Introduction

         Kokin Wakashu is an anthology of 1,111 Japanese poems (in the most widely circulated editions) compiled and edited early in the 10th century. The title, conventionally abridged in Japanese to Kokinshu, may be translated "Collection of Old and New Japanese Poems" or, perhaps more precisely, "...Ancient and Modern Japanese Poems." The collection begins with a Preface in Japanese and, in some editions, concludes with one in Chinese. The Japanese Preface, opening with the famous words "Japanese poetry takes as its seed the human heart," was long regarded as a model of classical prose, and line for line is undoubtedly the most heavily commented secular prose text of the Japanese tradition. The poems are divided into twenty scrolls or books (maki) each of which bears a title referring to conventional poetic topics (the seasons, love, parting, mourning, miscellaneous or "mixed" topics, etc.) or to genres ("acrostic" poems, "mixed" or miscellaneous forms, and poems of the "Bureau of Song"). The great majority of poems in the collection (all but 9, in fact) are in the form today usually called tanka (literally "short poem or song") but traditionally referred to as waka ("Japanese song / poem") or simply as uta ("song, poem") because this was the predominant canonical form of Japanese poetry from perhaps the 8th century until the late 19th century.

         The poems of Kokinshu can be roughly divided into three periods, which also reflect certain broad stylistic differences: anonymous poems of the early to mid-9th century or before; those of the age of the "Six Poetic Geniuses," the mid-9th century; and poems by the editors and their contemporaries, from the late 9th and early 10th centuries. Well over half of the poems are attributed to nearly 130 known (or named) poets, mostly of the late 9th century. Of the approximately 450 anonymous poems, many are believed to derive from oral traditions of folk song, though some Heian and medieval commentaries assert, plausibly enough, that the editors deliberately identified as anonymous certain poems by those of the highest social rank, others by persons of very low status, some of those by the compilers themselves, and poems which tended to impinge upon various taboos.

         Prosodically, the waka is defined quantitatively (there being no basis for identification of "feet," no strict distinction between accented and unaccented syllables in Japanese prosodics) as consisting of 31 syllables grouped according to a pattern of 5 ku or measures of, respectively, 5, 7, 5, 7, and 7 syllables, each of which was also required to be grammatically independent in the sense that phrasal breaks in syntax regularly coincide with the divisions between successive measures. At a higher level of organization, the first three measures as a whole were traditionally called the kaminoku (upper verse or stich) or simply moto (base), and the last two the shimonoku (lower verse or stich) or sue (end). Although this was a purely quantitative distinction in principle, it reflects a tendency (increasingly evident from the late 12th century) to place a strong syntactic break after the third measure, and less often after the first, resulting in what is called a "7-5 rhythm" (shichigocho, as opposed to the alternative, a "5-7 rhythm" in which breaks occur after a 7 syllable measure). 7-5 rhythm is a characteristic which distinguishes the later poetry of Kokinshu from that of the 8th century anthology Manyoshu and much oral poetry, including many of the early anonymous poems of Kokinshu, which favored breaks after the second and fourth measures. The 7-5 rhythm became established as the dominant rhythm of waka by the time of the early 13th century anthology Shinkokinshu.

         Syntactically, a majority of the poems in Kokinshu can be parsed as a single compound sentence or as two simple sentences; in the latter case, these are often in a relation of question and answer, enigma or dilemma and (often inadequate) solution, or a generalization followed by a restrictive condition. Exceptions abound, and some of the most memorable poems of the anthology can be read as lyrical observations of how things are, but it is safe to say that a questioning or plaintive mood prevails, the poet asking why things must be as they are, or why does experience not better agree with either reason or imagination? This has earned Kokinshu (more precisely, the middle and later poems usually taken to typify the anthology) a reputation for ironic wit and ratiocination which in turn has, on a favorable interpretation, been read as evidence of a sophisticated awareness of the discrepancies between language and reality, or, on a less sympathetic reading, as indulgence in sophistry or sheer wordplay.

         Along a somewhat different axis, scholars and critics of recent years have debated whether this interrogative mood reflects the ironic affirmation of a recently acquired technical facility of poetic expression, or rather a sense of nostalgic resignation, even of despair, inspired perhaps by the dissemination of Buddhist doctrine, or by the increasing political hegemony of the Fujiwara clan, or by some combination thereof. It is up to the reader to decide, of course, but while individual poems do show a fairly wide range of tonal variation, few, especially those of the seasonal and love books (which combined make up more than two thirds of the collection) are completely free of irony, and only a rather small minority of Kokinshu poems can be counted as unqualifiedly celebratory or "pastoral" in an affirmative sense.

         The immense prestige of Kokinshu as the unrivalled canon of classical waka throughout most of the tradition, especially after its recanonization in the late 12th century, assured that the limits on the range of acceptable tonal (in an esthetic sense) variations on a given poetic motif, even more so than the rules of decorum governing choices of diction and topics per se, were assumed to have been fixed by the precedents of this anthology. One factor which contributed to the tenacity of this assumption was a consensus, expressed by some of the most influential poets of the late 12th century (including Shunzei, Saigyo, and Kamo no Chomei, as well as Teika), that the poetry of Kokinshu represented the whole range of acceptable styles for serious poetry, and then some. The limits on both tonal range and stylistic license were, under the direction of the Nijo Family (later the Nijo School), which dominated traditions of Kokinshu interpretation from the late 13th through the early 19th centuries, defined as falling between haikai, at the lower end of the spectrum, and on the other a negative version of the sublime which medieval commentators often called yugen, an elusive word roughly translatable as "mysterious and dark" or "subtle and profound," though especially in later medieval usage this often designated a minimalist or even privative kind of beauty. Between these limits lay an ideal often referred to as ushin ("mindful," literally, or serious, stately, etc.) poetry, which in Nijo School practice came to mean a kind of conventionalism in which every element of a poem was harmonized and the impulse to originality satisfied by a single slight inflection on precedent. The latter was embraced by Nijo poets as the "Proper Style" (shofu) for waka, sometimes described by them as hitofushi mezurashi ("a single phrase of invention"), and ridiculed by their rivals as gokushin (sincere, or naive, to a fault).

         Haikai poetry in Kokinshu, most though not all of which is gathered in a sub-section of Book 19, "Mixed Forms" (thus suggesting that, although prosodically identical with waka, it was regarded as a distinct genre), was apparently classified as such on the basis of unorthodox (colloquial or archaic) diction as well as obtrusive irony or wit (the term haikai seems to have meant "discordant" or "dissonant" as well perhaps as "comic" to the editors of Kokinshu). If we see haikai as marking the lower bound of decorum, then as its antithesis yugen represented, at least for medieval readers, the surpassing ideal which few poems in fact achieved.

         One of the qualities of Kokinshu which helped to define its authority as the paradigm against which subsequent imperial anthologies (there were 20 of these) were measured and on which they were modelled, and which may account for the number of years apparently expended on compilation of the anthology, is the manifest care devoted by its editors to the structural arrangement of individual poems in sequences within each book, most conspicuously in those of seasonal and travel poems. The many forms of such arrangements include temporal progression through the seasons and through the stages of a courtly love affair, sub-sequences of topical images with subtle variations in treatment, alternation of anonymous (older) poems with those by contemporary (new or modern) authors and of rhythmic forms (poems with a caesura after the second vs. those with one after the third measure), patterns in which strongly original poems are set off by more conventional verses, and so forth. Medieval exegetes were very attentive to these principles of arrangement, which they referred to as budate (structure of the book or section) and shidai (sequence) and regarded as more pertinent than presumed authorial intentions in reaching a satisfactory interpretation of individual poems in the context of the anthology.

         One consequence of the editors' concern with arranging the poems into larger esthetically tangible patterns was that they omitted a number of the finest poems at their disposal (most of which were eventually seized upon by later anthologists) and included many which have never since been estimated very highly. Although Kokinshu was complete enough in its mapping of the world of courtly poetic topoi that (supplemented, to be sure, by the two subsequent imperially-commissioned anthologies, Gosenshu and Shuishu) it was regarded for centuries as an indispensable guide for aspiring poets, it was evidently not meant to be a shukashu, that is, a treasury of all (and only) the best poetry of its age, and has rarely been taken as such. Instead, however, thanks to the care with which it is arranged, Kokinshu is an eminently readable anthology, a fact which has perhaps contributed as much to its endurance as the excellence of its finest poems.

         The early textual history of Kokinshu is obscure and complicated by a number of variant texts which apparently reflect different stages of the editorial process. The most painstaking and widely accepted modern research into extant early manuscripts, that of Kyusojin Hitaku, resulted in the identification of at least 5 states of the text prior to its final acceptance, depending on the criteria applied. Differences among the earliest of these are not entirely negligible, but readers interested in Kokinshu as it was known to almost everyone who read or cited it after its recanonization in the late 12th century can be advised to begin with one of the 18 or so versions believed to have been transcribed and edited by Fujiwara Teika from 1209 until 1237, four years before his death. There are relatively minor variations among these, and considerably larger variations between the Teika editions and many of the extant (mostly fragmentary) pre-13th century editions, but the prestige of Teika and his heirs assured that it was his editions and copies thereof which defined Kokinshu for all practical purposes from around the mid-13th century down to the present day. Fortunately, all of Teika's editions are based upon a textual line which derives, with slight variations, from the fifth and final state of the "public" draft of Kokinshu prior to its official submission for imperial review and endorsement.

         Teika's editions are in large part based on one received from and edited by his father Shunzei, which in turn is a collation of an edition, the Shin'in Gohon (Hanazono Safu-hon), used by Emperor Sutoku for formal readings, with that received by Shunzei from his teacher, Mototoshi. One of the clearest distinctions of the Teika editions is that eleven poems present in Shunzei's received text but omitted from the Shin'in Gohon, which were included but marked for deletion in Shunzei's collated edition, were extracted and placed in an appendix at the end of the text. Otherwise, Teika made a number of editorial changes, some of them evidently based on his esthetic preferences. In addition, his editions typically appear to have included numerous interlinear notations giving textual variants, proposed emendations, identifications of certain poets by alternative names or titles, etc.

         The specific text on which the Japanese Text Initiative e-text is based is one of only two Teika holograph manuscripts known to be extant and the only one available for scholarly use (in the form of facsimile editions). In 1938 when this manuscript was designated a "National Treasure" (later renamed "Important Cultural Property"), a collotype reproduction was prepared and titled the "Date Family Kokin Wakashu" because it had been presented to Date Masamune by Toyotomi Hideyoshi in the 16th century and remained in the possession of the Date family until recently. (Ownership has since changed but it is still widely referred to as the Date Family text.)

         The manuscript is exceptional in that the colophon contains no date, but Kyusojin Hitaku has offered a very credible argument, based on an analysis of the diction of the colophon, the handwriting, and an entry in Teika's diary, for surmising that the manuscript was prepared in the spring of 1227 for presentation to the eldest daughter of the former emperor Tsuchimidako. This would place it in close proximity with Teika's holograph manuscript of 1226, the Reizei Family text, and just a few years later than the edition of 1223, which became the standard edition for the Nijo Family and the basis of the most widely circulated versions of Teika's Kokinshu.

         There are a number of mostly minor discrepancies among these three texts, and a handful of what are presumed to be scribal errors in the Date Family text. But the fact that the latter is the only extant edition in Teika's hand which has been made public is a compelling reason for its choice as the source text for modern editions of Kokinshu, and in fact several such editions, including those of Takeoka, Kyusojin (Kodansha Gakujutsu Bunko edition, 1982) and the Shinpen Kokka Taikan, are based on the Date Family text. The Shinpen Kokka Taikan now serves as the de facto standard for referring to poems in Kokinshu, another reason the Date Family text was selected as the source text for the Japan Text Initiative's electronic edition.

         Two complete and meticulous English translations of Kokin Wakashu have been published in recent years: Kokinshu: A Collection of Poems Ancient and Modern, translated by L.R. Rodd with M.C. Henkenius (Princeton: Princeton University Press, 1984); and Kokin Wakashu: The First Imperial Anthology of Japanese Poetry, translated by H.C. McCullough (Palo Alto: Stanford University Press, 1984). The Princeton edition of Kokinshu is out-of-print, but has been reprinted by Cheng & Tsui.