A Hundred Gourds 4:1 December
current issue : haiku : tanka : haiga : haibun : renku : expositions : feature : submissions : editors : search : archives
Buson or Shiki: The
Confusing Authorship of the “Two Autumns” Poem
by Michael Dylan Welch
In reading poems translated from
Japan, it does not take long before one comes
across the following classic haiku, one that is
often attributed to Buson but is actually by
for me going
for you staying—
This poem is effective because of its memorable
wording and for its fresh notion of separation
indicated by the supposedly different autumns.
Yet, though two people are separating, perhaps
their autumns will be shared in spirit. It is this
oneness, despite separation, that gives the poem
appeal and resonance. R. H. Blyth says that “the
whole of life is given here, our meetings, our
partings, the world of nature we each live in,
different yet the same” (page xxx (sic), Haiku
The poem has had so much appeal, in fact, that in
1990 Garry Gay named the annual reading series of
the Haiku Poets of Northern California after this
poem. Every year HPNC publishes an anthology of
poems by the featured readers in the Two Autumns
reading series, and the first book was itself
named Two Autumns. HPNC’s press is also
named Two Autumns Press after this first book.
Garry found the poem in Harold G. Henderson’s An
Introduction to Haiku (Doubleday, 1958),
and wrote to the publisher to secure permission to
use the poem for the group’s purposes, even though
he chose to revise Henderson’s version. As of
2014, HPNC has published more than thirty books
with its Two Autumns Press imprint, including more
than two dozen for the reading series, featuring a
hundred poets since the series began, including a
retrospective twenty-fifth anniversary anthology
in 2014, titled One Song. With its
singular song of haiku, Two Autumns is, I believe,
the longest-running haiku poetry reading series
What’s interesting, however, is an error in this
poem’s attribution. This error is presumably of
particular interest to the Haiku Poets of Northern
California, given its emphasized association with
the “two autumns” poem, but also of interest to
anyone who has encountered the poem in numerous
books of haiku translation over many decades.
Henderson’s book identifies the author as Buson,
the second of Japan’s four great haiku masters,
who lived from 1716 to 1784. Yet Buson did not
write the poem. While the attribution error may be
minor, it has been a persistent one, and the
curious situation of who really wrote the poem—and
why incorrect attribution continues to
occur—emphasizes the vulnerability of
English-language haiku poets who must receive
their information on Japanese haiku through
translations. The vulnerability in this case may
be limited to a typo, or it may be more serious,
but the misattribution of the “two autumns” poem
illustrates how readers and translators of haiku
might be more cautious in their acceptance of
The “two autumns” poem has appeared in English in
a number of versions over the years. The preceding
rendition is what appears in the first Two Autumns
book. Garry Gay adapted it from the translation he
found in Henderson’s An Introduction to
Haiku, which, he says, was why he sought
Doubleday’s permission to use the poem, which they
granted. Henderson attributed the poem to Buson
and it also appeared in the first Two Autumns book
attributed to Buson. This poem also surfaces in
Robert Hass’s prominent book, The Essential
Haiku (Ecco, 1994). For comparison, here
are the versions by Henderson and Hass:
For me who go,
you who stay—
(Henderson, page 111)
(Hass, page 81)
Hass’s book also attributes the haiku to Buson.
However, contrary to these attributions, in A
History of Haiku, Volume Two (Hokuseido,
1964), R. H. Blyth attributes the poem to Shiki.
Specifically, in Chapter 37, entitled “Shiki: The
Haiku Poet,” Blyth clearly suggests by the
chapter’s context that the poem is Shiki’s:
“This was written,” Blyth comments, “in the 2nd
year of Meiji [1870; this reference is incorrect,
however, as will be explained], upon [the poet’s]
parting from Sōseki on the 19th of October, at
Matsuyama, when leaving for Tōkyō. It is a kind of
existentialism” (page 97). The reference to Sōseki
seems to definitively place the poem in the time
of Shiki, who lived from 1867 to 1902. Sōseki was
born in the same year as Shiki, and lived until
1916. Blyth’s reference is confusing, though,
because the Meiji era ran from 1868 to 1912;
consequently, in the second year of the Meiji
period, Shiki and Sōseki would have both been
toddlers. As much of a master poet Shiki may have
been, I rather suspect he was not writing haiku at
age two. We might conclude that Blyth’s anecdote
is in error and the poem is indeed by Buson, but I
think the error here is just limited to the
reference to the year.
A more trusted source on when the poem was
written, and by whom, is If Someone Asks . .
. Masaoka Shiki’s Life and Haiku, a book
of Shiki’s haiku from the Matsuyama Municipal
Shiki-Kinen Museum published in 2001. This book,
published in Shiki’s hometown, says that Shiki
wrote the poem at age 28, which would have been
around 1895. Here is this book’s version:
Another translation from Matsuyama, which I
believe to be by Kimiyo Tanaka, is also available
online at the Matsuyama University
website where the poem appears
as follows, attributed to Shiki:
In his book Masaoka Shiki: Selected Poems
(Columbia University Press, 1998), Burton Watson
also attributes the poem to Shiki, with the
accompanying explanation that matches the
anecdotal information Blyth provided:
I am going
two autumns for us
Taking leave of Sōseki (the
novelist Natsume Sōseki (1867–1916), who at this
time was a middle school teacher in Matsuyama.
Shiki was leaving Matsuyama for Tokyo).
What is to be made of the various contradictions
in attributions? We may still wonder who really
wrote the poem. Because Blyth’s and Watson’s
translations are so assertively linked to a
specific person who lived during Shiki’s life, how
could the poem have been written in Buson’s time a
hundred years earlier? Who really wrote this
haiku, and when?
For me, who go,
for you who stay behind—
Perhaps the romaji versions of the poems might
reveal that the two poets merely wrote similar
poems (a phenomenon I’ve referred to as déjà-ku).
In comparing the romaji, however, the poems are
not just similar but identical (except for
capitalization), and nowhere is there any
discussion about the amazing coincidence of two
poets independently writing identical poems.
Occam’s razor would have us believe that this is
simple an attribution error. Hass’s book does not
include the romaji, but Henderson’s, Blyth’s,
Watson’s, and the Shiki Museum’s books do, and
their being identical suggests, with other
evidence, that the attribution to Buson is simply
Yuku ware ni todomaru
nare ni aki futatsu
nare ni aki
yuku ware ni / todomaru nare ni /
yuku ware ni todomaru nare ni aki
Even for those who do not read Japanese, it is
easy to see that, where provided, the Japanese
characters are identical, too. If the poem is
Shiki’s, how did Henderson come to attribute the
poem to Buson? Henderson died in 1974, so perhaps
we will never know the answer to that question.
Hass apparently used Henderson’s book as a source
and merely repeated the error. In May of 2010, I
wrote to Professor Hass about this issue, and this
is what he wrote in response:
Thanks, Michael, for your note. I’m
aware of the problem. I did my work on haiku
over a stretch of years, so I don’t remember
exactly how I came to perpetuate this confusion.
I think I must have come across the poem in
Henderson and had its provenance confirmed by
Blythe [sic]—when he was attributing the poem to
Buson. I haven’t got definitive textual
confirmation from a Shiki or a Buson scholar,
but I am pretty sure the poem is Shiki’s. So I
intend—regretfully, it is such an extraordinary
poem—to cut it from The Essential Haiku
in the next printing when there is one. Hope you
However, there is more to the mystery. R. H. Blyth
not only attributes the poem to Shiki, as already
mentioned, but also attributes the poem to Buson,
as alluded to by Hass. In the fourth volume of
Haiku (1952), it appears as follows,
including Blyth’s introductory sentence (page xxx
The following five verses are by
Buson, showing the humanity of the artist-poet,
much greater than usually supposed:
ware ni todomaru
nare ni aki
Notice that the romaji is identical as on page 97
of Blyth’s A History of Haiku, Volume
Two, but that the translation differs. What,
indeed, are we to make of this?
When I asked William Higginson for his opinion on
this puzzle, he said that because a translator as
esteemed as Burton Watson had so assertively
connected the poem to a contemporary anecdote and
explanation, he considered the matter to be at
rest—that the poem must indeed be Shiki’s. Janine
Beichman’s biography of the poet, Masaoka
Shiki (Kodansha International, 1982),
does not mention the “two autumns” poem. When
Donald Keene’s biography, The Winter Sun
Shines In: A Life of Masaoka Shiki
(Columbia University Press), was published in
2013, I wondered if it might address the issue, or
quote the poem, but unfortunately it does not.
However, the friendship between Shiki and Sōseki
is made abundantly clear. Surely the matching
anecdotal explanations given by Blyth and Watson
carry enough weight to lend credence to
Higginson’s conclusion. Indeed, when I contacted
representatives of the Shiki Museum in Matsuyama
about this poem, they expressed indignation that
anyone would misattribute Shiki’s poem in English.
Their book of Shiki’s best and most appealing
poems features just 115 out of more than 23,600
haiku that Shiki wrote, so their indignation at
the misattribution of so prominent a poem is
I believe we can safely conclude that Blyth’s
original attribution of the poem to Buson is
simply in error (although how he did that is still
a mystery), and that Henderson and Hass repeated
the same error, seemingly using the erroneous
Blyth attribution as a source rather than the
original Japanese. But I still remain curious how
Blyth could have made the error in the first
place. In 2005, I invited R. H. Blyth’s daughter
Harumi M. Blyth to speak at the Haiku North
America conference in Port Townsend, Washington. I
asked her if she might shed some light on this
mystery, but she said she knew nothing about it,
or her father’s translation process, and also
suggested that the bulk of her father’s papers
were not preserved. Blyth wrote much of his work
on haiku while interned in a relocation camp for
foreigners in Japan during World War II, so his
access to original source material was limited
(for more on this context, read Robert Aitken’s
brief memoir, “Remembering Blyth Sensei,” in
Tricycle, Spring 1998, also available
online, with photos and the title of “Remembering R. H. Blyth”.
However, given the accuracy and breadth of the
remainder of Blyth’s books, his misattribution of
the Shiki poem seems uncharacteristic—greatly in
the minority. Yet this misattribution has
proliferated not only in books but now on numerous
websites as well.
Harold Henderson’s initial repetition of this
error strongly suggests that Henderson
specifically consulted Blyth’s first translation
of this poem (published in 1952) when he wrote his
Introduction to Haiku (published in 1958).
Henderson’s earlier book on haiku, The Bamboo
Broom (Houghton Mifflin, 1934), does not
contain the “two autumns” poem, so Henderson’s
first translation of this poem would appear to be
from 1958. It seems reasonable to assume that
either Henderson perpetuated Blyth’s original
error or that both Henderson and Blyth consulted
the same original but erroneous source—an
über-source that I can find no evidence for.
Blyth’s A History of Haiku (1964) is
more recent than the four-volume Haiku set
(1949–1952), so readers can view Blyth’s most
recent book to be more reliable, and that Blyth
took this later publication as an opportunity to
The discovery of this attribution error and the
contradiction even in Blyth’s hallowed writings
illustrates the fallibility of translators and
their publishers, though of course readers should
be forgiving. Because the poem is clearly Shiki’s,
what this discovery means to the Haiku Poets of
Northern California is simply a small correction
to the history of the organization and its
long-running reading series.
What this discovery means to other readers of
haiku translations, however, is a caution to be
ever vigilant and wary—and to question even the
most reliable translators. There seem to be a
variety of problems to be wary of. For example,
Keisuke Nishimoto’s Haiku Picturebook for
Children (Heian International, 1998)
incorrectly attributes an Issa poem to Matsumoto
Takashi (I have written about this at ‘Two Books for Children’
; the original Japanese version of this book does
not have the error). Two other translators I know
of who have perpetuated the error of attributing
the “two autumns” poem to Buson are Hiag Akmakjian
in his book Snow Falling from a Bamboo Leaf:
The Art of Haiku (Capra Press, 1979) and
Naomi Wakan in her book Haiku: One Breath
Poetry (Pacific Rim Publishers, 1993).
Though they provide their own versions of the
poem, one wonders whether they translated from an
authoritative original Japanese source or if they
might have just reworked a previous translation,
such as Blyth’s or Henderson’s—and now seem to be
caught out in such a practice by the repetition of
the attribution error (this reminds me of
cartographers who deliberately add “errors” to
their maps, sometimes called “trap streets,” to
prove if they’ve been copied). One assumes,
naturally, that a true translation would be made
from the original text, but perhaps we should
never make that assumption. At least Robert Hass
admits in his book that he consulted other
translations to make most of his “versions.”
Here, too, is another version, much earlier this
time, from Harold Stewart’s A Net of
Fireflies (Rutland, Vermont: Tuttle,
1960, page 84), complete with Stewart’s ponderous
title and inimitable rhyme, also attributing the
poem to Buson:
SINCE IT MUST BE SO . . .
You must remain, I must depart,
Two autumns falling in the heart.
In a brief “Research Note” in Frogpond 35:1,
Winter 2012, Charles Trumbull says that this poem
was attributed to Buson in two editions of X. J.
Kennedy’s college textbook, An Introduction
to Poetry (1995, seventh edition, page 73,
citing Henderson as the translator; and 1998,
ninth edition, page 100, citing Hass as the
The attribution error is perpetuated in other
ways, too, not even counting many online
misattributions. In Lonnie Hull Dupont’s
publication, The Haiku Box (Journey
Editions/Tuttle, 2001), she quotes Hass’s
translation, attributing the poem to Buson. Thus
the ripples spread out. Similarly, a search of the
Internet reveals numerous repetitions of the Buson
attribution error, seemingly all by Westerners.
One wonders where they got their information—or
maybe it’s obvious that they consulted Hass,
Henderson, or Blyth. At ahapoetry.com,
Jane Reichhold actually attributes the same poem
to Bashō, but that is probably an unrelated lapse
in scholarship, yet I’ve seen this error
perpetuated elsewhere online, too, usually citing
Reichhold as the translator, but believing the
poem to be Bashō’s. I suppose someone might as
well attribute the poem to Issa, too, so that all
the great haiku masters might be credited. In
contrast, Japanese sources I’ve explored on the
Internet have uniformly attributed the poem to
Shiki, and do not make the same attribution error
that seems to have begun with Blyth.
The problem goes beyond the attribution of the
Shiki poem. For example, David Lanoue has written
in Modern Haiku (Volume XXXI, Number
2, Summer 2000) about Sam Hamill’s repetition of
translation errors from Nobuyuki Yuasa, calling
into question the method that Hamill asserts to
have used: working from the original Japanese
texts. Apparently, though, Hamill did not use
original Japanese texts—at least not all the time.
Little harm may come from many attribution or
translation errors themselves, but the ease with
which such errors are perpetuated in haiku should
give us pause regarding other information we
receive in English about haiku. English-language
haiku poets are indeed in a vulnerable position in
that we receive our perceptions of Japanese haiku
largely through translators. If the translators
are in error, we can too easily perpetuate the
error, for we have few other sources of
information—and sometimes no other sources at all.
It is therefore not just the translators who
should be vigilant with haiku, but readers also.
Hiag Akmakjian. Snow Falling from a
Bamboo Leaf: The Art of Haiku. Santa
Barbara, California: Capra Press, 1979.
Janine Beichman. Masaoka Shiki.
Tokyo: Kodansha International, 1982.
R. H. Blyth. Haiku. 4 volumes.
Tokyo: Hokuseido, 1949–1952.
R. H. Blyth. A History of Haiku.
2 volumes. Tokyo: Hokuseido, 1964.
Lonnie Hull Dupont. The Haiku Box.
Boston: Journey Editions/Tuttle, 2001.
Robert Hass. The Essential Haiku.
Hopewell, New Jersey: Ecco, 1994.
Harold G. Henderson. An Introduction to
Haiku. New York: Doubleday, 1958.
Harold G. Henderson. The Bamboo Broom.
New York: Houghton Mifflin, 1934.
Donald Keene. The Winter Sun Shines In: A
Life of Masaoka Shiki. New York:
Columbia University Press, 2013.
X. J. Kennedy. An Introduction to Poetry.
New York: Longman, 7th edition, 1995; 9th
David Lanoue. “Translating Translations:
A Disturbing Trend,” in Modern Haiku
XXXI:2, Summer 2000, pp. 53–58.
Matsuyama Municipal Shiki-Kinen Musuem.
If Someone Asks . . . Masaoka Shiki’s Life
and Haiku. Matsuyama, Japan:
Matsuyama Municipal Shiki-Kinen Musuem,
Keisuke Nishimoto. Haiku Picturebook
for Children. Torrance, California:
Heian International, 1998.
Harold Stewart. A Net of Fireflies.
Rutland, Vermont: Tuttle, 1960.
John Thompson, ed. Two Autumns.
San Francisco: Two Autumns Press, 1990.
[poems by Pat Donegan, Eugenie Waldteufel,
Michael Dylan Welch, and Paul O. Williams]
Naomi Wakan. Haiku: One Breath Poetry.
Victoria, British Columbia: Pacific
Rim Publishers, 1993.
Burton Watson. Masaoka Shiki: Selected
Poems. New York: Columbia University
expositions contents :