by Keiko Imaoka
The following essay was first published in my journal Woodnotes (#29, Summer 1996, pages 27–33) after being posted online to the Shiki haiku discussion list on 30 December 1995, where it received immediate and enthusiastic praise. I consider this essay to be essential reading for anyone interested in learning haiku, especially if they are under the impression that haiku needs to be in a pattern of 5-7-5 syllables in English. Please also note the new postscript at the end, which explains why this version of the essay is the definitive version compared with other versions also available online. Also at the end is Keiko’s short “Beginner's Mind” essay, also first published in Woodnotes, which explains how she got involved in haiku poetry. Keiko Imaoka passed away in April of 2002.
Japanese haiku have been traditionally composed in 5-7-5 syllables. When poets started writing English haiku in the 1950s, they adopted this 5-7-5 form, thinking it created a similar condition for English-language haiku. This style is what is generally considered “traditional” English haiku.
Over the years, however, most haiku poets in North America have become aware that 17 English syllables convey a great deal more information than 17 Japanese syllables, and have come to write haiku in fewer syllables, most often in three segments that follow a short-long-short pattern without a rigid structure. This style is called by some “free-form” haiku. In this essay, I will discuss the linguistic circumstances that necessitate shorter English haiku to be more loosely structured than Japanese haiku.
The 5-7-5-syllable rhythm in Japanese haiku is not the matter of arbitrary choice that it may appear to be to a non-Japanese haiku writer. Various combinations of 5 and 7 syllables have dominated the Japanese literary scene for most of its history, tanka (5-7-5-7-7) being the most prominent example. To most Japanese, words phrased in these configurations have a remarkably mnemonic—at times haunting—quality, so much so that many war and political propaganda have used this form:
hoshigarimasen (7) katsumadewa (5):
“we want nothing till we win (the war)”
kono dote-ni (5) noboru-bekarazu (7) keishichou (5):
“Do Not Climb This Levee—The Police Department”
Likewise, many Japanese aphorisms and proverbs as well as song lyrics, including translations, take similar forms:
owariyokereba (7) subete yoshi (5):
“All’s well that ends well”
hotaru-no hikari (7) mado-no yuki (5):
“the light of fireflies, snow by the window”
[lyrics corresponding to “should auld acquaintance be forgot (8)
Because of these rhythmic structures, Japanese haiku and tanka can be memorized with little or no effort, which is one of the major reasons for the popularity and longevity of these literary forms. On the other hand, there is no such inherent mnemonic quality to 5-7-5 English haiku, which are indeed relatively difficult to commit to memory. Moreover, there is no discernible rhythmic structure to such an arrangement, due to the disparate length of English syllables. (The mnemonic quality of 5-7-5 Japanese phrases is much closer to that of metered rhymes in English.) These factors combined with the fact that English carries significantly more information per syllable than Japanese, indicate that using the 5-7-5 form does not necessarily provide an analogous condition for writing haiku in English.
This is not to say, however, that all who write in 5-7-5 syllables should stop doing so. I believe that 5-7-5 English haiku as a derivative of Japanese haiku has its place in the world of poetry, just as 5-7-5 Chinese haiku is another such derivative, seemingly containing about three times as much information as a Japanese haiku.
Today, many bilingual poets and translators in the mainstream North American haiku scene agree that something in the vicinity of 11 English syllables is a suitable approximation of 17 Japanese syllables. This length conveys about the same amount of information as well as the brevity and the fragmentary quality found in Japanese haiku. As to the form, some American poets advocate writing in 3-5-3 syllables or 2-3-2 accented beats. While rigid structuring can be accomplished in 5-7-5 haiku with relative ease due to a greater degree of freedom provided by the extra syllables (in English), such structuring in shorter haiku will have the effect of imposing much more stringent rules on English haiku than on Japanese haiku, thereby severely limiting its potential.
Two major linguistic factors make the Japanese language more flexible—and thus easier to fit into a rigid form such as 5-7-5. Both of these factors derive from the fact that the grammatical units in Japanese are largely independent, and are relatively free to move about within a sentence.
The English language owes much of its grammatical simplicity to the fact that the word order plays a major role in determining the relationships between words and phrases (subject, object, and so on). In such a language, words and phrases cannot be moved about freely without changing the meaning of a sentence. For example, within a sentence such as “Mother gave it to the kitten,” the words cannot be rearranged without altering the meaning.
In the Japanese language, however, because of the presence of grammatical particles (joshi) that are suffixed to nouns and mark their syntactic relationships, word units become independent and can be moved about more freely within a sentence or a clause without affecting its meaning. As a result, the preceding sentence can be rearranged in many ways in both spoken and written Japanese without altering its core message.
1. haha-ga koneko-ni sore-o ageta mother/to the kitten/it/gave
2. haha-ga sore-o koneko-ni ageta mother/it/to the kitten/gave
3. sore-o koneko-ni haha-ga ageta it/to the kitten/mother/gave
4. sore-o haha-ga koneko-ni ageta it/mother/to the kitten/gave
5. koneko-ni sore-o haha-ga ageta to the kitten/it/mother/gave
6. koneko-ni haha-ga sore-o ageta to the kitten/mother/it/gave
Furthermore, if I were to add the word “yesterday” to the sentence, the available options in English would be “Mother gave it to the kitten yesterday,” and “Yesterday, mother gave it to the kitten.” In each of the above six Japanese versions, however, “yesterday” (kinou) can be inserted anywhere there is a space (including the beginning of the sentence), except at the very end, without significantly altering the meaning, thus multiplying the number of options by a factor of four. Therefore, “Mother gave it to the kitten yesterday” can be expressed in Japanese in 24 (6 x 4) different ways. Some of the alternatives are smoother than others, of course, but the least smooth is still as natural as “Yesterday, mother gave it to the kitten.” Here, I will only elaborate on #1 to demonstrate the four possible alternatives to “Mother gave it to the kitten yesterday.”
1. haha-ga koneko-ni sore-o ageta + kinou:
kinou haha-ga koneko-ni sore-o ageta
haha-ga kinou koneko-ni sore-o ageta
haha-ga koneko-ni kinou sore-o ageta
haha-ga koneko-ni sore-o kinou ageta
In other words, noun phrases and many adverbs in the Japanese language can be placed almost anywhere within a sentence except that the verbs or verb phrases (including negations) must come at the end. In more complex sentence structures than the one in the above example (for example, two or more sentences joined by conjunctives), the same degree of freedom holds within each clause. While one option may be superior to others in poetic terms, having a multitude of alternative expressions at one’s disposal allows a Japanese haiku poet far more freedom within the 5-7-5 structure than is available in English.
Moreover, in Japanese, some of the grammatical particles and the subject of a sentence can be omitted depending on the context, which gives the language further flexibility.
This relative independence of grammatical components also results in the ease in dividing a phrase into 5-7-5. In the above example, the six Japanese versions of “Mother gave it to the kitten” can be segmented anywhere there is a space. Therefore, if they were to occur within haiku or tanka (though very unlikely, since they are so unpoetic!), there are three equally plausible locations where each sentence can be divided, whereas, in English, “Mother gave it to the kitten” offers fewer options. Likewise, in the case of “Mother gave it to the kitten yesterday,” each of the 24 possible Japanese sentences can be divided wherever there is a space (four locations).
Thus there are more places where a Japanese phrase can be divided without disrupting its meaning. If English had the same degree of segmentation flexibility as Japanese, the following haiku of my own,
across the arroyo
of a joy ride
could be rewritten to approximate the 3-5-3 form as
arroyo, deep scars
of a joy ride
without affecting the meaning. As it is, doing so sacrifices too much in the natural flow of words and interferes with the image. Because Japanese haiku are written in a single (vertical) line, with no spacing (breaks) between the segments, there is no danger of disrupting the flow in this manner. It is merely an artifact borne of the linguistic differences between the two languages and of the three-line convention of English haiku that makes the former appear as if it does not have a classical form. The type of unnatural line breaks seen in the latter is a problem associated with the 3-5-3 (or other short) form, whereas the 5-7-5 form is long enough to accommodate natural line breaks dictated by the English grammar, due to a greater degree of freedom provided by the extra syllables.
Thus we are in a bind, a Catch-22. If one wishes to have the brevity and the fragmentary quality of Japanese haiku in English haiku, 17 syllables is too long. On the other hand, if one desires a rigid structure, 11 syllables is too short. One must choose between the two. The choice depends on which of the two factors a poet considers more important to haiku. The majority of contemporary English-haiku poets have let go of the tight forms in favor of brevity to develop the mainstream North American haiku.
As demonstrated previously, 5-7-5 segmentation is not a division based on content as we think of it in English. Strictly in terms of content, classical Japanese haiku are composed of two major parts of varying lengths, such as 5-12, 12-5, 8-9, 9-8, 7-10, and 10-7, in the generally decreasing order of prevalence, with the first two being the most prevalent. Here are some examples from the great masters (with my literal translation):
yuku haru-ya (5) tori naki uo-no me-ni namida (12)
birds cry, tears in the eyes of fish
neko-no meshi shoubansuru-ya (12) suzume-no-ko (5)
sampling the cat’s food—
a baby sparrow
ware-to kite asobe-ya (9) oya-no nai suzume (8)
come play with me—
you motherless sparrow
uguisu-no naku-ya (8) chiisaki kuchi akete (9)
[uguisu: a nightingale-like bird]
with the small mouth open
A close observation of “free-form” English haiku reveals that they are composed of two major segments. The majority of them are divided after the first or the second line and the rest near the middle, and thus they are in accord with the underlying structures of classical Japanese haiku.
In writing short English haiku, the decision as to where the division falls is based mainly on the dictates of English grammar and the poetic merits of given expressions. To limit short haiku to those that can be fitted into a rigid three-part structure is to severely limit the type of ideas that can be expressed in this style.
Besides the two linguistic considerations and the varying underlying structures already described, yet another factor adds freedom and flexibility to contemporary Japanese haiku. Although a majority of contemporary Japanese haiku are still written in the classical 5-7-5 form, a significant number of them make use of what is called ku-matagari (segment-straddling), where a word straddles two segments. Many haiku that appear to be 5-7-5, and can be read aloud as such, are actually 7-5-5, 8-4-5, 5-9-3, 5-8-4, or some other variation. This technique is more frequently used by the poets in the avant-garde schools, and only those schools seem to allow more than one straddling within one haiku.
Although the popular use of ku-matagari is a relatively recent phenomenon, I have come across some examples in a comprehensive collection of Issa’s haiku. Here is one example:
dou owaretemo (7) hitozato-o (5) watari-dori (5)
migrating birds still
fly over towns
In English haiku, it is simply not conceivable to break a line in the middle of a word just to serve the form. The reason why ku-matagari is feasible in Japanese haiku is because the 5-7-5 rhythm is rooted so deeply in the Japanese psyche that the readers are able to keep track of the form despite the lack of a break where one is expected, as well as because of the aforementioned fact that the Japanese haiku are written in a single line with no spacing (breaks between the parts). If they were written in three lines as English haiku are, an awkward situation would result when having to split a word across two lines.
In short, writing within the rigid structure of Japanese haiku is made possible by the remarkable malleability and redundancy of the Japanese language that allows for a multitude of options in expressing a single thought. In languages such as English and its relatives whose grammars depend heavily on word order, haiku must and will take a much different form from haiku in Japanese. By concerning ourselves too much with the outward form of haiku, we can lose sight of its essence.
The issue of Woodnotes in which Keiko Imaoka’s essay first appeared in print (#29, Summer 1996) was distributed in early July of that year. On 25 July 1996, Keiko sent me the following email message:
I just got the summer issue of Woodnotes yesterday. What a delight to receive a hand-written note from you! Yes, I was absolutely thrilled to see my article among the fine poems and commentary. Thank you so much for giving it space in your magazine, and for your high regard for its significance on the subject. It never occurred to me to seek publication before or after I wrote the article. All I was concerned with was to convince Shiki folks of the validity of the “free-form” [that is, non-5-7-5] English haiku. It quadruples my joy if my article could contribute in a small way towards freeing the mainstream haiku poets of excessive concern over form. Thanks again for your generosity.
P.S. I enjoyed your notes on “Japan-Think, Ameri-Think” on page 44. It gives me a sense that we are moving on to the more important issues (than forms) in thinking about English haiku.
In her P.S., Keiko is referring to a page of comments I made about cultural differences raised by the book Japan-Think, Ameri-Think, by Robert J. Collins (Penguin, 1992). I was very pleased to publish her essay in Woodnotes, and consider it to be the most important essay I ever published in all of the journal’s 31 issues. In my “Note from the Editor” for issue #29, I wrote the following (page 2):
I want to draw your special attention to Keiko Imaoka’s excellent article on “Forms in English Haiku” with its cogent explanation of differences between English and Japanese that show the limitation of so-called “traditional” form in English. I consider this a definitive article on the subject and hope you will find her article—and indeed, this entire issue—a pleasure to read.
An additional comment I would make about this article is that the author uses the phrase “English haiku,” but it would be more appropriate, I believe, if she had said “English-language haiku.” This is an edit I regret not suggesting before the essay appeared in Woodnotes. Keiko lived in Arizona, but was born and grew up in Japan, and may not have been sensitive to the distinction, sometimes asserted in the worldwide haiku community, that “English-language haiku” refers to all haiku written in English, whereas “English haiku” can be taken to refer to haiku written by people in England.
Please also note that the preceding essay appears in at least three locations online, reproducing essentially the author’s original and unedited posting to the Shiki discussion list. The version published in Woodnotes, reproduced here, contains a number of my edits, approved by Keiko Imaoka, making this the author’s final approved version. These edits, among others, include the refinements of saying “classical” instead of “classic” and “fragmentary” instead of “fragmented,” the revisions of which both have subtle but significant differences of meaning.
The following is my memorial poem written for Keiko when she died, in 2002, by her own hand:
a wrenching in my chest—
the white peony
pulled from the garden
You can read additional memorial poems for Keiko Imaoka, as well as samples of her work (many of which I compiled from Woodnotes) at the Aha Poetry site.
—7 January 2011
I cannot be sure when I first became aware of haiku and
tanka in my childhood in Japan. They seemed to have existed for a long time in
the perimeter of my awareness, undifferentiated from proverbs, mottoes,
aphorisms, and song lyrics that were phrased in similar forms. Sometime during
my grade school years, Ogura Hyakunin-Isshu (Ogura Collection of One Hundred
Tanka, edited by Teika Fujiwara around 1235) became known to me as a New
Year’s card game, in which players compete to capture shimonoku cards
(100 cards on each of which the last half of a verse is printed, spread out on
the floor in front of the players) that finish the verses being read aloud. At
abacus school, where we played this game at every new year’s party, my prowess
in the game improved dramatically when I was in the sixth grade, after I had
memorized all the poems with my tenth-grade sister who was required to do so in
her archaic grammar course in school. I had learned to recall each poem by the
first few syllables, which enabled me to locate the last parts of the verses
 Although these proverbs correspond almost word for word, there were most likely independently conceived, since the idea expresses a universal human sentiment.
 Here, two of the Japanese syllables span two notes. These are the first words of the lyrics sung to the tune of Auld Lang Syne by Japanese school kids at the time of graduation. The lyrics are inspired by an old Chinese poem. Incidentally, we all grew up thinking this and many other Western tunes were written and composed by Japanese, including Home Sweet Home, Glory Glory Hallelujah, Oh My Darling Clementine, I’ve Been Working on the Railroad; the list goes on. The Japanese lyrics tend to be totally different in content from that of the original. This is because the lyrics have been entirely rewritten because literal or even figurative Japanese translations simply cannot be fitted into the tune. The Japanese lyrics are well integrated into the tunes, and are sung with one or more syllables per note.
 kaku-joshi (case particles): -wa, -ga, -mo, (-no): subjective case; -wa and -ga are often abbreviated -o, -ni, -e, -to, -kara: objective case; -o is often abbreviated -no: possessive case. Example using kare: kare-wa: he; kare-ga: he; kare-mo: he too; kare-no: he (in archaic grammar); kare-o: him; kare-ni: him; to him; by him; kare-kara: from him; by him; kare-e: to him; kare-to: with him; kare-no: his.
 Ku-matagari is used more extensively in contemporary Japanese tanka.
Further Reading >