I have always attempted to develop rules (for myself anyway) governing the art of translating poetry, but few fit all cases, so I often throw them out and play it by ear.

The Art of Translation


A number of years ago several of our Japanese-related journals carried an ongoing debate on the art or techniques of translating the prose literature of Japan. Some of these manifestos and arguments often degenerated into a subtle, or not so subtle, academic name-calling. But two distinct groups did emerge. One was called, by the other perhaps, the “literalists” (those who strove as much as possible to remain faithful to the original syntax), and the other side was branded as the “libertines” (those who moved away from the original syntax in an attempt, they felt, to present the reading public a work in English that conformed to their own standards of literature).

But when dealing with poetry today, not much has entered the journals in the way of academic argument. There are, however, probably two or three hundred books on the market or in libraries today that present collections of Japanese poetry in English translation. These books range from Peter Pauper Press collections of haiku to published dissertations on specific poets or works of poetry. Many of these books contain, in the preface or introductions, some statement of the translator’s view of translation. By looking at both the statements of the compilers and at the translations themselves, we find it relatively easy to apply the “literalist/ libertine” models to a number of the collections. But for the most part, they seem to fall in the middle of the road. Most translators of poetry attempt to remain faithful to both the original poetry and their own views of English language verse. Yet faithfulness in any human endeavor can be a relative thing. One person’s view of English-language poetry may be limited to rhyme or even to thumping couplets of iambic pentameter. Another translator may insist upon rendering even the most structured poem in the original in free verse statements. Let us look at a few translations of a single well-known haiku. Matsuo Basho’s (1644-1694) famous frog poem reads in Japanese as:

Furuike ya
kawazu tobikomu
mizu no oto

This has been translated in a number of ways:

The old pond;
A frog jumps in,—
The sound of the water.’

old pond—
and a frog-jump-in
water sound.2

The old pond aye! and the sound of a frog leaping into the water.3
Into the calm old lake
A frog with flying leap goes plop!
The peaceful hush to break.4

The old pond!
A frog lept into—
List, the water sound!5

A lonely pond in age-old stillness sleeps. .
Apart, unstirred by sound or motion. .
Suddenly into it a lithe frog leaps.6

Into an old pond
A frog took a sudden plunge,
Then is heard a splash.7

Old garden lake!
The frog thy depth doth seek,
And sleeping echoes wake.8

An ancient pond!
A frog leaps in;
The sound of the water!9

An old-time pond, from off whose shadowed depth
Is heard the splash where some lithe frog leaps in.10

The ancient pond,
A frog has plunged—



An old pond—
The sound
Of a diving frog.14

The old green pond is silent; hear the hop
Of a frog plumbs the evening stillness: plup!15

An old pond
A frog jumps in—
Sound of water16

old pond
a frog in-leaping

The old pond—
A frog leaps in
And a splash;18

mossy pond;
frog leaping in—

When the old pond
Get a new frog
It’s a new pond.20
[“. . .to express the concept of satori.”]

The bungling frog
Leaped for the pond, but landed
in Basho’s brain.21
[“… translated as a Zen koan.”]

But let me become more personal and speak of my own views of translation. In the two decades or so that I have been involved with the translation of Japanese poetry, I have found myself evolving slowly from the libertine position to the attitude of the literalist. I am not quite all the way yet. The logical conclusion of the literalist movement, I believe, would be pidgin English, and I hope that I have not gone quite that far. But as I now view my earliest attempts at translation, I find that my more libertine views were not based so much on any kind of philosophical or literary argument, but simply on a lack of skill. Not that I was ever one of the worst of the libertines. Since Japanese poetry does not rhyme, I did not add rhymes. I did not add a lot of my own adjectives to a translation thinking the poem would be considered too stark without them. I did not add passages of my own poetry to my translations in hopes of improving a good thing. I did not repeat one or two “good” lines over and over in an attempt to give a translation some kind of refrain that did not exist in the original. I did not take a highly structured poem like the haiku and expand it into four, five, or six lines of Victorian or even free verse.

In the early 1960s I translated and published a lot of Hagiwara Sakutaro, one of the founding fathers of modern Japanese poetry. I presently prefer Hiroaki Sato’s versions to my own older ones. Hagiwara Sakutaro often worked in long flowing lines of free verse, an example would be:

Aogeba takaki matsu ga eda ni koto kakenarasu

A single line that could be translated something like:

Gazing up I see a koto being played in high pine branches.

Of course, I have added the personal pronoun “I” to this line since the verb aogeba (looking or gazing up) has no subject in the original. I have taken some liberties with the matsu ga eda by rendering it as “pine branches” since there is no indication as to number in the original, as is the usual case with most Japanese nouns. But in the earlier years of my career, I had difficulty in maintaining any sort of rhythm in such a long line. I found it easier to break long flowing lines into shorter choppier ones. So consequently, one of my earlier translations was published as:

Looking up I see
In the high branches of a pine tree
A lute that’s being played.

(I see that I have taken liberties, too, with that marvelous Japanese koto and magically turned it into a lute.) But I am no longer very happy with my choppy unfaithfulness. Throughout the sixties, however, this lack of fidelity was rewarded in a strange way. Shortly after I began publishing such translations in literary magazines and journals, I found that I was being paid by the line. Oh, sometimes as much as fifty cents or one dollar a line. Weakness or not, I found that I could get maybe twelve dollars out of an eight-line poem in the original. More than one bottle of sake was purchased out of a stretched-out poem of Hagiwara Sakutaro. (But I feel he would have understood.)

All in all, I have found no ideal solution to the literalist / libertine problems confronting the translation of Japanese poetry. I stress in my translation classes at Antioch that translations of poems are basically opinions: the first opinion involves the original poem (What do I think this poem is all about anyway?), and the second opinion involves a personal aesthetic judgment as to what that poem should look or sound like in English. I try to be both literal and poetic (according to my ear anyway) at the same time. But being faithful to two principles is like being faithful to two lovers. At any one moment, you are bound to be a little too close to one side to suit the other. It is a case of being damned if you do and damned if you don’t. The scholars always complain that you are too far removed. The poets nag you about lack of affection.

Over the years I have always attempted to develop rules (for myself anyway) governing the art of translating poetry, but few, if any, of my rules have fit all cases, so I often throw out the rules and play it by ear. I do, however, have several guidelines that I follow, for the most part:

• I attempt to totally understand the original poem. Now this may seem like a silly guideline. “Of course,” you may say, but remember that a number of the “best-selling” translators of Japanese poetry do not read a word of the Japanese language. These poets work for the most part from literal drafts provided by Japanese scholars. But I read a poem over and over until I feel that I have actually gotten inside the poem from a Japanese point of view. If something is fuzzy for me, I ask a native speaker, preferably the poet himself or herself if I happen to be in Japan. This does not always help, however. Once when discussing a difficult line with Tamura Tyuichi, he looked at me puzzled and said, “I thought I knew what that line meant when I wrote it, but I don’t understand it myself now.” Also, I ask poets like Tanikawa Shuntaro to tape readings of their poetry for me. This aids me greatly in trying to establish moods, nuances, and rhythms of the original poem.

• I never add any more to a poem than is absolutely necessary to make sense in English. I add subjects to verbs, for example.

• I try to follow the original line order when translating any poem. The poet had a reason for presenting the images of a poem in a certain order, and I try to respect that order. It is much harder, of course, to try to repeat the word order in a given line due to the S.O. V. (Subject, Object, and Verb) syntax of the Japanese language as opposed to the S.V.O. structure of English. But I do try to keep the line order intact. Some poems, especially the more traditional forms like the haiku, are, in structure, similar to jokes. You just do not give away the punch line early when telling a joke, and it is not good to give the last line away in translating poetry either.

4. This guideline is used only when working with traditional verse forms. It deals with syllables. When I translate a seventeen-syllable haiku or a thirty-one-syllable tanka, I try to capture the original syllables of the poem. I say try. This is something akin to trying to work international crossword puzzles, and I find myself rather harmlessly elated when I get one to work, or nearly work. A number of my tanka translations have four- or eight-syllable lines rather than the perfect five or seven. Take one of my translations from the Man’yoshu:

To love someone
who does not return that love
is like offering prayers
Back behind a starving god
within a Buddhist temple.22

This poem has the required thirty-one syllables, as does the original, but the first line in my translation has merely four, and the third line has, according to my count, six. To be a perfect syllabic translation, each of these lines should have five syllables. But I say this is close enough. To get it any closer I would have to take away a number of things I do like in the poem. Anyway, from this translation you can see that the original poem was a tanka and that Lady Kasa was not, as others may lead us to believe, writing either free verse or iambic pentameter in eighth-century Japan.


• R.H. Blyth, Japanese Life and Character in Senryu (Tokyo: Hokuseido, 1960).
• Harold G. Henderson, An Introduction to Haiku: An Anthology of Poems and Poets from Basho to Shiki (Garden City, N.Y.: Doubleday, 1958)
• Basil Hall Chamberlain, Japanese Poetry (London: John Murray, 1911)
• William N. Porter, A Year of Japanese Epigrams (Oxford: Oxford University Press, 1911)
• Yone Noguchi, The Spirit of Japanese Poetry, Widsom of the East Series (London: J. Murray, 1914).
• Curtis Hidden Page, Japanese Poetry: An Historical Essay with Two Hundred and Thirty Translations (Boston and New York: Houghton Mifflin, 1923)
• Asataro Miyamori, comp. and trans., An Anthology of Haiku Ancient and Modern (Tokyo: Maruzen, 1932), p. 132. As translated by Inazo Nitobe.
• Ibid. As translated by Hidesaburo Saito.
• Ibid. As translated by Minoru Toyoda.

• Ibid., p. 133. As translated by Clara A. Walsh.
• Ibid., p. 130. As translated by A. Miyamori.
Japanese Haiku: Three Hundred and Thirty Examples of Seventeen-Syllable Poems by Basho, Buson, Issa, Shiki, Sokan, Kikaku, Ransetsu, Joso, Yaha, Boncho, and others, in New Translation (Mount Vernon, N.Y.: Peter Pauper Press, 1956)
• Peter Beilinson and Harry Beh, Haiku Harvest (Mount Vernon, N.Y.: Peter Pauper Press, 1962).
• Kenneth Rexroth, One-hundred Poems from the Japanese (New York: New Directions, 1964).
• Harold Stewart, A Net of Fireflies: Japanese Haiku and Haiku Paintings with Verse Translations and an Essay (Tokyo: Tuttle, 1960).
• Geoffrey Bownas and Anthony Twaite, eds. and trans., The Penguin Book of Japanese Verse (Baltimore: Penguin Books, 1964).
• Maeda Cana, Monkey’s Raincoat (Tokyo: Mushinsha, 1973).
• Makoto Ueda, Matsuo Basho, Twayne’s World Author Series (New York: Twayne, 1973).
• William Howard Cohen, comp., To Walk in Seasons: An Introduction to Haiku. . . (Rutland, Vt.: Tuttle, 1972).
• Ibid.
• Ibid.
• Harold Wright, trans., Ten-Thousand Leaves: Love Poems from the Manyoshu (Boulder: Shambhala, 1979).


Comments are closed.