Dew on the Grass : Translating the Masters



WHCessay – WHF2000: Daniel Gallimore

WHC/WHF2000 Celebrations
‘dew on the grass’ : translating the masters
sound, image, logic: an approach to haiku translation

Daniel Gallimore
Oxford, UK

When I look at my copy of Lewis MacKensie’s translation of Issa poems – a Kodansha paperback published in 1984 with the title of The Autumn Wind – I see a pretty haiga painting of a persimmon on the bough with a persimmon-coloured border on front and back. It is a small volume (about 11 by 18 centimetres) with the poems crammed in, from three to six to the page depending on notes, and printed on the same low-grade paper as a Penguin Classic might be. This edition seems to say that yes, haiku are short, but there’s a lot in them, and you would like them even if they were written on toilet paper (which could itself make a worthy subject of haiku)! Looking at the blurb on the back, the following remarks by renowned critic Ueda Makotohit the eye: ‘Intense personality … vital language … shockingly impassioned verse’. And then, from the Oriental Economist, ‘Issa has … the strength not to romanticise or sentimentalise. Mackensie’s translations and comments have the same quality.’ What is clear is that translators and their publishers have to work as hard as anyone to sell their product, and this is doubly true of the British book market which is notoriously resistant to translated literature.

The literary translator Gregory Rabassa writes that:

Translation is a disturbing craft because there is precious little certainty about what we are doing, which makes it so difficult in this age of fervent belief and ideology, this age of greed and screed.

Rabassa is not (I suppose) referring to Islamic fundamentalism, but to a universal and sometimes pathological need for fixed identities, one which inevitably conflicts with attitudes of ambiguity and doubt. Translators are put in the unenviable schizophrenic position of having to coordinate their own motives for translation – which may include such factors as inward investment by Japanese companies and the popularity of the food – with what they understand the poet to have been saying, and on top of that come numerous difficult decisions regarding translating style and the treatment of individual words and phrases.

The haiku translator obviously has fewer words to contend with but, as I have already mentioned, the ambiguities of expression and varieties of context, above all the need to make each and every one of those little poems stand on its own, they all call for particular skills. Faced with such difficulties as a translator myself, I wish to offer two reassurances. The first comes from some advice given by Shiki to his disciple Kyoshi at the turn of the last century:

If you examine a moonflower closely your previous mental images will completely disappear, leaving only a new, shasei [or depictive] appreciation in your mind.

It seems to me, by analogy, that if you read a Japanese haiku closely enough it will imprint itself on your mind, and it is that imprint which constitutes the raw material of translation. Sensitivity is the key, and sensitivity is a quality we can all manifest so long as we are able to free ourselves from distractions.

Likewise, I wish to demythologise a little the act of making haiku. Kitahara Hakushu, who was a major tanka poet of the first half of the 20th Century and pioneer of modern Japanese poetry, wrote that:

In Japanese poetry there is a shade between each phrase and each word, an odour; the rhythm floats on the surface. How to achieve this fusion, this subtlety, can be easily enough understood.

I would add that if it can ‘be easily enough understood’, then one can attempt to reproduce it in translation. One mistake is to try to do too much at one time. Any poem can be understood according to the three Aristotelian dimensions of melopoeia (musicality), phanopoeia (symbolic beauty) and logopoeia (logical beauty), and yet it is rare even in the source for these three dimensions to come together in an integral whole. There are examples of integration, but before looking at one I would like to consider each of these three dimensions in turn with regard to some classic haiku by Basho and their translations.


Basho is renowned – according to Shirane and other Japanese scholars from the Meiji Era onwards – for the musicality of his style, and so it is to Shirane’s translation of a poem written by Basho in the autumn of 1694, shortly before he died, that I turn for an example of melopoeia (i.e. musicality). Shirane cites this poem as an example of the return to the low style which characterises Basho’s last years. In the first part of his career the poet had sought to transcend his humble origins through study of classical, medieval and Chinese poetics but (in the words of Shirane), he returns in his last years ‘to the exploration of various aspects of Tokugawa commoner life and language.’ What Shirane does not mention, however, perhaps because it is a commonplace of both the high and low styles, is the poem’s remarkable musicality:

Aki fukaki
Tonari wa nani wo
Suru hito zo

Autumn deepening –
My neighbour
How does he live, I wonder?
This is a haiku which can survive even the worst of translations, which Shirane’s certainly is not. For even if we do not know its context we can immediately appreciate the implicit image of the poet reaching out for neighbourly warmth as the days get shorter and colder. That is one way of reading the poem, an instinctive one perhaps, but in fact the poem is a good deal more subtle. A simple contrast of cold and warmth would be enough to constitute a phrase in some extended lyric, but we know that good haiku – especially those by Basho – offer more than simple antitheses, and this is a point which is particularly important to translators trying to render some of that musical complexity.

The first phrase is phonologically closed: the rhyme on aki, the crisp k and delicate i sounds, describe the sweetly relentless onset of autumn and (to admit the contextual metaphor) of Basho’s declining years. The two na sounds are clammy, moist; the poet weakens. But the wo at the turn of the line is a very different, majestic sound that is repeated in the emphatic particle zo. In other words, the solution to that invasive, get-you-down clamminess is not necessarily to visit his neighbour but to go on a journey – as (in a sense) he has been doing throughout his career – to wander, to guess, to allow the poetry to justify his existence. What better way after all to face old age than to carry on using one’s mind?

The final zo ends the poem on a note of triumph, telling the world that he is still a haijin after all, and in fact the poem was submitted as the hokku for a poetry session which Basho was too sick to attend. The English language, on the whole, lacks the capacity of Japanese for compression of sounds, which pushes Shirane to the other extreme of opening up the spaces between the words and foregrounding their denotative meanings. The diphthongs in the first line (‘au-’ and ‘dee’) establish the contemplative pace. The detachment of ‘my neighbour’ puts the neighbour in mind (makes him the object), since this is not an antisocial poem, and then those four monosyllables – ‘how’, ‘does’, ‘he’, ‘live’ – offer a third aspect, communicating the mystery of the neighbour’s existence. The phrase is also an effectively ambiguous version of nani wo suru; both questions could refer to a multiplicity of activities. Shirane does not reproduce the sound values of the original but he does maintain the tripartite diction.


For a rather surprising example of phanopoeia, Dorothy Britton’s translation of a poem from Oku no hosomichi, this one written at Zensho-ji temple in modern-day Ishikawa Prefecture:

Akikaze kiku ya
Ura no yama

All through the night
I listened to the autumn wind
In the lonely hills

It is a poem about listening and the loneliness of sound (oto no sabishisa) and yet the strange thing is that it brings a quite symbolic, visual image to mind. This is partly because of a vivid personal memory I have from my days in the school cadet corps. We went on a night manoeuvre on the Berkshire Downs where I was requisitioned to wait in the dark by a hedge for three hours until someone found me. I well remember the stars and the wind, the emptiness of the place and, of course, the cold. The time of year was late autumn and so reading this poem I feel (as perhaps some of you do) that I know exactly what Basho meant.

Phanopoeia is something of a hit-and-miss affair for the haiku translator as he seasons the few syllables allowed him with little hints which he hopes will trigger off somewhat larger memories. Britton has the special advantage, no doubt, of an unusual lineage: born in Japan, half-English and half-American, spent most of her adult life in England married to an air force officer; she can view all three of those cultures with a privileged detachment. With regard to this poem and to Basho’s musical style, it is also worth knowing that she is a professionally trained musician. To paraphrase Takiguchi’s analysis, perhaps the most distinctive feature of her style is the way she splices minimalist expressions (that seem quite literal translations of the Japanese) with a more ‘English’ idiom that often contains echoes of English poetry and popular songs. That first line ‘All through the night’ manages both, in fact: it is a literal translation of the source phrase yomosugara and is also the name of an old Welsh song (‘Ar hyd y nos’), of which the first verse reads as follows:

While the moon her watch is keeping,
All through the night;
While the weary world is sleeping,
All through the night;
O’er thy bosom gently stealing,
Visions of the light revealing,
Breathe a pure and holy feeling,
All through the night.

The undercurrent of the first two lines is one of deep content: ‘I am snug in bed, comfortable in my pew or wherever, and the wind is whistling around me.’ Unlike Basho, Britton does not insert a cutting-word (kireji) and does not need to, because the rhythm creates an equivalent rhetorical effect. The momentum of the first two lines is broken by the return to the shorter syllabic of the last and by this unusual word ‘lonely’. Britton’s use of the word is in danger of being labelled a pathetic fallacy – how can hills be lonely? – but I believe its very strangeness is sufficient for the haiku moment, the emotional arrest. Basho’s phrase ura no yama refers to the cultural backwardness and relative isolation of ura Nihon, the Japan Sea-facing ‘back of Japan’, both of which in haiku terms connote loneliness, and the Welsh association of Britton’s ‘All through the night’ might have a similar connotation.


Logopoeia, that is to say the logical lucidity of poetry, might be regarded as a misnomer when applied to the haiku, when – both in Japanese and English- the logical connections between objects, events and phenomena are usually vague and unstated. The question that haiku raises is not ‘Did the earth move for you?’ but rather two disjointed statements, ‘A rumple of bedclothes. An earthquake.’ In the first volume of his haiku anthology (in the section on haiku and poetry), Blyth gives examples of haiku which he considers to supply the stimulus by which the mind is encouraged to make the effort to overcome the difficulty of uniting what God has put asunder.

These are examples of haiku where the real is transcended romantically, and the possibility of a brighter vision made real. I would agree with Blyth of the necessity of making this kind of distinction if we are going to talk of a sub-genre of haiku love poetry, for example. This is one of his two examples from Basho:

Neko no koi
Yamu toki neya no

The loves of the cats;
When it was over, the hazy moon
Over the bed-chamber.
One of the devices used to facilitate logical continuity in English poetry is enjambment (corresponding with the haiku device of kumatagari), and we see the same process at work in Basho’s poem. The first phrase crosses over the five up to toki; the second, contrasting phrase crosses over the seven onto oborozuki. Logical beauty is allied with syntax, since it is largely through syntax that sentences make sense, but working against or compounding such logic are the purely phonological, alliterative effects on ko, ya and ki. The phonological complexity complements the semantic ambiguity which you will already have noticed. For it is never quite clear who has been doing what in the bed-chamber (neya) and indeed what the hazy moon (oborozuki) has to do with things; therein lies the rhetorical logic of this poem. We are of course adult enough not to need to be told what has been going on and to make the symbolic connection between ‘the loves of the cats’ and ‘the hazy moon’, which is used to evoke erotic love in traditional Japanese poetry. In other words, the poetic logic takes us halfway there and the rhetoric as far as we want to go. Blyth does his best to convey the ambiguity with a comic discrepancy between the pluralised ‘loves’ (as in ‘Cats do it without pawsing’) and the singular ‘it’ (in ‘When it was over’), unless that is that Blyth (and Basho) have naughtier imaginations than we give them credit for …

Crystallisation: ‘Fire and Ice’ in Shiki and Basho

All three of these poems by Basho manage each in their way to work on the auditory, logical and visual imaginations, but for one where the processes are quite clearly deliberate a haiku written by Shiki in 1902, during the last summer of his life:

Bara wo kiru
Hasami no oto ya

The sound of scissors
Clipping roses –
A clear spell in May.
Knowing that Shiki was confined to his sickbed with tuberculosis, we realise the poignancy of this image of this dying man grabbing whatever opportunity he could to bring a little beauty into his life, but what I would like to consider here is how the image is constructed. Shiki’s poem does not, at first reading, seem a particularly musical or onomatopoeic poem. It is true that the succession of seven open a vowels give the poem a certain unity but apart from that there is nothing very meaningful to catch the eye or ear. Actually, there is, as the open, graceful bara (‘rose’) and the clear-cut kiru (‘cut’) are combined within one word in the final phrase, satsukibare. This is what cut roses and fine weather in May mean to each other at this moment in Shiki’s life: the sounds are almost identical to the picture so that when the two work together like this the logic works very fast indeed. ‘Death concentrates the mind wonderfully’, and it is not the clipping of scissors which is figured, as one might expect, but the beauty of those flowers; in other words, the sound values work in relation to the logic of the poem but are not subordinate.

The translator, Burton Watson, similarly evokes this feeling that the poet is more interested in the significance of the clipping than the ominous sound itself. The word ‘clipping’ is separated by enjambment from ‘scissors’ such that this terse sound – with all its freshness and its transience – belongs as much to the roses as to the scissors, especially as it alliterates with ‘clear’ in the final line. The alliteration also accounts for the logical development of Watson’s version; as with the source, we seem to see, hear and experience it all at once. This phenomenon is surely what was meant by Yasuda’s term ‘crystallisation’: ‘a crystallised haiku is held together by the organic, emotional force of the experience’.

Shiki’s poem moves so quickly that it can be read as an example of that momentary rhetorical fusion of subject and object which distinguishes great haiku: very difficult to pull off but when it works, powerful indeed. For an analogy with poetry written in English, I like Robert Frost’s poem ‘Fire and Ice’ (1921):

Some say the world will end in fire,
Some say in ice.
From what I’ve tasted of desire
I hold those who favour fire.
But if it had to perish twice,
I think I know enough of hate
To say that for destruction ice
Is also great
And would suffice.
The measured diction and rhymes seem to fuse the two simplest of metaphors – fire and ice – in a way that their metaphorical potential is unleashed. In my mind at least, they are metaphors for the condition of poetry lying between entropy on the one hand (‘ice’) and manic self-destruction on the other (‘fire’), and in Basho’s career that tension becomes a tension between the desire of body and soul to travel and imagine and a desire to sit still and quite literally lose oneself in the nothingness of the moment:
Atsuki hi wo
Umi ni iretari



The river Mogami
Has drowned the hot, summer sun
And sunk it in the sea!

Pouring the hot day
Into the sea –
Mogami River.
Basho’s poem plays with a question which has been with us ever since we first saw the sun setting over the sea as children – what would happen if the sun really did collide with the sea? It is that unanswered mystery which the poem evokes. The contrast of sun and water is clear and refreshing enough but what gives the poem its continuity is the central image of pouring: sunlight into the water, the river into the sea. The image plays no tricks on the reader; it does not allow the reader to distance himself from the experience.

Dorothy Britton’s bold statement has a Shakespearean quality serving both to individuate and to generalise the experience, perhaps even alluding to some event in the human world: ‘Have you heard? The river Mogami has drowned the hot, summer sun and sunk it in the sea?’ Shirane’s minimalism, by contrast, lets him down here. The two phrases are not quite sufficient to give the line its grandeur, although the translation is clear enough.

the rhetoric of haiku translation

It is all very well mastering the poetics of Basho, but would people actually want to buy your translation? In 1965, the translation theorist Hugo Friedrich made a vital ‘distinction between language as reality (Gegebenheit) and language as act (Tat), that is, style’, and we can say that both these functions of language are functions of haiku as well. One quality that unites Basho’s haiku with Shiki’s is the primacy of immediate reality. Conversely, it is the sense of immediacy, ‘the presentation of such a ‘complex’ instantaneously which gives a sense of sudden liberation’ (to quote Ezra Pound), the active, stylistic function of haiku.

We all know that the relationship between language and reality is unstable but that poetry can effect a temporary illusion of stability. I would describe the haiku ‘moment’ as one such illusion and extend it with reference to the haiku ‘resonance’ or yôj; yôj is that lingering difference which the haiku poem can make to our moods and even behaviour. Goethe said of the translator’s art that ‘We are led, yes, compelled as it were, back to the source text’, which I suppose must be the highest compliment one can pay to a literary translator. One also thinks of Leslie Downer retracing the steps of Basho for Channel 4 in the 1990s and (on a less ambitious scale) of those of us who take the trouble to read books about the haiku masters, who want to know what they were like as people.

The irony of the effective translation is that it is all too aware of its limitations and does not try too hard to persuade us otherwise. It leaves a space for the reader to explore at will, having been left with no doubt in his mind of the poet’s intrinsic interest, and this space is what may be meant by Walter Benjamin’s concept of ‘pure language’ (reine Sprache), that hidden presence within the source that transcends both source and target cultures. Irrespective of whether or not the translator chooses to make deliberate reference to his own culture, the final test of a haiku must be whether or not the haiku ‘moment’ rises lotus-like in the reader’s mind and enables him to see another side. To do so will require careful attention to technique, of which some concluding words.

Reading the prosody: 5-7-5

The magic of the fixed form or yuki teikei, the 5-7-5 syllabic format, has nothing (one should hasten to add) to do with any magical qualities associated with those numbers in Eastern and Western cultures. First and foremost, 5-7-5 gives the haiku a definite shape – note that I do not say ‘the’ or even ‘its’ shape – which is registered phonically and often on the written page as well. Form is intrinsically pleasing and even more so when it supports a symmetry or progression of meaning. Translating a Japanese haiku into strict 5-7-5 in English is not impossible but is not easily justified poetically. What is more important is that the translator read the rhythms of the source effectively. The most typical pattern is for the rhythm to be broken by a cutting-word at the break of the poem after the seven, which is certainly a device worth trying in English, but the translator must also read those subtle internal rhythms (or naizairitsu), which are determined mainly by syllabic boundaries (threes and fours are common) but also by pitch accent. These rhythms contribute to tone and register and should be read as such.

In The Japanese Haiku, Yasuda discusses the prosodic potential of threes and fours, showing how it is quite possible in English for variations in syllabic length to match semantic modulations. It is only our inevitable bias toward stress accent as native English speakers which blinds us to the possibilities of syllabic metre. Yasuda’s book was published in 1957 but in an essay published last year and now available on the Internet, two professors working in Kumamoto, Richard Gilbert and Judy Yoneoka, measure out some common ground between Japanese and English by arguing that haiku in both languages can be fitted into a broad 8-8-8 template. They produce convincing evidence that native speakers read haiku both between and within lines so that even traditional 5-7-5 haiku can attain a longer syllabic. This is an idea which haiku translators may do well to consider, although I would qualify it with the criticism that the authors do not make an adequate distinction between rhythmically sensitive and rhythmically insensitive readings. I do not believe that haiku should be read flat: they do contain latent rhythms waiting to be exploited.

My own experience of translating haiku has suggested that the 5-7-5 syllabic will more likely be arrived at by accident than by design. Here is one such happy coincidence, my translation of a haiku by a young woman called Ishikawa Tomiko, composed shortly after World War II:

Jopu tome
Orishi momiji ya
Nogiku kana

A stationary
Jeep. Fallen maple leaves and
Wild chrysanthemums.
In my mind at least, the poem recalls a rendezvous between a young woman (not necessarily the author) and some GI in a secluded field. The ‘fallen maple leaves’ might refer to GIs taking their chances while they can or to parents or other authority figures killed in war and the wild chrysanthemums to those untamed young women with no one around to keep them in order. The fact that the translation is in 5-7-5 probably makes little difference to the rhetorical effect, especially as the moraic values differ between Japanese and English, but for the translator it is itself a kind of template, a structure on which to build the translation.

There is an ebb and flow in haiku – a pithy phrase in the first line, followed by an expulsion in the second before settling into some final essence in the last – and this is a movement which will usually be felt in English as well with the greater number of both syllables and stress accents in the middle line. By way of contrast, here is another haiku in 5-7-5 by Ishikawa Tomiko, which I rendered in a 3-6-7 format:

Chokan no
Koishi ni hisomu
Kokani kana

Morning frost.
A little crab, I guess,
Hidden among the pebbles.
Here the three-line structure allows me to stress the word ‘hidden’, which relates to the tight mood established in the first line.

Cutting words

Tomiko’s haiku is not unusual in containing a cutting-word (kireji), which Blyth defines as a kind of ‘poetical punctuation’, and as such, they usually do require a more literal treatment than shichigo cho. Kireji mark a definite break in the poem but can also alter its mood significantly and should therefore be translated in context. Kana, one of the more common, tends to give a poem karumi (lightness of touch), to let it fade gently away. In conversation with Japanese speakers I sometimes hear them end their sentences with this phrase which can literally mean ‘or something like that’, but is translated poetically as ‘I guess’ or ‘I wonder’, and I wonder indeed whether they know of its use in haiku. The emphatic zo and yo, however, are usually rendered with full stops or exclamation marks, the softer keri as a comma.

Translating number

At the other extreme is the problem of number: how to quantify nouns when the absence of plural endings in Japanese makes it difficult to determine whether there is one or more than one. This is a problem singled out, as it were, by Blyth and made even harder in haiku translation by the lack of context. Here is my translation of another haiku by Ishikawa Tomiko by way of illustration:

Yamiyo nimo
Sabisho sakeri

Even on this dark
Night forlorn, the evening
The question is whether there is more than one evening primrose. Does its loneliness refer to its singularity, popping up alone amidst foliage, or to the fact that it is unusual among flowers in flowering only in the evening? The answer is more likely the latter as the first phrase ‘Even in this dark night’ specifically denotes its unusual flowering schedule, and yet it has to be admitted that the first possibility is suggested at least by the classical mood of the poem. Sabisho contains the archaic -sho inflection and tsukimi in the flower’s name means ‘moon-viewing’, which is a byword for romance in classical poetry. Isn’t there another story going on: the story of a young woman who has loved too late and sees her image in the pale yellow flowers? One of the pleasures of haiku is that through these real, immediate images we are able to glimpse an array of other possibilities, and in this case to read the past. They are possibilities which can be suggested by the slightest of stylistic touches. To pluralise the flower – ‘evening primroses’ – strikes me as unnatural. Why not make it a metonym, let it stand both for evening primroses and for unrequited loves for all time.

conclusion: haiku translation for all

The Meiji critic,Tsubouchi Shoyo, who was himself the pioneer of Shakespeare translation in Japan, once dismissed haiku as ‘the pastime of dilettantes’. Yet it can hardly be doubted that the globalisation of haiku in the 20th Century has demonstrated its seriousness as a poetic form both outside Japan and within. Shoyo’s agenda was to develop a national literature sufficiently serious to counter the weight of Western tradition. He believed that the haiku form was too trivial to make a difference; Shiki and Shiki’s disciples proved him wrong. Events such as World Haiku Festival would have little meaning if its participants did not in some sense ‘believe’ in haiku. Yet it goes without saying that only a fraction of this process of intercultural transmission would have been possible without translation.

Haiku translation deserves to be taken seriously as an art in itself and for the impact that haiku translations can have on their target cultures. The reception of haiku translations is not something covered in any depth in this paper. What I have tried to do instead is to uncover some of the problems and processes of translation. Translation is a process as reading is a process. I know from my experience of the handful of haiku that I have translated, how not only is one’s knowledge of Japanese language and culture enhanced by the act of translation, but in taking on board those haiku ‘moments’ and those haiku images for several moments longer than one usually does than when reading someone else’s translation off the written page, that one’s appreciation of the haiku form is deepened. I do not expect every haiku poet to become a haiku translator as well, but I do believe that the globalisation of haiku can only be advanced by the teaching of haiku translation in Japanese language and literature courses outside Japan. Haiku translation is also an issue between other languages (Bulgarian and Italian, for example) but Japan remains the home of haiku culture, and so exchange between that and other haiku cultures will be rooted in the teaching of the Japanese language.

Daniel Gallimore is studying for a doctorate at Linacre College, Oxford, on Japanese translations of Shakespeare’s A Midsummer Night’s Dream. He also teaches Japanese part-time at Oxford Brookes University. He is interested in the therapeutic value of haiku and recently led a haiku writing group at Acorn, a MIND day centre in Oxford.

This paper was read at WHF2000, London/Oxford Conference, August 2000.

This entry was posted in Classics, Haiku, Vol 1-3 November 2001 and tagged , , , , . Bookmark the permalink.

Make a comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s