The first page of Shiki’s “Rhyme Finder,” beginning with words that end in “a.”

Anthony and I drove to Boston last week so he could check on a construction project. I brought back a carload of books that I’ll need for teaching in the spring and picked up a semester’s worth of mail at the department.  Several issues of the Shiki Kaishi were there waiting for me, kindly sent by the Matsuyama Shiki Society’s president Karasudani Teruo. As I was scanning them to upload here, I took a break to read an essay that appeared in the July issue by Imamura Takeshi, a Shiki scholar whom I admire very much. I had been feeling a little down, but reading Imamura-san’s article cheered me up, so I thought I would share its contents here.

Imamura-san begins his article, which appeared in issue #167 (July, 2020) of the Shiki Kaishi, by reporting that he happened to take down from his bookshelf a magazine issue celebrating the start of construction of the Shiki Museum in 1979.  The article mentioned a few of the future museum’s most significant holdings, among which was the original manuscript of the Rhyme Finder (「韻さぐり」, a rhyming dictionary which Shiki began compiling in January of 1897. Thinking he would “get some energy from Shiki” by writing about this extraordinary manuscript, Imamura-san thought of visiting the Shiki Museum to have a look at the object itself, but then realized that this would not be possible because of restrictions due to the coronavirus. So he asked the curators there to let him know its dimensions and to send him images of the cover and the first page. They were happy to oblige and Imamura-san included the images in his article along with the results of his research on the manuscript, some of which I summarize below along with a few thoughts of my own.

Shiki was thirty years old when he began compiling the Rhyme Finder. His haiku revolution had begun to take hold and he was starting to branch out into other genres including “new style poems,” or shintaishi. Most forms of Japanese poetry do not use rhyme but some shintaishi do, and Shiki seems to have hit on the idea of creating a rhyming dictionary to promote and facilitate the use of rhyme in this new poetic genre.

Looking at the research timeline I’ve been keeping on Shiki, I see that 1897 was also the year when Shiki mostly stopped writing poems in Chinese (kanshi), abandoning a genre he had practiced more or less consistently since age ten. In Chinese, kanshi rhyme. But because the characters are pronounced differently in Japanese,  Japanese poets use rhyming dictionaries to compose them, only for the rhymes they have so painstakingly found to disappear again when the poems are read back in Japanese. For Shiki, the effort to find rhymes was still worth it because the constraint could make for better poems. Imamura-san suggests that Shiki compiled the Rhyme Finder in order to promote the use of rhyme in shintai-shi because of his experience with rhyming in kanshi. It’s as if Shiki’s interest in rhyme transferred from kanshi to shintaishi in this year as he set out to experiment with the potential of rhymed poetry in Japanese.

Characteristically, Shiki went about his new interest in rhyming shintaishi systematically and on a scale a that seems almost superhuman, especially given the state of his health in 1897. Imamura-san reports that the Rhyme Finder contains approximately 11,500 words, written in Shiki’s hand and grouped by final vowel and total syllable count to make it easy to find an appropriate rhyme. The book is hand sewn in the Japanese style and measures 249 mm from top to bottom and 164 mm across. It is 16 mm thick and includes a total of 256 pages.

Because verbs and adjectives in Japanese have a limited number of grammatical endings, rhyming has never really taken root in Japanese poetry. It’s just too easy. But nouns offer a more diverse collection of final syllables, so one way around the problem of the too-easy rhyme is to write poems in which the lines end with nouns.  Imamura-san surmises that this is why 80.8 % of the 11,500 words in Shiki’s Rhyme Finder are nouns.

Imamura-san cites and discusses several examples of Shiki’s rhyming shintaishi. Here is the opening of “On Visiting Kohaku’s Grave,” a new-style poem Shiki wrote to mark the second anniversary of the death of his cousin Fujino Kohaku, who committed suicide in the spring of 1895.

Why didst thou desert the world

When the world did not leave thee

And nor, my friend, did we?

Twas thee, alas, who deserted me.


Nani yue nare wa  / yo wo suteshi

Ukiyo wa nare wo / sutezaru ni

Warera wa nare wo / sutezaru ni

Nare wa ware wo zo / mi-sute-nishi


As you can see, in Japanese, every line ends in an “i” sound (pronounced “ē”). I tried to approximate the rhyme and the poem’s pseudo-archaic register in the translation. Rhyming poetry like this did not catch on in Japan, despite Shiki’s efforts. But what I find interesting is less his success or failure at promoting rhyme than the openness this chapter of his career shows toward different genres, forms, and poetic techniques. Shiki published this poem in the Kohaku-ikō, an 1897 collection he compiled with other friends in Kohaku’s memory. It included Kohaku’s own writings along with a biographical essay on Kohaku by Shiki and eulogies in poetry and prose by many of their friends.

Imamura-san ends his article with a few “gleanings” he has harvested while reading through Shiki’s Rhyme Finder.  In one of these, he notes that Shiki includes the word “Western Dog” (洋犬) among the rhyming words. “Western dog” is usually pronounced “yōken,” but Shiki has glossed it to be pronounced “kamé” and put it in the section for words ending with the syllable “mé.”   Why can “Western dog” be pronounced “kamé?” According to one theory, when Meiji Japanese saw English-speaking dog owners call their dogs by saying “Come here!” the sound they heard was “kamé-ya.Shortened to “kamé” in due course, the word came to refer to any non-Japanese dog.

Shiki reveled in linguistic variety and recorded all kinds of words in the “Rhyme Finder” and all sorts of other glossaries and lexicons he created. Together they are a valuable record of Meiji-period language usage and pronunciation. Thinking about  Shiki compiling this vast collection of words from his sickbed and of the passion that continues to inspire scholars like Imamura-san to write about Shiki’s work with such care and attention, my spirits lifted.



1 Comment

  1. Sophie Sheng

    Fascinating read. I never thought of how the phonetic patterns like Hyosoku would translate into Japanese Kanshi. When we were learning Tang poems in modern mandarin, we would read certain characters differently too. Chinese phonetics changed so much during the years and we would borrow from dialects like Cantonese or Shanxi accent to fit the rhyme. One of the examples that come to mind is 杜牧《山行》,

    Yuan shang han shan shi jing xie/xia
    Bai yun shen chu you ren jia
    Ting che zuo ai feng li wan
    Shuang ye hong yu er yue hua

    Here the 斜(xie2) is often taught as Xia2 to match its old pronunciation of Zsia for the rhyme.
    These were so confusing when I learned them as a child. I wonder how Japanese poets like Shiki navigated all these complications.


Submit a Comment

Your email address will not be published. Required fields are marked *