now the flower blooms.

    In Naniwa Bay, now the flowers are blossoming. After lying dormant all winter, now the spring has come and those flowers are blossoming.

    The poem is traditionally attributed to the scholar Wani, who supposedly came to Japan from Korea in the 3rd century. It was included in the Kokinwakashu 古今和歌集 collection published in 905.

  6. Chris Kern said,

    November 29, 2015 @ 10:02 pm When would that be? Maybe I'm misunderstanding you, but even in Edo-period woodblock prints, the kanji are still usually written in cursive forms in Japanese texts. I don't think it was until Meiji-era printing that you finally got the full, clear separation between kanji and hiragana in terms of the cursive vs. non-cursive styles. (And even then, the cursive forms persist in handwriting.)

  7. Matt said,

    November 30, 2015 @ 12:21 am

    Checking the Japanese language sources, the claim seems to be that this is the first mokkan found that contains the entire Naniwazu poem. I'm not sure if that is a sensible thing to claim or not. (i.e. has it really not been found in its entirety on a mokkan before? Seems possible, sure.)

    I know that in 2007 there was a lot of excitement because a fragment of mokkan had been found to have the first half of "naniwazu" on one side and the first half of "asakayama" on the other, making it reasonable to suppose that the fragment was the top end of a mokkan that, in its original state, had held the full text of both poems. This was a big deal because "asakayama" is also quoted in the Kokinshu preface as one of the progenitor-poems of the waka tradition, AND it appears in the Man'yoshu, AND the MYS text is not in Man'yogana whereas the mokkan was, which suggests some things about how mokkan were used — for example, because parts of the Naniwazu poem appear on so many mokkan fragments, often alongside other bits and pieces of poetry, one theory is that the Naniwazu poem was used for writing practice — but Inukai Takashi has argued that even if this is so the phonographic nature of the rarer "full-poem mokkan" suggests that they were used to display the actual poems themselves, perhaps even as a sort of primitive karaoke lyrics display for festivities.

    Re the "missing link" thing, what Bob Ramsey says is true, you can't draw a sharp line with "cursive man'yogana" on one side and "hiragana" on the other. But that doesn't mean that new examples of cursive script from certain periods can't shed light on exactly how things proceeded. For example, the writing on the mokkan pictured in the news reports linked here is MUCH more like hiragana than the writing on the mokkan that made news in 2007 that I mention above which is a bit cursive but each kanji is still quite identifiable. So maybe this new mokkan discovery suggests something new about the interaction between "pre-hiragana cursive" (if you like) and "consciously writing kanji that can still be read as kanji", for example. I imagine that would be a fairly technical matter for paleographers though ("this characterform was not thought to have developed until 175 years later! this changes everything!") and not really amenable to newspaper reportage.

  8. Ross Bender said,

    November 30, 2015 @ 2:41 pm

    I just dug up Joshua Frydman's abstract for his paper at the 2014 (not 2012) AAS conference in Philadelphia:

    Spell, Song or Salutation? The Naniwazu Poem on Inscribed Objects

    This paper presents a group of inscribed objects from the 7th through 9th centuries that challenge the primarily manuscript-based understanding of early Japanese literary development. From the Asuka (538-710) through early Heian Periods (784-1185), the rapidly centralizing Imperial court found the solution to a wide range of its needs in mokkan, wooden tablets inscribed with ink. Over 400,000 mokkan have been discovered so far, used for everything from marking tax shipments to inter-bureau government memos. A small subset of about 40 objects contains waka, poems in Japanese, most of which predate the Man’yōshū, the earliest extant textual source for Japanese poetry. Intriguingly, almost half of these poem-bearing objects feature fragments of the same waka, a piece known as the Naniwazu Poem. The Naniwazu Poem also appears inscribed on pottery shards, roof tiles, and even the roof beams of pagodas dating from the same time span. However, the poem does not appear in the Man’yōshū, and its first citation is not until the 10th century, where it is mentioned as being basic to literacy education. Scholars have proposed various theories as to why the Naniwazu Poem appears so frequently on material objects, ranging from ritualized to educational uses. I aim to present these inquiries, which challenge narratives of literary history that eschew the relationship between texts and material culture, and use them to explore new methods of exploring the development of early poetry and the technologies of writing which enabled it.

  9. Sharon Goetz said,

    November 30, 2015 @ 4:00 pm

    David Lurie's Realms of Literacy (2011) seems relevant here: publisher's page.

  10. Ross Bender said,

    November 30, 2015 @ 5:18 pm

    Yes, Lurie p. 262 notes that the "earliest currently known version" of the Naniwa Port poem was found on a mokkan at the Kannonji site in Shikoku, and may date back to the reign of Tenmu (672-686) or even earlier. This mokkan has only the beginning lines:


    na ni ha zu ni sa ku ya ko no ha na

    His discussion of the poem is on pp. 261 and ff

  11. Nathan Myers said,

    December 1, 2015 @ 5:09 am

    Should the rise (and even institutionalization, in Unicode) of emojis be taken as a sort of kanjification of English?

RSS feed for comments on this post