SOURCE: "Association and Progression: Principles of Integration in Anthologies and Sequences of Japanese Court Poetry, A.D. 900-1350," translated by Robert H. Brower and Earl Miner, Harvard Journal of Asiatic Studies, Vol. 21, December, 1958, pp. 67-127.

[In the following excerpt, Konishi demonstrates that poems in the Shinkokinshu were inspired by and developed from poetry in the Kokinshu.]

… The meaning of the title, Shinkokinshū, is "New Anthology of Poems Ancient and Modern"—in other words, the "New Kokinshū." In> giving their anthology this name, the compilers were consciously expressing a neoclassical ideal and were specifying the source of their inspiration—the Kokinshū, or "Anthology of Poems Ancient and Modem," the first collection of Japanese poetry compiled by imperial command early in the tenth century. The Kokinshū remained, despite fundamental changes in poetic theory and practice, the almost universally accepted standard of propriety in poetic diction and, to a lesser extent, technique, throughout the history of the Japanese classical tradition. In choosing the name for their anthology, and in raising the image of the older collection through echoings, the compilers of the Shinkokinshū were giving expression to their ideal of recreating in their own age—an age, significantly, of political and social decline for the court aristocracy—the poetic achievement and, by implication, the social brilliance of that more happy period of the Kokinshū. We may demonstrate the ways in which these overtones are conveyed by comparing poems 10 through 15 in the Shinkokinshū (which have been already quoted above) with poems 17 through 22 in the first book of the Kokinshū, a book which is also naturally devoted to spring.

KKS I: 17

 … Be kind to us,
And do not burn today the withered fields
  Of ancient Kasuga,
For like the young grass sprouting underneath,
She hides there, and I there by her side.
Anonymous

KKS I: 18

 … O guardian of the fields
Of Tobuhi in ancient Kasuga,
  Come out and look
And tell how many days I still must wait
Until the joyous time to pick young shoots.
Anonymous

KKS I: 19

 … Deep in the mountains,
Even the snow that fell upon the pines
  Has not begun to melt,
But in the Capital, the fields are thronged
With courtiers gaily picking the young shoots.
Anonymous

KKS I: 20

 … Today the rains of spring
Spring on us with the suddenness
  Of a far bent bow—
If only they will fall once more tomorrow
So we may soon go out to pick young greens!
Anonymous

KKS I: 21

 … It was for you
That I went out to the fields of spring
  To pick young shoots,
Though all the while the falling snow
Piled without surcease upon my sleeves.
Emperor Kōkō (830-887)

KKS I: 22

 … Do those girls set forth
On an excursion for young shoots,
  That they so gaily wave
Their white linen sleeves in beckoning
Towards the fields of ancient Kasuga?
Ki no Tsurayuki (d.?945)

A comparison of these six poems with poems 10 through 15 of the Shinkokinshū will show that such complexes of images as "picking young shoots in the fields of ancient Kasuga" appear in both sets. It was perhaps inevitable, given the popularity of such images and the annual observance of this social rite, that there should be poems on young shoots in both anthologies. However, the number of poems in each anthology is the same and, what is more important, the same kind of progression is followed in each case. The first poem in each group (KKS I: 17 and SKKS I: 10) is not properly speaking on the topic of young shoots at all. Each poem is rather an introduction to the series, each prepares the way for those that follow by treating only the young grass which had begun to spring up on the fields of Kasuga. The next three poems in each case develop the progress of the actual picking; and the fifth poem in each sequence (KKS I: 21 and SKKS I: 14) treats the topic in terms of a speaker who has gone out and gathered greens and then makes a present of them to someone who did not go. Finally, the speaker of the sixth poem in each set is a person who, for some reason, did not or could not go on the outing himself.⁶

This evidence seems convincing enough to show that the compilers of the Shinkokinshū were consciously attempting to raise the image of the Kokinshū at this point in their anthology, and at times this echoing becomes very complex. We must recognize, for example, that Ki no Tsurayuki in the last poem of the group in the Kokinshū (KKS I: 22) has in turn raised the image of an earlier age with his "fields of ancient Kasuga" and his use of such an old poetic technique as the "pillow word" (makurakotoba), or conventional attribute, in the phrase shirotae no sode, or "white linen sleeves." This setting and this technique evoke the life and the poetry of the seventh and eighth rather than the late ninth and early tenth centuries when Tsurayuki lived. Thus the reappearance of the "fields of ancient Kasuga" in the Shinkokinshū, in a context which echoes by other means the corresponding poems of the Kokinshū, means that the readers of the thirteenth century were reminded of the age of the Kokinshū, but that this allusion itself alludes to an even more remote and romantic era. Tsurayuki's former evocation of the past mingles with the new evocation, harmonizing three ages of past and present. The identity of the poets within the sequences is also important. The "anonymous" poems, as well as the one by Emperor Kōkō, in the Kokinshū set are old, in the sense that they employ techniques more "primitive" than those employed in the tenth century—these poems are mostly in the declarative mode characteristic of Japanese poetry of a hundred years or more before the generation of Tsurayuki. The correspondence is not exact, but the poems in the Shinkokinshū on the topic of "young shoots" are also for the most part by poets who lived from several hundred years to a generation or two before the time when this anthology was compiled. The fourth and sixth poems in the set (SKKS I: 13 and 15) do not fit this rather loose definition, however, and perhaps indeed we are straining a point here. We shall content ourselves, therefore, with simply mentioning the facts and suggesting the possibility that this kind of echoing (which occurs beyond question elsewhere in the anthology) might have been consciously, if imperfectly, attempted in this group of Shinkokinshū poems as well.

Despite these astonishing similarities, however, we must be fully aware of the differences which exist between the two sets of poems in the older and the newer collection. The principles of association and progression are already evident in the Kokinshū, for example, but not in as thorough or consistent a way as in the Shinkokinshū. Thus the first poem in the Kokinshū set on "young shoots" (KKS I: 17) is principally concerned with the burning of withered fields at winter's end, and may be said to be somewhat too remote from the designated topic. Again, the third poem (KKS 1: 19) treats the topic in such a way that the courtiers are represented to be already out on their excursion, whereas in the fourth poem (KKS I: 20) the young shoots are not yet ready to pick. In other words, these poems are not arranged according to a logically developed time sequence. But in the Shinkokinshū, the six poems are all (except the first) clearly on the topic of "young shoots," and the time progression is completely logical and in harmony with the external world. In brief, the poems in the Shinkokinshū are far more carefully and consistently arranged than those of the Kokinsh ū. The compilers of the later anthology certainly used the earlier one as their model and their guide, but there is a further significance in the title "New Kokinshū": the compilers were not satisfied with mere slavish imitation of their ideal; instead, while following the general outlines laid down in the Kokinshū, they obviously attempted to create something "new" and something better than the anthology to which they looked for inspiration.

The "newness" is evident not only in the much greater care taken with the associations and the progression from poem to poem; it is perhaps nowhere more clear than in the appearance in this set of Shinkokinshū poems of a personality which is hardly realized at all in the Kokinshū group. This is the figure of the "man of elegance," who appears most conspicuously in the second and third poems of the Shinkokinshū set (SKKS I: 11-12). The notion of "elegance" was perhaps the dominant ideal of both art and life among the aristocrats of Japan in the Heian period (794-1185), and it certainly was already widespread by the time of the Kokinshū. It was, however, only a century or two later that the concept became realized to such an extent that it was translatable, so to speak, into the terms of poetry. It is not surprising, therefore, that the ideal should be more vividly realized in the Shinkokinshū.⁷ Like most aristocratic ideals, this one was at once amateurish and esoteric, a glorification of the dilletante with social and philosophical overtones; it is very like Renaissance European ideals of the courtier. Such a man would rope off his fields so that he and his friends might indulge in the elegant, if primitivistic and romantic, activity of picking spring greens, undisturbed by the unwelcome intrusions of the inelegant commoners. Therefore, although this group of poems in the Shinkokinshū echoes the similar group in the older anthology, it does so with a difference—an elegance that the compilers have invited us to contrast with that of the anthology on which they patterned theirs but which they were obviously determined to surpass.

Such subtleties in the association and progression of the seasonal poems of the Shinkokinshū perhaps make Tristram Shandy, Ulysses, or Ezra Pound's Cantos seem less uniquely Western or strangely modern than they often seem to be. These same subtle techniques of association and progression are employed in other sections of this anthology such as the love poems (Books XI-XV) or the travel poems (X). There is no need to analyze these sequences to show how they employ the same techniques as the seasonal poems, but they are worth examination for what they show of new techniques.

The love poems are of particular interest in that they are organized in accordance with a kind of time progression which is based entirely upon human concerns and which is more dramatic than that of the seasonal poems; and the poems on travel illustrate still another kind of progression. We shall begin with the love poems, taking as our initial example the first four poems in Book XI.

SKKS XI: 990

 … Can it be that I
Shall never see at closer range
  The pure white snow
That glimmers in far Kazuraki
On the peak of Mount Takama?
Anonymous

SKKS XI: 991

 … Only in story
Had I heard of the waterfalls
  Of beauteous Yoshino,
Until this day when my own sleeves
Are moistened with their spray.
Anonymous

SKKS XI: 992

 … My love is like
The smouldering fires they tend
 Beside their huts
 To frighten deer from ripened grain
 In the fields of distant hills.
Kakinomoto Hitomaro (fl. c. 680-700)

SKKS XI: 993

 … Must I go on,
My love shut up within my breast
 Never to show forth
As do the ears of grain at Wasada
In Furu, land as ancient as the gods?
Kakinomoto Hitomaro

It may seem surprising that a volume of love poems should begin with what is apparently a seasonal poem on a winter topic. But when this first poem is read in the context of the following verses, its use of the allegorical mode becomes obvious: the speaker of the poem is a man, and the white snow on the peak is his beloved. The poem tells us, further, that although the speaker has seen this woman at a distance, and has fallen in love, he has not yet met her, and we assume that she on her part is unaware of his tender feelings. Such a one-sided beginning was the rule in affairs of the heart at the Japanese court and became a fixed convention in the literary treatment of first love. Each of the poems quoted is, in fact, a variation upon the fixed topic of "love before the first meeting." All of the poems in Books XI and XII of the Shinkokinshū are on this, or virtually identical topics, which required a more or less uniform treatment and tone. This shared topic gives these poems a thematic unity. At the beginning of Book XIII, however, there is a progression in the dramatic development of the human relationship, as can be seen in the first poem of this book:

SKKS XIII: 1149

 … Since it can scarcely be
That you will remember this road of love
 To the end of our life's journey,
I wish that death would take me now
On this day of new-found happiness.
The Mother of Gidō Sanshi (fl. c. 980)

This poem was composed on a topic which a Japanese would probably have called "love after the first meeting"—the affair has begun and is in its happy early stages. At the same time, there is foreboding in the poem. Whether or not it was always true in life, it was conventional in the literary treatment of love in this period that the affair should gradually cool and the man become less and less attentive and regular in his visits. At the Japanese court, where polygamy was the norm, it was the common thing for a court noble to keep two or three mistresses or concubines in addition to his consort or lawful wife, and he might also carry on a number of secret affairs as well. These secondary relationships might be formed and broken by the man with more or less casualness, and it was of course the woman who suffered most from the consequences of such an affair. The customs of the day demanded that high-born ladies live in guarded palaces, hidden from the eyes of all men except their husbands or enterprising lovers, and they could seldom leave their cloistered apartments. A woman might be abandoned at any time by her fickle lover, and while he might move on to new and fairer flowers, she must continue her secluded and now empty life until perhaps discovered by some other man, whereupon a new affair might begin its inevitable course.

Such, at least, was the conventional treatment of love in Japanese literature, and it is therefore natural that it should be reflected in the dramatic progression of the love poems in the Shinkokinshū. The last half of Book XIV and all of Book XV in fact are devoted to poems which express the woman's suffering as her lover's visits become less frequent. Book XV begins with the following famous poem.

SKKS XV: 1336

 … The white sleeves covering us,
Glistening with dew and sparkling with our
 tears,
 Are parted by the dawn,
And as we dress, shake in the autumn wind
Which blows its pale color through our
 hearts.⁸
Fujiwara Teika (1162-1241)

We may assume that the parting treated in this poem involves a temporary separation, and that the lover will return. At the same time, the autumn wind is a conventional symbol in Chinese and Japanese poetry for the death of love, and the chill which it blows into the heart is a foreboding of the doom of eternal separation. This theme is developed through Book XV with increasing intensity, and the last poem is an expression of bitter resentment and despair at the lover's infidelity.

The arrangement of the love poems in the Shinkokinshū thus follows dramatically the progression of a typical love affair from the first glimpse of the beloved through the successive stages of a passionate courtship, marriage or liaison, disenchantment, separation, and final despair and loneliness. This provides a clearly defined plot structure which would be inappropriate or impossible for poems on subjects other than love. There are types of association and progression in the poems on travel in Book X, for example, which follow quite different principles, and it is to these poems that we now turn.

Although the poems in Book X do not appear to be arranged according to any common topical element, this is not by any means to say that the arrangement is haphazard, for instead of the time progression of the seasonal poems or the plot development of the love poems, the sequence of the poems on travel appears to have been designed to show—incredible as it may seem—the historical development of Japanese poetry through the centuries. That is to say, close study reveals that the poems are arranged in several large groups which represent in chronological sequence the four major periods in classical poetry down to the age of the Shinkokinshū. Specifically, of the total of ninety-four poems in Book X, the first six (SKKS X: 896-901) are by poets of the so-called "Man'yō period" (c. A. D. 500-750), the early great age of literary Japanese poetry. Following this group is a single anonymous verse which might be called transitional in that it is in a style characteristic of the late eighth and early ninth centuries, the age between the two great periods of the Man'yōōshu on the one hand and the Kokinshū on the other. The next group of five poems (SKKS X: 903-907) is from the age of the Kokinshū proper, and this is followed by a sequence of twenty-four poems which range in date from the mid-tenth to the late eleventh centuries; in other words, they belong to the period spanned by six imperial anthologies, beginning with the second anthology, the Gosenshū (c. 951), and ending with the seventh, the Senzaishū (c. 1188). The last fifty-eight poems (SKKS X: 932-989) are contemporary, which is to say they are by poets of the late twelfth and early thirteenth centuries, the age of the Shinkokinshū itself.

There is no doubt that the audience of the Shinkokinshū period who read through this volume was aware of this chronological sequence by literary periods, and that this sequence contributed to their pleasure. But there are also in this volume other techniques of association and progression which involve topics, treatment, and imagery. The twelve poems with which the volume begins illustrate these more detailed techniques.

SKKS X: 896

 … If I abandon my village
Of Asuka, where birds are said to soar,
 For a new capital,
Will it be that I shall nevermore
See you present by my side?
Empress Gemmyō (661-721)

SKKS X: 897

  … Longing for my love,
I gaze forth across the pines
  Of Waka's forest,
And see that over the strand of Shiohi
The cranes fly off with mournful cries.
Emperor Shōmu (701-756)

SKKS X: 898

  … Come along, lads!
Let us quickly make our way home
  Towards the Rising Sun—
The pines on the shore at Mitsu in ōtomo
Must wait impatiently for our return.
Yamanoe Okura (?660-?733)

SKKS X: 899

 … As I come rowing
Over the long sea-path from wilds
 Distant as the sky,
Through the Straits of Akashi
The Isles of Yamato come into view.
Kakinomoto Hitomaro (fl. c. 680-700)

SKKS X: 900

  … Because I come
From parting with the wife I love,
  The leaves of rough bamboo
Seem to fill these mountain depths
With their mournful rustling sound.⁹
Kakinomoto Hitomaro

SKKS X: 901

  … Having come this far,
I ask, "Where is Tsukushi now?"
  It seems to lie
Back to the West beyond those hills
Where the white clouds trail away.
Ōtomo Tabito (665-731)

SKKS X: 902

 … Do you walk
Alone along the narrow path
 Across the peaks,
Your robe not yet dried out
From its drenching in the morning mists?
Anonymous

SKKS: X: 903

 … How can the smoke
That rises from the Peak of Asama
 In this country of Shinano
Fail to strike the people far and near
With amazement at the sight?
Ariwara Narihira (825-880)

SKKS X: 904

 … Not in a reality
As real as these hills of Utsu
 That rise in Suruga,
Nor even in the unreal world of dreams,
Can I meet face to face with her I love.¹⁰
Ariwara Narihira

SKKS X: 905

 … The evening wind
Binds with cold the roadside grass
  That pillows me—
If only I might ask for shelter in that house
That sounds with mallets fulling cloth.
Ki no Tsurayuki (d.?945)

SKKS X: 906

 … I hope this day
That I shall cross the distant bridge
 That hangs across
The gulf between those mountain peaks
Now draped in streamers of white clouds.
Ki no Tsurayuki

SKKS X: 907

 … Am I then doomed
To pass my life in low estate,
 While you rise high
To lofty peaks concealed by clouds
Like Mount Saya on my Eastern way?
Mibu no Tadamine (fl. c. 877-922)

Although it may not be immediately apparent, the first six of these poems share a common topic which is more specific than mere "travel"—a topic which might be paraphrased as "longing for absent loved ones." At the same time, these six poems are linked still more closely through associations of images from poem to poem, just as in the spring poems which we have already examined. To begin with, the phrase "soaring bird" in the first poem is a pillow-word associated in this case with the place name Asuka. So dulled had this and other such phrases become through long usage that by the thirteenth century they had lost most of their imagistic freshness, and the technique of the pillow-word appears in the poetic practice of the time largely as a neoclassical device for creating a heightened solemnity of tone. But such an expression, no matter how dulled by familiarity, is potentially an image; and when we read the second poem, with its image of cranes flying across the bay, the "soaring bird" of the first poem springs to renewed poetic life, and at the same time provides the association that links the two poems.

The image of the pine forest in the second poem leads us smoothly to the pine-fringed beach in the third, and, further, the movement suggested by the "soaring bird" in the first, and the flight of the cranes in the second poem, provides the impetus to "carry" the reader across the sea to China, which is the setting of the third poem. We should note, too, that in the first three poems we have progressed from an inland setting to the seacoast, and then across the sea. Our direction is at the same time reversed, however, for the movement back to the shores of Japan already begins in this third poem. The overt link between the third and fourth poems is the appearance in each of a poetic term for Japan—the Rising Sun in the third, and Yamato in the fourth—but there is also the imagery of a ship and the sea which, while only implied in the situation of the third poem, is raised in the fourth with its image of rowing. In addition, such images as "strand" (kata), "beach" (hama) and "straits" (to) in the second, third and fourth poems are related in terms of what might be called a category of "sea phenomena."¹¹

The change in the situation of the speaker from sea to land is foreshadowed in the fourth poem, with its image of the "isles of Yamato." We may imagine that the speaker of this poem sees from his boat the outlines of the mountains of Japan against the sky, and that there is in this poem submerged or implied imagery of mountains. This provides a link with the fifth poem, where mountains are specifically mentioned. The mountain image of course recurs in the sixth poem, and provides the association between this and the preceding one.

The next seven poems (SKKS X: 901-907) are all obviously associated through the imagery of mountains, except for the tenth, which to the Western reader must appear to lack the requisite imagery. However, if the Japanese text is compared with that of the preceding poem (SKKS X: 904), it will be seen that the verb utsu "to full cloth" in the tenth poem is homophonous with the place name Utsu, the mountainous region of Suruga province mentioned in the ninth poem. In the context of the arrangement of poems in the Shinkokinshū, therefore, the verb utsu in the tenth poem is a "pivot word" (kakekotoba): it carries two meanings, one of which is applicable to the part of the poem which precedes it, and the second to the part that follows. Consequently, the house in the tenth verse is really "that house in Utsu that sounds with mallets fulling cloth," and the place name thus suggested also raises the image of the mountains with which the place is associated.

There are also other, less sustained patterns of images by which the association from poem to poem is strengthened. The image of clouds (kumo) in the sixth, and mist (kiri) in the seventh poem are words which were traditionally associated by the Japanese poets as members of the same "category of phenomena." This kind of association was particularly important in the technique and practice of the renga, or "linked verse," which began to flourish some hundred years or so after the Shinkokinshū was compiled, and names came to be given by the renga poets to the different categories which they recognized. Therefore, to borrow a term from the technical vocabulary of this later poetic genre, the images of clouds, dew, and the like are "rising phenomena" (sobikimono)—moisture or vapor which rises into the air and dissipates itself. Once this category is recognized, the association between the "mist" of the seventh poem and the "smoke" (keburi) of the eighth can be accepted, since smoke, like clouds and mist, is a "rising phenomenon."

There is a further link between the seventh and eighth poems. This association is not immediately apparent because it involves a conventional symbolic value attached to a given expression but one which had not been evoked by the poet for the immediate purposes of this particular poem. Specifically, the image of the wet robe (nurenishi koromo) in the seventh poem is frequently used as a conventional symbol for the indignation of a lover falsely accused of infidelity. Such is not its function in this poem, but looked at in connection with the verb mitogamu "to regard with amazement" in the eighth poem, its symbolic meaning is inevitably raised for an instant, because the verb mitogamu is a traditional engo or word association for the "wet robe" in its metaphorical sense. This kind of association is quite different from that of the related images of mountains, clouds, fog and the like which we find in this group of poems, for the symbol of the indignant lover has no function either in the poem in which it is used or in the developing pattern of associations in this sequence of poems. As a purely mechanical device for linking the two poems, however, it must be recognized, and although we moderns may find such a technique rather forced, the audience of the thirteenth century probably felt it to be very clever and interesting.

The eighth and ninth poems share the central term, "people/person" (hito), although the association is not close because of the great difference in relationship between the people mentioned and the speaker of each poem; the "people" in the eighth poem are strangers, whereas the "person" in the ninth is the speaker's beloved. A much closer association is found in the use in each verse of a famous mountain—the Peak of Asama in the eighth and the hills of Utsu in the ninth poem. These and other names of famous mountains, rivers, and the like came through tradition to possess a quasi-imagistic status, and were bound to evoke in the minds of poet and audience similar associations of romantic beauty. It will be noted further that in each of these poems the name of a province is mentioned—Shinano in the first, and Suruga in the second—and this provides an additional association of related categories.

The ninth and tenth poems are associated, as we have said, through the device of the pivot-word on Utsu/utsu; there is an additional link in the conventional word-association of "dream" (yume) in the ninth and "pillow" (makura) in the tenth. The same device links the tenth and eleventh verses, where the word-association is found in "wind" (kaze) and "streamers" (tanabiku, lit., "to trail"). The association between this latter pair of poems, however, is one of total situations as well as discrete elements, and there is a progression from the one to the other in this respect. The time of the tenth poem is evening, and the speaker has lain down upon the ground to spend the cold night under the open sky; in the eleventh poem, it is early morning, and by accepting the speaker as the same person in both cases, we imagine the traveler rising from his bed of grass and gazing out across the mountains in contemplation of the day's journey that lies ahead.

The images of clouds and mountains provide the principal points of detailed association between the eleventh and twelfth poems. It should be pointed out finally, however, that the ninth through the twelfth poems are associated in terms of shared rhetorical techniques as well as those other elements which have been discussed. That is to say, the first two lines in the ninth poem are a "preface" (Jo or jokotoba) for the word utsutsu "in reality," and serve to anticipate this word through the identity of sound in the first two syllables of the latter and in the place name, "Utsu," in the "preface." The tenth poem is related on the same basis of rhetorical technique. Kusamakura, "pillow of grass," suggests the latent meaning in the yū ("to bind up"—the grass for a pillow) of yūkaze, "evening wind."

The third line of the eleventh poem is a pillow-word, which is similar in technique and feeling to the preface, though more conventional and usually shorter. Like the preface, it is characteristic of early Japanese poetry, and in the context of this sequence of poems it sustains both the romantic atmosphere of association with a bygone age and the effect of unity achieved through the use of similar rhetorical devices in successive poems. We discover another preface in the first two lines of the twelfth poem, where the word sayaka in the third line is "introduced" through the identity of sound in its first two syllables and in the place name "Saya."

The foregoing analyses of sequences from the seasonal, love, and travel poems will serve to show that although the overall unifying principles of progression may differ in the individual books of the Shinkokinshū, depending upon the subjects of the poems and the appropriateness to these of different types of oiganization, the association in detail of images and rhetorical techniques is a constant principle which governs the choice and arrangement of all the poems in the anthology. The labor involved in such a careful and painstaking attempt to achieve an overall unity of structure and harmony of detail among disparate elements must have been prodigious. Such effort would not have been expended unless association and progression answered to the desire of the age for techniques which would create long lyric sequences from the individual poems written by the poets of the age or inherited from a valued literary tradition. The problem of the motivation behind this desire can best be solved by showing how this literary principle was developed and gradually refined in the earlier imperial anthologies before the Shinkokinshū, and by tracing it to its ultimate origins.

III Integration in the Imperial Anthologies Preceding the Shinkokinshū

With its status as the first and most influential of the imperial anthologies, the Kokinshū is clearly the place to begin. To facilitate comparison, we shall again begin by quoting the first few poems from the first book—"Spring"—to see how it exemplifies association and progression.

KKS I: 1

 … The Old Year not yet gone,
The longed for spring has come at last
  Yet brought confusion—
For are we now to say "last year,"
Or should we rather say "this year?"
Ariwara Motokata (888-953)

KKS I: 2

 … Will the wind
That gently blows on this first day of spring
 Melt perhaps the ice
To which was changed the water of the stream
That wet my sleeves in summer when I drank?
Ki no Tsurayuki (d.?945)

KKS I: 3

 … Where does it rise,
The haze that is the sign of spring?
 In lovely Yoshino,
Yes, here upon the hills of Yoshino,
The winter snows still fall.
Anonymous

KKS I: 4

 … Amidst the snow
The long-awaited spring has come—
 Will they melt today,

Those tears shed by the warbler crying
And turned to ice in winter's cold?
Empress Takako (fl. 858-882)

KKS I: 5

 … Although the warbler,
Who is the harbinger of spring,
 Has already come
And perches singing in the plum tree
Amongst the branches, the snow still falls.
Anonymous

As these first poems indicate, the seasonal poems in the Kokinshū set the pattern for a progressive topical development based on the passage of time. In the first two books of the anthology, the spring topics follow one another in harmony with external nature from the beginning of the season (poems 1-16), through "picking young shoots" (17-22); "new growth in the fields and hills" (23-27); "spring birds" (28-31); "plum blossoms" (32-48); "blossoming cherries" (49-68); "falling cherry blossoms" (69-118); "wisteria in bloom" (119-120); "blooming of the yellow mountain rose" (121-125); and the "passing of spring" (126-134). The same kind of arrangement is also to be found in the poems on summer in Book III, autumn in books IV and V, and winter in Book VI.

An examination of the love poems, which are found in Books XI through XV of the Kokinshū, also yields similar results: the poems are arranged by similar topics in accordance with the development of an affair from "love yet undeclared" (469-551), through such phases as courtship (552-615), love after the first meeting (616-704), the lover's growing coolness (705-746), and the ending of the affair in bitterness and misery (747-828). It would be possible to divide these larger topical categories into several lesser ones, but what has been said is probably sufficient to show that the love poems in the Kokinshū are arranged in a kind of dramatic plot structure very much like that later realized in the Shinkokinshū.

At the same time, although there appears on the surface to be no marked difference between the Kokinshū and the Shinkokinshū, at least as far as the seasonal and love poems are concerned, there are two important points to be made. The first is that a smooth association and transition from poem to poem in terms of images, rhetoric, and the like is not nearly so carefully contrived in the Kokinshū One example of this lack of attention to such matters is to be found in the first two in the series of spring poems which have just been quoted. Because these poems are both on the same topic—the arrival of spring—they can hardly fail to have at least that much in common, and we therefore cannot say that they are completely unrelated. Nevertheless, if it had been the intention of the compliers to provide a really smooth and harmonious shift from the first poem to the second, with its images of "sleeves," "water," "wind" and the like, it would have been possible to find without great difficulty a more suitable poem to use in place of the first one. The fact that this was not done leads us to conclude that the compilers of the Kokinshū were not in this case very deeply concerned with a harmonious association of images from the first poem to the second. The same observation holds true for the relationship between the second and third poems, and could be made repeatedly concerning given sequences of poems throughout the anthology. On the other hand, the succession of images from the third through the fifth poems ("haze," "mountains," and "snow" in poem 3; "snow," "warbler," and "plum" in poem 5) is relatively smooth, so that we cannot say that no effort at all was made to produce this effect of harmony. In other words, although a progression in terms of a logical sequence of topics was consciously carried out in the Kokinshū, the association of images from poem to poem was attempted only spasmodically, and no effort was made to give unity to entire books of poems through the consistent application of this principle.

The second point that must be brought out is that even development in a progressive sequence according to poetic topics is actually to be found only in the seasonal poems (Books I-VI) and the love poems (Books XI-XV) in the Kokinshū In the other volumes of the anthology, whether for example those devoted to congratulatory poems (Book VII), those to poems on parting (Book VIII), or those to travel poems (Book IV), there is no evidence at all of any attempt to arrange the poems according to any set pattern of topical development. The reason for the relative lack of attention given to these books by the compilers may perhaps be that they were considered of secondary importance. It should be pointed out in this connection that the arrangement of the twenty books of the Kokinshu reflects a clear distinction between different grades of what may be called "formal" and "informal" poetry. Formal poetry was intended for the eyes or the ears of a relatively large audience, and required a greater degree of technical polish; informal poetry was, ostensibly at least, a mode of private communication between a poet and his mistress or friends, and might be written with less complex techniques. In the Kokinsū, the first ten books are devoted to formal, and the second ten to informal poetry; and within these two major categories the seasonal poems appear to have been considered the most important variety of formal, and the love poems the most important variety of informal verse. Less care was expended for the arrangement of poems on other subjects and no attempt was made to arrange them according to a topical progression or time sequence.¹² It is true that a few scattered groups of poems in these "secondary volumes" show evidence of some attempt to achieve an associational progression of related images, but it is clear that no great store was set by the result. In other words, the principles of association and progression so carefully followed throughout the Shin-kokinshū are only partially and rather carelessly applied in the Kokinshū. At the same time, however, the earliest attempt to apply these principles to an anthology of Japanese poetry, albeit in a rather rudimentary fashion, can be traced to this first imperial collection.

A study of the topical arrangement of poems in the six imperial anthologies that fall between the Kokinshū and the Shinkokinshū reveals a situation very similar to that which we have found in the Kokinshū. That is, a topical progression according to the passage of time is limited to the seasonal and love poems in these collections, and no attempt is made to give any such overall unity to the other volumes or the anthology as a whole. On the other hand, we observe an increasing concern with the problem of achieving a smooth transition from poem to poem through associations of images, rhetorical techniques, and the like. Such an attempt, while by no means so overriding a consideration as it was to become in the Shinkokinshū, is made with increasing seriousness and consistency in each successive anthology—the later the collection, the more conspicuous this phenomenon becomes. The first six poems in the Senzaishū (1188), the seventh imperial anthology and immediate predecessor of the Shinkokinshū will illustrate the point.

SZS I: 1

  … As I gaze far out,
I see that spring has come this morning,
  For today the haze
Begins to rise across the moor
Of Ashita and its morning fields.
Minamoto Shunrai (d. 1129)

SZS I: 2

 … At Mount Mimuro
Has spring now at last arrived
 Deep in the valleys?
The water melting underneath the snow
Taps impatiently against the rocks.
Minamoto Kuninobu (1069-1111)

SZS I: 3

 … Even to Yoshino
The spring has come where winter drifts
 Blocked up the path
Leading steeply to my mountain village,
And no visitor's footprint marked the snow.
Lady Taiken Mon in Horikawa (fl. c. 1130-1145)

SZS I: 4

 … I hated them,
For they would block the visitor's path
  To my mountain village,
But still their melting fills me with regret,
These snowdrifts of the year just past.
Ōe no Masafusa (1041-1111)

SZS I: 5

 … Now that spring begins,
From underneath the snowdrifts
 Streams of water flow,
And the warbler in the valley
This moment breaks into his song.
Fujiwara Akitsuna (1029-1103)

SZS I: 6

 … Is it because
The warbler knows that spring has come
 To this mountain village,
That even before the haze appears
He sings atop my brushwood fence?
Minamoto Takakuni (1004-1077)

Topically, these poems seem to follow a regular progression from what might be called the "onset of spring" into "early spring," and the association of images seems to develop smoothly from poem to poem as the setting shifts from the plains to the mountains and from poem country to an isolated village. On closer examination, however, it may strike us that the position of the fifth poem is somehow wrong. First, the subject of this poem is the first day of spring, whereas in the fourth poem we have already moved beyond this point in the time progression. Second, and more important than this, the description of streams of water from melting snow in mountain valleys is very similar in the second and fifth poems. Therefore from both considerations—time sequence of topics, and close association of imagery—the most logical position for poem 5 would seem to be between the second and third poems. On the other hand, we can find some justification for the placing of this fifth poem in the order to which the compiler assigned it: its imagery of mountains and snow associates it with poem 4, and it introduces the warbler, which becomes the principal image in poem 6. No doubt it was these considerations which led to placing this poem in the position in which we find it today, but it must still be admitted that the topical sequence and progression of images is somewhat rough at this point.

In a situation of this kind, the compilers of the Shinkokinshū would have laid primary emphasis upon smoothness of association and progression, and would even have chosen a less "outstanding" poem in place of this fifth one if thereby a more felicitous progression could have been effected. However, when the Senzaishū was compiled, the principles of association and progression had not yet been accepted by the poetic elite to be more important even than high quality in each individual poem, and the compiler would not have felt free to reject a good poem simply because another one, less good, would make for a smoother sequence of images. Thus, although the same kinds of principles are followed in the arrangement of poems in both the Senzaishū and the Shinkokinshū, these principles are applied with much greater consistency and thoroughness in the Shinkokinshū; in many instances they outweigh even those considerations of high individual quality which down to this time had been the principal standard for the inclusion of poems in an imperial anthology.

IV. The Origins and Development of Association and Progression

Since, as we have seen, the principles of association and progression are found as early as the first imperial anthology of Japanese poetry, the Kokinshū, we may ask whether the compilers of this collection did not themselves borrow these principles from some earlier poetic source or develop them by analogy with other art forms. All available evidence suggests that the Kokinshū was indeed the first such anthology, for neither in the Man yōōshū nor in the anthologies of China do we find any such arrangement of poems. We do discover in the field of painting, however, conventions which are very similar in effect. By this we mean the techniques employed in the horizontal picture scroll (emakimono), in which a given subject is developed continuously from scene to scene in chronological sequence. Sometimes, to be sure, narrative or descriptive passages were inserted at various points to explain the incidents depicted in the scrolls, but frequently the events were portrayed entirely through a continuous series of pictures. In either case, one of the outstanding characteristics of this art form was the continuity of the total sequence. Even though the flow might be sometimes broken by written passages, most of the individual scenes were not set off by frames or other devices, but blended into one continuous linear progression; clouds, mountains, or other natural scenery and even scattered human figures provided a unifying thread connecting the more concentrated scenes of separate incidents in the narrative.

The similarity between the chronological development of scenes in the picture scrolls and the temporal progression in the Kokinshū and other imperial anthologies is obvious, but a further analogy suggests itself between the continuous flow of scenery and figures in the scrolls and the association of images and rhetorical devices that link the separate poems in the anthologies. It is tempting to conjecture that a familiarity with the conventions of the picture scroll may have suggested to the compilers of the Kokinshū the desirability of applying these techniques to an anthology of poems. The great age of the picture scrolls, especially those which treat secular subjects, extended from the eleventh through the sixteenth century, however; and since almost no examples survive from an earlier period, we do not know the extent to which such scrolls may have existed and been appreciated by the court nobility of the early tenth century, when the Kokinshū was compiled. At the same time, one famous example, the Kako Genzai Ingakyō, or "Sutra on Causality Between Past and Present," has come down from the eighth century. This religious scroll is different from the later genre of the emakimono in that it is divided horizontally into a continuous written text along the bottom half and a continuous series of pictures illustrating episodes from the text along the top. However, although it is regarded as a somewhat crude ancestor of the great lay scrolls of subsequent centuries, this illustrated sutra shows in its graphic portion the basic parrative conventions of the form: a given number of episodes in the legendary life of the Buddha depicted in a continuous series of pictures in chronological sequence. There is little doubt that the secular emakimono developed from such illustrated religious texts as this, and even though the first lay scrolls may have been painted somewhat later than the Kokinshū age, numerous illustrated Buddhist scriptures, no longer extant, were undoubtedly available to the aristocracy of the time. Through familiarity with such scrolls, the compilers of the Kokinshū may have become accustomed to the convention of narration through a chronological sequence of tableaux which, though separate, are depicted in a single continuum.

Whether or not the poets of the tenth century derived the notions of progression and association from the emakimono, their application of these techniques to a series of poems was quite original and unprecedented. Furthermore, as we have seen, it was primarily the techniques of progression through time that were used in the arrangement of the seasonal poems and the love poems in the Kokinshū—the various techniques of association through images and rhetoric had apparently not yet been developed to the point of consistent application. Consequently, the consistency as well as the increasing variety and complexity with which the techniques of progression are combined with techniques of association in the Shinkokinshū leads us to look to other sources outside the successive imperial anthologies for elements that contribute to this process of refinement.

It is probable that the increasing attention given to the linking of successive poems in the imperial anthologies through image and other associations reflects a growing concern with such matters in other kinds of sequences than those of the great collections. The genre which immediately suggests itself is the hyakushuuta, or "hundred-poem sequence," a series of tanka composed on a given number of topics by an individual poet. This genre was a formal one, in that it was intended to be read and appreciated by an audience of peers, and it shows in its most usual form the influence of the imperial anthologies in the sequence and kind of topics. But the important characteristic of the hundred-poem sequence was that it was composed and judged as a single artistic unit: praise or criticism was accorded a given sequence not on the basis of the merits of the individual poems, but in terms of the overall effect of harmony, beauty, variety, and smoothness conveyed by the sequence as a whole.

The practice of composing hundred-poem sequences certainly existed by the middle of the tenth century, and we find in the personal collections of the poet Minamoto Shigeyuki (fl. c. 970-1000) and Sone no Yoshitada (fl. c. 985) examples of the genre.¹³ However it was not until the twelfth and thirteenth centuries that the hundred-poem sequence became really popular among the aristocracy, and the beginning of this later vogue may be traced to the so-called Horikawa Hyakushu, or "Hundred-Poem Sequences Submitted by Command of the Ex-Emperor Horikawa," which were composed between 1099 and 1103 by sixteen of the most prominent poets.¹⁴

The first poet to apply the techniques of association of images in the hundred-poem sequence may well have been Sone no Yoshitada, for in both of his two surviving sequences we find such techniques. Yoshitada was, however, a poet who was unappreciated and even scorned in his own day, and perhaps because the other poets of his time had not yet come to appreciate the esthetic possibilities of association, we do not find it employed in other hundred-poem sequences until the end of the twelfth century. The reappearance of the technique in this later period was probably not due to a "discovery" and imitation of Yoshitada's technique, but rather to the elaboration of topics and categories in certain kinds of sequences.

The most usual kind of hundred-poem sequence was a kind of miniature imperial anthology in its arrangement and distribution of topics: ordinarily it began with twenty poems on spring, followed by ten on summer, twenty on autumn, ten on winter, twenty on love, and twenty on so-called miscellaneous topics. Towards the end of the twelfth century, however, sets of poems with much more detailed topical divisions began to be composed. We find, for example, a sequence composed in 1200 by the Ex-Emperor Go-Toba (1180-1239) which consists of five poems each on the following specific rather than the usual more general topics: haze, the warbler, cherry blossoms (spring); the hototogisu, early summer rain (summer); flowers and plants, the moon, autumn foliage (autumn); snow, ice (winter); Shintō, Buddhism, dawn, dusk, mountain roads, the seaside, the Imperial Palace, entertainments, annual observances, and felicitations (miscellaneous).¹⁵ All of these specific topics are, it will be noted, either images in themselves, or require, as in the case of "annual observances," that a common imagery of setting be conventionally used in their treatment. A natural and almost inevitable result is that every poem in each group of five would be associated with the other members of the group through the same or related images. Such sequences as this appear to have enjoyed a kind of vogue from the end of the twelfth through the beginning of the thirteenth century, and many examples of them have been preserved.¹⁶ Although they represent only one of several possible methods of organizing and classifying the necessary hundred poems, it may still be conjectured that they influenced other varieties, and that the result was a re-emergence of those techniques of imagistic association that were at first perhaps unconsciously, but later consciously, employed.

Such an assumption is borne out by a study of the hundred-poem sequences of the late twelfth and thirteenth centuries classified according to the usual more general categories of spring, summer, autumn, and so on. For although practice varies with individual poets, we find that in most cases the same poet has composed some sequences in which no attempt at association is made and others in which the poems are linked through deliberate associations of imagery.¹⁷ There is, however, one significant exception—the work of the Ex-Emperor Go-Toba. In all of his extant sequences, each successive poem is linked with the preceding one through association of images.¹⁸ This fact seems of particular importance because of Go-Toba's relation to the Shinkokinshū: although this anthology was nominally compiled by a group of five courtiers headed by Fujiwara Teika, their function was in reality only that of assistants or advisers to Go-Toba, and it was the Ex-Emperor himself who was the chief compiler and had the final say in the selection or rejection of poems. Therefore we may conclude that the application of techniques of association in such thorough fashion in the Shinkokinshū was a reflection of Go-Toba's own taste and preference for these techniques.¹⁹

In addition to the techniques of association, there was another feature of the hundred-poem sequences which probably exerted a considerable influence upon the way in which the progression was handled in the imperial anthologies, and again particularly in the Shinkokinshū. Because the hundred-poem sequence was intended to be appreciated as a single artistic whole, the overall effect of harmony and balance, variety, and contrast was therefore of greatest importance. In producing the desired impression, a conscious effort was made to vary the pace and avoid monotony within the progression by creating a certain number of high and low points. The high points were individual poems which were striking or remarkable for technical or other reasons, and the effect of such poems might be considered to last longer in the minds of the audience if they were placed next to more mediocre poems which would create no strong impression. In other words, the poet would deliberately include a certain number of bland or "easy" poems at crucial points in his sequence so as to enhance the effect of the more interesting ones and create a general impression of sinuous, undulating flow. By analogy with a piece of woven material, the "easy" verses were called ji no uta or "background poems," and the more striking ones were known as mon no uta or "design poems": just as the effect of beauty in a piece of material is made more striking when a pattern is contrasted against a plain or neutral background, so with a sequence of poems.

Just when this esthetic principle began to be consciously applied to the hundred-poem sequence is not certain, but we find the term ji no uta, or "background poem," employed without definition in one of the poetical treatises of the Ex-Emperor Go-Toba. This is probably some indication that the principle had been current for some time, and that the terms "design poem" and "background poem" were assumed by Go-Toba to be meaningful to his readers. Therefore, although Go-Toba's treatise was presumably written some years after the Shinkokinshū was compiled, it is perfectly reasonable to assume that the principle had been known and more or less generally accepted before this anthology was completed.²⁰ At any rate, such an assumption is helpful in explaining the apparent anomaly of finding in the Shinkokinshū a number of poems which could only have been considered bland or mediocre by the prevailing standards of the age. In other words, the inclusion of a number of such poems in this collection was not simply the result of a dearth of suitable materials, nor was it simply an inevitable result of placing paramount importance on the achievement of a smooth transition from poem to poem. On the contrary, the technique appears to have been deliberately used in order to create variety, contrast, and alterations of pace within the progression, and, on the model of the hundred-poem sequences, to apply the principle of "design" and "background" even at the expense of uniformly high quality in an imperial anthology.

Unfortunately, there are no extant treatises on poetics in which the theory of design and background in sequences of Japanese poems is discussed, and there remains the question of how such a concept came to suggest itself to the poets of the twelfth and thirteenth centuries and why it was accepted as a valid esthetic principle. The answer probably lies in part at least in a natural response to the problem of monotony or sameness inevitably posed by a series of tanka intended to be read as a single unit. On the other hand, it may be suggested that a considerable influence in shaping the principle was exerted by the theory and practice of a very different literary genre, the formal prose essay. This esteemed prose genre was highly complex, first because all formal essays and official documents had to be composed in Chinese, and secondly because the proper style of composition required the mastery of numerous rules and techniques. Nevertheless, like the French of English courts of law or the Latin of the humanists, the Chinese language was the only medium considered proper for official documents, both secular and religious, from the eighth to the middle of the nineteenth century in Japan; and this difficult medium had to be learned by court officials and priests—in other words, by the same group who composed hundred-poem sequences and compiled imperial anthologies.

During the four hundred years from the tenth to the thirteenth centuries, it was the official written style of the T'ang dynasty (618-907) in China which was followed as the model for Japanese state papers. This style was particularly complex, and both in China and Japan a number of handbooks were compiled as aids to composition and study. Though none of these has survived in China, at least one Japanese copy of a late ninth or early tenth-century Chinese rule book exists, and a number of manuals written by Japanese were in existence by the twelfth century, of which the Sakumon Daitai, or Essentials for Composing Formal Prose, by Nakamikado Munetada (1061-1141) is a typical example.²¹ The most striking aspect of the practice of the formal essay, and that which seems to relate it to the integration of poetic sequences, was the combination of two distinct kinds of style in a single essay. The first of these consisted of a rhetorically ornate, rhymed, parallel prose, which was reserved for the most important sections of the essay, while the second style, used in the less important passages, was what might be called ordinary prose, in that meter, rhyme, antithesis, and the like were not used. Both styles were employed alternately in the same essay, and a conscious effort was made to distribute the passages in ordinary prose in such a way as to enhance the impression created by the more elegant and ornate sections.

Obviously this practice of alternation of styles resembles very closely the effect of background and design aimed at in the hundred-poem sequences and finally in the imperial anthologies, and it is quite possible that in the techniques of the formal essay we have the original concept which was gradually adapted to the poetic sequences. It should also be pointed out that the Japanese literati became more than ever scrupulous about strict adherence to the rules of formal essay-writing just at the time—around the beginning of the twelfth century—when the hundred-poem sequences began to come into vogue. The lack of a written poetics for the sequences makes it impossible to state categorically that the inclusion of mediocre verses in the earlier ones was not due simply to lack of talent on the part of the poet. However, the historical facts as they have been described indicate that this is highly unlikely, and it would appear, rather, that the most probable historical development was an introduction of the technique of "design" and "background" through the stylistics of the Chinese essay, followed by a conscious adaptation of this esthetic to the hundred-poem sequence, and finally its application to an imperial anthology like the Shinkokinshū.

The notion of improving the overall quality of an anthology of poems by including the bad along with the good may seem doubtful practice to the Westerner—surely the consistent application of absolute standards of quality should yield better results. But the integrated beauty of the Shinkokinshū speaks for itself, and to the Japanese of the thirteenth century, a long tradition of Buddhist teaching had made clear that absolute standards were an illusion. The central concept of the Tendai sect—the sect whose teachings were most influential among the early medieval aristocracy—was that no phenomenon exists independently in and of itself, but only in complex relation to all other phenomena. This was true of abstract values as well as of anything else; "good" and "bad" did not exist as absolutes, but only in terms of specific situations, and a "bad" act might be "good" if it proved to be a cause that brought the individual closer to final enlightenment. This does not mean that such a general philosophical outlook was a direct cause for the development of the particular literary practice we have been considering. But it helps explain how the medieval Japanese poets could accept without any sense of anomaly the notion that a "bad" poem could be "good" in a given context if its effect was to enhance the overall effect of the interrelated parts of a poetic sequence.…

Notes

…⁶ While it might be thought that the compilers of the Shinkokinshū may have had in mind as a pattern one of the similar sequences to be found in all of the six anthologies that intervene between the Kokinshū and this eighth anthology, examination shows that either the number of poems is different, or where, as in the Gosenshū, the number is the same, the images and patterns of progression differ from those in the Kokinshū and Shinkokinshū.

⁷ The Japanese esthetic ideal, modeled in considerable measure on similar Chinese concepts, came to have an importance in Japan comparable to that of the Confucian ethical ideal in China. The elegant Japanese courtier was artistically and aristocratically conceived: he was expected to have achieved a good hand with the brush, compose poetry, perform reasonably well on one or two instruments, appreciate a beautiful woman, and hold his wine like a gentleman. The hero of The Tale of Genji is an embodiment of this and other ideals. In his Japan: A Short Cultural History (rev. ed., New York, 1943), Sir George Sansom has defined the tastes of Heian courtiers and the artistic accomplishments of this "almost entirely aesthetic" culture (pp. 235-41), and in his latest work, A History of Japan to 1334 (Stanford, 1958), a very interesting chapter—aptly entitled "The Rule of Taste"—is devoted to an analysis of these aspects of Heian court life (pp. 178-96).

⁸ The reference to sleeves in this poem suggests not only tears but the custom of lovers, lying together under cover of their removed garments.

⁹ Like many of the other poems from the Nara period taken into the Shinkokinshū, this one has been altered somewhat to fit new tastes or because the readers of a later age had difficulty in parsing the Chinese characters used in varying ways in the Man'yōōshū Cf. MYS II: 133. With the next poem in the text, cf. MYS IV: 574.

¹⁰ This poem is the tenth and the preceding one the eighth poem in the Ise Monogatari, the famous collection of tales combining poems with prose contexts and attributed to Narihira. Sees the Kōchū Nihon Bungaku Taikei (Tokyo, 1936), II, 40 and 41.

¹¹ Such categories as "falling phenomena" to designate rain, snow, and the like were identified by later renga poets. But we see here a technique which implies these categories and indeed which led to the use and naming of them by renga poets. This is one of the respects in which the practice of integrating tanka influenced the formation and practice of the renga. See also the discussion which follows in the text and section V. below.

¹² Professor Konishi has discussed the arrangement of poems in the Kokinshū in his edition, Kokinwakashū, for the Shinchū Kokubungaku Sōsho, XLIV (Tokyo, 1949), 30-31. The distinction between formal and informal poetry is related to, but different from, that between public and private poetry. All public poetry (e. g., Pope's Dunciad) is formal, but not all formal poetry is public (e. g., Donne's "Good-Morrow"). All informal poetry (e. g., Swift's "bagatelles" to Sheridan and others or Dryden's verse letter to Etherege) is private, but not all private poetry (Donne's poem) is informal. The distinctions are largely those of subject, tone, technique, and esthetic distance. Poems on public and social themes tend to disappear from Japanese poetry after the Nara period, except for semi-ritualistic poems on such "congratulatory" occasions as the New Year or those addressed to Emperors on their accession. From about the ninth through the eleventh centuries, the Japanese developed and esteemed a wide range of private poetry, both formal and informal. Gradually formal poetry rises in estimation to the point of almost excluding informal poetry as a valued art; and by the latter eleventh or twelfth century, most of the poetry thought worthy of preservation in official anthologies and family collections is formal private poetry.

¹³ Shigeyuki's one sequence, composed by command of the Crown Prince Norihira (950-1011; reigned 967-969 as Emperor Reizei), is printed in Shink Gunsho Ruijū (Tokyo, 1928), XI, 507-509. Yoshitada's two sequences may be found in Kōchū Kokka Taikei (Tokyo, 1929; hereafter Kokka Taikei), XIII, 53-71.

¹⁴ Printed in Shinkō Gunsho Ruijū, VIII, 65-107.

¹⁵Kokka Taikei, X, 56-64.

¹⁶ See, for example, the seventeen sequences by six major poets of the late twelfth and early thirteenth century in Kokka Taikei, X, 507-514 (Fujiwara Shunzei); 543-550, 625-632, 655-663, 670-673, 679-686, 716-724, 746-758 (Priest Jien); and XI, 7-15, 15-20, 35-43 (Fujiwara Yoshitsune); 309-316 (Priest Saigyo); 381-388, 447-455, 660-669 (Fujiwara Teika); 754-762, 814-822 (Fujiwara letaka).

¹⁷ Two sequences by Priest Jien (1155-?1225) may serve as illustrations. In a hundred-poem sequence composed in 1187 there is no attempt to employ association, while in another, composed in 1190, association is employed to integrate the sequence. See Kokka Taikei, X, 612-625 and 670-679, respectively.

¹⁸ Go-Toba's personal collection, printed in Kokka Taikei, X, 25-180, contains three sequences of 100 poems, one sequence of 500 poems (a kind of grandiose version of the hyakushuuta), and seven sequences of thirty poems (an abbreviated version of the hyakushuuta).

¹⁹ For a detailed discussion of Go-Toba's dominant role in the compilation of the Shinkokinshū, see Kojima Yoshio, Shinkokinwakashū no Kenkyū, II (Tokyo, 1946), 1-48.

²⁰ See the Go-Toba-In Kuden in Sasaki Nobutsuna, ed., Niholn Kagaku Taikei, III (Tokyo, 1941), 3. Go-Toba uses the term in criticizing the hundred-poem sequences of Fujiwara Yoshitsune (1169-1206), which, he says, give the impression of containing too few "background poems." Since Yoshitsune is referred to in this passage as the "late Regent," we assume that the treatise was written some years after his death, which, it will be noted, took place the year after the Shinkokinshū was first compiled.

²¹ An early thirteenth-century Japanese manuscript of a Chinese rule book entitled Fu P 'u, or Rules of Composition, is in the private collection of Mr. Goto Keita of Tokyo. Two unpublished manuscripts of the Sakumon Daitai—not to be confused with the fourteenth-century work of the same title in Shinkō Gunsho Ruijū, VI, 488-504—are in the Higashiyama Library of the Imperial Palace at Kyoto.…