Three Hokku by E. E. Cummings

This essay first appeared in the Haiku Society of America’s Frogpond XVI:1, Spring–Summer, 1993, pages 51–56. Please also note the more recent postscript at the end of this essay, which describes how Cummings’ three hokku have been set to music. For a more thorough analysis of haiku’s influence on the poetry of E. E. Cummings, see “The Haiku Sensibilities of E. E. Cummings.”       +       +

Readers of twentieth-century American poetry may have wondered about the impact of haiku on the imagistic poems of E. E. Cummings. In the work of other poets as diverse as Ezra Pound, Amy Lowell, and Adelaide Crapsey, the influence of haiku and related Japanese forms is directly traceable. Further influence on Imagism may be seen in the short poems of Carl Sandburg, William Carlos Williams, and others. One has only to consider the station of the metro, little cat feet, and the red wheelbarrow to know the sway of haiku in modern American Imagism. Indeed, so much does depend upon it.

        But does the poetry of E. E. Cummings depend on haiku? A search of Katharine McBride’s concordance to Cummings’ poetry reveals no use of haibun, haiku, hokku, renga, renku, senryu, or tanka.[1] Yet, for all the evidence of haiku’s influence found directly in the poetry of E. E. Cummings (a topic also worthy of discussion, but beyond present scope), no evidence has been readily available to prove that the Cambridge poet knew of haiku, or, more importantly, had written in the form himself.

        Until now. The recent publication of the revised, corrected, and expanded edition of E. E. Cummings: Complete Poems 1904–1962, edited by George J. Firmage, includes three such poems under the heading “Hokku.”[2] The corrections and revisions that augment the previous edition of the book (1972) make this new collection (1991) the most dependable source of Cummings’ poetry as he intended it. Yet of paramount interest here is a group of 36 previously uncollected poems, said to have been published in sometimes obscure periodicals and other books and anthologies between 1910 and 1962. While no sources are listed that might indicate the year of composition or publication, the three hokku together are numbered 25 among the 36 uncollected poems in that period [they date from the April 1916 issue of Harvard Monthly, shown below as they originally appeared]. They are worth considering for their significance to the history of haiku in English:

I care not greatly

Should the world remember me

In some tomorrow.


There is a journey,

And who is for the long road

Loves not to linger.


For him the night calls,

Out of the dawn and sunset

Who has made poems.

        As can be seen, all three poems use the 5-7-5 syllable count, and are surprisingly weak on images. Much of Cummings’ poetry is rich in image, yet the poet has often been thought of as a verbal poet since in many of his poems he attempts through nonstandard usage to make abstract parts of speech function concretely (consider, for example, various lines of “anyone lived in a pretty how town”[3] and numerous other poems). In the verses above, the second and third speak only of a “long road” and “dawn and sunset,” and the first hokku contains no image at all. One could argue that “the world” is an image of sorts, albeit a poor one, but in this poem he means in metonymic fashion the people of the world, not the earth itself. At any rate, since haiku has been described as the poetry of the noun—that is, the thing, visible, touchable, even turnable in one’s hands—these three haiku seem poor by today’s image-centered standards.

        Yet haiku poets stand at a different vantage point, with greater experience in English-language haiku than Cummings had when he penned these poems. In English, the counting of syllables to fit the 5-7-5 format has been largely abandoned by today’s best poets.[4] In Cummings’ defense, none of the three poems has awkward line breaks nor seems padded to fit the 17-syllable pattern. But the pattern he used was no doubt accepted for haiku at the time, since many early translators chose (incorrectly, I maintain) to equate the always short Japanese onji with the widely variable length of the English syllable [the term “onji” is now considered obsolete; the correct term is “on,” pronounced as own, but quicker]. The counting of syllables is a relatively superficial aspect of haiku, however. Our vantage point also includes greater experience with matters of subject, juxtaposition, language, suchness, present tense, seasons, internal comparison, and many other aspects as they relate to the crafting of haiku. Note that none of Cummings’ three poems contains a direct seasonal reference. A “long road” may suggest autumn, but this would be an arbitrary, personal inference. Since Cummings was playing with the Japanese form, did he not know of the tradition of including seasonal references? And if he did know, why did he leave them out? We may never learn the answers to these questions, nor the context in which Cummings composed his three hokku. We are left simply with the text of the poems themselves, and they remain intellectual and somewhat abstract, not haiku-like at all.

        To be specific, Cummings’ first hokku immediately starts with the ego, “I . . . ,” and then pronounces a judgment, a personal ego-assertion: “I care not . . . .” His poem grows imprecise in referring to the entire “world,” and even more abstract with the third line: “some tomorrow.” In substance, the poet states that he is not concerned about being remembered in the future. But in a Zen way does Bashō not say the same thing with his most famous poem?


Old pond . . .

   a frog leaps in

      water’s sound.[5]


        In Bashō’s poem, the ego is absent, as it usually is in successful haiku. The persona is present—someone apparently hears the water’s sound—but that is all. While many yardsticks may be used to measure a haiku, consider some of James W. Hackett’s “Suggestions for Writing Haiku in English”[6] as they apply to the first of Cummings’ hokku: Is the poem centered in the present moment? No (“tomorrow” is not now). Is it about nature? No. A close observation of nature? No. Interpenetration and identification with nature? No. A poem of solitude, quietness, and reflection? Perhaps (reflection and thoughtfulness, yes, but this is intellectually told rather than implied or shown). Is the verse about nature just as it is, without judgment? No. Are words chosen carefully? More or less. Are verbs written in present tense? Yes. Seasonal reference? No. Common language? Yes. Three lines of approximately 17 syllables? Yes. Avoidance of end rhyme? Yes. Does the poem express “lifefulness,” rather than just beauty? Unclear. Is it intuitive rather than abstract or intellectual? Definitely not. Humorous? No (but this is not necessary). Carefully polished? Yes, it seems so, at least verbally (if not content-wise), given what Cummings may have known about haiku. Is the subject or object of the poem overcome by the language used to describe it? Yes, since the poem expresses a thought, rather than portraying an unjudged image. Does it honor the senses with awareness? No; the poem offers an intellectual awareness, but no awareness through the senses. Clearly, the spirit of haiku is lacking. Many other questions could be asked of the poem, but, in short, the reader can only conclude that the first of these three poems is not a haiku.

        Nor are the second and third. They may be slightly more imagistic, but they also fail against most of the standard yardsticks applied to haiku. Again, to be specific, the second poem expresses a judgment (that the traveler journeying over a long road loves only to travel, and does not linger). A true haiku would simply depict the traveler not lingering over chrysanthemums, say, thereby implying that the true traveler loves only to travel. Also, the somewhat vague image of the long road is not juxtaposed with any other image, and offers no resonance through internal comparison with another object or setting. Similarly, the third hokku expresses an abstraction, that somehow night calls to the poet. How? What specifically happens to create this impression? That is where the haiku may be found. The general images of dawn and sunset are not enough, and, in fact, they express two widely separated moments rather than one present moment, as is the norm of successful haiku. The third poem asserts less judgment than the first two hokku (a factor of the poem’s focus on the night, rather than on the poet and how he feels about it), but it still fails to satisfy as a well-crafted English-language haiku.

        From my experience as an editor of a haiku journal and as a judge of haiku contests, all three of Cummings’ haiku are typical of what any beginner might write. Most neophytes initially concern themselves with just the formal aspect of haiku. If they are taught that the haiku is a 5-7-5 nature poem, then that is what they will start writing, yet they are also likely to assert their own egos, opinions, and judgments of nature unless they are instructed otherwise. Without study, instruction, or practice, no beginner can be blamed for this all-too-common shortcoming of asserting judgment and intellectualization. In that sense, E. E. Cummings should also be forgiven. A greater case for the imagistic impact of haiku on the other poetry of E. E. Cummings is the subject of a larger article. But beyond these three examples, Cummings gives no evidence of exploring the haiku form, let alone mastering it, even to the limited extent in the mainstream poetics of Richard Wright, Jack Kerouac, Gary Snyder, A. R. Ammons, or Etheridge Knight.

        Perhaps these three hokku might be considered as a set, as a nine-line whole. Without knowing the poet’s intent, this is only an unlikely guess. Biographer Richard S. Kennedy refers to the “pattern of near-haiku stanzas” that Cummings wrote to a lover in 1923,[7] but in the example cited of four three-line verses,[8] only an approximate and ultimately superficial syllable count bares any resemblance to traditional haiku. I will not take space to quote them here, but the set of three-line verses clearly flows as a larger whole, even though they are haiku-like. So, are E. E. Cummings’ three hokku meant to be a single poem? I would suggest not. The only factor that suggests any unity to the nine lines is the self-reference to the poet in the first verse, and the direct identification of one “who has made poems” in the last verse. The poet may also be seen to be on a journey, but this theme is tenuous at best. Rather, it seems clear that they are independent. Indeed, because they are identified as “Hokku” at the top of the page, and due to the fact that the singular also functions as the plural in Japanese (one haiku; many haiku), and since all three together could not be a single hokku due to excess combined length, it seems obvious that each verse is offered as an independent experiment in the hokku form as Cummings understood it.

        As such, it is gratifying to haiku poets to know that E. E. Cummings did in fact experiment with the haiku form. By today’s standards it seems that he failed. He may have succeeded to some degree by whatever standards existed in the time when he wrote his hokku, but his attempts seem overly Westernized and too intellectual. They may be enjoyed regardless of labels for their sense or meaning, but in the end, their primary (and minor) significance in the development of English-language haiku is historical, and not literary. The influence of haiku upon the majority of imagistic poets of the early twentieth century is as clear as a sharply rung school bell. At the very least, that bell now rings with concrete evidence for E. E. Cummings.


[1] McBride, Katharine Winters, ed. A Concordance to the Complete Poems of E. E. Cummings. Ithaca, New York: Cornell U. P., 1989. Note that this concordance is based on the 1972 edition of E. E. Cummings: Complete Poems 1913–1962 (compare with Firmage, below).

[2] Firmage, George J., ed. E. E. Cummings: Complete Poems 1904–1962, rev. ed. New York: Liveright, 1991. p. 875.

[3] Firmage, p. 515.

[4] A survey of major English haiku journals quickly reveals that most haiku are approximately 10 to 14 syllables, especially in the following journals: Frogpond, published by the Haiku Society of America and edited by Sylvia Forges-Ryan, Modem Haiku, edited by Robert Spiess, Brussels Sprout, edited by Francine Porad, and Woodnotes, published by the Haiku Poets of Northern California, edited by Michael Dylan Welch. The fact that most of the best English haiku are fewer than 17 syllables is also demonstrated (and recommended) in Cor van den Heuvel’s The Haiku Anthology (Simon & Schuster, 1986 and Fireside, 1991), Bruce Ross’s Haiku Moment (Tuttle, 1993), and William J. Higginson’s The Haiku Handbook (McGraw-Hill, 1975 and Kodansha, 1989). These books are especially recommended for anyone interested in learning about the growth and practice of haiku in English; the first two books are anthologies, and the third is an excellent introduction to the arts of reading, writing, and teaching haiku.

[5] Sato, Hiroaki. One Hundred Frogs: From Renga to Haiku to English. New York: Weatherhill, 1983, p. 161. William J. Higginson, trans.

[6] Hackett, J. W. The Zen Haiku and Other Zen Poems of J. W. Hackett. Tokyo: Japan Publications, 1983. pp. 255–256. James Hackett was one of earliest poets writing in English to concentrate on the haiku form. In 1964 he won the first international Japan Air Lines haiku contest out of over 30,000 entries.

[7] Kennedy, Richard S. Dreams in the Mirror: A Biography of E. E. Cummings. New York: Liveright, 1980. p. 87.

[8] Kennedy, p. 88.


In 2007, Gary Bachlund set Cummings’ three hokku to music, for voice and piano, as shown above. You can download the sheet music for Bachlund’s composition online, where you can also read the composer’s comments about the hokku. In describing the poems, Bachlund says that “The traditional 3 line form of alternating 5-7-5 syllable lengths is not quite exact for the last strophe,” presumably meaning that “Who has made poems” in the third hokku does not count as five syllables. This observation is incorrect, however, because “poems” does indeed count as two syllables. One need not be a linguist to know this fact—the raised dot in “po·em,” which appears in every standard dictionary (such as, indicates that the word is two syllables. Consequently, Cummings’ last line does contain the five syllables that Bachlund presumes it should contain, but thinks it doesn’t. Bachlund’s score presents the word as a single note, as if pronounced as “pomes,” but a correct two-syllable pronunciation would easily fit his composition as it stands, either by elision or by repeating the same note for the second of the two syllables. For more information on syllables in haiku, please read “What Is a Syllable?”

        In addition, while Bachlund does rightly say that “The form . . . is of less consequence than the message therein,” I believe he still misses two points. The first is that he misunderstands the form for haiku in English, which has been widely mistaught as requiring 5-7-5 syllables in English. Japanese haiku do not count syllables but mora (this is a Western linguistic term; the term for what the Japanese count is on, pronounced as “own,” but quicker). As a result, linguists, translators, and leading haiku poets almost uniformly assert that seventeen syllables in English is significantly longer than the seventeen sounds they count for Japanese haiku and that about 10 to 14 syllables is equivalent to the 17 sounds in Japanese. To illustrate the difference between the two languages, the word “haiku” is two syllables in English, but is counted as three sounds in Japanese. Not all words exhibit this difference, but enough do to make it a sort of urban myth that haiku should be written in 17 syllables in English. The second point is that, in haiku terms, the “message” in Cummings’ three hokku is actually poor—at least in terms of the haiku tradition, for reasons that the essay presented here attempts to explain. I should also add that “hokku” is not really an “older spelling for haiku,” as Bachlund claims. Rather, “haiku” was Shiki’s new (invented) term for an independent verse form that grew out of hokku; hokku is actually the “starting verse” of longer linked-verse compositions first known as “renga” and later as “renku.” The act of renaming hokku as haiku had the dramatic effect of energizing this verse form, a literary reformation that has spread around the world only in the last hundred years since Shiki's death.

        Bachlund also states that “the use of three of these singular forms in one poem is also a variant [of haiku].” This statement exhibits a misunderstanding of the word “Hokku,” used to title Cummings’ three poems. That’s because “hokku” (like “haiku”) is both singular and plural, which means that Cummings correctly understood each of these three poems to be independent poems, and does not suggest that all three poems together are somehow one hokku (for what it’s worth, haiku poets, translators, and scholars consider it incorrect to ever say “hokkus” or “haikus,” and such a treatment is nearly always the mark of an uninformed neophyte, but in this case the lack of a concluding “s” had the effect of making Bachlund not realize that the three poems were being referred to plurally rather than singularly). I have never seen any evidence to suggest that Cummings meant these three poems as one single poem, and examples of many additional unpublished haiku or hokku in the Cummings archives at Harvard University indicate that each three-line poem is intended as an independent poem, as I believe they should be. Cummings also read Blyth, where each three-line poem was clearly independent (other than when appearing in linked-verse collaborations known as renga or renku). The problem is that Bachlund incorrectly apprehends the three hokku as one single poem, thus making a puzzling composition, at least lyrically, akin to reading any other three poems by one author as one single poem. In the case of these three poems, I find that they do not make much unified sense when interpreted as if they are a single poem, so the composition seems flawed in its apparent assumption that they are intended as a single poem. Nevertheless, it is pleasing to know that Cummings’ three hokku have been set to music. I do not play music well enough to be able to perform this piece myself, but I’ve heard others play it and I’m pleased to hear how Gary Bachlund’s music sounds.

—8 November, 1, 24 December 2010