The Beat of Different Drummers

The Beat of Different Drummers:
English Translations of Hokku from Matsuo Bashō’s Oku no hosomichi

Mark Jewel (Waseda University)

(The following paper was first published, without the Appendix, in The JASEC Bulletin, vol. 10, no. 1, September 2001, pp. 1-15. A few very minor corrections (almost all of them typographical) have been made in this online version, the scanned orginal of which can be downloaded here. The Japanese Association for Studies in English Communication (JASEC) is based at Waseda University in Tokyo. A separate online version may be viewed at The World Haiku Review.)

Haiku is without question Japan’s most successful literary export. Indeed, along with judo in the field of sports and, more recently, anime and video games, haiku is one of only a handful of Japanese cultural products that can be said to have acquired an international following of any significant size. Haiku in English boasts a history in translation of over one hundred years, and an active “haiku community” of original poets that dates back at least as far as the first regularly published magazines of English haiku in the 1960s. As one indication of just how popular English haiku has become in the past quarter century, it may suffice to point out that more than ten single-volume anthologies of haiku in English have been published since the first such anthology—Cor van den Heuval’s The Haiku Anthology (Garden City, N.Y.: Anchor Books)—came out in 1974.[1]

Small wonder it may seem, then, that the poetic travel diary Oku no hosomichi, by Matsuo Bashō (1654-1694), which contains fifty of Bashō’s hokku, has been translated into English more frequently than any other major work of Japanese literature, with no fewer than eight complete published versions.[2] Part of the purpose of this paper is to suggest that, in fact, eight different versions cannot be called an overabundance in this case. But before turning to an examination of some of the hokku from Oku no hosomichi to help justify this assertion, I think it will be helpful to review the changing fortunes of haiku in English over the past hundred years, for the current high regard in which Bashō’s poetry is held by both translators and English-language haiku poets by no means reflects its reputation among the first serious foreign students of Japanese literature. A brief historical survey should allow us both to identify some of the basic problems attendant upon the translation of this quintessentially Japanese literary form, and also to remark on the existence of a productive dialectic in English between translation and original composition that has already influenced both and promises to lead the genre in new directions in the future.

Makoto Ueda has identified Lafcadio Hearn as the earliest translator of Bashō’s hokku into English.[3] The famous poem about the frog jumping into an ancient pond, for example, appeared in Exotics and Retrospectives in 1898. Later works by Hearn also include a significant sprinkling of  hokku, with the Japanese arranged into three lines and a one- or two-sentence English translation placed in brackets underneath. While showing a sympathetic appreciation for the genre, Hearn (in the “Insect Studies” section of Kwaidan) admits that it must be considered an “acquired taste.” British scholar W.G. Aston, on the other hand, writing at about the same time in the first complete English history of Japanese literature, is much more direct in his criticism: “It would be absurd to put forward any serious claim on behalf of Haikai to an important position in literature.”[4] Aston goes on to confidently assert that with the appearance of longer, Western-inspired poetic forms, “the day of Tanka and Haikai seems to have passed. These miniature forms of poetry are now the exception and not the rule.”[5]

The kind of cultural arrogance that lies behind Aston’s pronouncements is really quite astonishing when viewed from our post-Second World War multicultural perspective. Even his compliments are backhanded ones:

Can it be imagined that when a religion is presented to [the Japanese] which alone is adapted to satisfy far more completely all the cravings of their higher nature, the Japanese, with their eminently receptive minds, will fail in time to recognise its immense superiority? They have already accepted European philosophy and science. It is simply inconceivable that the Christian religion should not follow.[6]

But the basic argument that Aston makes is surely one that any advocate of hokku or modern haiku must address—that haiku is an essentially trivial form unsuited to dealing with the intellectual and emotional complexities of modern life. This argument was made to even more devastating effect by another British scholar, Basil Hall Chamberlain, whose reputation among the Japanese themselves has been eclipsed by that of his more congenial contemporary, Hearn. Chamberlain, a scholar of immense erudition, discusses Bashō in a detailed paper presented at the Asiatic Society of Japan 1902.[7] He concludes that compared with the “Palaces of Art” constructed by Tennyson, Japanese hokku resemble a “litter of single bricks, half bricks in fact,”[8] and remarks that the hokku “appears, now as a tiny herb or flower on our path, now as some brilliant insect which hovers for a moment, and, ere we have noticed it, flits away out of sight and memory.”[9] Like Hearn, Chamberlain provides transliterations of the Japanese divided into lines of 5-7-5 syllables, but for translation uses an epigrammatic style that does not follow any set formal pattern (although caesuras and exclamations are often indicated typographically by means of punctuation).

Given this rather dismal early assessment of the value of hokku, what happened to turn the situation around? In terms of the history of English poetry itself, a major turning point came with the modernist revolt instituted by Imagists Ezra Pound and Amy Lowell beginning in around 1912. The poetry of the Imagists resulted in a general (and lasting) preference on the part of practicing poets for patterns of clearly defined images rather than narrative, a preference that was informed by a sympathetic if not necessarily well-informed understanding of Chinese and Japanese poetry, including haiku.[10] In short, haiku now seemed strikingly compatible with the modern mode of perception being advocated by such poets as Pound, Lowell, William Carlos Williams, William Butler Yeats, and T.S. Eliot.[11]

In the field of translation, the transition from lukewarm acceptance to ardent approval was accomplished largely through the work of two men: R.H. Blyth and Harold G. Henderson. Blyth was an Englishman who lived in Japan for more than thirty years until his death in 1964 and spent the Second World War interned in Kobe; Henderson was an American acquaintance of his who published the first (very short) book on haiku in English in 1934 and helped found the Haiku Society of America in 1968. Although the two were close friends (at least initially), they held somewhat different views on haiku. Blyth, who had studied and practiced Zen Buddhism in Korea before arriving to Japan, emphasized the Zen aspect of haiku: an intuitive sort of immediacy that points the way to enlightenment.[12] His translations, which started appearing just after the war, were read by and influenced the Beat poets of the 1950s. Although now considered rather déclassé among many specialists in Japanese literature, Blyth provided the direct inspiration among poets and readers in English for taking haiku seriously as an art form, and his spirit informs the work of such current translators as Lucien Stryk, co-translator of The Penguin Book of Zen Poetry. Blyth’s translations typically give both Japanese (in the original and transliteration) and English, with the latter arranged into three lines of no fixed syllabic length, but with the first and third lines indented so as to give visual prominence to the second line. For haiku in English, Blyth advocates a three-line form that consciously avoids rhyme, with a 2-3-2 accented-beat rhythm that is, however, neither regularly iambic nor anapestic.[13]

Henderson, who taught at Columbia University, revised his earlier book on haiku in 1958, and in 1967 also wrote a book called Haiku in English, which was published specifically in response to a growing demand in the 1960s by teachers, readers, and practitioners for a detailed explanation of what haiku is and how to write it or teach others to write it.[14] No doubt partly because of their brevity and accessibility, these two books greatly influenced the first few postwar generations of Japanologists and the general public as well. In Haiku in English, Henderson concisely reviews the basics of haiku, formulating four “general rules” for traditional Japanese haiku: the use of a 5-7-5 syllable count; the insertion of a conventional reference to nature (the kigo, or “season word”); an emphasis upon particularity rather than generality; and a focus on the present time rather than on the past. These rules are then discussed in connection with writing English haiku, including a short discussion of the use of rhyme (a technique almost invariably and yet subtly employed by Henderson himself). No hard-and-fast conclusions are drawn about the applicability of Japanese models, and Henderson is at particular pains to discount the need for unvarying observance of  the 5-7-5 syllabic pattern in English. Instead, the emphasis is placed on conveying by means of suitable imagery what has come to be known as “the haiku moment”: the simple, direct expression of an emotion evoked by some particular natural event or aspect of nature. This approach, although not inherently antithetical to Blyth’s more transcendental, Zen-based approach, does seem to end up being rather more modest in its ultimate goal. And if Blyth’s versions carried greater philosophical weight (especially among the Zen-inspired Beat poets of the 1950s), Henderson’s approach helped to ingratiate the form with the general reading public and facilitated its adoption by American school curricula in the 1960s and 1970s. This, then, is the period when haiku can be said to have entered the poetic mainstream, at least in the United States.[15] And in spite of Henderson’s own reservations, it is probably when the 5-7-5 syllabic pattern came to be widely regarded as a model for composition in English as well.

Currently, the leading proponent of the second, more nature-centered approach to English haiku described above is the American poet and translator William J. Higginson, who asserts in his 1985 The Haiku Handbook that Western haiku poets “concentrate on capturing the kinds of moments—the sudden intimate seeings—that they wish to remember themselves and share with others.”[16] To my mind, this bears an uncomfortable resemblance to the “Kodak moment” extolled in television commercials by the well-known American manufacturer of film and cameras, and runs the risk of re-trivializing or perhaps simply confirming the trivial nature of modern haiku (a question that has by no means been settled). Nevertheless, in its emphasis upon the central role of nature and seasonal change, it does appear to be the approach now followed by the majority of haiku poets writing in languages other than Japanese. Higginson follows Blyth in his preference for a 2-3-2 accented-beat rhythm in English, claiming that this results in a better approximation of the length of Japanese haiku when read aloud than does the 5-7-5 syllabic form. But it should be noted that this pattern does adopt the basic short-long-short rhythmic model of the Japanese, and that Higginson also uses a fairly standard three-line format in his translations. In other words, even while rejecting the authority of the traditional Japanese syllabic count in determining the form of haiku in English, Higginson implicitly acknowledges the importance of both the original rhythm and a three-part organizational scheme. Furthermore, Higginson has recently argued for the usefulness of an international saijiki, or “haiku almanac” categorized by season word, as a guide in composing haiku in Western languages.[17] Acknowledging that the choice of season words to be included in such an almanac must take into account different geographic locales and that provision should be made for a larger “no-season” category than in a Japanese saijiki, he nevertheless holds that this traditional sort of poetic manual fosters the sort of seasonal awareness he views as essential to good haiku in any language.

In this way, Higginson’s attempt to reconcile traditional Japanese hokku / haiku conventions with a nascent set of English conventions can be said to be characterized by a certain amount of expediency and compromise. But rather than criticize Higginson for a lack of logical consistency, it seems best to recognize that expediency and compromise are inherent in any such undertaking, and to regard his example as pointing to the key role played by cross-cultural mediation in the development of this relatively young English literary genre. Indeed, it seems to me that the efforts of both translators and original poets to work out a hybrid set of conventions are a clear indication of the vitality of haiku in English. That is, even as translators have contributed to the development of haiku in English by appealing to the authority of Japanese models, their own practice has been influenced by the work of other translators and by original haiku in English. It may well be that this is the only field in Japanese literature where specialists feel compelled to take into account the work done by those who may themselves have only a very modest background in the Japanese language and the study of Japanese poetry.

Precisely as a result of this quasi-collaborative process, the translations I propose to discuss here can be expected to reveal a surprising diversity of approaches to the problem of translating what is surely one of the most rigidly defined of poetic forms. Of course, reasons that are purely linguistic are also involved—even within the range of seventeen syllables, there is enormous room for variation in syntax and diction. Yet the large amount of variation also reflects conscious choices on the part of the translators about how to handle form and images, and these choices have, in turn, often been influenced by earlier translations, by an awareness of the conventions of English poetry, or by the rejection of solutions adopted by previous translators. The task of the attentive reader is to take note of the methods used in each case and, quite simply, try to decide how successful they are.

To simplify that task in this paper, I have chosen to examine five hokku out of the fifty composed by Bashō for Oku no hosomichi as rendered by the eight translators mentioned in the second endnote. The major criterion for selection was personal preference, guided to some extent by an eye toward the problems of translation. Despite the relatively small size of the sample, I believe that it can be considered representative—the translators tend to be consistent in their methodology, and increasing the number of examples would not change my basic conclusions. I want, first of all, to use the translations to point out both the strengths and weaknesses of each translator and to make a number of specific comments about hokku translation in English. In this context, the first two hokku are discussed in some detail with regard to each translated version; the remaining three are then used to review and qualify a number of the points already made. After this analysis, and rather immodestly perhaps, I intend to offer a final judgment about the general effectiveness of each translator’s approach, which the reader is free to accept or reject as he or she sees fit.

There are two initial points to be made concerning the original Japanese versions. The first is that of Bashō’s fifty hokku, forty-seven follow the standard 5-7-5 syllabic pattern. Bashō was a master of this structural pattern, and a significant amount of internal rhythmic variation is to be found in these hokku. Nevertheless, it seems obvious that at least for Oku no hosomichi, Bashō decided to adhere very closely indeed to the standard pattern.[18] The second point to note is that many of the verses rely for their effect on the use of the technique of the juxtaposition of images that is usually held to be one of the defining characteristics of Bashō’s mature style. It would therefore seem logical to assume that the ordering of the images is intended to produce a specific effect in each case, and that tampering with this order in translation risks altering that effect.

Using these two preliminary observations as our starting point, then, let us turn to the hokku themselves:

1.  no o yoko ni / uma hikimukeyo / hototogisu

Yuasa: Turn the head of your horse  /  Sideways across the field,  /  To let me hear  /  The cry of the cuckoo.
Corman  /  Kamaike:  across the fields  /  head the horse  /  hototogisu
Miner: Cutting across the moor,  /  Draw still the horse you lead along—  /  Hear the wood thrush again
Britton: Turn across that moor,  /  O horseman, for I hear  /  A cuckoo singing there!
McCullough: A cuckoo song:  /  please make the horse angle off  /  across the field.
Sato: Turn the horse round across the field, cuckoo
Keene: Lead the horse sideways  /  Across the meadows—I hear  /  A nightingale.
Hamill: The horse turns his head—  /  from across the wide plain,  /  a cuckoo’s cry

To begin at the level of interpretation, it should of course be remembered that all of the hokku in Oku no hosomichi are placed in a specific narrative context. In this case, Bashō is being led on a horse to the famous Killing Stone (Sesshōseki) in present-day Tochigi Prefecture when the man who is leading the horse asks him for a poem. The quoted hokku is Bashō’s response. Taking the translations in order, we see first of all that Nobuyuki Yuasa has chosen to translate in four lines.[19] To me, this seems a very misleading method. One reason is that it often forces the translator to fill out the lines with extra material, here meaning the entire third line of the translation. The second reason is that it gives the reader the wrong idea about the type of rhythmic balance that is created in hokku: an asymmetrical three-part balance that is ill-served by Yuasa’s first-half, second-half symmetry. The rhythm in translation is created simply by dividing the English into semantic units, yielding a total syllable count of twenty-two. The 6-6-4-6 pattern used here is repeated just once in Yuasa’s other translations, and indeed, no syllabic pattern is repeated more than once in any of Yuasa’s versions.[20] All in all, the translation comes across as somewhat stilted, although Yuasa does succeed in preserving the order of images in the original, which has the intended effect of emphasizing the (call of the) cuckoo as both the inspiration for and goal of the poet’s proposed detour across the field.

Cid Corman and Susumu Kamaike’s translation offers a stark contrast to Yuasa’s. First, no notice is paid to English conventions such as capitalization, ending punctuation, or even normal syntax. It is also hard to discern any consistent use of rhythm, either in terms of syllable count or even accented beats (although the number seldom exceeds three in any one line). This makes it appear as though Corman and Kamaike want above all to maintain a certain imagistic fidelity to the Japanese even at the risk of violating the usual rules of English usage. In this context, it should be noted that Corman himself (Corman is the one responsible for the final English form of the translations) is a modernist poet of distinction, and that his practice in composing English poetry is no different from his practice in translating Japanese. This may, therefore, be a case in which the perceived similarity in style (the use of concrete, fragmentary images) has intentionally—and somewhat misleadingly—been allowed to take precedence over formal regularity as a principle of fidelity. The treatment accorded “cuckoo” (hototogisu), too, may on one level be said to reflect an insistence on paying attention to the importance of the concrete image—a hototogisu may be a member of the cuckoo family, but it is not exactly the same bird English speakers know, and should not be treated as if it were. Yet one cannot help feeling that the translator’s responsibility as a communicator of meaning is being slighted here, for a reader not already familiar with the bird called hototogisu by the Japanese will have no idea just what is being referred to here. Insistence upon the uniqueness of the image does run the risk of obscurity. Taken to its extreme, the refusal to paraphrase or accept any substitutes would simply result in a word-for-word repetition of the Japanese—the very antithesis of translation. Corman and Kamaike are not really quite so extreme, but the desire to make interpretation more challenging, and hence more rewarding, by disrupting conventional expectations in this manner is a distinctly modern approach. Perhaps another reason for leaving hototogisu untranslated is the effective use made of alliteration in the English version. The translation contains twelve syllables (arranged in a 4-3-5 pattern) which, when read aloud, have a pronounced rhythm attributable in large measure to the repeated H sounds. Corman, naturally enough considering his experience, has an acute ear for rhythm in short poetic forms, at least in English, and this particular hokku seems to me to be one of the more successful translations in the Corman-Kamaike version of Oku no hosomichi.

The third translation is that of Earl Miner, who along with Robert H. Brower wrote the book when it comes to translating and analyzing waka.[21] This twenty-syllable version is very nearly as long in three lines as Yuasa’s is in four, and I must confess that it seems quite wordy to me now in a way it did not when I first read it more than twenty-five years ago. Miner does follow a short-long-short syllabic pattern, but it is telling that only six of his translations from Oku no hosomichi actually fall below eighteen syllables in English. Two other problems exist here. First, Miner interprets the situation differently from the other translators, creating the impression that the poet is already crossing the field on the horse, which is surely mistaken. It is an unusual slip for him. Second, although the major break follows the Japanese in coming at the end of the second line, the imperative verb in the third line shifts the attention of the reader from the cuckoo to the horse driver, which is both repetitious (one command has already been given) and blurs the focus, which should be on the image of the bird (or, more conventionally, its song). Miner’s use of “wood thrush” is precisely the sort of vague substitution that Corman and Kamaike appear to disdain. It may be possible to acknowledge its usefulness as an interpretive crutch for non-specialists, but one must finally admit the incongruity of a bird native to North America attracting the attention of a seventeenth-century Japanese poet. This translation thus labors under the disadvantages of being both misleading and drawn out.

Dorothy Britton’s translation comes the closest so far to the standard seventeen-syllable count (the majority of her versions actually fall into the nineteen- or twenty-syllable range). But the addition of extraneous information in the second and third lines—the direct address to the horseman and the explanation of the reason for the poet’s request—lowers the tension achieved in Japanese by keeping the last line semantically separate from the first two. The one-sentence format makes for smooth reading, but one almost feels that it is too smooth, that the juxtaposition of images should have a more forceful impact upon the reader. The conventionality of the English is reinforced by the use of capitalization at the beginning of each line, the rather archaic form of direct address (“O horseman”), and an attempt to match sound values at the end of the first and third lines (a technique that tends toward the rhyming versions Britton often produces). A tight formal unity is achieved, but that unity derives solely from English conventions and disguises the way the translator has rearranged the order of the images. Ease of reading alone can hardly be considered the hallmark of a faithful translation.

The next version comes from Helen McCullough, a translator who can arguably be said to have translated more classical Japanese into English than any other person.[22] When she translates waka, McCullough tends to follow a 5-7-5-7-7 syllabic pattern quite closely, but apparently she finds the hokku form too constraining for such a rigorous level of consistency: although none of her versions of Bashō’s hokku exceed nineteen syllables or fall below fourteen, only seventeen—or about thirty percent—actually follow a 5-7-5 pattern (Keene, with fifteen, is the only other translator even to approach the same level of consistency). McCullough prefers full-sentence English syntax (noun phrases are allowed to stand as whole poems, but otherwise subjects and verbs are clearly stated), which means that her translations typically contain participles and prepositional phrases, end with periods, and use capitalization only at the beginning of a new sentence. She thus follows the pattern set down by Yuasa, Miner, and Britton, but with the important difference that she is more concise than the first two and less given than Britton to applying traditional rhyming techniques and standards of diction. On the other hand, McCullough reverses the position of the first and third lines of the Japanese version. I suspect that she did this in order to avoid adding the sort of explanatory material added by Britton. But, of course, moving the concrete noun to the beginning of the hokku reduces the force of the ending, so that the English version appears to trail off weakly. Perhaps the damage is not as great as it might be with a hokku more obviously dialectic in effect, but the loss of focus is not negligible. Granted that McCullough seems to have found it necessary to compromise in this case, hers seems to be a careful, scholarly approach that draws its strength from its reliability.

The next version, by Hiroaki Sato, is as radical in its own way as that by Corman and Kamaike. Sato advocates one-line English haiku on the basis that Japanese haiku are written and printed as one line and that, when read aloud, the duration of an English haiku should approximate the duration of a Japanese haiku.[23] In terms of arguing for duration as a standard of both translated and original haiku in English, his position is close to that of Higginson, whose suggestion of a 2-3-2 accented-beat pattern (with a total length of about twelve syllables) has been noted above. However, while Higginson continues to write in three lines, Sato takes the additional formal step of joining the lines together. The potential disadvantages of such an approach are amply in evidence here. I admit to being confused about how one can turn a horse “round across” a field, and my first instinct is to take “cuckoo” at the end as a direct address, so that the poet is telling the bird to do something with the horse (and in another hokku from Oku no hosomichi, Sato uses “cuckoo” in precisely this fashion). These misreadings follow directly from the format chosen by the translator, which in my view argues strongly against the applicability of this translated form.

Donald Keene’s translation follows the long-short-long rhythmic pattern he takes as his basic model (next to McCullough, Keene has the most translations in “standard” 5-7-5 form—fourteen in all). Cast in sentence-pattern syntax, it seems intended to be as clear as possible in meaning: like Britton, Keene adds “I hear” in the second line to make explicit a logical connection that is allowed to go unstated in Japanese. Keene apparently feels that without such an explanation, the motivation will not be sufficiently clear to the inexperienced reader. But in the attempt both to maintain a syllabic count approximating the 5-7-5 Japanese pattern and to retain the original image order, he has broken the second line in the middle, in effect creating two halves rather than a two-line, one-line division. This has the effect of disrupting the rhythm of the original and slightly drawing attention away from the cuckoo (which Keene inexplicably translates as “nightingale”).[24] When one also considers that leading the horse “sideways” creates the potential for comic confusion, it must be concluded that this is not one of Keene’s more convincing efforts.

Finally, Sam Hamill’s translation must be considered quite wide of the mark in terms of accuracy. Not only does Hamill change a Japanese command into a descriptive phrase in English (“the horse turns his head”), he makes it seem as though the cuckoo’s call has caused the horse’s reaction and that that ends the implications of the poem. As Japanese commentators invariably explain, the situation is that having heard the cuckoo call out once in the distance, Bashō is telling the horseman to lead the horse nearer the spot so that he (or they) can hear it again. In other words, there is an implied purpose to the poet’s command that is simply ignored in this version. Hamill is an experienced poet and translator, but this kind of carelessness (or eccentricity) appears with distressing regularity in his versions of hokku from Oku no hosomichi. In terms of syllabic count here and throughout, Hamill comes close to Keene in observing a regular short-long-short rhythm (without, however, matching McCullough’s level of consistency), yet he manages to achieve greater directness through the more frequent omission of ending punctuation. This attempt to better match the immediacy of the Japanese, however, seems inadequate compensation for the problematic rendering of meaning.

2.  oi mo tachi mo / satsuki ni kazare / kaminobori

Yuasa: Proudly exhibit  /  With flying banners  /  The sword and the satchel  /  This May Festival Day.
Corman  /  Kamaike:  chest too and sword  /  in May hoist high as  /  paper standards
Miner: The pannier and sword:  /  Use them to decorate the Boys’ Festival  /  Along with carp streamers.
Britton: What a proud display!  /  Chest and sword and paper carp,  /  For Boy’s Festival Day.
McCullough: Paper carp flying!  /  Display pannier and sword, too,  /  in the Fifth Month.
Sato: Display both casket and sword in May with paper carps
Keene: Sword and altar both  /  Display on Boy’s Day in May  /  When paper banners fly.
Hamill: Sword, chest, and wind-carp  /  all proudly displayed  /  on Boys’ Festival Day

The situation, as described in Oku no hosomichi, is when Bashō arrives at the temple where stand the graves of the two wives of Satō Tsugunobu and Satō Tadanobu, loyal followers of Minamoto no Yoshitsune. Bashō is moved to find that the temple has on display both Yoshitsune’s sword and the pannier carried by Yoshitsune’s famous retainer, Benkei.[25] Yuasa again translates as if the hokku had a basic structure of four parts, although in this case his reversal of normal English sentence order allows for a three-line, one-line division that can be said to approximate the two-phrase, one-phrase division in Japanese. Still, the fact that the last, relatively independent line refers to the month rather than to the paper carp streamers (which Yuasa has confusingly called “flying banners”) certainly detracts from the concreteness of the image. Extra information has been provided in English with the use of “proudly,” a subjective judgment that is best left to the reader to make. The translation also makes a problematic reference to Satsuki as May, which gives the impression that this is some kind of spring festival. In fact, the reference is to the fifth lunar month, which corresponds to the greater part of June under the modern calendar (the season word kaminobori belongs to the “summer” category). If the seasonal reference is to be considered central to the effect of hokku, it will simply not do to substitute spring for summer. Finally, the English translation seems to call on the listener to display all three items together, while the grammar of the original uses the preexisting image of the paper carp as the basis for suggesting what to do with the other two objects. The Japanese, in other words, more clearly reveals the imagination of the poet at work, even in a hokku that is not especially serious in intent.

Corman and Kamaike puzzlingly (in view of their previous treatment of hototogisu) repeat the misleading English reference to the fifth lunar month. Furthermore, the command to “hoist high” the sword and chest results in (for me) the rather bizarre image of the two objects dangling unceremoniously from ropes. Clearly, placing the semantic elements in the same order as in the Japanese does not by itself make for appropriate translation.

Miner substitutes “Boys’ Festival” for “Satsuki,” attempting to avoid the calendar problem while relying on the reader’s knowledge to locate the festival in its proper season. The potential gain in clarity, however, once again comes at the cost of a certain verbosity (the second line alone contains eleven syllables). “Carp,” too, is a more specific image than “paper banners” or “standards,” signaling Miner’s basic policy of making concessions to the needs of non-Japanese readers. Miner does not seem to add subjective elements as Yuasa does, but otherwise his style of rendering English in fairly complete semantic units produces a similar impression of bulkiness (Miner arranges his translation in just three lines, but it actually contains one more syllable than Yuasa’s four-line version).

Britton’s translation is tightly unified by rhyme this time. It appears to be a technique adopted from Henderson, who justifies the practice on the basis of personal preference and the need to keep hokku from seeming fragmentary.[26]  Since the Imagists and other modernists have prized just this fragmentary aspect of hokku, the general reluctance of other translators and haiku poets to adopt rhyme is perhaps only to be expected. Here, especially, the result is a sing-song quality that is positively distracting. Not only that, the imagery itself has again been rearranged to achieve the rhyme. Instead of one set of images (“pannier” and “sword” in the first line) set in juxtaposition to another image (“carp streamers” in the third line) with the second line used to mediate between the two, we have all three images lumped together in the second line. This is a distortion of the basic technique for which Bashō is most justly famous, and in this case it trivializes the poem.

McCullough’s translation, too, rearranges the order of the images, transporting the paper carp to the first line, moving the pannier and sword down a line, and ending with the reference to the Fifth Month. Although, as before, a juxtaposition of sorts is maintained, the mediating function of Bashō’s second line is lost, and the result is an undue emphasis on the time of year. Thus, even while McCullough’s translation can be called faithful in that it adds no extra interpretive material and takes note of the lunar calendar by the expedient of capitalizing “Fifth Month,” the effect is by no means the same as when the hokku is read in Japanese.

Once again, Sato’s translation seems almost to flout English standards of common sense in its determination to match the presumed one-line format of the Japanese. Since the English is in one line, the phrase “with paper carps” may at first be taken as modifying “in May” rather than the verb “displayed.” It might perhaps be argued that recognizing a 2-2-2 accented-beat rhythm in the English helps to avoid that misreading by cutting off “paper carps” from the immediately preceding phrase; but if that rhythm is to be taken as the basis for a semantic yoking, then a different problem arises in the separation of “casket” from “sword” and in the subsequent linking of “sword” with “May.” In addition, Sato makes the unfortunate decision to translate the Japanese “oi” as  “casket,” a word too readily associated with the image of a coffin. “Carps,” while technically correct, is a relatively uncommon plural form that calls undue attention to itself here. And apparently Sato intends the inaccurate use of “May” to be justified on the basis of contemporary custom rather than tradition. It is interesting to note that, as a rule, Sato’s most successful translations resemble the epigrammatic forms used long ago by Hearn and Chamberlain. Sato is, however, less concerned with observing standard English grammar, and when he departs from it in his desire to establish a fixed rhythm (of sorts), the gain in rhythmic regularity can be outweighed by an increase in semantic confusion.

Considering Keene and Hamill together, we notice that although Keene has also unaccountably decided to translate “Satsuki” as “May,” he retains both the order and placement of the original images. Hamill has invented a new English word in “wind-carp,” combined the three central images in the first line, and inserted a subjective judgment with the addition of the word “proudly.” Hamill’s version may have a bit more sparkle and a better sense of English rhythm than Keene’s, but Keene’s does Bashō the service of preserving his characteristically synthetic method of constructing poetic meaning.[27]  As before, blandness seems a modest price to pay for Keene’s more consistent level of fidelity to the original.

3.  natsukusa ya / tsuwamonodomo no / yume no ato

Yuasa: A thicket of summer grass  /  Is all that remains  /  Of the dreams and ambitions  /  Of ancient warriors.
Corman  /  Kamaike:  summer grass  /  warriors  /  dreams’ ruins
Miner: The summer grasses:  /  The high bravery of men-at-arms,  /  The vestiges of dream.
Britton: A mound of summer grass:  /  Are warriors’ heroic deeds  /  Only dreams that pass?
McCullough: A dream of warriors  /  after dreaming is done,  /  the summer grasses.
Sato: Summer grass: where the warriors used to dream
Keene: The summer grasses­––  /  Of brave soldiers’ dreams  /  The aftermath.
Hamill: Summer grasses:  /  all that remains of great soldiers’  /  imperial dreams

This hokku, one of Bashō’s most famous, refers to the ill-fated members of the Fujiwara clan at Hiraizumi who perished at the sword of Minamoto no Yoritomo. The order of images in the Japanese is summer grass-warriors-remnants of dreams, again reflecting Bashō’s characteristic three-part organizational process. A major break comes after the cutting word at the end of the first line, resulting in a common variation of the standard 5-7-5 syllabic pattern. Yuasa is the only translator to depart conspicuously from this pattern, although McCullough once again reverses the first and last lines, thereby failing to suggest the correct location of the break while still separating the image of summer grasses from the other images. Corman and Kamaike are the most “literal,” stripping the English down to its bare essentials, but awkwardness results from using “warriors dreams” possessively as a compound noun and breaking it in half at the same time. Britton’s question is, of course, rhetorical, but the Japanese does not even imply the trifling doubt of a rhetorical question. It is perhaps misleading for Hamill to refer to the “imperial” dreams of soldiers who lived at a time when the imperial court was no longer the seat of real political power. Other than these relatively minor quibbles, however, this hokku seems to offer the reader an excellent chance to form a preference for any translator purely on the basis of style. And if that decision seems a hard one to make (at last with respect to more than one translator), then I think a good case has already been made for translating a poet like Bashō repeatedly. In translation, “definitive” is not a word to be used lightly, and different approaches can succeed in illuminating different aspects of the same work. I certainly would hesitate to choose any of these versions as the definitive English version of Bashō’s hokku; rather, I enjoy having the chance to read and consider them all.

4.  shizukasa ya / iwa ni shimiiru / semi no koe

Yuasa: In the utter silence  /  Of a temple,  /  A cicada’s voice alone  /  Penetrates the rocks.
Corman  /  Kamaike:  quiet  /  into rock absorbing  /  cicada sounds
Miner: In seclusion, silence.  /  Shrilling into the mountain boulder,  /  The cicada’s rasp.
Britton: In this hush profound,  /  Into the very rocks it seeps—  /  The cicada sound.
McCullough: Ah, tranquility!  /  Penetrating the very rock,  /  a cicada’s voice.
Sato: Quietness: seeping into the rocks, the cicada’s voice
Keene: How still it is here­––  /  Stinging into the stones,  /  The locusts’ trill.
Hamill: Lonely stillness—  /  a single cicada’s cry  /  sinking into stone

This is a personal favorite of mine and another of Bashō’s most famous verses, in which poetic meaning is once again generated by a synthetic process based upon the juxtaposition of images. The text of Oku no hosomichi makes it clear that the reference is to Risshaku (or Ryūshaku) Temple in present-day Yamagata Prefecture. The translations by the various translators essentially fit the stylistic patterns that have already been identified. Yuasa adds an extra line; Corman and Kamaike are cryptic; Miner is wordy; Britton uses rhyme; McCullough is reliable; Sato uses a single line (here, however, clearly demarcated semantically); Keene uses an Americanism to translate “semi,” and the alliteration does not suggest the sound of cicadas very well; Hamill is the only translator to take the (largely unwarranted) liberty of altering the original order of images. What makes this hokku especially interesting, however, is the various ways the translators have rendered “shizukasa,” “iwa” and “semi.” Corman and Kamaike are the only ones to use a single English word for the first line of the Japanese, while McCullough adds an exclamation to account for the cutting word “ya.” The other translators all try to specify the quality of the silence by adding modifiers. Five of the translators use a countable word for “iwa,” the other three use an uncountable word. A majority (five) prefer the idea of one insect, one translator hears more than one, and two translators finesse the issue by using a compound noun in which “sound” becomes the key word (although that, too, can be countable or uncountable).[28] Here, it seems to me, is another hokku where the existence of different variations in English translation tends to amplify the meaning of the original rather than disperse it. How many insects should we hear? What is the precise quality of the silence? Is it, in fact, necessary to provide a definite answer to these questions? I think it is a tribute to Bashō’s skill that the Japanese encompasses all of the possibilities suggested by the translators (it is not common to find such wide discrepancy in the use of singular and plural among experienced translators of Japanese), and at least for this hokku, I would say that the answer to the last of my three questions is no.

5.  hamaguri no / futami ni wakare / yuku aki zo

Yuasa: As firmly cemented clam shells  /  Fall apart in autumn,  /  So I must take to the road again,  /  Farewell, my friends.
Corman  /  Kamaike:  clam  /  shell and innards parting  /  departing fall
Miner: Parting for Futami Bay  /  Is like tearing the body from the clam-shell:  /  Autumn goes to its end.
Britton: Sadly, I part from you:  /  Like a clam torn from its shell,  /  I go, the autumn too.
McCullough: Off to Futami,  /  loath to part as clam from shell  /  in waning autumn.
Sato: A clam  /  separates lid  /  from flesh as autumn departs
Keene: Dividing like clam  /  And shell, I leave for Futami—  /  Autumn is passing by.
Hamill: Clam ripped from its shell,  /  I move on to Futami Bay:  /  passing autumn

This is the last hokku in Oku no hosomichi, composed as Bashō, after being greeted by disciples and friends at the end of his journey, is preparing to set off in a boat to offer prayers at Ise Shrine. Of the hokku being considered here, it probably represents the greatest technical challenge for the translator because of the wordplay surrounding “futami” (both the name of the bay that is Bashō’s destination and a phonetic combination meaning “shell” and “body”) and “wakare-yuku” (referring to the separation of a clam from its shell, the departure of the poet, and also to the passing of autumn). Yuasa simplifies matters greatly by dropping the place name and the reference to the end of autumn, and yet even then he requires twenty-seven syllables to make his translation. What is more, he has the poet address his friends directly, something not warranted by the Japanese. Corman and Kamaike also pass over the geographical reference (and so the poet’s reference to his own departure), but the whimsical “parting departing” combination is surely the most effective treatment of “wakare-yuku” to be found among these versions. Miner fits in all of the elements, but takes twenty-four syllables to do it. Britton’s use of rhyme seems less intrusive here than elsewhere, and this can probably be counted among her more successful translations. However, she, too, is unable to find room for the geographical reference, and the use of “sadly” adds an unnecessarily sentimental note to the verse. McCullough rises to the challenge nicely by explaining the meaning of the puns in a well-turned 5-7-5 translation. Of course, to explain wordplay in this manner is also to diminish its effectiveness as play, so perhaps it cannot be helped that her version (like almost all of the others) does not convey the lightness of Bashō’s original. Sato gives the distinctly mistaken impression that the clam is somehow dividing itself from the shell and, of course, he also omits the geographical reference. He has, however, translated this hokku in three lines, noting that “at least two” of the three manuscript traditions also do so.[29] This seems an oddly literal affirmation of the formal constraints imposed by the original when Sato has no compunctions about ignoring Japanese conventions regarding syllable count. Keene ends up committing a grammatical error in attempting to accommodate all the elements involved in the wordplay: starting the first line with a participial clause, he has “I” dividing like clam and shell, apparently intending this to refer to Bashō’s separation from his friends. The effect, however, is inadvertently humorous. Hamill’s version seems well done in this case, although the violence of the separation (not inappropriate to the actual act of shelling a clam) is stronger than that implied by the intransitive form of “separate” used by Bashō. Wordplay may be difficult or even impossible to render adequately in translation, but it goes without saying that in a verse obviously meant to embody the playful spirit of haikai, a translation that works will itself be lighthearted. This is why, despite its omission of one key element, Corman and Kamaike’s translation finally seems best here, and why the other translations from this relatively unknown version of Oku no hosomichi can at least stake a valid claim to the reader’s attention.

Having surveyed a total of forty versions of five hokku, we now seem very much in danger of running afoul of the law of diminishing returns. Let me therefore conclude this discussion of specific examples by making a few short (and admittedly opinionated) comments on the overall merit of each translator’s approach. Yuasa’s complete translation of Oku no hosomichi was the first to appear, and since it is included in the Penguin Books series, a certain amount of prestige has accrued to it. His versions of hokku are often the ones readers encounter first. Yuasa’s understanding of the meaning is reliable enough, and the prose sections of his translation of are quite competent, if rather pedestrian. But I have already expressed my strong reservations about the four-line model he has chosen as the paradigm for his translations of hokku, and I would say that his versions have now served their purpose and should be retired to the back shelves of the library. Penguin Books needs a fresh edition of Bashō.

Corman and Kamaike, as we have seen, take a deliberately modernist approach that also insists on following the original image order as closely as possible. William J. Higginson, for one, has pronounced this the best translation available in English.[30] I am not so sure. Granted that even in Japanese a certain amount of external knowledge is required to identify allusions and untie the syntax of many hokku, the original still strikes me as being much more conventional in expression than Corman and Kamaike’s English would lead the reader to believe. Their choppy and sometimes cryptic style often makes it almost as much of a challenge to get through the English as to read the Japanese, and the rhythm -- at least for me -- is a very different kind of rhythm. It is an interesting experiment with some notable successes, but still very much experimental in nature. I would not choose to send a class of American students to it first.

Miner’s translation is a relatively early scholarly version, and as such it admits of more substitution and approximation in its hokku than most scholars (and other translators) would now feel comfortable with. Although Miner always observes a three-line format in English, the wordiness of his translations sometimes comes close to defeating the purpose in doing so. Yet I myself did not really notice this verbosity until after I had had the opportunity to read other translations, and comparing Miner’s version with those of several later academic translators provides a useful index to the extent the modern preference for brevity and unexplained, juxtaposed images has taken hold even among specialists in Japanese literature.

The hokku translations contained in Britton’s translation of Oku no hosomichi are clearly intended for the general reader. They suffer to some extent from being removed from their prose context, the smoothness of which makes her version quite accessible to those who do not want to wade through copious notes and explanations. Unfortunately, the obtrusive reliance on rhyme and the resulting distortions of poetic structure mean that Britton’s hokku often have a different specific gravity, so to speak, than Bashō’s originals, and her style is unlikely to appeal to many future translators.

McCullough typifies the diligent, scholarly approach of many academic translators. At her best, she combines attention to detail and precise diction with a disciplined respect for the syllabic rhythms of the original, without, however, insisting on unwavering obedience to those rhythms. If there is a risk to this kind of scrupulousness, it is the risk of dryness and the willingness to subordinate other aspects of poetic structure—such as image order—to a syntactic order based on syllabic and syntactic considerations. Like most other academics, McCullough diligently includes transliterations as a way of allowing the reader to verify the appropriateness of her methods. All told, McCullough would seem to be an excellent choice for serious students of Japanese literature or for those enrolled in university survey courses.

In the introduction to his translation of Oku no hosomichi, Sato declines to justify his method of translating in single lines, stating that it is an experiment that must stand or fall on its own merit. This is surely true, but personally I tend to agree with Edwin Cranston when he says that the tension and interplay between visually separate lines offer valuable opportunities unavailable to a translator such as Sato.[31] I have remarked on the awkwardness of Sato’s hokku translations, and this awkwardness also characterizes the prose sections of his version of the travel diary. I must confess I am quite puzzled when, in the introduction, Cor van den Heuval pronounces this the “most accessible” English version to be found. That is hardly the case. While admitting the value of the experiment, I have to say that I think the experiment fails. This is the most heavily annotated of the translations, by the way, and the wealth of information is very useful. But providing so many notes in a version like this (which also contains transliterations of the Japanese) seems to constitute an invitation to the reader to regard them as a sort of corrective to the excesses of translation, and one gets the impression that the real goal is a synthesis of the two aspects. I do not think the translation stands up very well on its own.

I have heard it said that Donald Keene has read more Japanese literature than any other person alive. Such scholarship is not to be taken lightly. Still, as a translator Keene has never been the stylist that contemporaries like Edward Seidensticker have proved to be. Keene’s translations almost always leave the impression of being extremely competent but plain. This plainness is no doubt deceptive to some degree, for I know from experience how hard it is to demonstrate the same level of competence. Keene’s translation of hokku, too, is always competent and in places quite skillful, but I find it hard to say that his versions are in any sense definitive. Indeed, the complete translation of Oku no hosomichi from which they are taken seems rather unsure of its own audience, for the information on the end-flaps is written in Japanese, and Japanese translations are included even for the English notes, which themselves can be puzzling in the sort of information they do or do not offer. Perhaps if a style-minded editor had gone over the English carefully with the translator, or if Keene had designed the book more specifically for a knowledgeable English-speaking readership, more sparkle would have resulted.

The versions offered by Hamill are also disappointing in my view, for reasons that have already been stated. Hamill simply misleads the reader too often regarding the basic meaning (“too often” is a relative term, and I do not mean to imply a very high statistical frequency). I was, quite frankly, startled in the prose section of Oku no hosomichi to find him refer to an honest innkeeper nicknamed Hotoke Gozaemon in Japanese as “Joe Buddha,” and the dropped-subject, diary-like style favored by Hamill in an attempt to convey the compression of Bashō’s prose seems to me to misrepresent the stylistic polish of the original. Still, while the prose deserves consideration as an attempt to come to terms with the stylistic implications of the Japanese, the hokku translations lack authority.

What, then, are some of the general conclusions that can be drawn about translated hokku by Bashō? First of all, as I have already noted, it is important to recognize that definitive versions of individual hokku are very hard to come by. In spite of the criticism I have leveled at some of them, most have something to recommend them and most also have drawbacks. In this sense, the brevity of hokku actually works to the reader’s advantage because it allows different versions of the same verses to be compared without making unreasonable demands on one’s time. It is a luxury not available to many other literary genres. The richness of meaning that has accrued to Bashō’s hokku in English is very much the product of the different versions published by different translators over the years. And the fact that the truest appreciation for the Japanese emerges after reading multiple versions is not so much an ironic comment on the value of translation as it is an indication of the potential of the hokku form as realized by its first great master in Japanese (the recognition of this potential, beginning with the work of Blyth and Henderson, counts as a permanent change in the prestige of the genre in English).

A second point is that the issues raised by the earliest translated versions remain quite volatile today, both among scholars and within the so-called haiku community of poets in English. There is still no unity, for example, over the most appropriate format to use when writing in English. This is so despite the fact that certain methods (Yuasa’s four-line method and Britton’s preference for rhyme, for instance) have failed to win the support of the majority of either translators or original poets. One- or two-line epigrammatic forms, 5-7-5 three-line schemes, rhymed lines, and accented-beat patterns -- all have attracted and continue to attract supporters, for experimental purposes at least. The interesting thing is that these different formats have not developed in isolation but have had to compete with and accommodate each other in a way that has resulted, for example, in Sato’s combination of epigrammatic form with the modernist technique of syntactic disruption and an English-based rhythmic pattern that nevertheless appeals to one aspect of traditional Japanese practice for its authority. Even scholarly translators have been affected by the actual composition of haiku in English to the extent of modifying their own procedures.[32]

Both of the foregoing points are signs of the lively atmosphere surrounding the production of haiku in English (and other languages) today. And all this activity would seem to indicate that we are presently in an age of interaction and convergence, where the tension created between traditional versions, translated versions, and foreign-language versions of hokku / haiku is coming to inform our awareness of the achievements, limitations, and possibilities of each category. It will be interesting to see where this interaction takes haiku in the future.



[1]   The third edition of this anthology was published by W.W. Norton in 1999, and includes about 850 poems in more than 400 pages. Among the other anthologies, the Red Moon Press series, under the editorship of Jim Kacian, is notable for having published a new volume of original English haiku each year since 1996.

[2]   The translations to be discussed here are those contained in the following books, which are listed in order of original publication: Nobuyuki Yuasa, trans., Narrow Road to the Deep North and Other Travel Sketches (Harmondsworth, England: Penguin Books, 1966); Cid Corman and Susumu Kamaike, trans., Back Roads to Far Towns: Bashō’s Oku-No-Hosomichi (1968; Hopewell, N.J.: The Ecco Press, 1996); Earl Miner, trans., Japanese Poetic Diaries (Berkeley, Calif.: University of California Press, 1969); Dorothy Britton, trans., A Haiku Journey: Narrow Road to a Far Province (1974; Tokyo: Kodansha International, 1980); Helen Craig McCullough, ed. and trans., Classical Japanese Prose: An Anthology (Stanford, Calif.: Stanford University Press, 1990); Donald Keene, trans., The Narrow Road to Oku (Tokyo: Kodansha International, 1996); Sam Hamill, trans., Narrow Road to the Interior and Other Writings (1999; Boston: Shambhala Publications, 2000). Because of the many variations in the translated titles, I will be referring to the travel diary here solely by its Japanese title.
      It should be noted that a distinction is being observed here between the words “haiku” and “hokku.” It is now widely recognized in the English-speaking world that the hokku (literally,“starting verse”) was originally not an independent form of poetry but simply the first of a sequence of linked verses (renga or renku) that typically went on until a conventional length—thirty-six, fifty, or one hundred verses—was reached. The first poet would begin by composing a verse in 5-7-5 syllabic form, and the second poet would add a verse in 7-7 form to cap the first verse and produce a complete 31-syllable tanka. Then a third poet would add another 5-7-5 verse which, when added to the previous 7-7 verse, constituted a second poem formally independent of the first and yet related to it by a fairly complex set of linking rules as well as by the shared lines. Although the importance of the opening verse meant that poets in Bashō’s day often composed hokku separately, there was always the expectation that hokku would (or at least could) be used to start a complete sequence. The term “haiku,” as used to refer to an independent poem in 5-7-5 syllabic form, was popularized by the modern poet Masaoka Shiki (1867-1902). Following what is now becoming standard practice, I therefore use “hokku” to refer to the “starting” verses composed by Bashō, and “haiku” to refer to modern verses that are intended to be read as independent poems.

[3]   Makoto Ueda, Bashō and His Interpreters (Stanford, Calif.: Stanford University Press, 1991), p. 5.

[4]   W.G. Aston, History of Japanese Literature (London: William Heinemann, 1899), p. 294. “Haikai” is the general term for the genre of “playful” linked verse that emerged in the Muromachi period and in which Bashō worked; the hokku was the opening verse of a haikai sequence. Aston’s book remained the only reasonably complete history of Japanese literature in English until Donald Keene’s version began to appear in 1976. It is still available from Tuttle.

[5]   Aston, p. 395. Aston translates in three lines of English, but maintains no regular syllable count in the very small number of hokku he offers (he disposes of the genres haikai, haibun, and kyōka in the space of ten pages).

[6]   Aston, p. 399.

[7]   Basil Hall Chamberlain, “Bashō and the Japanese Poetical Epigram,” Transactions of the Asiatic Society of Japan, vol. 30 (1902), pp. 243-362. Reprinted in Japanese Poetry (London: John Murray, 1910), pp.145-260, and also in the final volume of Early Japanology: Aston, Satow, Chamberlain, 4 vols. (Tokyo: Yūshōdō Press, 1997), pp. 305-426.  See also the “Poetry” section of Chamberlain’s Things Japanese, 5th ed. (London: J. Murray, 1905), pp. 374-382, where Chamberlain refers to hokku as the “limit of the little” in poetry.

[8]   Chamberlain, “Bashō and the Japanese Poetical Epigram,” p. 307.

[9]   Chamberlain, “Bashō and the Japanese Poetical Epigram,” p. 309.

[10]   Pound’s 1913 “In a Station of the Metro” is sometimes put forward as the first English haiku: “The apparition of these faces in the crowd; / Petals, on a wet black bough.” This assessment is not universally acknowledged.

[11]   For a convenient, if brief, discussion of haiku in English starting with the Imagists, see Haruo Shirane, Traces of Dreams: Landscape, Cultural Memory and the Poetry of Bashō (Stanford, Calif.: Stanford University Press, 1998), pp. 44-51. More detailed information can be found in chapters 4 to 6 of William J. Higginson, with Penny Harter, The Haiku Handbook: How to Write, Share, and Teach Haiku (1985; Tokyo: Kodansha International, 1989), pp. 49-83. Also see George Swede, “Haiku in English in North America,” publication date unknown, 15 March 2000 <>, which apparently combines articles previously published in Haiku Canada Newsletter, vol. 10, no. 2, January 1997, and vol. 10, no. 3, March 1997.

[12]   In his preface to the first volume of Haiku, 4 vols. (1949-1952; Tokyo: Hokuseidō, 1981-1982), Blyth states flatly that “haiku are to be understood from the Zen point of view.” For biographical information about Blyth, see James Kirkup’s introduction to The Genius of Haiku: Readings from R.H. Blyth (Tokyo: Hokuseidō, 1994); in Japanese, see Yoshimura Ikuyo, R.H. Buraisu no shōgai: zen to haiku o aishite (Tokyo: Dōhō Shuppan, 1996).

[13]   R.H. Blyth, A History of Haiku, vol. 2 (Tokyo: Hokuseidō, 1964), p. 351. Quoted in Harold G. Henderson, Haiku in English (Rutland, Vermont: Charles E. Tuttle Company, 1967), p. 32.

[14]   Harold G. Henderson, An Introduction to Haiku (Garden City, New York: Doubleday Anchor Books, 1958), and Haiku in English, cited above.

[15]   The first English haiku contest sponsored by Japan Air Lines in 1964 is often mentioned as a watershed. About 41,000 entries were submitted.

[16]   Higginson, The Haiku Handbook, p. 96.

[17]   Higginson has written two books on the subject: The Haiku Seasons: Poetry of the Natural World (Tokyo: Kodansha International, 1996) and, with Meagan Calogeras, Haiku World: An International Poetry Almanac (Tokyo: Kodansha International, 1997). The first book makes the argument on a more theoretical basis; the second represents an actual attempt to create such a seasonal almanac.

[18]   Of the three that stray from the basic pattern (all are in 6-7-5 syllabic form), two appear to do so mostly for linguistic reasons. One of these starts with the phrase oi mo tachi mo (literally, “pannier and sword and”), which consists of two nouns joined in a parallel construction; the other begins with Atsumiyama ya, which is a place name followed by a “cutting word” (kireji) conventionally used when place names appear in the first line of a hokku. The extra syllable in the opening line of the third exception (tsuka mo ukoke / wa ga naku koe wa / aki no kaze) seems to reflect an intent to add emotional emphasis, which makes it a unique case.

[19]   Again for reasons of space, line divisions in the translations are indicated by virgules. The divisions shown in the Japanese originals are matters of convention and convenience rather than a reflection of the practice of the poet. For Yuasa’s (finally unconvincing) reasons for adopting a four-line method, see The Narrow Road to the Deep North, p. 48.

[20]   Yuasa repeats syllabic patterns a total of only four different times.

[21]   Robert H. Brower and Earl Miner, Japanese Court Poetry (Stanford, Calif.: Stanford University Press, 1961). This was also my introduction to the field, and as Edwin Cranston observes when making the same point in the introduction to A Waka Anthology: The Gem-Glistening Cup (Stanford, Calif.: Stanford University Press, 1993), it may be that one’s own preferences as a translator are determined by one’s first encounter with a translation, for I have always regarded the five-line translations of tanka by Brower and Miner as basic models for translation. With regard to hokku, however, other translations have succeeded in making me feel the need for greater brevity.

[22]   Among her translations (in addition to the anthology of prose quoted earlier) are complete versions of Ise monogatari, Kokin wakashū, Heike monogatari, Taiheiki, Eiga monogatari (with W.H. McCullough), and Gikeiki.

[23]   Sato argued for one-line versions of tanka as early as the 1987 article “Lineation of Tanka in English Translation,” Monumenta Nipponica, 42:3 (Autumn 1987). His translation of Oku no hosomichi appears to signal an attempt to extend the same principle to hokku.

[24]   Although Keene’s full version of Oku no hosomichi was published only recently, he did translate selections for his 1955 Anthology of Japanese Literature. It may be that he was working from previous notes and inadvertently repeated an earlier, immature error (in Japanese, both hototogisu and kankodori, or kakkō, are cuckoos, so he may have wanted to draw a distinction between them). Still, “nightingale” is both factually inaccurate and the usual prewar translation for uguisu (now “bush warbler”), so it should not appear in this translation.

[25]   This hokku serves to demonstrate that the current haiku “rule” about emphasizing the present moment ignores the actual practice of Bashō, for whom the past was a constant preoccupation (Oku no hosomichi itself explicitly invokes the experience of past poets such as Sōgi and Saigyō, present already in the travel diary’s famous opening lines). For further consideration of this point, see Haruo Shirane, “Beyond the Haiku Moment: Bashō, Buson, and Modern Haiku Myths,” Modern Haiku, XXXI:1 (Winter-Spring 2000), pp. 48-63.

[26]   Henderson, An Introduction to Haiku, pp. ix-x.

[27]   Space prevents extensive citation of the kind of misleading carelessness I have already pointed out in Hamill’s translations, but in one hokku, for example, Hamill has young girls making dye when they are in fact dyeing cloth; in another, he describes a famous Chinese beauty as “wrapped in sleeping leaves” when a comparison with mimosa drooping in the rain is intended; and in a third, he translates a line as “Tremble, oh my grave,” when the Japanese obviously refers to another person’s tomb.

[28]   Ueda, in Bashō and His Interpreters, adopts the plural form in his translation of the same hokku; Shirane, in Traces of Dreams, prefers the singular. Ueda used the singular in an earlier translation published in Matsuo Bashō (1970; Tokyo: Kodansha International, 1982), but has apparently changed his mind since then. My own experience with cicadas suggests to me that plural is best, but I must admit that the choice is timorously made.

[29]   Sato, p.132.

[30]   To verify this, it will be necessary for the reader to locate the book on’s Web site and then read Higginson’s review, the URL for which is too long to include here.

[31]   Cranston, A Waka Anthology: The Gem-Glistening Cup, p. xix.

[32]   Both Ueda and Shirane address the issues raised by English “haikuists,” and in Ueda’s case it is instructive to compare the versions of translated hokku that appear in Matsuo Bashō with those that appear in Bashō and His Interpreters. Ueda and Shirane, however, also revert to the practice initiated by Henderson in An Introduction to Haiku in providing word-for-word translations along with the final polished version. The attempt to have things both ways is a recognition of the provisionality of the translation no less than an exercise in scholarly diligence. On a related note, it can be seen that the commonly advanced complaint that a 5-7-5 syllabic count constitutes an arbitrary constraint in English certainly cannot be laid at the door of translators, who as a group have never been dogmatic about form in English.


Syllable counts for eight English translations of all 50 hokku by Bashō in Oku no hosomichi. The arrangement is based simply on the length of the first "line," with the length of subsequent lines in turn determining the order when the first line is the same length (there is no correlation with the sequence of poems, and using total syllable count would result in a somewhat different ordering). Except in the case of Sato, shading is used for multiple instances of the same syllable count.

Yuasa Corman Miner Britton McCullough Keene Sato Hamill
3-4-5-5 1-5-2 3-8-5 4-8-5 3-7-5 3-7-5 10 3-6-6
4-2-3-4 1-6-4 4-7-5 4-8-6 4-5-5 4-6-5 10 3-6-6
4-4-4-4 2-3-4 4-7-5 5-6-6 4-6-5 4-6-5 10 4-5-6
4-5-6-6 2-3-4 4-8-5 5-7-4 4-6-5 4-7-5 10 4-6-5
4-5-6-7 2-4-4 4-8-6 5-7-5 4-6-6 4-7-5 10 4-6-5
4-5-6-7 2-4-5 4-8-6 5-7-5 4-7-4 4-7-6 11 4-6-5
4-5-9-8 2-4-5 4-9-5 5-7-5 4-7-5 4-7-6 11 4-7-4
4-6-3-5 2-5-2 4-9-6 5-7-5 4-7-5 4-7-6 11 4-7-5
4-6-6-5 2-5-4 4-9-7 5-7-5 4-7-7 5-5-4 11 4-7-5
4-6-11-6 2-5-4 4-10-4 5-7-5 4-7-7 5-5-5 12 4-7-5
4-7-8-5 2-5-5 4-11-5 5-7-5 5-6-3 5-6-4 12 4-7-5
4-10-5-7 2-6-3 4-11-7 5-7-5 5-6-4 5-6-5 12 4-8-5
5-4-4-2 2-6-3 4-12-6 5-7-5 5-6-4 5-6-5 13 5-4-5
5-5-5-6 2-6-4 5-6-6 5-7-5 5-6-4 5-6-5 13 5-5-5
5-5-6-6 2-7-3 5-8-4 5-7-5 5-6-5 5-6-5 13 5-5-5
5-5-6-6 3-2-2 5-8-5 5-7-6 5-6-5 5-6-5 13 5-5-6
5-6-5-6 3-2-3 5-8-5 5-7-6 5-6-6 5-6-5 13 5-6-4
5-6-6-7 3-3-5 5-8-5 5-7-6 5-7-4 5-6-5 13 5-6-4
5-6-7-5 3-5-5 5-8-6 5-7-6 5-7-4 5-6-5 13 5-6-4
5-6-7-6 3-6-2 5-8-6 5-7-6 5-7-5 5-7-4 13 5-6-4
5-6-7-6 3-6-4 5-8-7 5-7-6 5-7-5 5-7-4 13 5-6-5
5-7-5-6 3-6-5 5-9-5 5-7-6 5-7-5 5-7-4 13 5-6-5
5-8-6-7 3-7-5 5-9-5 5-8-5 5-7-5 5-7-5 13 5-6-5
5-8-7-7 4-3-4 5-9-5 5-8-5 5-7-5 5-7-5 13 5-6-5
5-9-7-7 4-3-5 5-9-5 5-8-5 5-7-5 5-7-5 13 5-6-6
5-10-7-8 4-4-2 5-9-6 5-8-5 5-7-5 5-7-5 13 5-6-6
6-2-9-5 4-4-3 5-9-7 5-8-5 5-7-5 5-7-5 14 5-6-6
6-4-7-5 4-4-3 5-10-5 5-8-5 5-7-5 5-7-5 14 5-6-7
6-5-5-5 4-4-3 5-10-5 5-8-6 5-7-5 5-7-5 14 5-7-4
6-5-9-5 4-5-4 5-10-6 5-8-6 5-7-5 5-7-5 14 5-7-4
6-6-4-5 4-5-4 5-10-6 5-8-6 5-7-5 5-7-5 14 5-7-4
6-6-4-6 4-5-5 5-10-6 5-9-6 5-7-5 5-7-5 14 5-7-5
6-6-4-6 4-5-5 5-10-7 6-7-5 5-7-5 5-7-5 14 5-7-5
6-6-4-7 4-6-3 5-10-7 6-7-5 5-7-5 5-7-5 14 5-7-5
6-6-5-5 4-6-3 5-10-7 6-7-5 5-7-5 5-7-5 14 5-7-6
6-6-7-6 4-6-5 6-7-5 6-7-5 5-7-5 5-7-5 14 5-7-6
6-7-6-7 4-6-5 6-8-5 6-7-5 5-7-6 5-7-6 14 5-7-6
6-7-6-8 4-6-5 6-8-5 6-7-6 5-7-6 5-7-6 14 5-7-7
6-7-8-5 4-8-3 6-8-6 6-7-6 5-7-6 5-7-6 15 5-7-8
6-10-4-8 5-3-2 6-8-6 6-7-6 5-7-6 5-8-5 15 5-8-4
6-10-8-6 5-4-2 6-9-5 6-7-7 5-7-7 5-8-5 15 5-8-5
7-4-5-4 5-4-3 6-9-6 6-7-7 5-8-4 5-8-5 15 5-8-5
7-5-4-7 5-4-4 6-10-7 6-8-5 5-8-5 5-8-5 15 5-9-5
7-5-7-5 5-4-4 6-11-5 6-8-5 5-8-5 5-8-6 15 6-5-5
7-6-4-7 5-5-3 6-11-6 6-8-5 5-8-5 5-8-6 15 6-6-4
7-9-7-5 5-5-5 7-9-6 6-8-5 6-7-5 5-8-7 16 6-7-3
7-11-7-5 5-6-2 7-9-7 6-8-6 6-7-5 6-7-5 16 6-7-5
8-8-7-9 5-6-3 7-10-6 6-9-5 6-7-6 6-7-5 17 6-7-5
8-3-6-7 5-6-4 7-10-6 7-7-5 6-7-6 6-8-5 17 6-8-4
8-6-9-4 5-7-4 7-11-6 7-9-6 6-8-5 7-6-4 17 8-10-5

Jlit Home