Primary Sources

Allen, Laura W. 1995. “Images of the Poet Saigyo as Recluse”. Journal of Japanese Studies, 21 (1): 65 – 102

Aston, W. G. trans., 1972. Nihongi: Chronicles of Japan from the Earliest Times to A.D. 697. Vermont: Tuttle publishing.

Bardo, Jonathan, 2002. “Picture and Witness at the Site of the Wilderness”. In Landscape and Power, edited by W. J. T. Mitchell, pp. 291 – 317. Chicago: University of Chicago.

Barnes, Gina L. 2007. State Formation in Japan: Emergence of a 4th-century Ruling Elite. Abingdon: Routledge.

Batten, Bruce L., 2003. To The Ends of Japan: Premodern Frontiers, Boundaries, and Interactions. USA: University of Hawai’i Press.

Brown, Jan, 1993. Exploring Tohoku: A Guide to Japan’s Back Country. New York, Whetherhill Inc.

Bunn, David, 2002. “Thomas Pringle’s African Landscapes”. In Landscape and Power, ed. W. J. T. Mitchell, pp. 127 – 173.

Chamberlain, Basil Hall, 1981. The Kojiki: Records of Ancient Matters. Singapore: Tuttle Publishing.

Cranston, Edward A., 2006. A Waka Anthology: Grasses of remembrance, Volume II. Stanford University Press: California.

Friday, Karl 1997. “Pushing Beyond the Pale: The Yamato Conquest of the Emishi and Northern Japan”. Journal of Japanese Studies 23 (1 winter): 1-24

————– 1988. “Teeth and Claws: Provincial Warriors and the Heian Court”. Monumenta Nipponica 43 (2 Summer): 153-185

Goodrich, L. Carrington and Tsunoda Ryusaku 1951. “Japan in the Chinese Dynastic Histories”. Perkins Asiatic Monograph 2: 38-40

Heldt, Gustav, 2008. The Pursuit of Harmony: Poetry and Power in Early Heian Japan. USA: Cornell East Asia Program.

Honda, H. H. trans., 1971. The Sankashu. Tokyo: The Hokuseido Press.

Hudson, Mark J., 1999. Ruins of Identity: Ethnogenisis in the Japanese Islands. USA: University of Hawai’I Press.

———————, 1999. “Ainu Ethnogenisis and the Northern Fujiwara”. Arctic Anthropology 36 (1): 73-83

Hurst, G. Cameron, 2007. “Kugyo and Zuryo: Center and Periphery in the Era of Fujiwara no Michinaga”. In Heian Japan, Centers and Peripheries, ed. Mikael Adolphson, Edward Kamens and Stacie Matumoto, pp. 66 – 101. USA: University of Hawai’i Press.

Kamens, Edward, 1997. Utamakura, Allusion, and Intertextuality in Traditional Japanese Poetry. New Haven, Yale University Press.

Inoue Tatsuo, 2006. “The Hitachi Fudoki and the Fujiwara”, interpreted by Michiko Aoki. In Capital and Countryside in Japan, 300 – 1180, ed. Piggott, Joan R., 99. 109 – 127. USA: Cornell University East Asia Program.

Keene, Donald (trans.), 1996. The Narrow Road to Oku. USA: Kodansha International Ltd.

Keene, Donald, Tanabe, George, Varley, Paul, and WM. Theodore de Bary eds., 2001. Sources of Japanese Tradition: From Earliest Times to 1600 Volume I Second Edition. New York: Columbia University Press.

Koyama Yasunori, 2006. “East and West in the Late Classical Age”. Interpreted by Bruce L. Batten in Capital and Countryside, ed. Piggott, Joan R., pp. 366 – 401.

Lafleur, William R., 2009. Awesome Nightfall: The Life, Times, and Poetry of Saigyo. USA: Wisdom Publications.

Marra, Michele, 1991. The Aesthetics of Discontent: Politics and Reclusion in Medieval Japanese Literature. USA: University of Hawaii Press.

McCullough, Helen Craig 1964. “A Tale of Mutsu”. Harvard Journal of Asiatic Studies 25: 178-211

——————————–, 1985. Kokin Wakashu: The First Imperial Anthology of Japanese Poetry. USA: Stanford University Press.

McKinney, Meredith (trans.), 2006. The Pillow Book. England: Penguin Classics.

McKinney, Meredith (trans.), 1998. The Tale of Saigyo (Saigyo Monogatari). USA: University of Michigan.

Morita Tei, 2006. “Towards Regency Leadership at Court”. Interpreted by Joan R. Piggott in Capital and Countryside, ed. Piggott, Joan R., pp. 209- 226.

Morris, Mark 1980. “Sei Shonagon’s Poetic Catalogues”. Harvard University of Asiatic Studies 40 (1 June): 5-54

Mostow, Joshua S., Royall Tyler, 2010. The Ise Stories: Ise Monogatari. USA: Hawai’i Press.

Newhard, Jamie L., 2013. Knowing the Amorous Man: A History of Scholarship on Tales of Ise. USA: Harvard University.

Okada, H. Richard, 1991. Figures of Resistance: Language, Poetry, and Narrating in The Tale of Genji and Other Mid-Heian Texts. USA: Duke University Press.

Rexroth, Kenneth, 1964. One Hundred Poems From the Japanese. USA: New Directions.

Rodd, Laurel Rasplica, Henkenius, Mary Catherine, 1996. Kokinshu: A Collection of Poems Ancient and Modern. USA: Cheng and Tsu Company Inc.

Saigyo, Poems of a Mountain Home. Translated from Japanese by Burton Watson, 1991. USA: Columbia University Press.

Shirane Haruo, 2012. Japan and the Culture of the Four Seasons: Nature, Literature and the Arts. New York: Columbia University Press.

—————-, 1998. Traces of Dreams: Landscape, Cultural Memory, and the Poetry of Basho. California: Standford University Press.

Shizuki Tomotaro, eds. 1952. Ise Monogatari: Tempuku-bon. Japan: Koten Bunko No. 64.

Stockdale, Jonathan, 2015. Imagining Exile in Heian Japan: Banishment in Law, Literature and Cult. USA: University of Hawai’i Press.

Takahashi Tomio, 2006. “The Classical Polity and its Frontier”, translated by Karl Friday. In Capital and Countryside, ed. Piggott, Joan R., pp. 128-145.

Takamitsu, Konoshi, 2000. “Constructing Imperial Mythology: Kojiki and Nihon shoki”. In Inventing the Classics: Modernity, National Identity and Japanese Literature, ed. Shirane Haruo and Suzuki Tomi, pp. 51 – 67. California: Stanford University Press.

Yoshikazu, Shinada, 2000. “Man’yoshu: The Invention of a National Poetry Anthology”, translated by Kevin Collins. In Inventing the Classics, ed. Shirane and Suzuki, pp. 31-50.

謌枕名 / 吉田幸一, 神作光一, 橘りつ 」、古典文庫 , 東京 : 古典文庫, 昭和 49-51 ,  [1974-1976]

Internet Sources

I.J. Parker, (accessed 10th September, 2015)

(Author unknown) (accessed 11th July 2015)

(It goes without saying that this work is my own. It is copyrighted both to myself and to SOAS University; any personal use of this material must be authorised by myself first. Thanks.)



In the late 17th century, Haiku poet Matsuo Basho took up his stick and followed Saigyo’s footsteps through Michinoku. He tells us in his haibun work Oku no Hosomichi (Narrow Road to the Interior) that at Sanekata’s grave he found no pampas grasses, nor the hanagasumi (iris flowers) which, poetry tells us, once adorned Asakayama.1 At Tohoku’s utamakura sites today however, you can witness the ‘correct’ keepsakes once allotted to their place-names by waka, as people have replaced them: the traces of poetry are still embedded in the landscape, and are still a source of tourism. It has been said that “the wilderness picture … casts a retrospective and nostalgic aesthetic in the landscape art that forgets the trauma of history”,2 and looking at Michinoku / Tohoku today, it is easy to agree. I believe this investigation has laid the groundwork in exploring the significance of Michinoku in the early pre-modern Japanese process of state-building; as such, I hope it may inspire further enquiry, and further answers.


Figure 4.1 Replanting iris flowers at Asakayama, Fukushima-ken: 2013 (photo taken by the author)


1 Donald Keene (trans.), 1996. The Narrow Road to Oku. USA: Kodansha International Ltd.

2Bardo, Wilderness, p. 310

(It goes without saying that this work is my own. It is copyrighted both to myself and to SOAS University; any personal use of this material must be authorised by myself first. Thanks.)


Shifting Contexts, Shifting Boundaries

Ise Monogatari

The Ise monogatari (Tales of Ise), written between ca. 850 – 1011 and by many hands, demonstrates how mid-Heian authors treated kegai spaces in accordance with the politically tumultuous 9th, 10th and 11th centuries. A “poetics of exile”1 was appearing in contemporary works which removed fictional characters from the capital, and allowed authors to critique their social order from afar.2 In contrast to Heian poetry then, it develops into a counter-ideological text, and we can see this development through its treatment of the periphery.

The tale is a semi-fictive biographic of the early-Heian poet Narihira no Ariwara (825-880), and loosely unites poems (uta) through a basic narrative (monogatari) of episodes (dan). The composition of the text divides into three periods: the first third of episodes is thought to pre-date the Kokinwakashu (905), and contains poems which are (foremost accurately) attributed to Narihira.3 The middle section is an expansion of the tale, dating around another imperial poetry anthology the Gosenwakashu (Later Collection of Poetry, 956). The final instalment, in which Narihira is explicitly named the tale’s hero for the first time4 is a rewrite of the first, and by featuring poems which pre- and post-date Narihira by decades its authors make little effort towards historical accuracy. I argue that ‘truth’ was not a primary concern for the Ise authors who intended to create a counter-ideological text in response to unwelcome political changes at court.

Following the Kusuko Disturbance in 810 in which Narihira’s grandfather Emperor Heizei (r. 806 – 809) attempted a political coup against his brother Saga (r. 809 – 823), the courtier Fujiwara Fuyutsugu of the Northern Fujiwara family was made Imperial Regent by Saga for his support. The Northern Fujiwara thereafter proceeded to gain control over the imperial line through marriage and by demoting / exiling their rivals from the capital, including Heizei’s family line and the equally threatening Ki and Minamoto clans. Descendants of these families were left to establish literary careers at court, and poetry was becoming a means of expressing political dissatisfaction rather than state harmony.5 Michele Marra attributes the Ise’s authorship to generations of such disillusioned imperials, who constructed an aesthetics of miyabi – courtly elegance – as a counter-ideology to Fujiwara control.6

Michinoku’s inclusion in Ise complements this theory. In a ‘poetics of exile’ trope, Narihira is sent to Azuma. Reminiscent of our shirakawa no seki poets, he becomes nostalgic for home, as unfamiliar geomantic features remind him of familiar ones back in the capital: Mount Fuji brings to mind the deer at Kaguyayama shrine in Heian-kyo through its “fawn-dappled flecks of white” (kanoko madara).7 Such comparisons show Narihira transporting his aristocratic cultural heritage into an uncultured space, making Azuma a stage for exile which does not express Narihira’s physical distance from the capital so much as his cultural proximity to it. The azuma-kudari (Journey to the East) frames kegai landscapes within a kinai rhetoric to elevate the hero as truly aristocratic.

By episode 14 he has reached Michinoku. A Michinoku woman, thinking Narihira very fine indeed (“medurakaniya”),8 composes a poem in which she refers to herself as a silkworm (kuhako).9 Modelled on Man’yoshu imagery, the kuhako motif is dismissed by Narhira who calls it inelegant (“hinabi”).10 By leaving her in the middle of the night, Narihira also ignores the courtly decorum of staying with a lover until dawn. This slight – as well as his jibe at her ‘not being human enough’ to take back to the capital (“hito naraba” – ‘if you were a person)11 – eludes her, and she composes another disastrous waka. Although Narihira has encountered rustic ascetics and pushy provincial governor parents in Azuma, this is the only time he meets someone so uncultured, and whom he treats so inelegantly. If Ise is the counter-ideological text Marra prescribes, and represents ‘true’ imperial nobles as inherently miyabi, then this episode surely works against Narihira’s integrity; if not, then we must assume that to behave in such an un-courtly manner in Michinoku does not count as un-courtly, aptly expressing the province’s inferiority.

Our second Michinoku female in episode 15 causes the Ise authors, translated by Royall Tyler, to ask “what could he [Narihira] possibly have gained from looking into a heart as mean and crabbed (ebisu kokoro) as that?”12 Although Tyler rightly points out that the literal translation of ebisu signifies ‘prawn’13 (and thus “crabbed”), we have already seen the term adopted by state histories in a manner less forgiving: he might more accurately have written ‘the heart of a northern foreigner’. The poem itself contains no derogatory terminology:

shinobu yama Ah, Mount Shinobu,
shinobite kayofu
I would gladly find a way
michi mogana
in secret to frequent her,
hito no kokoro no
that I might further explore
oku mo mirubeku
the recesses of her heart.14

The poet opts for “hito no kokoro” (that person’s heart / sensibility). This could be attention to proper waka aesthetics which would disallow ‘ebisu’, but I think it is more probable that for this poem, the location of Shinobuyama is insignificant to the poem’s meaning. ‘Shinobu’ is a homonym: as a noun, it translates to ‘fern’ (); a verb (忍ぶ) means ‘to conceal / to endure’, and another (偲ぶ) ‘to be nostalgic for’. As a proper noun, Shinobu is a town near Fukushima City. Not only is this utamakura’s wordplay potential rich, but a famous shinobu poem was already in circulation during Narihira’s lifetime. Minamoto no Toru (822 – 890), a contemporary of Narihira who was also an imperial prince-turned-commoner under the Northern Fujiwara, composed this waka referenced to in the opening dan:

Michinoku no like Michinoku
shinobu michizuri cloth, printed with tangled ferns,
dare yue ni my mind is disordered
midare somenishi because of you,
ware naranaku ni but my love is not.15


The place-name here is significant more for its familiar wordplay and allusive potential than its geographic location, and therefore doesn’t hold much significance to our investigation.

Already an existing signifier for otherness to the readership, ‘Michinoku’ becomes in early Ise an extended utamakura which, through a poetics of exile, displaces ‘true’ nobility from the centre to the provinces; and could even imply through inelegant Michinoku women that the Northern Fujiwara were imposters in court who attempt but fail to re-enact Narihira’s miyabi. All the while, these episodes reveal that aristocratic attitudes towards Michinoku have little changed since Yamato and Nara times.

The latter third of Ise, written centuries later, features Michinoku in a far more favourable light. Our other exile Minamoto no Toru is once again featured in the narrative with Michinoku. He hosts a poetry party in his garden which is constructed on the appearance of a scenic bay in Miyagi, Shiogama. The late Ise authors accompany this story with a poem which romanticizes Shiogama’s beauty:

Michinoku wa In Michinoku
izuku wa aredo
Anywhere is where to be,
shiogama no
but Shiogama
ura kogu fune no
Where the boats row on the bay
tsunade kanashi mo
Enchants with its lines of ropemen.16

Toru’s implantation of a kegai landscape in the capital reverses Narihira’s earlier nostalgic mourning in Azuma: the periphery has been brought into the centre space rather than the other way around. The later Ise authors add to this revearsal of ideological boundaries: Episode 115 reverses the narrative of episode 14, in which a Michinoku woman referred to as onna (woman) is bearable enough for Narihira to live with (“otoko onna sumikeri”),17 and when he leaves she composes an incredibly skilful verse:

woki no wite sharper than the pain
mi wo yaku yori mo
of live coals laid on the skin
kanashiki ha
miyakoshimabe no
burns by Miyako-shima
wakare narikeri
this parting into my heart.18

Phonetically, “woki no wite” insinuates the semantic clause “to put [on] coals”; “mi wo yaku” tells us that it is herself (mi mo) being burned (yaku); “Miyako-shima” literally translates to ‘Capital Island(s)’, whose namesake is where Narihira is headed. The poem’s quality is proved by its having been attributed to Ono no Komachi, one of the most celebrated female poets of the Heian period. The comparison to episode 14‘s disastrous waka is evident.

So why the ideological shift? Just as much anti-Fujiwara sentiment was felt in the 11th century as in the 9th19 (in dan 101, wisteria (fujiwara) flowers “cover now more than before”),20 but perceptions of the north-east were changing. In 995, a courtier on the wrong side of the fearsome Northern Fujiwara regent Michinaga is exiled by Emperor Ichijo (r. 986 – 1011) to Michinoku. Fujiwara Sanekata, admired by Sei Shonagon as “famous for his poetic skill”,21 is ‘demoted’ to Middle Captain (chujo) of Mutsu Province. Ichijo consoles Sanekata by telling him to “go off, and have a look for yourself at all those utamakura”.22 Through his comment we see the extent to which the court now considered Michinoku part of the realm. Exile as state punishment had long been practised in Japan, and in 724 Emperor Shomu announced that destinations for exile would depend on the severity of the crime. Various locations were defined as near, medium, or distant from the capital as measure:

Near Exile: Echizen, Aki
Medium Exile: Suo, Iyo
Distant Exile: Izu, Awa, Hitachi, Sado, Oki, Tosa23

Thus in the Nara period, the ‘distant’ northern destinations only stretched to Sado and Hitachi, but by Sanekata’s demotion two centuries later, Michinoku is included.24 For a nobleman of Sanekata’s standing to be sent to Mutsu would not only have been insulting, but dangerous. By the 10th century, the northern frontier had reached the Kitakami Basin and Koromo River in present-day Iwate.25 It would not be until 1095 when another Fujiwara family would establish the first permanent ‘Japanese’ city in Michinoku: nicknamed the ‘Mirror of Kyoto’ as a ‘reflection’ of the centre,26 Hiraizumi in the Kitakami Basin was finally proof of northern pacification – or, as we shall explore later, accepted emishi autonomy.27 A century before this, Sanekata was thus entering an unresolved conflict: Mihcinoku was ‘taken’, but not pacified. Beyond the abstract spatial constructs of imperial proximity and Heian socio-political hierarchy, it was Sanekata’s physical presence in the north which determined the outcome of his real, not political, life: he died in Natori in 998. The literary representation of Narihira’s exile in comparison with Sanekata’s real fate could not be more different. Having “roamed all the way to … Michinoku”,28 Narihira follows the Kokinshu’s cyclical notions of travel by returning to Heian-kyo,29 whilst Sanekata cannot. In reality, Michinoku is not a point of comparison for the capital, but its own geographical place. The political contexts both at court and in kegai lands during Ise‘s long authorship evidently effected their textual treatment of Michinoku; but as the differences between Narihira and Sanekata’s experiences of Michinoku demonstrate, even by the 11th century when Ise was completed, Michinoku had maintained its symbolic signification as ‘the other’, unrelated to its geo-realities.

Saigyo Monogatari

Not so, by the end of the Heian period. Saigyo monogatari’s hero, a recluse Buddhist poet-monk travelling through Natori in the late 12th century, came across Sanekata’s grave. Arriving in winter, and seeing the poet’s resting place covered with dead pampas grasses and snow, Saigyo Hoshi composes this poem:

kuchi mo senu One part of him
sono na bakari o
escaped decay – his name,
still around here like
kareno no susuki
this field’s withered grass:
katami mi zo miru
my view of the relic he left.30

In this verse, Saigyo omits and contradicts Natorigawa’s usual utamakura associations of ‘losing a name’ by claiming that Sanekata’s is the “one part” of him which remains. This places Sanekata himself as the poem’s topic rather than the expected place, and demonstrates Saigyo’s preference for historical concreteness over the reliance and reuse of old ideas – a theme which carries throughout his life as well as his verse. Like Ise, Saigyo monogatari was compiled by unknown authors31 who semi-fictively portray the hero’s life through the poem-tale genre. It is a multi-authored expression of political dissatisfaction that utilizes the readership’s expectations of Michinoku in order to thwart them. Reciting a real Saigyo’s real visits to Michinoku, undertaken in his thirties and later in his sixties, the narrative is relatively reliable: Saigyo was the first noble in Japanese history who his contemporaries knew for sure had journeyed to Michinoku voluntarily (unlike both Narihira and his Noin, Saigyo’s influential predecessor, who had allegedly suntanned in his back garden instead of passing Shirakawa Barrier).32 His many journeys were inspired by his distrust for the contradictions he saw in the social institutions around him. A poem composed on his quitting the military service of Emperor Sutoku (1119-1164) and taking the tonsure at the age of twenty-three reflects his disdain for a model of rulership which put power in the hands of retired emperors rather than the child emperors they placed upon the throne (insei):

Written when I was petitioning the insei Emperor Toba to grant me his permission to leave secular life:
oshimu tote
so loath to lose
oshimarenubeki what maybe should be loathed:
kono yo kawa one’s place in the world;
mi o sutete koso we maybe rescue best the self
mi o mo tasukeme by simply throwing it away.33

Upon finding the same preoccupations with rivalry for power at his Buddhist temple, however, he soon left there too:

sutetaredo to think you’ve thrown
kakurete sumanu
the world away and then
hito ni nareba
still live unhidden is
nao yo ni aru ni
to be like any other worldling
niaru narikeri
still dwelling in the world of men.34

As a great-grandson of a northern elite provincial warrior35 and a Buddhist monk distrustful of institutional integrity, Saigyo wrote of Michinoku in a manner which shows a refreshing change to earlier Heian centre/periphery ideologies. His Michinoku poems are recorded in his Sankawakashu (Poems from a Mountain Home) as well as in this tale, and show Saigyo reworking the centre’s associations with the region through an application of Buddhist aesthetics and realistic depiction:

After passing the Barrier of Shirakawa he came to a place called Shinobu, and thinking of Priest Noin’s journey:

After starting from the capital
until I crossed
Osaka Hill
oft did I think
when should I pass
the Shirakawa Barrier.36

utamakura in this verse omit Saigyo’s actual location (Shinobu) whilst including an utamakura disassociated with Michinoku (‘Osaka Hill’ – a homonym with ‘Meeting Hill’). Coupling the capital with Shirakawa Barrier demonstrates Saigyo’s literary knowledge, but he denies the usual Shinobu associations (tangled hearts, Michinoku cloth used by Narihira and Toru) by instead relying on utamakura to narrate his subjective experience of the landscape: his imaginary ‘meeting’ with Noin who, Saigyo believed, had undertaken the same journey. Through selecting these utamakura, Saigyo connects his waka to the trodden earth of the Michinoku landscape rather than to its name.

Passing into Hiraizumi, Saigyo stands at the Koromo River bank which had for centuries been the frontier of Ainu and emishi resistance to central colonisation:

The writer arrived at Hiraizumi on october 12. it was a snow-stormy day. He stood on the banks of the Koromo, and saw the residence of the Fujiwaras with admiration. Feeling cold at the frozen water’s edge:

Freezing is the day.
I see the hall where lived the Fujiwaras
from the banks of the Koromo,
and am moved with admiration.”37

Koromogawa was not an established utamakura; Saigyo’s admiration of the place relates to its historical rather than allusive context. Three families had established Hiraizumi: the Abe, Kiyowara and Northern Fujiwara. They were all descendants of fushu (‘surrendered barbarians’) who, no longer ‘enemies’ to the state, had developed almost full autonomous rule over the Kitakami Basin since the decline of a central presence in the north.38 Archaeological evidence of singular religious practices taking place in Hiraizumi, and of frequent trade with Ainu communities in Ezo (Hokkaido) and the Sakhalin Islands, demonstrates their cultural autonomy.39 Nevertheless, Hiraizumi influenced the centre: after its erection, terms began to appear in late Heian and early medieval literature which distinguished between the different ethnic groups in Michinoku for the first time. Whilst “ebisu” continued to signify the genetically ‘Yayoi Japanese’ who practised a northern culture, “ezo” came to signify the ‘Ainu Japanese’.40 Not only does Saigyo admire descendants of the characters who were represented so negatively in the early Ise, but contemporary authors in the capital were no longer identifying those descendants in direct opposition to the state.

The authors of Saigyo Monogatari follow suit. They include the following poem in the Michinoku section of their tale:

Though not as in the capital, here too people seemed to be vying with one another in bustling about with year-end preparations. Moved, he wrote:

yamagatsu no As marker for their boundary
kataoka kakete
on the long slope
shimuru no no
the poor mountain folk
sakai ni tateru
have planted
tama no oyanagi
the jewels of these willows.41

Yama-gatsu translates to ‘mountain’ – ‘lowly’ / ‘few’, and Saigyo’s pairing of these ‘lowly’ people with willow-tree ‘jewels’ (tama) suggests they have obtained a richness not through possessions but through nature. Through this Buddhist aesthetic which defies material attachment, Saigyo’s representation of common folk contrasts greatly with Ise’s ‘ebisu kokoro’. The Saigyo authors comply by admitting to sharing a culture with yamagatsu whilst inferring no inferiority for their differences in style. The poem and narration allow a peripheral space to represent a new kind of harmony suited to new times, much in the same way that early Heian authors signified their notions of harmony in the capital. Saigyo’s actual presence in Michinoku, however – his recalling literary associations within a landscape but choosing not to read the landscape through them – distinguishes Saigyo’s journey to the east from that of Ise’s Narihira.

So what changed the classical notions of harmony, miyabi and the capital into a preference for solitude, detachment and the periphery? Through Sanekata’s exile and Toru’s garden we have observed that the centre was increasingly associating Michinoku with the state in reality rather than in poetry; Hiraizumi was altering how the centre interpreted the periphery; the return of state subjects to the region was once more being realized by Saigyo‘s presence, as he united the landscape with its poetic associations after decades of imperial absence: the notions of centre which had so defined pre-modern Japanese concepts of state identity were coming undone, and by the time Saigyo monogatari was published in the Kamakura Period, they had been completely overhauled. The Genpei Wars of 1186 – 1189 had destroyed Heian-kyo. A provincial clan, the Taira, had taken control of the imperial line decades before, but were usurped by another, the Minamoto from the east. Japan’s first civil war in over four-hundred years was followed by famine, plague and pestilence, and contradicted any possible notion of centre-centred harmony. Minamoto Yoritomo, heralding himself the shogun (governing warlord) of Japan and usurper of the empirical line, moved the capital city to eastern Kamakura in old Azuma territory; Buddhist monks called it the start of the end of Buddha’s salvation (mappo), and composed numerous advisory parables for enlightened living; the political rhetoric of poetry was now obsolete, and early Kamakura poets still in Heian-kyo turned to a new, grounded aesthetics to express their disillusionment with old Heian ideals. Meanwhile, Saigyo became an example for monks and poets alike: the waka anthology Shinkokinwakashu (New Collection of Poems, Ancient and Modern, 1206) includes ninety-six of his waka, and Buddhist works like Kamo-no-Chomei’s Hojoki (The Ten Foot Square Hut, early 13th century) not only endorse his reclusive lifestyle, but critique the same aristocracy which he abandoned decades before. Saigyo died just two years before Yoritomo removed his government to the east, symbolising the end of an age which the monk appeared to have seen coming. The end of the Heian period had resulted in a remapped country and displaced centre, a new ruling class and the abandonment of an ideology which had, in a sense, predicted all along that danger would come to the centre from the “harmful” east. There was no longer a centre, and thus, no longer anything for Michinoku to symbolically oppose. Saigyo’s authors were a world away from Ise’s: Japan was being rewritten.


1Jonathan Stockdale, 2015. Imagining Exile in Heian Japan: Banishment in Law, Literature and Cult. USA: University of Hawai’i Press, p. 44

2Stockdale, Exile, p. 42

3Jamie L. Newhard, 2013. Knowing the Amorous Man: A History of Scholarship on Tales of Ise. USA: Harvard University, p. 20

4Michele Marra, 1991. The Aesthetics of Discontent. USA: University of Hawaii Press, p. 37

5Heldt, Harmony, p. 29

6Marra, Discontent, p. 40

7Joshua S. Mostow and Royall Tyler, 2010. The Ise Stories: Ise Monogatari. USA: Hawai’i Press, p. 35

8Tyler, Ise, p. 7

9Tyler, Ise, p. 46

10Tyler, Ise, p. 49

11Shizuki Tomotaro (ed.), 1952. Ise Monogatari: Tempuku-bon. Japan: Koten Bunko No. 64, p. 14

12Tyler, Ise, p. 49

13Tyler, Ise, p. 49

14Tyler, Ise, p. 49

15 Kenneth Rexroth, 1964. One Hundred Poems from the Japanese. USA: New Directions, p. 80

16Edward A. Cranston, 2006. A Waka Anthology: Grasses of remembrance, Volume II. Stanford University Press: California, p. 10

17Shizuki, Ise Monogatari, p. 101

18Tyler, Ise, p.233

19Marra, Discontent, p. 49

20Tyler, Ise, p. 212

21McKinney, Pillow Book, p. 90

22Kamens, Utamakura, p. 29

23Stockdale, Exile, p. 91

24G. Cameron Hurst, 2007. ‘Kugyo and Zuryo: Center and Periphery in the Era of Fujiwara no Michinaga.’ In Heian Japan, Centers and Peripheries, ed. Mikael Adolphson, Edward Kamens and Stacie Matumoto. USA: University of Hawai’i Press, p. 93

25Helen Craig McCoullough,1964. ‘A Tale of Mutstu.’ In Harvard Journal of Asiatic Studies 25, pp. 179-211

26Mark J. Hudson, 1999. ‘Ainu Ethnogenisis and the Northern Fujiwara’ in Arctic Anthropology 26 (1 /2), p. 77

27Hudson, ‘Ainu’, p. 77

28Tyler, Ise, p. 234

29Tyler, Ise, p. 234

30William LaFleur, 2003. Awesome Nightfall: The Life, Times, and Poetry of Saigyo. USA: Wisdom Publications, p. 24

31 Laura W. Allen, 1995. ‘Images of the Poet Saigyo as Recluse.’ In Journal of Japanese Studies, 21 (1): 65 – 102

32 Kamens, Utamakura, p. 152

33 LaFleur, Nightfall, p. 12

34LaFleur, Nightfall, p. 18

35Friday, Claws, p. 168

36 Honda, The Sankashu,, p. 203

37 Honda, The Sankashu, p. 204

38 McCoullough, Mutsu, p. 182

39 Hudson, Ethnogenisis, p. 72

40Hudson, ‘Ainu’, p. 79

41 Meredith McKinney (trans.), 1998. The Tale of Saigyo (Saigyo Monogatari). USA: University of Michigan, p. 61

(It goes without saying that this work is my own. It is copyrighted both to myself and to SOAS University; any personal use of this material must be authorised by myself first. Thanks.)


Waka as State-Building

If a harmonious cosmos was the measure of righteous rule, then a state seeking recognition from their continental cultural superiors required it. Aristocrats of classical Japan could reflect and create a harmonious environment through literature and the arts, and primarily through waka (‘harmony’, and‘song’ = Japanese poetry). Waka allowed courtiers to express their personal proximity to the Emperor (and thus high position in the socio-political hierarchy) and demonstrate their literary prowess, thus becoming a means of identity formation. With the important aim of creating a harmonious state, waka aesthetics were under strict regulation. Whilst elegance and sensibility (aware) became the acceptable courtly rhetoric,1 anything compromising the ideal was omitted: the textual creation of the centre was also under ideological fabrication.

Similar to the fudoki and other state literatures, early waka anthologies such as the Man’yoshu (Collection of Ten Thousand Leaves, 759) and the Kokinwakashu (A Collection of Poetry Ancient and New, ca. 905) made frequent use of the landscape to depict the breadth of empirical reach. The inclusion of kegai territory within this political rhetoric was a method of claiming it,2 and a poetic lexicon of place-names soon developed. At first, place-names represented lands ‘belonging’ to Japan; throughout the Nara and early Heian periods, they developed an explicitly heteroglot signification which referenced not only geographic place, and also the coded narratives which were being allotted to them through wordplay. With many homonyms existing in spoken Japanese, and with place-names being written with Chinese ideographs (kanji) which carried their own meanings, poets were offered many opportunities for allusion through the semantic and phonetic qualities of place-names. This lexicon was frequently developed and enlarged as poets built upon the works of their forebears, applying new associations to past allusions and creating an intertextual poetic device 歌枕utamakura. The kanji for ‘song’ and ‘pillow’ make up the word ‘utamakura’, and it signifies words upon which readers can rest their heads in comfort. Having such a rich intertextual history of reuse, utamakura were like codewords: by the Heian period, place-names had come to refer more to emotions, themes or topics than they did to physical places. Over time, utamakura became increasingly disparate from their geo-spatial significations. One Heian poet writes that

It is of no practical value to remember which provinces these places are in … though I do not attempt to memorize such things, I have come to know that Yoshino is in Yamato province.3

To the extent which utamakura were disassociated from the physical landscape is best demonstrated in Mark Morris’ investigation of Sei Shonagon’s Makura no Zoshi (The Pillow Book, ca. 1002). Accompanying Sei’s account of her life in Empress Teishi’s service (977 – 1000) are many dan: cataloguing lists of various topics. Many include utamakura. In the style of utamakura zoshi (mid-Heian handbooks for poets, showing how best to employ utamakura), Sei’s lists are seemingly free of narrative. Dan 10 is simply titled ‘mountains’: “Mount Kase … Mount Konokure. Mount Iritachi. Mount Wasurezu. Sue no Matsu Mountain … Mount Itsuhata. Mount Kaeru”.4 The locations of these mountains vary greatly: Sue no Matsu Mountain is in Michinoku; Mount Kaeru is located in Tsuruga, Fukui – but Morris suggests that Sei’s ordering of these mountains is related to the grammar of utamakura, not geography. Translating their names shows us how. Mount Konokure translates to ‘the dark of the trees’; Mount Wasurezu is ‘Not Forgetting Mountain’; ‘itsu fata’ from Mount Itsuhata asks ‘When, when?’; Mount Kaeru is a mountain which ‘returns’: here Morris reads a conversation between lovers who ask in the darkness of uncertainty, whilst promising not to forget one another, when their lover will return.5 Other such catalogues follow as Sei reworks repositions and even invents6 geomantic elements of the landscape for the sake of coded literary narrative. As Sei writes her Makura no zoshi in the late Heian period, we can see how the courtly imagination was removing place from place-names: utamakura was but another poetic device with which Japanese poets could endorse and create their harmonious state, and their individual places within it.

Michinoku in Waka

We have seen how early Japanese histories depicted Michinoku as a barbaric wilderness as “part of a declaratory apparatus for the constituting of territory”;7 given the inelegant nature of a frontier, the court’s appropriation of Michinoku in a poetic rhetoric which aimed to omit inelegance is difficult to understand. An analysis of Michinoku utamakura in waka reveals a medley of contradictory textual representations of colonial realities and courtly ideologies.

‘Michinoku’ appears in a number of Man’yoshu poems grouped together as ‘Eastern Songs’. Meiji literary scholars searching for evidence of a ‘poetry of the people’ (kokumin-shi) read these azuma uta (Eastern Songs) as representing the true voice of the countryside: proof that poetry was an intrinsically ‘Japanese’ trait unhampered by unlearning or distinctions of class, but something that was in the blood (a more recent state-building ideology). However, as modern scholar Yoshikazu Shinada points out, the majority of these azuma uta “employ a poetic form identical to that used by aristocrats … [and] exhibit too great a formal regularity to have been orally composed by illiterate commoners”.8 Even so, in Japan today there still lingers an association of the anthology with a national poetic spirit. By analysing azuma uta against the context of state-formation, however, I argue that these poems were likely to have been written in the provinces, but not by people from them.

Poem no. 3807 in Book 16 of the Man’yoshu features a mountain in Michinoku:

Asakayama                The love I bear you
kage sae miyuru        is not like the shallow pool
yama no i no              mountain spring water
asaki kokoro wo        holding the mirrored image
oga omoha naku ni 
of Asaka Mountain itself.9

Asakayama sits in present-day Fukushima prefecture near ancient Michinoku’s southern border. As the Man’yoshu was compiled during the Nara period, it follows that we analyse the state’s involvement with Asakayama during this time to understand why this Michinoku location features in the anthology. The Asakayama border underwent many bouts of re-placement in the Nara period. The hitachi fudoki details that in 718 the districts of Iwashiro and Ihawi (modern Fukushima districts Iwashiro and Iwaki) were moved out of Michinoku and into neighbouring Hitachi,10 only to be switched back again in 721.11 Asakayama was thus in a position of almost-annexation into the Japanese state proper, demonstrating its relative familiarity to the court. True to aesthetic regulations, the poem mentions no inelegant militaristic expansionism, but uses the utamakura ‘Asaka’ to imply the verb asaku (asai in modern Japanese): to be shallow / superficial. A love narrative is thus attached to a place-name which is unrelated to its geomantic object in the poem. Meanwhile, a modern utamakura zoshi12 shows that place-names from particular regions in the north were used more frequently in early poetry than others. Compared to the numerous appearances of ‘Michinoku’ or ‘Mutsu’ are those rarer uses of north-western ‘Dewa’. Dewa province was only ‘officially’ distinguished from its eastern neighbour of Mutsu in 712. Dewa’s relatively late creation by the state tallies with its earlier literary absence, showing that the relationship between state action in the Michinoku region tallies with the court’s appropriation of its place-names within poetry:13 wherever members of the Japanese state were moving in Michinoku, so those places became part of their poetic lexicon.

Michinoku’s colonial realities are evident in the Shoku Nihongi (791), which states that in 702, “Tsukushi and Echigo were made to select female attendants and bodyguards and to send them to the palace. But Mutsu not to do so [my emphasis added]”.14 The Nara court found the kegai provinces of Tsukushi and Echigo quite troublesome in their assimilation process, and encouraged their subjugation through tributary ‘trades’. Although we do see the court receive presents from Mutsu in 697,15 their peace negotiations had evidently not reached the extent to that of other kegai regions. Through this evidence, I believe it unlikely that Michinoku communities were contributing poetry to the state’s anthology. Rather, that the contents of azuma uta are often aligned with the state’s militaristic movement of the period and have a suspiciously similar structural make-up to the rest of the anothology, they were surely written by colonisers as a tool of colonialism rather than by the kegai peoples being colonised.

Moving forward a few centuries, and the romanticism of ancient sakimori (barrier guards) and governors writing far from home was instigating an association of kegai provinces with the themes of distance, separation and grief in the Heian period. The Kokinwakashu’s travel section features many poets who employ ‘far-off’ place-names to tell of their separation from loved ones (and of course, from the emperor). This becomes contradictory, however, when the state army is dissolved at the end of the Nara period in 792. The Man’yoshu poets were men who really were in the distant provinces they were colonising; the lack of a state army after this, however, throws into question how Heian poets could be writing from the same experience.

Following the state campaigns against emishi in 801 and 810, it was pronounced that Michinoku was officially pacified. Give that state subjects were demonstrating their fatigue at the high taxes which funded these campaigns, the announcement of emishi pacification – “notwithstanding the still-substantial autonomy of surrendered barbarians”16 – was more likely a financial rather than militaristic choice;17 further campaigns in 878 and 879 and the many rebellions of emishi and fushu chieftains which where conducted throughout the Heian period contradict the court’s claim. Even so, the state began replacing sakimori and the conscripted state army with provincial governors ‘promoted’ to rule the countryside in their stead.18 There was no longer any reason to send high-ranking courtiers – our Kokinwakashu poets – to the northern frontier,19 and yet Heian literary representations of ‘Michinoku’ in poetry far exceed those of their predecessors.

Michinoku and her poesy became the lexicon of choice in expressing topics of distance, travel and grief in the Kokinwakashu. All poets featured in the travel section were supposedly “either exiled [officials] or serving his sovereign away from court”,20 and their longing to return to the capital reflects the new aesthetic associations of travel: courtiers only left the capital with the intention to return.21 This reiterated the harmonious superiority of the emperor’s domain, fulfilling the requirements of waka aesthetics whilst the poet expressed their own rightful place being in it. They employed Michinoku place-names within an aristocratic cultural context which was thoroughly removed from the geographic realities of those places: Miyagi prefecture’s 名取川 Natorigawa (literal translation: Name-taking River) became interwoven with the grief of losing a good name in aristocratic society for having a secret love affair discovered; Fukushima prefecture’s 阿武隈川 Abukumagawa (Let-us-be-together River) is appealed to for its assistance in concealing such an affair; Miyagi’s 末の松山 Sue no Matsuyama (Pining / Waiting Mountain at Sue) is too tall and too reliant as a waiting lover for ocean waves to crest it. Through utamakura, Heian poets transformed Michinoku into ‘Michinoku’: an imagined space rather than place.

Michinoku is thus represented as opposite to the capital in the Kokinwakashu; it appears only in the last section, and therefore in structural opposition to the opening topic of ‘spring’: a seasonal motif which expressed the everlasting reign of a new emperor. This positioning mirrors the centralised map of the state, and demonstrates the court’s political intention for the Kokinwakashu to reflect – and claim – the realm. This treatment of Michinoku was consistent, even up to modern times: Fujiwara no Teika’s layout of painted screens (fusuma) depicting meisho (famous places) at Retired Emperor Gotoba’s Saishoshitennoin monastery in 120722 designates the northernmost rooms to display Michinoku fusuma. He struggles over the layout, deliberating that

the mysterious Michinoku sites … would have to be handled with care. Their exotic flavours … [should] be hidden away in rooms … in which their “strange” forces might be released without deleterious effect to the inhabitants23

Modern utamakura zoshi editors also follow the pre-modern tradition of positioning capital-related lexicons first, kinai lexicons next and kegai lexicons last.24 It seems the ideological removal of Michinoku from the emperor’s divinity persists.

白河のShirakawa no seki (Shirakawa Barrier) is the most prolific Michinoku utamakura which explicitly portrays the region as a binary opposite to the capital. The utamakura nayose (1971), a modern utamakura zoshi, includes no other place-name which is twinned so frequently with  miyako (the capital).25 The Shirakawa Barrier was a gateway to the realm constructed under order of Kotoku to guard the capital against ‘eastern barbarians’, and it sits today along the ancient Michinoku border in Tochigi Prefecture. As an utamakura, it appears in the verses of the most celebrated pre-modern poets, including Taira no Kanemori in Shuiwakashu (Collection of Gleanings, 1005), monk-poet Noin Hoshi (Goshuiwakashu: Later Collections of Gleanings, ca. 1075), Fujiwara no Suemichi (Senzaishu, 1188) and Saigyo Hoshi (Sankashu, before 1190). In its earliest appearances, 白河の shirakawa no seki is twinned withmiyako for the purpose of contrast:

tayoriaraba                          If only I had a messenger,
ikade miyako ni
                   I certainly would send word to the capital
                  To let them know that today
shirakawa no seki wa I have crossed
koenu to
                        The Shirakawa Barrier26
– Kanemori

miyako wo ba                 I left the capital
yuki totomo ni
             along with the spring haze
tachi shikado                  
but now the autumn wind
aki kaze zo fuku             
is blowing across
shirakawa no seki         The Shirakawa Barrier27
– Noin

Whether figuratively or in reality, these poets claim to have crossed the boundary between culturally transformed lands and a conceptual wilderness, and their recalling the capital implies that the contrast of place has inspired their nostalgic mourning for home. Emphasising the importance of this boundary in the Heian imagination, Kanemori uses the verb suffix ‘koenu’, turning “to cross” (“koeru”) the barrier into a definitive, undoable action. The rest of his verse is full of apprehension, as hypotheticals (“araba; tsukenaramu”) and questions (“ikade”) frequently appear; whilst he is unsure as to what lies beyond the realm, he knows it will be different. Noin demonstrates his nostalgia by employing seasonal motifs, contrasting autumn winds to the capital’s spring in communicating the spatial, temporal and cultural distance he has travelled. Repeating Kanemori’s pairing of geographic locators helps establish them as binary opposites. By the time late-Heian poets were writing however, this binary pair was so familiar that including them both had become unnecessary:

mite suguru                since no-one passes
hito shi nakereba
      without looking
unohana no ha
           at the shrub fence
sakeru kaki ya           
blossoming with white deutzia
shirakawa no seki
     it must be the Shirakawa Barrier.28
– Suemichi

Upon his pilgrimage to Michinoku on a moonlit night, Saigyo stopped at the Shirakawa Barrier, and remembered the priest Noin’s verse:

shirakawa no            Shirakawa Barrier house
wo tsuki no    guarded by the moon –
moru kage wa           
its light leaking in
hito no kokoro wo
     holds captive
tomuru narikeri
.       The human heart.29

– Saigyo

Suemichi contests the earlier Heian tradition in his poem with unohana, a flower which blooms in the 4th month (May) and contradicts Noin’s autumnal imagery. He also semantically contrasts with Kanemori by filling his verse with certainties and concrete clauses (“hito nashi; kakine ya”). Saigyo pairs the barrier with a past poet rather than its utamakura associations, choosing to ground the place in historical rather than allusive narrative. That his moon is the barrier’s only guardsman highlights the absence of sakimori, and expresses the time passed since the barrier’s erection which, no longer manned, is no longer a frontier. From the imagery these later poets employ, it is evident that earlier boundaries of centre/ periphery, kinai / kegai were changing. An analysis of our two uta monogatari will show the Michinoku-as-other ideology shift even further until the turbulent end of the Heian period, when established aristocratic identities begin to collapse.


1Shirane, Seasons, p. 14

2Gustav Heldt, 2008. The Pursuit of Harmony: Poetry and Power in Early Heian Japan. USA: Cornell University East Asia Program, p. 43

3Shirane, Seasons, p. 68

4Meredith McKinney (trans.), 2006. The Pillow Book. England: Penguin Classics, p. 16

5Mark Morris, 1980. ‘Sei Shonagon’s Poetic Catalogues.’ In Harvard University of Asiatic Studies 40 (1 Jun), p. 16

6Morris,Catalogues’, p. 18

7Jonathan Bardo, 2002. ‘Picture and Witness at the Site of the Wilderness.’ In Landscape and Power, ed. W. J. T. Mitchell (Chicago: University of Chicago, 2002), p. 309

8Yoshikazu Shinada, 2000. ‘Man’yoshu: The Invention of a National Poetry Anthology’, trans. Collins, Kevin. In Inventing the Classics, ed. Shirane and Suzuki, pp. 31-50

9 Helen Craig McCullough, 1985. Kokin Wakashu: The First Imperial Anthology of Japanese Poetry. USA: Stanford University Press, p. 9

10Inoue Tatsuo, 2006. ‘The Hitachi Fudoki and the Fujiwara’, interpreted by Michiko Aoki. In Capital and Countryside, ed. Piggott, Joan R., p. 105

11Inoue, ‘Fudoki’, p. 115

12 謌枕名 / 吉田幸一, 神作光一, 橘りつ 」、古典文庫 , 東京 : 古典文庫, 昭和 49-51 ,  [1974-1976]

13 Aston, Nihongi, p. 89- 90

16Friday, Pale, p. 1

17Friday, Pale p.22

18Friday, Pale, p. 23

19 Karl Friday, 1988. ‘Teeth and Claws: Provincial Warriors and the Heian Court’. In Monumenta Nipponica 43 (2 Summer), p. 115

20 Heldt, Harmony, p. 164

21 Heldt, Harmony, p. 165

22Kamens, Utamakura, p. 218

23Kamens, Utamakura, p. 179

24謌枕名, p. 129


26Kamens, Utamakura, p. 154

27Kamens, Utamakura, p. 155

28Haruo Shirane, 1998. Traces of Dreams: Landscape, Cultural Memory, and the Poetry of Basho. California: Standford University Press, p. 330

29H. H. Honda, 1971. The Sankashu (Tokyo: The Hokuseido Press), p. 203

(It goes without saying that this work is my own. It is copyrighted both to myself and to SOAS University; any personal use of this material must be authorised by myself first. Thanks.)


‘State’ Building and ‘Other’ Building

Michinoku Province – “The Province Beyond the Realm”1 – was an incredibly significant place and idea to the western-based ancient and classical Japanese polity. Honshu’s most north-eastern region may have been geographically disparate from the pre-modern capital cities of Yamato (modern Asuka prefecture, ca. 250 – 710 C.E.), Nara (710 – 794) and Heian-kyo (modern Kyoto, 794 – 1185), but from the court’s first written piece of legislation in 604 C.E. until the Meiji period in 1868, it was inseparable from notions of home in the aristocratic imagination. For centuries, ancient Japan had been a member of China’s East Asian tributary network, submitting to Chinese rule in exchange for its protection and a share in its cultural establishments.2 Seeking equality with China, the Yamato court installed the same political institutions at home from which they were trying to break free. Prince Shotoku (573 – 621 C.E.), the nephew of Yamato’s Empress Suiko (r. 592–628), devised a native ruling framework in the form of Japan’s first legal constitution, and its seventeen articles of law, modelled upon Buddhist and Confucianist ideologies, enabled a process of Japanese state-building which mapped a socio-political structure onto the archipelago both horizontally and vertically: a vertical hierarchy situated the imperial ruler as the child of God, their imperial palace and aristocracy as the sky’s clouds, and ‘common’ subjects as the land below; a horizontal hierarchy established the Emperor’s body as the realm’s divine centre (geo-spatially symbolised by his capital city), and the lands which dispersed beyond its borders were measured in inferiority by the degree of their distance from it.3

Map 1.1 Early Provinces of Japan, ca. 645 – 800 C. E.4

Forty years later, Emperor Kotoku furthered the imperial claim on the land through his Taika Reforms (645), which installed a bureaucratic administrative system centralised in Kyoto and implanted his claim that “under the heavens [the emperor] there is no land which is not the king’s land. Among holders of land there is none which is not the king’s vassal”.5 He built guarded barriers around the home provinces (kinai), nationwide highways to connect the far reaches of the realm, and commissioned fudoki (provincial gazeteers) which mapped, named and brought in tax from the provinces.6 His territory stretched from northern Tsukushi (modern Kyushu) in the south, Shikoku in the west, and to the northern borders of Hitachi (Ibaraki), Shimotsuke (Tochigi) and Echigo (Niigata) provinces in the north. Beyond their northern borders was ‘Michinoku’, a region which belonged to the state only in name: ‘The Province Beyond the Realm’. This contradictary name at once signified Michinoku’s otherness to the state (“beyond”) whilst still being claimed under sovereign rule (“province” of Japan). From the moment it was ‘created’, Michinoku was a paradox of place.

The land was not all that needed identifying through naming. The throne also commissioned literatures in an attempt to unify the elite, for the elite were the only literate class and could together identify themselves through state-based literatures as belonging to that state: as ‘Japanese’.7 Two mytho-historical works, the Kojiki (Records of Ancient Matters, 712) and Nihongi (Chronicles of Japan, 720), narrate the state’s origin story. A semi-mythological emperor Jinmu (r. 660–585 B.C.E.) quits heaven to establish an earthly domain (Japan).8 His pacification of the people served to demonstrate his militaristic prestige, and many chapters of both works are dedicated to their subjugation and his creation of provinces. Japan was created out of the people’s submission to divine rule. However, the Nihongi claimed that:

the remote regions d[id] not yet enjoy the blessings of imperial rule. … it is true that the frontier lands are still unpurified and that a remnant of evil is still refactory … but in the Central Land, there is no more wind and dust.9

The people ‘beyond’ this frontier – a boundary which moved further and further north throughout the pre-modern period – appear in the chronicles as 蝦夷 ’emishi’ or ‘ezo’ (‘prawn brutish foreigner’). Although there were many provincial rebels who were pacified in these narratives, only those located in Michinoku were called emishi. A later state history, Shoku Nihongi (770) consistently names rebels from other eastern provinces as ‘ebisu‘ (‘prawn northern foreigner’), whereas rebels in Michinoku remain 奥蝦 (Michinoku emishi).10 Moreso than anywhere else, the eastern landscape itself appears dangerous, as serpents eat deities and Japanese heroes die far from home.11 As we shall see, later literatures follow this early tradition of equating the north-east with danger. Given that the purpose of these works was to construct a statehood identity, this treatment of the north-east as quintessentially different from the rest of Japan, both in landscape and people, is significant.

So why did they do so, and why the north-east? Bruce Batten argues that since pre-historic times, natural boundaries have helped form a cultural borderland between western and eastern Honshu.12 During the Yayoi period (ca. 300 B.C.E. – 300 C.E.), the ‘Yamato Japanese’ migrated from the Korean continent into south-west Japan, bringing with them wet-agriculture which flourished in the warmer climate. These are the communities who would become the subjects of the Japanese state. The ‘Ainu Japanese’ however descended from the native Jomon people of Japan, and practised a hunter-forager lifestyle which could endure the colder, northern climate. The ethnicity, cultures, languages and ways of living between these people were different, which could indicate why the ancient and Heian Japanese state sought to distinguish themselves. As Batten points out, however, a similar disparity of culture and climate also existed in southern Tsukushi,13 where the central Japanese state had formed another frontier against the Hayato (Okinawans). In contrast to Michinoku though, this Tsukushi frontier is relatively absent from state histories, suggesting that cultural/ genetic differences weren’t all that inspired the state’s othering of Michinoku. Furthermore, the separateness between the Yayoi Japanese in the centre-west and Ainu Japanese in the north-east were not clear cut: archaeological evidence from Ishinomaki (modern Miyagi prefecture) shows that a “grey boundary” stretching from Shizuoka to Niigata contained communities in which both the Ainu Japanese and Yamato Japanese lived side-by-side with a mix of language and culture.14 It was an idea more than reality, then, that differentiated the centre from the north-east. Japanese scholars may have claimed that in the north-east “lay the land of the hairy men”,15 but early Chinese annals also describe ‘eastern barbarians’ this way.16 As Classical Chinese literature was an enormous influence on Yamato and Heian statehood, it is probable that this blanket description of all Michinoku peoples suited their borrowed notions of geo-spatial rulership.17 The only distinctions the state did draw between Michinoku peoples is shown during a visit to Tang China in 659:

These missionaries took with them a pair of Emishi from Mutsu [synonymous with Michinoku] … The envoys explained that the “land of the Emishi” lay to the Northeast … and that it was inhabited by three groups, the “tsugaru”, “ara [wild] Emishi”, and the Nigi [tame] Emishi.18

The early Japanese court thus defined Michinoku peoples not in relation to their ethnicity, language or ways of living, but to their degree of state assimilation: their ‘closeness’ to the divine emperor.

The Nihongi further demonstrates how Tang theories of rulership influenced Japanese notions of geography and the state. Tang cosmologists proclaimed that ‘the centre’ was but one direction out of five which effected the cosmos; a harmonious cosmos reflected the emperor’s benevolence, and thus it was imperative for the state to ‘balance’ these five directions.19 The Nara court paid particular attention to “avoiding calamities”, an unbalance, which might put their ruler’s benevolence into question:

Planet        Element    Direction    Season       Signs of the Zodiac
Jupiter         wood        east             spring            tiger, hare
Mars            fire             south        summer        serpent, horse
Saturn          earth         centre        solstices    dog, ox, dragon, sheep
Venus          metal           west         autumn        monkey, cock
Mercury       water         north          winter         boar, rat

According to the theory of five phases or elements, the two elements bordering any particular element were beneficial to it, whereas the two separated elements were harmful. Thus both wood and earth were beneficial to fire, but metal and water were harmful.20

Figure 1.1 Cosmological theories in Nara Japan

The northern and eastern directions are shown to be ‘separated’ from, and thus ‘harmful’ towards the centre: even in cosmology, the north-east symbolised a threat.

Centuries of colonizing the east would dictate how much of the region was considered “harmful”: the frontier lands described in the Nihongi shrank, as the state took control of the eight Bando Provinces (Azuma, modern Tokyo area)21 before reaching the Michinoku border by the end of the Nara period. During this frontier process, some emishi were forcibly displaced throughout the country and re-branded fushu (‘surrendered barbarians’), whilst others made up the forces which would oppose state authority for centuries to come. Regardless of having no static spatial territory, the Togoku (eastern region) would still remain in the aristocratic imagination “a qualitatively different world that was subordinate to the kinai and the western provinces”22 up until the end of the Heian period.



1Takahashi, Frontier, p. 137

2Gina L. Barnes, 2007. State Formation in Japan: Emergence of a 4th-century Ruling Elite. Abingdon: Routledge, p. 27

3Donald Keene, George Tanabe, Paul Varley, Theodore de Bary eds. 2001. Sources of Japanese Tradition: From Earliest Times to 1600 Volume I Second Edition. New York: Columbia University Press, p. 60

5Keene (ed.), Sources, p. 51

6W. G. Aston (trans.), 1972. Nihongi Chronicles of Japan from the Earliest Times to A.D. 697. Vermont: Tuttle publishing, pp. 206 – 209

7Bruce L. Batten, 2003. To The Ends of Japan: Premodern Frontiers, Boundaries, and Interactions. USA: University of Hawai’i Press, p. 91

8Aston, Nihongi, pp. 109 – 111

9Aston, Nihongi, p. 112 – 135

11 Chamberlain, Kojiki, p. 73

12Batten, Ends, p. 61

13Batten, Ends, p. 62

14Batten, Ends, p. 62

15Tsunoda Ryusaku and L. Carrington Goodrich, 1951. ‘Japan in the Chinese Dynastic Histories’. In Perkins Asiatic Monograph 2: 38-40

16Takahashi, Frontier, p. 131

17Karl Friday, 1997. ‘Pushing Beyond the Pale: The Yamato Conquest of the Emishi and Northern Japan.’ In Journal of Japanese Studies 23 (1 winter), p. 13

18Aston, Nihongi, p. 261

19Shirane Haruo, 2012. Japan and the Culture of the Four Seasons: Nature, Literature and the Arts. New York: Columbia University Press, p. 208

20Aston, Nihongi, p. 67-8

21Friday, Pale, p. 6

22Aston, Nihongi, p. 369

23Shirane, Seasons, p. 14

(It goes without saying that this work is my own. It is copyrighted both to myself and to SOAS University; any personal use of this material must be authorised by myself first. Thanks.)



Various English-language scholars working within the fields of Japanese geo-historiography and literature argue that the authors of ancient and classical Japanese literature defined ‘foreign’ cultures as ‘other’ in order to define the Japanese aristocratic state as ‘self’.1 It is generally agreed that Japan’s first written legal constitution (ca. 604 C. E.) and its later amendments (the Taika Reforms, 645) were the earliest seeds of a statehood-as-self ideology in Japan, and were modelled on the polity of contemporary imperial Tang China: the ‘middle kingdom’ of Sinocentric East Asia to which Japan had long been a tributary state.2 By appropriating Tang China’s frameworks of rulership, the ancient Japanese Yamato court (250 – 710 C.E.) attempted to break away from this tributary identity, and be recognized by China on equal terms. Adopting the Tang geo-spatial boundaries of self as the imperial centre (kinai) and other as ‘beyond’ (kegai)3 allowed the Yamato and Nara (710 – 794 C.E.) courts to justify their territorial expansion into kegai lands as a process of state-building.4 They not only defined peoples and lands abroad as kegai, but any which were beyond their centre and surrounding “transformed” – culturally and ethnically similar – locales (kanai):5 the further people lived from the divine geo-spatial reach of the Japanese Emperor, the more inherently inferior they were believed to have been.

Particularly recipient to this kegai signification was the most distant region from the capital, north-eastern Michinoku. This enormous region’s various communities who rebelled against Yamato and Nara colonisation for centuries became ‘the other’ in pre-modern Japanese literature: a ‘barbaric’ collective in contrast to who the Japanese could elevate their own Chinese-styled civility. Whilst court-commissioned narratives detail the assimilation and subjugation of their north-eastern ‘barbarians’ co-ordinated by the divine-led Emperor, a plethora of historical and archaeological evidence contests them. This contradiction between Japan’s classical notions of state as expressed through literary representation and a different reality proven by archaeology is our pre-requisite to understanding how the classical Heian court (794 – 1186), our court of interest, came to view and write Michinoku into their literature through an internalized self/other binary. I have chosen to focus mostly on the Heian period’s courtly ideas about Michinoku because, as we shall see, a great many political developments took place at this time which greatly altered the Japanese state’s centre/ peripheral relations; although English-language studies have explored classical Japanese textual treatments of continental peripheries,6 still no such exploration for Heian Japan’s literary representations of Michinoku yet exists.

I hope to begin addressing this gap. First, I will demonstrate how Japanese state-building prior to the Heian period formed early Michinoku-Japanese relations, and how historical evidence contradicts the ancient state’s literary representations of them. Next, I will discuss the topographic poetic device utamakura (poetic word-play created through the use of place-names) and its uses towards state-building and othering. I will compare how two uta monogatari (poem-tales which employ utamakura), the Ise monogatari (ca. 850 – ca. 1011) and Saigyo monogatari (post 1185), utilize Michinoku as a setting for their narratives to demonstrate how over time, the symbolic ‘Michinoku’ was used first to critique the periphery and later the centre state, in response to changes in rulership throughout the late Heian and early Kamakura periods. Given the long periods and multiple authorships of both these works, they offer us a wider breadth of Heian attitudes than others might.

To analyse here all of the existing Heian literary representations of Michinoku is a task too large; nevertheless, I hope to give this new topic a good grounding, and enable further enquiry. As such, I hope this study will make a relevant contribution to English-language scholarship on pre-modern Japanese literature.


1Takahashi Tomio, 2006. ‘The Classical Polity and its Frontier’, trans. Friday, Karl, in Capital and Countryside, ed. Piggott, Joan R., pp. 128-145

2Mark Edward Lewis, 2006. The Construction of Space in Early China. USA: State University of New York Press.

3Takahashi, Frontier, p. 130

4Mark J. Hudson, 1999. Ruins of Identity: Ethnogenisis in the Japanese Islands. USA: University of Hawai’i Press,p. 194

5Takahashi, Frontier, p. 130

6Edward Kamens, 2007. ‘Terrains of Text in Mid-Heian Court Culture’. In Heian Japan, Centers and Peripheries. ed. Mikael Adolphson, Edward Kamens, Stacie Matsumoto. Honolulu: University of Hawai’i Press, pp.129 – 152

(It goes without saying that this work is my own. It is copyrighted both to myself and to SOAS University; any personal use of this material must be authorised by myself first. Thanks.)


This dissertation explores pre-modern Japanese literary representations of the northern Honshu region, Michinoku (modern Tohoku) in relation to pre-modern Japanese state-building processes. By reading ‘Michinoku’ as represented in state literatures against the realities made available through historical evidence (albeit unavoidably biased historical evidence), it unites two narratives hitherto treated separately by English-language scholarship in order to discover what Michinoku meant to the early Japanese, and why.

First, it discusses how classical Chinese geo-spatial notions of rulership influenced the pre-modern Japanese state’s ideological perception of and military action against Michinoku. Next is an in-depth literary analysis of Heian waka (poetry) and politics which demonstrates how a Michinoku-as-other ideology changed over the classical era in accordance with changes at court. Contradictions are shown to have arisen between courtly perceptions of Michinoku as a geographical space, and their symbolic notions of an imagined ‘Michinoku’ created through poetry. It becomes evident that courtly literary ideologies of Michinoku had a direct impact on the region, and that this impact can still be traced in the Michinoku / Tohoku landscape today.