1

Introduction: writing, literacy, and the

origins of Japanese literature
david lurie

Numerous problems of definition and scope confront any survey of the

beginnings of Japanese literature. We obviously have no direct access to
the stories and songs that circulated before the advent of writing. Some
features of this preliterate world can be extrapolated from later sources, but
this is difficult to do with any confidence because the writers of many early
texts deliberately engineer an impression of orality. In poetry – both verna-
cular (uta) and Chinese-style (shi) – it is also difficult to separate the mid to
late eighth-century anthologies (the Kaifūsō and the Man’yōshū) from the
historical milieux in which the poetry they collect was first composed and
appreciated. Scholars are interested in the unfolding of particular genres,
motifs, and techniques, but these anthologies themselves were shaped to
present their own selective and tendentious versions of such literary histories.
Similar difficulties pertain to prose, but in that case there is also the larger
problem of delineating the literary from other types of writing. Literate elites
of the eighth century devoted as much or more time to studying, composing,
and commenting on Confucian, Buddhist, technical, and legal writings as
they did to appreciating the rather small subset of prose works that are now
considered to be part of the canon of Nara period classics: the Kojiki (Record
of Ancient Matters, 712), Nihon shoki (Chronicles of Japan, 720), and the fudoki
gazetteers.
For all periods of premodern Japanese literature – and indeed, for all
premodern literatures – what survives is only a portion of the writings that
were produced, but this situation is more extreme for the Nara and early
Heian periods than for any subsequent point in Japanese history. The
circumstances of the fudoki make this abundantly clear. Five survive as
integral texts, only one of which is complete (that for Izumo Province).
The early eighth-century order that called for the production of these
works was directed at all of the provinces, of which there were then about
sixty, and quoted fragments survive from gazetteers for around forty of them.

15
 david lurie

If the lower figure reflects the number actually composed, only about
12 percent of the fudoki survive; actually it is probably closer to 10 percent.
Such high attrition is connected to the uncanonized status of these texts in
Heian and medieval Japan, but similar proportions of other genres met the
same fate. The Man’yōshū (Collection of Myriad Leaves, c. 759) refers to older
poetry collections as sources (citing a half dozen by name), but none survives;
the prefaces to the Kojiki and the Kaifūsō (Florilegium of Cherished Airs, 751)
mention lost works, as does the Nihon shoki; and the content of the Nihon
shoki itself shows that it drew on various sources, none of which is extant.
Considering the broader situation down through the end of the Heian period,
approximately two thirds of the titles mentioned in the Honchō shojaku
mokuroku (a late thirteenth-century bibliography) no longer exist. Statistics
like these remind us that there is ample reason to be skeptical of literary-
historical generalizations based on extant works.
But such limitations, again, pertain to any premodern society, and com-
paratively the literature of early Japan is rather better known than that of
many other ancient traditions. Extensive works like the Nihon shoki and
Man’yōshū survive intact, and, to the best that we can ascertain, the extant
sources are representative of the range and variety of early writings. One
reason for the relative accessibility of ancient Japanese literature is the speed
with which it emerged: only about three generations separate the advent of
widespread literacy, in the mid seventh century, from the composition of the
oldest extant works in the early eighth century.
The first appearance of writing in the Japanese archipelago was much
earlier: inscriptions in Chinese characters on imported artifacts (mostly coins
and mirrors) are found starting around the last century BCE, in the late Yayoi
period. The first substantial inscriptions that were domestically produced
date to the fifth century CE, in the Tomb period, but there is no evidence that
significant numbers of people were able to read or write. Until the mid
seventh century literacy remained the province of specialist scribes –
migrants from the Korean peninsula and their descendants – who were
employed by the Yamato Kings, rulers from the area of modern Nara and
Osaka who presided over a loose federation of local potentates spanning the
archipelago from Northern Kyushu to the Kantō region. The importation of
Buddhism in the mid to late sixth century introduced new kinds of texts and
new modes of literacy, but these too remained narrow, specialized pursuits.
Writing had little meaning for a population to whom it was still just a
talismanically powerful symbol, to the extent that it mattered at all.
(Subsequent myth-making by eighth-century ideologues, most prominently

16
 Introduction: writing, literacy, and the origins of Japanese literature

in the Nihon shoki, suggests an earlier and more vigorous adaptation of

writing in general and Buddhist textuality in particular, but there is little
archaeological support for this narrative, and doubt has been cast on the
dating of most of the inscriptions traditionally associated with it.)
The change, when it came, was part of a much larger regional transforma-
tion of East Asia that followed the reunification of China under the Sui and
Tang dynasties in the late sixth and early seventh centuries, which led to the
emergence or expansion of states in peripheral regions. In 645, as conflict
among the three kingdoms of the Korean peninsula heated up, the Yamato
ruler later known as Tenji engineered a coup against the Soga lineage group,
who had dominated the court for two generations. The following decades
saw rapid construction of a Chinese-style state apparatus, complete with
census-taking, more systematic taxation, and a better-organized military,
especially after Yamato forces became involved in the final defeat of their
ally Paekche in 663. The resulting crisis further spurred development of a
bureaucracy, at the same time that literate refugees from the Korean penin-
sula were fleeing to the Japanese archipelago. The archaeological record
shows that the widespread use of writing for everyday communication and
record-keeping emerges in these decades, just as it became both necessary
and possible to staff a government based on texts. Not coincidentally, works
like the Kaifūsō and Man’yōshū also suggest that Tenji’s court in his new
capital in Ōmi (on Lake Biwa near modern Kyoto) was a center of literary
composition in both Chinese and vernacular styles. In monumental inscrip-
tions, poetry composed at court banquets, and other forms of writing, early
Japanese ideology followed the classical Chinese linkage of a well-ordered
state with well-ordered, aesthetically pleasing texts.
One reason for the rapidity of the seventh-century transformation of
literacy was the flexible relationship between spoken languages and texts
written in Chinese characters, then essentially the only form of writing in
East Asia. As a primarily logographic script at this moment in their history,
characters were associated with Chinese words and morphemes, but they
could also be linked to Korean or Japanese words and morphemes of similar
meanings. This meant that texts that had originally been written in Chinese
could be read in Korean or Japanese, in a process called kundoku (literally,
reading by gloss). Conversely, it was possible for someone who did not speak
Chinese to write a text that could be read in that language, by following in
reverse the kundoku procedure for rearranging the syntactical order of
character-texts. Chinese-style writing thus provided a common medium for
communication across linguistic and political boundaries; it was also a

17
 david lurie

powerful source of prestige and authority. But kundoku could be used to

write logographic texts that departed from the orthodox Chinese style in
their ordering and character usage, and it was also possible to use characters
phonographically, spelling out syllables of non-Chinese languages without
regard for the meaning of the words that had originally been associated with
those graphs. In short, the system of writing that was adapted in mid seventh-
century Japan involved multiple principles and styles of inscription, and was
well suited to the various demands placed on it by an age of political
transformation and literary innovation.
Already from the mid seventh century, archaeological discoveries of
wooden tablets inscribed in ink (mokkan) show the range of available
styles, from Chinese-style logography, to more localized logographic
writing that could not be read in Chinese, to mixtures of logographic
and phonographic characters, to entirely phonographic texts. The details
remain unclear, but parallels with discoveries in Korean sites, and the
well-documented contributions of scribes and refugees from the Korean
peninsula, suggest that many of these techniques were imported. At any
rate they were all being used to write texts in Old Japanese by the
second half of the seventh century. A handful of these artifacts are
belletristic works – mainly vernacular poems (uta) written phonographi-
cally – but we must rely on eighth-century sources for a fuller picture of
the emergence of Japanese literature.
The political impetus for the creation of the earliest extant works was
provided by the state-building activities of Tenji’s successors. His death in 671
was followed by the Jinshin War, a brief conflict that pitted his brother, later
known as Tenmu, against Tenji’s son. The victorious Tenmu (r. 672–86),
along with his consort and successor Jitō (r. 686–97), embarked on a far-
reaching transformation of the nascent state. Among the developments of
their reigns were the country name Japan (Nihon or Nippon), the title tennō
(Heavenly Sovereign or Emperor), written law codes, new systems of court
rank and title, an expanded and more powerful central bureaucracy, greater
state control over religion, and eventually a new Chinese-style capital city
(Fujiwara, established in 694). These new institutions were matched by
literary innovations. The great poet Kakinomoto no Hitomaro and others
wrote soaring elegies and paeans to Tenmu, Jitō, and their princes that form
the core of the poetic canon established by the Man’yōshū. Sponsorship by
Tenmu and Jitō and their successors was responsible for the compilation of
the Nihon shoki, and, according to its preface, of the Kojiki. All of these works
were dedicated in their own way to the glorification – and at times even the

18
 Introduction: writing, literacy, and the origins of Japanese literature

deification – of the rulers who had established themselves as the first emper-
ors of Japan in the aftermath of the Jinshin War of 672.
In the early eighth century complete penal and administrative laws were
promulgated – the 701 Taihō code (revised in 757 as the Yōrō code) – and a
new capital city was established to the north of Fujiwara: the Heijō capital in
Nara, which with interruptions would remain the political center from 710
until 784. This was a period of great cultural dynamism, symbolized by the
construction of the enormous Tōdaiji temple at Nara and the country-wide
network of provincial temples (kokubunji) centered on it, and also by the
lavish art works and luxury products, many imported from Korea, China, and
the Silk Road, that are preserved in the Shōsōin depository. But the Nara
period was also marked by great political turmoil, with rebellions, conspira-
cies, and purges; there were also natural disasters like the great smallpox
epidemic of 735–7, which some scholars estimate killed as much as a third of
the population. This combination of brilliance and upheaval underlay the
literary production of the eighth century, including the composition of much
of the poetry collected in the Kaifūsō and Man’yōshū and also the compilation
of those anthologies themselves, the completion of the Kojiki and the Nihon
shoki, and the production of the fudoki gazetteers. All of these writings were
produced for the court, with official or unofficial sanction. More so than for
any subsequent era, the literature of ancient Japan is inseparably linked to its
political history.
The legitimacy of imperial rule by Tenmu’s and Jitō’s successors (their line
was supplanted in 770 with the accession of one of Tenji’s grandsons, but the
fundamental structures they established remained in place) was supported by
a mélange of symbols and rituals with complex origins. Similarly, early
Japanese poetry and prose drew on a wide range of sources, foreign and
domestic. But, as elsewhere in East Asia, the armature of this emergent
tradition was the literary Chinese canon. As reflected in the official university
curriculum outlined in the eighth-century administrative codes, the funda-
mental framework of learning and knowledge was provided by the Five
Classics and their commentaries: the Odes (shi), Documents (shu), Rites (li),
Changes (yi), and the Spring and Autumn Annals (chunqiu).
Early Japanese readers were also exposed to a surprisingly expansive corpus
of other works. The dynastic histories available in eighth-century Japan
included classics like the Shiji and Hanshu, and extended to those compiled
up to the early Tang. Allusions in works like the Nihon shoki, Kaifūsō, and
Man’yōshū, and scraps of text in wooden and paper documents, show that
poetry anthologies circulated widely. The most important was the Wenxuan (c.

19
 david lurie

526–30), but collections of individual authors and less well-known anthologies

like the Yutai xinyong (c. 545) were also influential. It is also clear that early
Japanese elites consulted a range of Taoist writings and technical manuals of
medicine, warfare, architecture, engineering and so on, in addition to enor-
mous quantities of Buddhist sutras, treatises, and commentaries. But the most
important imported texts were the classified omnibus works, both textbooks
and references, that served as the primary entry point into the world of literary
Chinese writings. Many of the classical allusions (and borrowings) in works like
the Nihon shoki were taken secondhand from such sources, which included
classified encyclopedias (leishu) like the early Tang Yiwen leiju and Chuxue ji.
Dictionaries like the Shuowen jiezi (c. 100 CE) and Qieyun (601) were also widely
consulted; the most important of these seems to have been the extensive sixth-
century Yupian, which served as both a dictionary and an encyclopedia.
By the eighth century it is clear that many domestic compositions had joined
the foregoing imported texts. In addition to poetry anthologies like the
Man’yōshū and Kaifūsō and prose works like the Kojiki and Nihon shoki, these
included family histories and biographies, accounts of temples and shrines,
dictionaries and glossaries, and commentaries on Buddhist texts. The legal
codes that were compiled during the reigns of Tenmu and Jitō, and possibly as
early as that of Tenji, were accompanied by substantial commentaries, even as
they generated further legal material in the form of kyaku (ordinances) and shiki
(statutory elaborations). Early commentary on the codes survives only in later
collections, the Ryō no gige (833) and Ryō no shūge (late ninth/early tenth
century). Government documents of the eighth century, often straightfor-
wardly expressed ordinances and statues, but sometimes of sufficient length
and elaboration to be considered quasi-literary works, are collected in the 927
Engi shiki and in categorized references like the Ruijū sandai kyaku (eleventh
century) and Ruijū fusenshō (late eleventh/early twelfth century).
We have only indirect evidence of what types of writing were considered
most valuable in ancient Japan, but it is surely anachronistic to treat the prose
of the Kojiki and fudoki along with the poetry of the Man’yōshū and Kaifūsō as
literary while excluding royal proclamations and reports to the throne in
elaborate Chinese-style parallel prose. Nonetheless, the writings surveyed in
the following pages include some of the most brilliant and engaging in the
Japanese tradition, establishing precedents for and anticipating features of later
works of poetry and narrative prose in both Chinese and vernacular styles.
In a prewar lecture, the influential scholar Orikuchi Shinobu (1887–1953)
raised a fundamental issue of early Japanese literature when he said the
Hitachi no kuni fudoki was “a gazetteer of imported modishness [haikara na

20
 Introduction: writing, literacy, and the origins of Japanese literature

fudoki]. That is, it gives the feeling of being thoroughly dominated by the
authority of the court. One could say it is the sort of work that has no dreams
at all – or rather, that if it does, they are dreams of China.” A subsequent
lecture expanded on this formulation: “To put this in contemporary terms,
the Hitachi no kuni fudoki was written by men of civilization [bunmeijin]
looking back at the world of the past, and therefore incorporates a cold,
indifferent attitude that is incapable of fully understanding that past.”1 The
use of words with Meiji resonances is deliberate, involving a parallel much
invoked by late nineteenth- and early twentieth-century intellectuals. Just as
the “civilization and enlightenment” discourse of the Meiji period strove to
leave behind traditional early modern culture, Orikuchi implies, the Sinicized
“civilization” of the eighth century was similarly opposed to a rich earlier
native culture. But this is a flawed analogy. While remnants of Edo period
culture were everywhere in evidence during the Meiji period, and indeed in
Orikuchi’s own day, the only traces of early Japanese literature are from
precisely this Sinicizing period. It is true that works like the Hitachi no kuni
fudoki or Nihon shoki, which rely on Chinese rhetoric and imagery, contrast
with “warmer,” apparently more “traditional” texts, such as the Kojiki or the
Izumo no kuni fudoki. But works of the latter type were in their own time just
as new-fangled and innovative as the more superficially Sinicized ones;
perhaps even more so, as they did not conform to the preexisting trans-
regional norm of Chinese-style writing.
Orikuchi limns a distinctive feature of the style and narratorial perspective
of the Hitachi no kuni fudoki. But we can accept this insight without the
baggage that has been loaded onto it. It seems unlikely that the authors and
readers of ancient Japan would have felt the need to choose between more
“modish” (if indeed that is what they were) Chinese-style writings and those
that, like the Kojiki, engineered new forms of distinctive local significance.
From the Man’yōshū to the Nihon shoki to the fudoki, eighth-century texts
demonstrate a delight in multiple accounts: variant narratives, alternate
attributions, differing local legends, and so on. The weighty authority of
the Nihon shoki, or the totalizing ambitions of the Kojiki, are an essential
feature of those works, but we should not allow the comparative scarcity of
surviving writing from this era to blind us to the fact that contemporary
readers would have experienced and appreciated them in the context of a
much wider world of diverse alternate accounts.

1
Orikuchi hakase kinen kodai kenkyūjo, eds., Orikuchi Shinobu zenshū nōto-hen, vol. 2
(Tokyo: Chūō Kōronsha, 1970), 215 and 231–2.

21