Wednesday, December 26, 2007

Further Thoughts on Lineation in Tanka

My article in the current issue of MET (#6), scratches the surface of the subject. I had not yet figured out for myself exactly what I was trying to get at when I wrote it, but went ahead with it on the assumption that if I stuck my neck out, somebody would respond by trying to chop it off, and that this would advance our understanding of the subject. I was right, somebody did, and it has. Here then is the response I made to him:

Japanese doesn't have meter; it does not stress or accent syllables. There is no variation in sound that determines line length, and in fact, spoken Japanese does vary from its written form. What the Japanese count are the kana, the written form. Thus, a word like 'desu' is written with two kana, and is counted as two kana, but is pronounced as 'dess', one mora. That the pronunciation varies from the written form is of little concern to the Japanese.

When dealing with classical waka, we never know what the originals are because old Japanese had many sounds that have disappeared from modern Japanese, and waka are always translated into strictly formal modern Japanese, which eliminates any irregularities. Thus, when we see Saigyo in Japanese romaji, we are seeing a modern Japanese translation of what Saigyo actually wrote. It is very rare -- and quite befuddling -- to confront the romaji inscriptions of classical waka. Translating from classical Japanese to modern Japanese is just as difficult and perhaps more so than translating from Anglo-Saxon to modern English. When we read Beowulf in English, we are not reading what the bard wrote. Yet if we are to confront the bard's word directly, we can make very little sense of it. So in reality, all our supposed adherence to the aesthetics of classical waka is complete poppycock: we are adhering to some translator's notion of what old waka were like. As far as I know, no one has ever made a direct translation from classical Japanese to modern English. Anything passed through a double layer of translation must depart far from the original.

Yet even in the strictest period of waka writing, under Ki no Tsurayuki and his influence, waka might vary by a syllable or two. When I was first writing tanka and adhering to the 5-7-5-7-7 format, I had a good friend from Japan who was a published poet. She constantly nudged me to not worry so much about counting syllables, telling me the Japanese didn't mind if it was off by a bit.

In the more flexible periods, eg, the modern era and the earliest eras, there was/is considerable flexibility. For example; the ancient form appears to have been alternating long and short lines, with lines varying from 3 to 9 syllables. When it was adopted by the courtiers, they seem to have settled on 5 and 7. Such a pattern is now seen as 'inherent' in the Japanese language, but it is long custom that has made it so. And yet, as noted, variation was still permitted.

I speculate that the Japanese settled on 5 and 7 for reasons having to do with Chinese influence, which saw the world made up of as sets of five -- five musical notes in the scale, five elements, five colors, five directions, etc, and natural harmonics, eg, the seven note scale, and so forth. Gagaku, the classical music of the same period as waka, is quite eerie to Western ears. It's based on the five note scale.

What sounded musical to a Japanese courtier of the Heian period will set the Westerner's nerves on end -- if you dislike bagpipe music, then gagaku will cause real suffering! I do not think we can import the Japanese classical musical system to our tanka and win any admirers. The Japanese model simply does not translate, either in syllables, rhythm, or sound.

See A Waka Anthology, Vol 1, for more details on ancient Japanese prosody.

Another thing to keep in mind is that tanka was not the only form, merely the most popular form. All kinds of forms were in used in ancient Japan, and the Japanese kept inventing new ones, like the ko-uta ('little song'). Thus, a thing that did not adhere to the tanka form might appear as a sedoka or ko-uta or Buddha's footstep poem or one of the many other forms. So the tanka poet was not confined to the 5-7-5-7-7 form; if a poem needed to be something else, it could be. Hell, even the acrostic survived in the courtly anthologies. When we fixate on tanka, we are ignoring the context which provided many more options to the Japanese poet than just 5-7-5-7-7.

Alas, we hear next to nothing about other forms. Thus, if I present a six line poem in a tanka context, the great majority of readers are not going to recognize it as a sedoka, they're going to think it's a defective or experimental tanka. If I use four lines, they aren't going to ask if I'm trying to do a Chinese quatrain or a ko-uta, they are going to see it as a radical tanka.
A Japanese reader would have no such problem.

Japanese rhythm is built on phrases, rather than words or lines. Generally speaking, waka have a single phrase per line. Japanese grammar being what it is, particles of various sorts demark groups of words as belonging together. In Japanese, these words are run together with no space between them, and are pronounced without pauses. This is very different from English with space and pauses between each word, no matter how important or unimportant they are. Thus a Japanese auditor can easily tell where the line breaks fall -- there is a pauses in the flow of sound. How many syllables are in that flow of sound doesn't matter. They aren't counting. They're registering the pauses between groups of syllables, along with certain grammatical markers, to know when a phrase ends and the next begins. English-language tanka is not built on phrases at all, which is why we can have enjambment and single words on lines, and other things. Our language is simply too different.

I could, if I was to imitate the Japanese rhythm, write tanka by eliminating the spaces between words. Thus the lineation would be completely irrelevant. Consider this:

thewoods thatseemedimpenetrable insummer arehollowin December'swind

Voila. A Japanese tanka in English. Written on one line with the breaks appearing at the ends of phrases. Counting syllables or morae is completely irrelevant; the breaks between phrases are obvious. Thus it becomes clear why the Japanese need not be too exacting when counting syllables. It also becomes clear that the need to take a breath is going to break the poem into 'utterances' if we can use that term to refer to the group of words sung between breaths. In poem of these size, two or three breaths is natural to an ordinary speaker; to utter a poem of this length in one breath or five is out of the ordinary, but perfectly feasible. To manipulate the breathing places is part of the poet's technique.

The above poem's breathing pattern is:

thewoodsthatseemedimpenetrableinsummer arehollowinDecember'swind

Such issues are given no attention by poets in English, except for a few who mention 'one or two breaths.' There is no discussion of the impact that such breaths give to prosody, or how such breaks might be exploited for literary benefit, or how they balance one another, or any such thing. The break at the end of L3 does have the virtue of dividing the poem into two roughly equal halves. Such a break is certainly easily accomplished by a singer with no special training. But why limit ourselves to what is simple and obvious?

When you get right down to it; the vast majority of tanka poets and readers working in English -- including most of the well known ones -- haven't a clue what really makes a Japanese tanka work, or what the equivalent is in English. As long as we keep on with our absolutely meaningless rubric of 'five phrases on five lines', we haven't a chance of learning anything about either Japanese tanka or English-language tanka. My article on lineation is primitive, I know, but I operate on the theory that sticking my neck out may provoke discussion, education, and improvement.

I thank you for raising the issue. I had not quite figured out for myself in any conscious way what I was getting at in 'Alternate Lineation,' my thoughts on the matter are much clearer now.



  1. Dear Kei:

    I read your article in ME6 on lineation and your blog post as well. If I understand it correctly, your blog post is a response to someone who wrote to you regarding your ME6 article, and your blog post is, in turn, a response to that person's critique. I haven't been able to locate the article to which the blog post responds. But here are a few observations anyway:

    It seems to me that English language poets interested in Japanese forms have mystifeid the Japanese language and created a host of obfuscations by concentrating on microaspects of the language. My basic view is that Japanese is just an ordinary language, that it is not unique or special and doesn't require new apparatus or analyses to comprehend it or its poetic culture.

    It is intriguing to me that in all these discussions about Tanka crossing cultures (from Japan to the English speaking world) there is very little (maybe no) attention paid to historical precedents for this kind of transmission. The movement of hexameters from Greek to Latin in the classical world is an instructive precedent, as is the movement of the Sonnet from Italy to other language communities. I think both of these are applicable and useful and could clear at least some of the air.

    Basically, Japanese poetics is syllabic based while English poetics is accentual-syllabic. French poetics are also primarily syllabic (again, Japanese poetics is not unique). Japanese poetics counts syllables alone, while English poetics counts feet and syllables; that is to say groupings of stresses as in "iambic pentameter"; five accentual units amounting to 10 syllables.

    There is such a thing, though, as syllabic rhythm; in English it is mostly defined by grammatical units and phrasing. It is perhaps here that a comparison with Japanese poetics could be fruitful.

    Just some occasional thoughts. I enjoyed both your article and your blog post as well.

    Best wishes,

    Jim Wilson

  2. Jim,

    Thanks for leaving a comment. In a nutshell, the argument over counting is a tempest in a teapot. However, it seems worthwhile to plumb the details as part of a general mastery of tanka. I think that it is not enough to wave our hands and say, "It's tanka because I say it is," which is the temptation at the other extreme.

    Email me if you'd like to correspond further. I know you and APA-Renga, and if you don't know my Bibliography of English-Language Tanka, I hope you'll look it up at and send me your additions and corrections. I have been lately considering adding tan-renga to the bibliography.