UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Ariake no wakare : genre, gender, and genealogy in a late 12th century monogatari Khan, Robert Omar 1998

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


ubc_1998-345610.pdf [ 21.65MB ]
JSON: 1.0099323.json
JSON-LD: 1.0099323+ld.json
RDF/XML (Pretty): 1.0099323.xml
RDF/JSON: 1.0099323+rdf.json
Turtle: 1.0099323+rdf-turtle.txt
N-Triples: 1.0099323+rdf-ntriples.txt
Original Record: 1.0099323 +original-record.json
Full Text

Full Text

ARIAKE NO  WAKARE:  GENRE, GENDER, AND GENEALOGY IN A LATE 12TH CENTURY MONOGATARI by ROBERT OMAR K H A N B . A . , Oxford University 1977 Diploma of Comparative Philology, Oxford University, 1980 M A . , The University of Texas at Austin, 1990 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF  PHILOSOPHY in  T H E F A C U L T Y OF G R A D U A T E STUDIES D E P A R T M E N T OF A S I A N STUDIES We accept this thesis as conforming to the required standard  T H E U N I V E R S I T Y OF BRITISH C O L U M B I A August 1998 © Robert Omar Khan, 1998  In presenting  this  degree at the  thesis  in partial fulfilment  University of  freely available for reference copying  of  department  this or  British Columbia, and study.  by  his  or  her  the  I agree  requirements that the  I further agree  thesis for scholarly purposes  publication of this  representatives.  may be It  is  thesis for financial gain shall not be  permission.  Department The University of British Columbia Vancouver, Canada  DE-6 (2/88)  of  that  for  an  advanced  Library shall make it  permission for extensive  granted  by the  understood  that  allowed without  head  of  my  copying  or  my written  Ariake no Wakare Genre, Gender and Genealogy in a late 12th century Monogatari Abstract Ariake no Wakare was thought to be a lost tale, but its unique manuscript was rediscovered in the early 1950s. Thirteenth-century references and internal evidence suggest a date of composition in the 1190s by an author in Teika's circle, and attest to Ariake's prominence in the thirteenth-century prose fiction canon. Thematically, it is virtually a 'summa' of previous monogatari themes woven together with remarkable dexterity and often startling originality. The term giko monogatari,  'pseudo-classical tales,' widely  used to describe such late Heian and Kamakura period tales, and the associated style term gikobun, turn out to be M e i j i era coinages with originally much wider and less pejorative connotations - a change perhaps related to contemporary language debates that valorized vernacular writing styles.  The use of respect language and narrative asides, and the interaction between the narration and the plot, evokes a narrator with a distinct point of view, and suggest she may be the lady-in-waiting Jiju, making the text more explicitly autobiographical, and perhaps accounting for aspects of the narrative structure. Statistical information about Ariake, and analysis of respect language and certain fields of the lexicon reveal that Ariake is linguistically much closer to the Genji than are the few other giko monogatari for which information is available, but there are also a few very marked differences. Similar analysis of other giko monogatari would clarify whether these differences are characteristic of the subgenre or peculiar to Ariake no Wakare. ii  Abstract  Ariake no Wakare critiques male behaviour in courtship and marriage, and explores female-to-male crossdressing; the male gaze (kaimami); incestuous sexual abuse; both male and female same-sex and same-gender love; spirit possession in a context of marriage, pregnancy, and rival female desires, and other instances of the conspicuously gendered supernatural; and the gendered significance of genealogy. The treatment of gender roles and sexuality focuses on the interaction of performance skill and innate ability or inclination, and presents the mysterious beauty of the ambiguously gendered and liminally human, while genealogy is celebrated as privileged female knowledge. The text simultaneously invites and resists modern modes of reading. Rather than merely imitative, Ariake's treatment of familiar elements with changed contexts and interpretations produces both nostalgia and novelty.  iii  Contents  Abstract  ii  List of Tables  vi  List of Figures  vii  Part I: Study  l  Frontispiece  2  Preface  3  I - Genealogies of the Text: Origins and Sources  5  II - 'Giko': Genealogy of a Term - from Style to Subgenre  34  III - Telling the Tale: Narrative Structure and Narrator  64  IV - The Language of Ariake no Wakare  95  V - Gender in Ariake no Wakare: Gender, Genre and Genealogy  123  Bibliography  176  iv  Part II: Translation  194  Frontispiece  195  Preface  196  Book I  200  Book II (Chapters XII & XIV)  352  Book III  373  Appendices  426  I  Intertextuality  427  II  Kokonchomonju and Teika  433  III  Use of 'giko' in Literary Histories  437  IV  Ariake no Wakare Chronology  440  V  Ariake no Wakare Characters  441  VI  Ariake no Wakare Promotions and Titles  450  V  List of Tables Table I  Top 50 Subjective Adjectives from Genji in Ariake no Wakare and Makura no Soshi  Table II  116  Top 50 Subjective Adjectives from Genji, in Ariake no Wakare Books I, II, and III  Table III  118  Adverbs and Interjections in Genji, Makura no Soshi, and Ariake no Wakare 122  vi  List of Figures Figure 1  Konkyomyo  Kyo (Me-nashi Kyo), showing underdrawings (Komatsu 1959,  frontispiece) Figure 2  Konkyomyo  2 Kyo, underdrawings of a couple, one of whom wears a straw  cape (Shirahata 1940, Plate V ) Figure 3.1  Konkyomyo  Kyo, courtiers playing biwas and flutes at a music party  (Komatsu 1960, Tokyu Bunko-bon N o . 8) Figure 3.2  Konkyomyo  Konkyomyo  Ariake  19  Kyo, courtier and lady watching a full moon (Komatsu 1960,  Tokyu Bunko-bon N o . 3) Figure 5  18  Kyo, courtiers watch figures descend on a cloud (Komatsu  1960, Tokyu Bunko-bon N o . 13) Figure 4  16  19  no Wakare manuscript, page 75*7, ika ni sen (I.xi.48) (Tenri  Toshokan Zenbon Sosho N o . 6, p. 300)  vii  194  Ariake no Wakare  Part One: Study  i  Fig. 1 Konkyomyo  Kyo (Me-nashi Kyo), showing underdrawings (Komatsu 1959) 2  Preface  This study and translation is based on the edition by Otsuki Osamu, Ariake Wakare: Am Danso no Himegimi  no Monogatari  no  [Partings at D a w n : The Tale of a  Princess in M e n ' s Clothing], (Soeisha, 1979, reprinted 1987) - for more details on editions of the text, consult the preface to the translation. The transcription scheme used for Japanese in the study is the Hepburn system but historical spelling is retained for quotations from Classical Japanese, in which 'hu' is used rather than ' f u ' and the voiced counterparts of 'tsu' and ' c h i ' are transcribed as 'dzu' and 'dji' to maintain correspondence with the original system. In the translation and excerpts from the translation which appear in the study, where Japanese transcription occurs, for example i n names, titles, and transcription that accompanies poems, the unmodified Hepburn system is used. A l l translations are my own unless otherwise credited. For criticism and comments on the study I would especially like to thank my supervisor Joshua Mostow, and all those who read and commented on parts at various stages or were particularly helpful in discussing the text with me, including Doris Bargen, L y n n Burson, Carolyn Haynes, Stephen M i l l e r , Susan Napier (largely responsible for my quite unforseen midlife entry into Japanese studies), Edward Pasanen, Gregory Pflugfelder,  3  Preface Edith Sarra, Paul Schalow, M i h o Shimada of B u k k y o University in Kyoto, Tatematsu K i k u k o of the Inter-University Center in Yokohama (while I was there with a fellowship generously provided by the Japan Foundation), D o n Teeuwsen, V i r g i n i a Skord, John Carpenter (who helped me with hentaigana so I could try to read some of the  Ariake  manuscript), and my students at Obirin University, the University of British Columbia, the University of Texas at Austin, and the University of Washington, Seattle. I alone am responsible for the undoubted errors that remain. Preparation of the manuscript of this dissertation was performed using Nisus Writer on a Macintosh G 3 computer. The spreadsheet tables in Chapter Four were produced with Claris Works. Roman text is in the Times Japan font, Japanese in Honmincho.  4  Chapter One Genealogies of the Text: Origins and Sources Surely few texts invite such unfettered speculation regarding both their origins and their purport as does Ariake no Wakare. Only the barest minimum can be established incontrovertibly regarding dating or authorship, even regarding the gender of the author. External evidence is so scanty as to offer only a plausible terminus ante quern. Internal evidence is so equivocal as to offer only a plethora of intriguing possibilities for authorship, dating, and sources. A s is so often the case with 'anonymous' works, perhaps more textual pleasure resides in imagining plausible contexts for such a free-floating text than in appreciating the known contextuality of some apparently more historically anchored work. T w o kinds of information can be used to attempt to trace the origins of the text. First there are specific external references that indicate at the very least by which date the text was written, but also offer some indications as to its reception, its ranking amongst other monogatari  1  in circulation at the times of the various external references. Second,  there are several different kinds of internal evidence: apparent borrowings and references - 'intertextual genealogy'; subject matter and topoi - 'thematic genealogy,' and finally the vexed question of generic affiliation - 'generic genealogy' which one should be able to address more confidently once the preceding two kinds of information are marshalled.  External Evidence There is one solitary piece of external evidence that may be used for dating the text, and thence for explorations of possible authorship. This is the single and brief 'In this context 'vernacular tales associated with the imperial court.' 5  Chapter I: Genealogies of the Text mention of a text by the name Ariake no Wakare in the very early twelfth century literary critical text Mumydzoshi.  It is an invaluable text, in the form of a dialogue between a  2  small group of aristocratic women who discuss their likes and dislikes in monogatari reading. One's delight at the survival of such a unique reception history document for extant classical monogatari is tempered with dismay at how many monogatari it mentions 3  that are now completely lost or known only in tiny fragments. Nearly twice as many are 4  lost as are currently known. The entire reference to Ariake no Wakare runs as follows:  ^^Ipf.^. A complete translation can be found in Michele Marra,  "Mumydzoshi:  Introduction and Translation" Monumenta Nipponica 39 (2) (1984): 115-145; (3): 281-305, and (4): 409-434. A n important recent study is by Thomas Rohlich, "In Search of Critical Space: The Path to Monogatari Criticism in the Mumydzoshi"  Harvard Journal of Asiatic  Studies 57 (1) (June 1997): 179-204. 3  Extant monogatari  Taketori Monogatari,  discussed or mentioned in Mumydzoshi  Utsuho Monogatari,  Sumiyoshi  Monogatari,  (10 monogatari): Genji  Sagoromo, Yoru no Nezame, Mitsu no Hamamatsu (Hamamatsu Chunagon  Monogatari, Monogatari?),  (Ima) Torikaebaya, Matsura no Miya (Monogatari), Ariake no Wakare. C o m p l e t e l y lost or very fragmentary monogatari (19 monogatari): Tamamo ni Asobu Gondainagon, Asakura,  (Ko) Torikaebaya,  Iwa Utsu Nami, Kawagiri,  Kakure-mino,  Kokoro  Takaki, Tdgu no Seji,  Ama no Karu mo, Sueba no Tsuyu, Tsuyu no  Yadori, Mikawa ni Sakeru, Uji no Kawa Nami, Koma Mukae, Odae no Numa, Yumegatari,  Namiji no Himegimi,  Ukinami,  Asajigawara no Naishi no Kami (unless the same as  Asaji ga Tsuyu). 6  Chapter I: Genealogies of the Text subete ima no yo no monogatari ha, furuki mikado nite, Sagoromo no ama no wotome, Nezame no uchishiki nado mo, ima sukoshi kotogotoshiku  ichihayaki  sama ni shi-nashitaru hodo ni, ito makoto shikarazu obitatashiki fushi-bushi  zo  haberu. Ariake no Wakare, Yume-gatari,  no  Nami-ji no Himegimi,  Naishi no Kami nado ha, kotoba-dzukahi yoshi omohite mi-motemakaw  Asagahara  nadaraka ni mimi-tatashikarazu,  hodo ni, ito osoroshiki kotodomo  nanigoto mo samuru kokochi sum koso, ito  ito  sashimajirite,  kuchi-woshikere  5  ' A l l the monogatari of our times are about the reigns of ancient Emperors, and like the heavenly maidens in Sagoromo,  6  and the feigned death in Nezame,  7  they seem overblown and carelessly written in a hurry, so there really are many features that are terribly unrealistic and exaggerated. W i t h ones like Ariake  no  Wakare, Yume-gatari, Nami-ji no Himegimi, and Asajigawara no Naishi no Kami,  8  5  K u w a b a r a H i r o s h i , Mumyozoshi,  Shincho N i h o n K o t e n  Shusei V o l . 7,  (Shinchosha, 1976), pp. 98-99. 6  # E 3 l Written by Rokujo Saiin Baishi Naishinno no Senji (71022-1092). The  next of the extant tsukuri-monogatari after Genji Monogatari. It is presumed to have been written between 1058 and 1092, probably between 1069 and 1072 (Princeton  Companion  to Classical Japanese Literature).  Presumably Yowa  no Nezame, ffi.'D^'M,,  anonymous,  but  presumed  to be by a  female author, from perhaps as as late as the twelfth century, though traditionally attributed to Sugawara Takasue no Musume, (b. 1008), author of the Sarashina Nikki. For translations, see Richard 1973, and Hochstedler 1979. V) , fe$&<D%M, W£tii^<Ofcfe. 7  The first two are no longer extant. The  Chapter I: Genealogies of the Text just when you're thinking the wording is smooth and not at all jarring to the ear, and they're rather good in fact, on closer inspection you find they're riddled with the most frightful details and then everything about them makes you feel disappointed. It's so annoying.' This, such as it is, constitutes the entirety of premodern critical commentary on Ariake no Wakare that has survived until now. Quite apart from the interest of the critical remarks themselves, the mere mention of Ariake no Wakare in this context provides a vital dating clue. The Mumydzoshi  is considered to have been written between the seventh  or eighth month of 1200 and the 11th month of 1201, thus establishing a terminus ante 9  quern of 1201 as the very latest possible date for Ariake no Wakare. Noting that the passage quoted above refers to the group of tales which includes Ariake no Wakare as an ima no yo no monogatari, 'monogatari of our times,' one can only concur with Otsuki that on the basis of Mumydzoshi  alone the most likely period of composition for Ariake  no Wakare is the last decades, i f not the last years, of the twelfth century, or the very beginning of the thirteenth.  10  The only other text from either the Heian or Kamakura periods to mention Ariake last may be the extant Asaji ga Tsuyu, Wkl^&M, which has also been edited by Otsuki Osamu, Asaji ga Tsuyu (Ofusha, 1975).  6tsuki 1979, p.  9  496 and Marra 1984, p . l 18, both following Higuchi Yoshimaro  in "Fukuro-zdshi, Mumydzoshi  no Seiritsu Jiki ni tsuite" Kokugo to Kokubungaku  (April 1970): 81-115.  6tsuki 1979, p. 499.  10  8  47 (4)  Chapter I: Genealogies of the Text no Wakare by name is the late thirteenth century waka Fuydwakashu.'  2  11  poetry anthology, the  This is another essential text for studying the reception of monogatari,  consisting, as it does, of a poetry collection derived entirely from the poems to be found in monogatari. It was ordered in 1271 by a consort of ex-Emperor Gosaga,  13  O m i y a In  Saionji Kitsushi (Yoshiko), and was presumably compiled by her ladies-in-waiting over 14  the next few years. The format follows closely the model for imperial anthologies, chokusenshu,  15  established by the Kokinwakashu  16  back at the beginning of the tenth  century. Eighteen books are extant out of an original twenty, covering the seasons (I-VI), Shinto and Buddhist poems (VII), partings and travel (VIII), grief (IX), felicitations ( X ) , love ( X I - X V ) , and miscellaneous ( X V I - X V I I I ) .  A s with Mumydzoshi,  the Fuyoshu text both gratifies and frustrates immensely,  providing a wealth of information both about extant and lost monogatari. Where Mumydzoshi  provides information on the critical reception and tastes of readers through  "31-syllable poems. M,M^UWiM-  >2  In Higuchi Yoshimaro, ed., Ochd Monogatari Shukasen, 2 vols.,  Iwanami Bunko N o . 700 (Iwanami, 1987, 1989). i t l i l l $ , r . 1242-1246.  1 3  1 4  7v'sffiH^^^.  Kameyama.  M o t h e r of the  Hill. It is possible that Lady  Emperors G o f u k a k u s a  WMM- and  Nijo (Go-Fukakusa-In Nijo, 1258- after 1307),  author of the Towazugatari, which begins in 1271, might have assisted in her project.  1  6  ^ ^ F l l M .  Commissioned probably in 905 by Emperor Daigo  930). 9  (HIM,  r. 897-  Chapter I: Genealogies of the Text explicit judgements and the space allocated to the different texts, the Fuyoshu constitutes something of an 'index of canonization' according to the number of poems it admits from the various monogatari. Furthermore, we can judge from the plot references in the headnotes, kotobagaki, whether any events or characters known to the Fuyoshu compiler differ from the versions of the monogatari as we know them. Whereas the Mumyozoshi  deals with  twenty-nine monogatari, of which ten are extant, the 1420 waka in the extant books of the Fuydwakashu  are culled from no fewer than 198 monogatari, and the complete text  probably contained 1563 waka from 220 monogatari, of which a mere twenty-three are 17  extant.  For lost monogatari, of course, these headnotes often supply all or most of our  18  scanty information about the plots and characters. H o w the Hellenistic poets of Alexandria would have appreciated the tantalizing evocations of absent narratives provided by these fragments of them in the form of poems.  The Fuyoshu contains twenty poems from Ariake no Wakare, which places it 17  O g i 1984, p. 4.  18  Extant monogatari from which poems are taken in Fuydwakashu  Genji Monogatari  (180 poems), Utsuho Monogatari  Tsurenaki (46), Iwade Shinobu Monogatari)  (110), Sagoromo  (33), Mitsu no Hamamatsu  (23 monogatari): (56), Kaze  (Hamamatsu  Chunagon  (29), Yoru no Nezame (24), Ariake no Wakare (20), Matsura no  Monogatari (IS),  Asaji ga Tsuyu (10), Mizukara  Torikaebaya (1), Sumiyoshi  Kuyuru (10), Ochikubo  Miya  (8), (Ima)  Monogatari (7), Waga Mi ni Tadoru (7), Iwashimizu  Taketori Monogatari (3), Shinobine Monogatari (3), Koke no Koromo (2), Shizuku Nigoru(2),  Mugura no Yado (2), Hodohodo no Keso (1), Tsutsumi Chunagon  (A):Ausaka Koenu (1), Kaiawase(l),  Haizumi (1), Hanasakura Oru 10  Chujd(l).  ni  (5), ni  Monogatari  Chapter I: Genealogies of the Text eighth out of the twenty-three extant monogatari, a very respectable showing nearly a century after its presumed time of writing and more favoured than Matsura no Monogatari,  19  Miya  and indeed far more so than monogatari that are rather better known nowadays,  such as Ochikubo,  20  Torikaebaya, ' 2  and Sumiyoshi,  22  to take monogatari of comparable  length and number of poems. This evaluation seems slightly at odds with the rather cursory mention in Mumyozoshi, Ima Torikaebaya.  23  especially by comparison with the attention given to  W e may perhaps conclude that the relative popularity of Ariake  Wakare improved rather during the couple of generations after Mumyozoshi  no  was written.  This concludes the explicit references to Ariake no Wakare to be found in premodern texts. Although most exiguous, the situation is hardly different for most other late Heian and Kamakura period monogatari. The major mid-Heian monogatari were the ones that continued to receive mention, above all Genji Monogatari,  which probably led later  critics to place these two groupings in distinct subgenre categories, as w i l l be addressed below in the discussion of the term giko monogatari, 'pseudo-classical monogatari.' 1 9  l & M l t ' ^ i n . Very plausibly attributed to Fujiwara Teika, both by the  24  Mumyozoshi  and by modern critics. See Lammers 1992 for a discussion and translation. $rW$)m-  W:  21  h  For a translation, see Whitehouse and Yanagisawa 1971.  fr^li^-^^M-  K n o w n to the Mumyozoshi  ladies as Ima Torikaebaya. For  a translation see W i l l i g 1984. 2 2 /  f i c f ^ f p . The extant monogatari is presumed to be a Kamakura work, but a  Heian version is mentioned in the Genji. For a translation see Parlett 1901. *SNKBS,  pp. 82-83, and Marra, pp. 411-412.  ^ ^ c ^ ^ f n - For a discussion of the evolution of this term, see Chapter II. 11  Chapter I: Genealogies of the Text G i v e n the small number of literary critical references or even mere listings of monogatari that can be used to establish some tracing of the genealogy of the genre, it is not inappropriate to remark on contexts where mention of Ariake no Wakare is conspicuous by its absence. Foremost among these is Fujiwara Teika's poem matching sequence, the 25  Go Hyakuban  Utaawase.  Fujiwara (Kujo) Yoshitsune commissioned this sequence in  26  27  the last decade of the twelfth century or the first years of the thirteenth, together with the Hyakuban  Uta-awase to  make up a set titled the MonogatariNihyakuban  28  Uta-awase  29  The matching challenge was to pair poems from Genji Monogatari with those from later monogatari. The Hyakuban  Uta-awase pairs one hundred poems from the Genji with a  matching number from Sagoromo. The Go Hyakuban Uta-awase pairs one hundred poems from the Genji with a matching number from ten other later monogatari.  Once again, the Go Hyakuban  Uta-awase is valuable as a tentative index of  canonization and source for dating, as well as for information on lost monogatari. The division into Sagoromo and 'the others' is surely also significant. 'The others,' in this instance, include two extant monogatari, two lost versions of extant monogatari, and six lost monogatari. O f special interest is the fact that the most poems go to the extant Yoru no Nezame (20) and Mitsu no Hamamatsu  (15), but a version of Torikaebaya  also  contributes six poems, some of which are mere minor variants of poems in the extant  B1MMU  25  2  6  (1162-1241).  ?&^#!fc-n.  27s  Also known as Shui Hyakuban  Utaawase  (1169-1206).  ' # J g g _ " S ' # | ^ ' o . Included with the Fuydwakashu in Higuchi 1987. 12  Chapter I: Genealogies of the Text version, ranking it sixth out of the ten monogatari.  30  This version of Torikaebaya  probably the non-extant older version known to the ladies of the Mumyozoshi,  is and  contrasted by them with the extant version which is presumed to be the monogatari they call Ima  Torikaebaya.  A s we shall see below under the consideration of candidates for possible authorship, Teika and his circle are certainly in the running, but i f Teika knew Ariake no Wakare, he did not judge its poems above Torikaebaya, although the compilers of the  Fuydwakashu  certainly did. The inclusion of Torikaebaya but no Ariake may be considered further encouragement for the very tentative hypothesis that Ariake was less highly regarded soon after its composition than it came to be a couple of generations later. Alternative hypotheses from this date are that Ariake was considered too similar to Torikaebaya for both works to be included in the same uta-awase, or that the author was someone too close to Teika's own circle to be decorously included in a collection put together by him. In this regard it is perhaps significant that Matsura no Miya  Monogatari,  very plausibly by Teika himself, is also absent from this list. 31  One last external document that may refer to Ariake no Wakare is the emaki 30  T h e others are Mikawa  ni Sakeru (15), Asakura (13), (Hidari mo Migi  mo)  Sode Nurasu (10), Kokoro Takaki (10), Tsuyu no Yado (5), Sueba no Tsuyu (3), and Ama no Karu Mo (3). The latter also appears to be a variant of the fragmentarily extant monogatari of the same name (Higuchi 1987). 31  F o r a full discussion of the attribution of Matsura no Miya Monogatari to Teika,  see Lammers 1992, Appendix B 'The Authorship of Matsura no Miya 13  Monogatari.'  Chapter I: Genealogies of the Text (picture scroll) known as the Me-nashi Kyo.  32  This is an early Kamakura work extant in  four scrolls out of an original five, with a colophon dated K e n k y u combines both the Konkdmyd of the Hannyarishu Hakubyd-e  Kyo.  35  Kyo  3<1  33  4 (1194). The text  (Golden Light Sutra) in three scrolls and one scroll  The full titles are Hakubyd-e  Rydshi Kongdmyd  Kyo and  Rydshi Hannyarishu K y o respectively. The interest of part,of the text is that 36  it consists of these sutra texts written over underlying illustrations, shita-e (see Figure 1 (frontispiece) and Figure 2), that may depict scenes from a monogatari episode involving a spirit possession or mono no ke . The figures in the illustrations are all lacking eyes or 37  noses, hence the name Me-nashi Kyo, 'Sutra without Eyes.' It has been suggested that the illustrations are for Genji Monogatari, since the construction resembles that of the famed 1=1  32  The earliest detailed scholarly examination of this text seems to be by  Shirahata Y o s h i , "Me-nashi Kyo ni tsuite," Bijutsu Kenkyu  1940 (105): 6-20. This article  still offers a useful and relatively comprehensive set of illustrations. 3 3  ^ ^ \ . For a discussion of this colophon, see Fujimoto K o i c h i , "Hakuby5-e  Shutsugen no Hito Shiten - Me-nashi Kyo no Okugaki no Kaishaku ni tsuite,"  Nihon  Rekishi 1982 (414): 82-89.  36  F o r more recent discussions of art historical aspects of this emaki, see Murashige  Yasushi 1978 and Murakami Harumi 1991. Neither of these articles addresses the question of the identification of the monogatari illustrations beneath the sutra text, but M u r a k a m i provides a wealth of illustration detail focusing on the faces, figures, and clothes.  14  Chapter I: Genealogies of the Text twelfth century Genji illustrated scroll, the quality of the work suggests that such copying 38  as an illustrated scroll was bestowed on a major text, and the scenes may be construed as following the narrative of the Genji. O n the other hand, it has also been suggested  39  that a  figure that appears many times seems to be wearing a straw cape (see Figure 2), suggestive of the possible kakure-mino possession scenes.  40  in Ariake no Wakare, a text which also has prominent spirit  There are also various other episodes in Ariake that may be reflected  in some of these sketches, as described below. Komatsu first draws together a number of illustrations involving mino, to show that the shapes appearing in these illustrations are likely to be mino, rather than merely shocks of hair, and examines what is known of the monogatari called Kakure-mino.^ the basis of Mumyozoshi's  sharp criticism of Kakure-mino,  he then posits Ariake  On no  Wakare as a much more likely choice for illustration at a time roughly contemporaneous with the Me-nashi Kyo. There are two problems with this attribution. First, it is by no means stated clearly that the protagonist of Ariake actually has a kakure-mino, 38  F o r a reproduction see S h i m i z u Y o s h i k o , Genji Monogatari  only that Go-ju-yo-jo  (Heibonsha, 1982). The Genji correspondences are examined in detail in Shirahata 1940. 39  F o r an article surveying this debate, see Komatsu Shigemi, "Menashi-kyo shita-e  to Ariake no Wakare monogatari," Parts I and II, Sansai. 121 (December 1959): 5-16, and 122 (January 1960): 16-23. 40  See Ariake no Wakare Book Il.xiv, and Book IILi.  "'Komatsu, Part I, pp. 7-10 and 15-16.  15  Chapter I: Genealogies of the Text  F i g . 2 Konkyomyo  Kyo, underdrawings of a couple, one of whom wears a straw cape (Shirahata 1940, Plate V , detail)  16  Chapter I: Genealogies of the Text he had a means of concealing himself like the kakure-mino.  42  Perhaps the complete  absence of further mention of an actual kakure-mino is mere stylistic restraint, but one cannot assert confidently that Ariake actually has one. Furthermore, several of the scenes depicting the kakure-mino  are interpreted as illustrations drawn from Books II and III,  such as the young Sadaijin's visit to the Former Emperor's, when he watches the Former Empress's ladies-in-waiting reading and looking at pictures (II.vii.24), and what looks like an exorcism scene. But these episodes take place long after Ariake has ceased to wander around invisibly, by virtue of kakure-mino or not, an ability that the protagonist seems to lose once the change of gender role is made in Book I, Chapter xiv. Nevertheless, it is tempting to try and see episodes from Ariake  i n these  pictures.Komatsu makes the following identifications:  T o k y u Bunko-bon N o . 8: courtiers with flute and biwa - music party after the Retired Emperor's Fortieth Birthday Celebrations (IILvi). See Figure 3.1. T o k y u Bunko-bon N o . 13: two courtiers look up at figures on a cloud - descent of heavenly maidens (IILvi). See Figure 3.2. T o k y u Bunko-bon N o . 11: court lady with child - Shijo-no-Ue bears the Sadaijin a son (III.v).  V^fc£>&££&<£  &tl$> *)%tct.&  (Book I.ii.6) 'and for whatever reason, the gods  provided a means of hiding oneself, just like we read about in tales like  kakure-mino.  With no private residences barred to him, Ariake wandered about and sneaked in everywhere he wished.' 17  Chapter I: Genealogies of the Text Jinko-in-bon N o . 6 / Tokyu Bunko-bon N o . 4: courtiers with plants - the Insect Contest (II.v). Jinko-in-bon N o . 14: court ladies looking at pictures - the Sadaijin visits the Former Emperor and finds the Former Empress and her ladies reading and looking at pictures (II.vii.24). T o k y u Bunko-bon N o . 3: courtier and lady under a full moon - Udaisho and the Sokyoden Lady under the Eighth Month full moon (I.xi.45, 50). See Figure 4.  Figure 3.1 Konkyomyo  Kyo, courtiers playing biwas and flutes at a music party  (Komatsu 1960, Tokyu Bunko-bon N o . 8)  18  Chapter I: Genealogies of the Text  Figure 3.2 Konkyomyd  Kyo, courtiers watch figures descend on a cloud  (Komatsu 1960, Tokyu Bunko-bon N o . 13) " i v  IE-  I *  i J  j  1 '  •  1  1  ill  Fig. 4 Konkyomyo  i ><+  Kyo, courtier and lady watching a full moon (Komatsu 1960, Tokyu Bunko-bon N o . 3) 19  Chapter I: Genealogies of the Text A s w i l l be readily remarked, scenes of courtiers playing music, ladies looking at pictures or with children, and couples enjoying the full moon are not exactly rare in monogatari. However, the juxtaposition of descending celestial figures and exorcisms is a provocative one, and Ariake is clearly the best candidate of the monogatari that have survived. Exactly what the shock-haired figure represents in so many of these illustrations is a real mystery. The three scrolls with these illustrations could correspond to the three-scroll structure of Ariake no Wakare. If these arguments were accepted, then the terminus ante quern for Ariake would be pushed back to 1194 at the very latest, i f one takes the date onthe colophon, so the illustrations would date from some time before then. The Me-nashi Kyo evidence is inconclusive, but given the paucity of any evidence relating to Ariake, it is of considerable interest. It is to be hoped that further research by art historians may confirm Ariake as the source for the illustrations, thereby allowing us not only to date Ariake with more assurance, but offering an interpretation of the text in another medium.  43  43  O n e can hardly forbear from observing that a palimpsest i n w h i c h the  predominantly female-associated hiragana text type of the monogatari is overwritten, or as it were, re-clothed, with the more male-associated Chinese characters of a sutra is a particularly interesting form for a set of illustrations for Ariake of all texts to take, with its themes of cross-dressing and gendered concealment (see frontispiece), even granted that women also copied out sutras written in Chinese characters. 20  Chapter I: Genealogies of the Text  Internal Evidence Intertextuality, in the sense of the conscious or unconscious recollection of a previous text or textual practice, is a particularly elusive kind of evidence to use in many contexts. The most obvious problems, once it is clear one is not dealing with an unambiguously direct quotation, are first, deciding what degree of similarity in wording constitutes evidence of actual borrowing (conscious intertextuality), and then, whether the immediate source is the earliest, 'original' source, or whether it might have been transmitted via an intermediary text. In the case of the late Heian and early Kamakura monogatari, as is the case with many texts in conservative literary traditions, there was conscious adherence to a circumscribed diction. The model was the diction of Genji Monogatari, much as waka poetry adhered to the diction of the Sandaishu.  44  This frequently  renders it most imprudent to make assured assertions about intertextuality as hard evidence of 'borrowing,' even regarding verbatim phrases, since these may well have passed into general currency. Noting similarities does however, at the very least, establish the likely existence of a web of genealogical relationships, without necessarily making claims of direct borrowing or influence. Nevertheless, in the case of a text as untrammeled by the bonds of external evidence as Ariake, the possibilities offered by internal evidence must be pursued as thoroughly as possible. Some responsibility for intertextual claims can be off-loaded onto ^HLJKM, the Kokinshu,  the first three imperially commissioned anthologies of Japanese poetry, the Gosenshu  ( H J I H C. 950), and the Shuishu.(taMM:  1010).  21  compiled c.  Chapter I: Genealogies of the Text other scholars by utilizing the intertextual suggestions that populate many of the footnotes to editions of the monogatari in question. The implications of 'cf.' and its synonyms cover many degrees of similarity which one collapses at one's peril, so the conclusions offered must necessarily be tentative. They do, however, seem to be of considerable interest. One can draw up and compare lists of intertextual references identified by various editors for different monogatari, starting with similarities in diction both from poetic and narrative sources, and come to some tentative conclusions that shed light on the broader genealogies of the texts concerned. Drawing up a list of intertextual references in the first book of Ariake no Wakare,  45  then grouping these according to the different sources, both  prompts intriguing suggestions regarding the distribution of these references in Ariake, and also implies patterns of connection with other texts. What is most noticeable regarding the distribution of intertextual references is how widely the number of plausibly identifiable references varies for different parts of the text. Even accounting for the fact that chapter divisions are not original (but follow Otsuki 1969), and taking into account a certain amount of variation in chapter length, one is struck by the clustering of references in Chapters I, I V , VIII, and X I . It might even be possible to note an alternation between groupings of chapters without an appreciable level of word-level intertextuality and these chapters that have noticeably more. Since one might well correlate denser intertextuality with 'higher style,' imbuing characters, scenes or events with a specifically refined aura at key moments, this distribution at least makes for a stylistically modulated narrative structure. 45  See Appendices l a and lb, 'Ariake Intertextuality,' based on this translation. 22  Chapter I: Genealogies of the Text Once one reorders the list of references to show clusters of favoured texts, additional patterns emerge. The overwhelming majority of references to poems from anthologies are, of course, to the Kokinshu, with nineteen references, to twelve distinct poems out of thirty-three references altogether, thus nearly two thirds of the total. Perhaps more interesting is the showing of the Goshuishu of 1086, poems. Gosenshu, Shuishu and Senzaishu  48  46  and even four references to  Man'yoshu  47  and sundry other poetry collections have only  one reference each. The distribution through the text of various prose sources, including uta-monogatari, seems random, but patterns emerge once they are collected together. Unsurprisingly, Genji corresponds to the place of the Kokinshu, with twenty-four references, and there are four to Ise Monogatari.  49  More interesting is the showing of Torikaebaya, with seven  references, hardly surprising given the subject matter, but there are also five references to Sagoromo, four to Utsuho, and one each to Yoru no Nezame (or Koke no Koromo), ga Tsuyu, kakure-mino,  Eiga Monogatari (possibly), and Matsura no Miya  Asaji  Monogatari.  The last is of special interest since it offers further information on the dating of the text. 50  This monogatari is thought to have been written 'within a year or two of 1190,' so with 51  immm.  46  M l i i l l l , compiled in the mid-8th century, including poems from at least the  4 7  preceding three centuries. ^ h i f c S completed 1188.  4 8  ffi$ty$)M, early tenth  49  century.  50  Chapter VIII, Section 35.  51  Lammers 1992, Appendix A 'Evidence on Dating Matsura no Miya  23  Monogatari,'  Chapter I: Genealogies of the Text this terminus a quo, one might be inclined to consider Ariake a product of the last decade of the twelfth century, thus a Kamakura period work (794-1185). Apart from this dating trouvaille, what one notices is the range of references to monogatari, and that only five of the Genji references are to the U j i chapters. This is surprising, given the popularity of the tone and standpoint of these chapters in many works which derive inpiration from the tone and standpoint of the U j i chapters of the Genji. A l l of this information is much more interesting when put into a comparative context, compiling similar charts for the intertextual references to be found in other monogatari. T o take Yoru no Nezame,  52  Torikaebaya,  53  and Matsura no Miya  Monogatari  54  for example, the following general characteristics can be identified, even i f the inventory of references is not taken to be unduly precise. First of all, Nezame, while sharing the predilection for the Kokinshu,  has remarkably few specific references to monogatari,  even Genji, which accounts for only a couple, together with isolated references to Utsuho and Sagoromo (possibly), Taketori and the lost monogatari Katano no Shosho which is also mentioned in Ochikubo, Makura no Soshi, Genji, and the Fuydshu. N o poetic sources other than the Kokinshu are particularly prominent. Torikaebaya broadly follows this pattern: there are three Genji references, three to Yamato, two to Ise, and one each to p.172. 52  See A p p e n d i x Ic, 'Yoru  no Nezame  I Torikaebaya  I Matsura  no  Miya  I Matsura  no  Miya  Intertextuality,' based on Richards 1973 (augmented). 53  See A p p e n d i x Ic, 'Yoru  no Nezame  I Torikaebaya  Intertextuality,' based on W i l l i g 1984 (revised). 54  See Lammers 1992, 'Index of First Lines of Poems Cited,' p.207. 24  Chapter I: Genealogies of the Text Sagoromo and Hamamatsu. Once again, no poetic sources other than the Kokinshu  are  particularly prominent. O n the other hand, Matsura no Miya has a small number of references to a variety of monogatari, Genji, Eiga Monogatari, Hamamatsu, Ise, Yamato, but the usual prominence of Kokinshu surpassed, by the Man'yoshu  Sagoromo,  (twelve references) is paralleled, even  (fourteen references), with no other collections making  notable showings, though this may be attributed in part to its Nara period setting (712-793). Can one hazard any overall genealogical inferences from these results? Ariake is notable for its number of monogatari references (unless we owe that to Professor Otsuki's greater diligence in tracking them down). Ariake is also noteworthy for its number of Man 'yoshu references  55  which set it apart from Nezame and Torikaebaya, but which are  even surpassed by Matsura no Miya. The attention to the Goshuishu seems to be another, though secondary, characteristic of the work's poetic intertextuality, and one which is peculiar to it. Overall, such data may help strengthen the sense that the author was someone close to Teika's circle, since even the Mumydzoshi  takes h i m to task for his  predilection for the Man 'yoshu evidenced in his Matsura no Miya. The Ariake evidence also reminds us that one aspect of the intertextuality of later monogatari is that monogatari themselves were increasingly the sources for references. Regarding Teika there are further intertextual features that bear consideration. The most cited of all the Kokinshu  poems is B o o k X I I I (Love III) 625, by M i b u no  Tadamine. Its distribution throughout the text and at especially prominent moments justifies referring to it as a 'leitmotiv' poem, and it is especially noteworthy that two external sources confirm that it was also one of Teika's most beloved poems. Not only does it 55  F o u r in B o o k I alone (see Appendix lb). 25  Chapter I: Genealogies of the Text appear in all eight of Teika's exemplary anthologies, but there is a remarkable anecdote 56  in the Kokonchomonju  which relates how, when Teika and Fujiwara no Ietaka were  57  required by the Inmeimon-in, a former empress, to cite a classic poem for a given topic, 58  they both wrote down this poem of Tadamine's. It should be added that another tale in the Kokonchomonju  mentions in passing Teika's interest in monogatari.  59  The kind of intertextuality just considered is word-level intertextuality based on specific diction that can be traced to specific texts. Another kind of intertextuality is based not on references at the level of diction but on theme. Some kinds of subject matter can be associated with a specific source text, other kinds are so pervasive that no specific antecedent is claimed, and these of course are known as topoi. In the case of Ariake, in addition to numerous characters, plot configurations and scenes that recall Genji Monogatari, a remarkable number of the text's seemingly unusual, sensational events have precedents in other, earlier monogatari. These elements constitute an aspect of the text's genealogy Teika  56  Jittei (1202-1213), Kindaishuka  (1209), Eiga Taigai (\2l6),Shuka  Daitai (after 1226), Hachidaishu Shuitsu (1234), Nishidaishu  (1235), Hyakunin  no Shuka  (1236), Hyakunin Isshu (1237). l^^fMWlM,  51  compiled in 1254 by Tachibana no Narisue. This tale is V o l u m e  V , B o o k V I , Tale 219 (78 of that book). For a translation see Appendix II, ' T w o Kokonchomonju 58  tales.'  Reishi, Consort of Emperor Tsuchimakado. She lived from 1185-1243 and became  Empress in 1205. 59  B o o k X I X , Section X X I X , Tale 662 (17 of that book). For a translation see  Appendix II ' T w o Teika Tales from  Kokonchomonju.' 26  Chapter I: Genealogies of the Text that is as important as its word-level intertextuality. Japanese editions of post-Genji monogatari generally investigate echoes of Genji quite thoroughly, both at the diction and thematic levels. Without going into considerable detail here, some of the more salient instances should be mentioned. In terms of characters and characterization, whereas the irogonomi, or libertine, characters of the first Sadaisho, his son Sanmi no Chujo, and the latter's unacknowledged son, the Sadaijin of the second and third books, all generically resemble the famous irogonomi characters such as Ariwara no Narihira, Heichu, and the young Genji, it is the troubled later Sadaijin who most recalls aspects of the Genji characters: Genji himself for his eroticised pursuit of a lost parent, and Kaoru in his troubled relationships with women, and his gradual uncovering of the truth of his own parentage. The overall structure of Ariake no Wakare itself seems to bear a strong resemblance to Genji, in that the text divides into the superficially autonomous but profoundly symbolically related stories of two successive generations of the higher echelons at court.  60  Just as in Genji there is a narrative 'gap' in which the death of the protagonist, (in the case of Ariake, the 'death' of the protagonist), is elided from the narrative. In Genji, Genji is last presented i n Chapter 41, Maboroshi,  'The W i z a r d , ' where he is still  contemplating taking orders and sad to be taking leave of his little grandson Niou: "Genji was more and more despondent as the N e w Year approached. N i o u scampered about exorcising devils, that the N e w Year might begin auspiciously. 60  F o r a broader treatment of the narrative structure of Ariake, see Chapter T w o ,  'Narrative Structure and Narrator.' 27  Chapter I: Genealogies of the Text "It takes a lot of noise to get rid of them. D o you have any ideas?" Everything about the scene, and especially the thought that he must say good-bye to the child, made Genji fear that he would soon be weeping again."  61  Chapter 42, Niou, ' H i s Perfumed Highness,' opens with the news that Genji is already dead and has left a void that cannot be filled:  ^  "The shining Genji was dead, and there was no one quite like h i m . "  62  In Ariake, B o o k I, Chapter XIII, Section 59 ends with A r i a k e ' s sad thoughts about his imminent disappearance, as he contemplates the child Tai-no-Ue conceived by Sanmi no Chujo whom he has acknowledged as his own child: " A s Ariake looked at what a beautiful child Wakagimi was growing up to be, even though he was not going to be very distantly separated from him, he had many sad thoughts indeed." Chapter X I V , Section 60 then opens with the announcement that, "During the storms of the Godless Month, Ariake's light was hidden from the world." In both texts the action then devolves onto a descendant of the original protagonist, principally Kaoru in Genji and the later Sadaijin in Ariake, though in both cases the new protagonist is neither as central, nor as idealized, as the initial one. Aspects of the narrative structure of Ariake are dealt with in greater detail in Chapter Three, below. Most noteworthy of the major motifs and themes in Ariake are the following: the theme of genealogy that motivates much of the plot; the 'tormented stepchild'; the 61  Seidensticker, p. 734.  62  Seidensticker, p. 735. 28  Chapter I: Genealogies of the Text protagonist's cross-dressing and dual gender roles; powers of invisibility; concern for women suffering the depredations of irogonomi men; and Ariake's supernatural musical abilities, which are all found in Book I; and ultimately her supernatural identity as a 'heavenly maiden,' tennin nyobo,  63  strongly hinted at earlier, and then virtually revealed  in B o o k III. Other significant motifs and themes are: marriage between a human and a supernatural being; forbidden love for a parent figure; the irogonomi's  search for the  ideal woman; the 'hidden flower' motif - a beautiful and refined woman living in obscurity; 'peeping,' kaima-mi:  M  female jealousy resulting in spirit possession of rivals (especially  during pregnancy); and the revelation of paternity secrets. A l l of these have their own genealogies i n previous prose texts, especially monogatari. A listing of major motifs, topoi, or themes occurring in court tales, ocho monogatari?  5  together with definitions and brief discussion of each can be found in the  1987 Kokubungaku types' (wakef),  Supplementary Volume on court tales,  66  in the section on 'narrative  including the following which correspond to those listed above in Ariake:  'divine descent of aristocrats' (kinin no korin,  6&  mentions Ariake,  under Amewaka-miko ), 69  which specifically  a narrative type in which aristocratic genealogies are traced to the  XXiZB.  •mmw*. 66'  'Fujii 1987.  'xmm^.  69  29  Chapter I: Genealogies of the Text descent to earth of divine beings; 'marriage between human and non-human'  (irui-kon'in ); 70  irogonomi, kaima-mi; 'musical instruments as narrative elements' (gakki to sono wakei ); 1]  the 'divine marriage' subtype of marriage quest (seikon I kyukon )  in which a mortal  12  pursues an immortal as the most idealized kind of marriage partner; 'heavenly maiden' (tennin nyobo); the 'tormenting of a stepchild' (mamako-ijime );  'secrets' (mono no  7i  magire); and 'dual gender (or hermaphroditism)' (ryosei guyu ). 74  In the Kokubungaku  Supplementary Volume listing, in addition to the definition  or characteristics of the motif, one or more examples are given. However, other than inclusion in the list itself denoting relative prevalence there is usually no indication of how widespread each element is. In fact the distribution varies from a small number of influential texts, to a large number of texts of many types. Since it is of interest to examine in greater detail the kind of texts to which these motifs link Ariake, it is worth trying to trace a larger number of them here. The results should give a broader indication of where Ariake stands in relation to its prose fiction (and other) forbears.  Cross-gendering: Torikaebaya; various setsuwa. Invisibility: kakure-mino?  5  7 5  Taketori; Makura no Soshi,  16  N o w lost, but discussed in Mumydzoshi.  found in the Fuyoshu. 30  Shuishu,  11  setsuwa (usually 78  11 poems with headnotes are to be  Chapter I: Genealogies of the Text constituting a kind of 'super-£a//na-mi'). Kaima-mi: passim, but especially Ise Monogatari, Ochikubo Monogatari, Irogonomi: passim, but especially Ise, Heichu Monogatari, Abduction of intended wife: Ise, Ochikubo,  Genji  Genji  Genji, Sagoromo,  Tsutsumi  Chunagon  Monogatari (Hanazakura Oru Shosho). Marriage Quest: passim, but especially Taketori, Utsuho (Atemiya). Obscured parentage, genealogy: Genji (Reizei, K a o r u ) , Hamamatsu Monogatari,  Torikaebaya, Matsura no Miya  Triangular affair: passim, but especially Genji,  Monogatari  Sagoromo.  Miraculous musicianship: Man 'yoshu, Utsuho Monogatari, 76  Chunagon  Yoru no Nezame, Matsura  D a n 4 5 , 100, 131.  " B o o k X V I I I , N o . 1192, Taira no Kinzane, Shinobitaru hito no mo to ni yokoshikeru kakure-mino, kakure-gasa wo mo ete shi ga na kitari to hito ni shirarezarubeku 'Sent to a person he was longing for: If only I had a cape or a hat to make me invisible Then my comings and goings would be completely unknown' 78  I n Konjaku  IV.24, three men use a medicine that confers invisibility, 'just as  though one had on that cloak of invisibility one hears of,' I M t l l t t V^*S* &Kfrfc'h fcfc \ :  <  ts % <D<D X  L " C \ to violate an Indian king's royal ladies. See also  Setsuwashu, 63.  31  Kohon  Chapter I: Genealogies of the Text no Miya, various setsuwa.  79  Female protagonist / 'feminist perspective': Genji (Uji chapters), Yoru no Nezame. Heavenly Beings on Earth: Fudoki, Man 'yoshu, Taketori, Sagoromo, Yoru no Nezame, Genji ' Y a d o r i g i , '  80  Konjaku Monogatari X X I V . 1.  81  Recovery from death or suspended animation (sosei):* Genji (Ukifune), 2  Sagoromo,  Ko Torikaebaya. This motif is especially popular in later monogatari: Asakura, Asaji ga Tsuyu. This is by no means an exhaustive list of the motifs and themes of Ariake, nor of the previous texts in which they appear. It should however serve to indicate what a large number of motifs and themes Ariake draws on, and what a large number of texts it draws on for them. F e w preceding texts other than Genji provide such a summa of the main themes and motifs of the court tale type of monogatari. There are yet other ways by which one could seek to establish a relationship between Ariake and other texts. In his study of Matsura no Miya  Monogatari, Wayne  Lammers notes that the dreamlike air with a strong suggestion of the mysterious and alluring supernatural is a close approximation to a narrative incarnation of Teika's ideals of yugen and yden, and indeed there are grounds for identifying characteristically Shinkokinshu'  l  diction, all of which he uses to bolster the argument for Teika's authorship.  Similar claims could be made for Ariake, but without the Mumydzoshi  attribution of  79  See also the Genji and Konjaku references under 'Heavenly Beings on Earth.'  80  Seidensticker p. 923.  81  F o r a translation, see Royall Tyler, Japanese Tales, p.61.  M±.  S2  32  Chapter I: Genealogies of the Text authorship to Teika that Matsura no Miya monogatari has, there is less incentive to do so, given the prevalence of such an aesthetic in Teika's circle. Unlike with Matsura no Miya, this line of reasoning is not likely to confirm any authorship attribution. In a genre characterized by a rather circumscribed set of narrative stratagems, Ariake clearly gathers together a remarkably large number of them. The most interesting questions concern what this monogatari does with these themes and motifs in terms of signification, and what kind of integration is attempted, i f any. This w i l l be addressed in subsequent chapters, but at this point further consideration of Ariake's  relationship to  other monogatari needs to address an additional question of generic affiliation, namely the subgenre of giko monogatari. What is understood by this term, where did it come from, and how has its denotation developed over time. In short, what is the genealogy of the term giko  monogatari!  33  Chapter Two 'Giko': Genealogy of a Term from Style to Subgenre 'The archaic period is winter. The norito, senmyd, and the Kojiki are the dense woods of pine and oak deep in the snow. The early Heian Taketori Monogatari,  Ise Monogatari, and Tosa Nikki are the budding of the first  plum blossoms in early spring. The period of the women writers, Murasaki and Sei Shonagon, is the season of genial spring breezes when masses of flowers are in full bloom. If Murasaki Shikibu is the cherry blossom, then Sei Shonagon is the red flowering plum. Izumi Shikibu is the peach blossom. The rise of wakankonkobun  in the Kamakura period is the fresh verdure of  early summer. The Muromachi period is like high summer festooned with crape myrtle and lilies. Coming to the Edo period with the sudden appearance of so many and various great writers, it is an autumnal landscape of the moors with their seven flowers, the hills with their yellow and scarlet leaves. The ripening of kanbun is that of the chrysanthemums and orchids made native to Japan. A n d the revival of wabun, that is to say gikobun, is the reflorescence of the cherry blossom.'  1  This encomiastic late-Meiji account of 'gikobun' as a highly valorized Edo period writing style comes as some surprise to the modern reader reflecting on the term 'giko monogatari.' Twentieth-century literary criticism of the monogatari genre makes fairly frequent use of the term 'giko monogatari' to refer to post-Heian monogatari, and especially 2  'Omachi Keigetsu, Nihon Bunshdshi, 2  1907, p. 1711.  E . g . , Tsuda Sokichi, Bungaku ni Arawaretaru Waga Kokumin 34  Shiso no  Kenkyu,  Chapter II - ' G i k o ' from Style to Genre those of the later twelfth and thirteenth centuries, which includes the period in which it seems most likely that Ariake was composed. The term itself appears to have emerged 3  very early in this century, and definitions that accompany its usage in literary histories are somewhat confusing and contradictory, varying both in terms of the period to which it applies, whether it constitutes a writing style or a genre, and whether it is favourably or unfavourably evaluated.  Accordingly, it is worth examining the characteristics attributed to the subgenre to see i f it can be defined in a principled way; to see i f such a definition corresponds to any subgeneric awareness perceptible in works from the original time, especially the Mumydzoshi; and to locate Ariake in relation to this subgenre. A s part of this process it w i l l be instructive to try and trace the genealogy of the term itself, in order to be fully alert to different associations it may have had at different periods. The term giko, 'pseudo-classical,' 'classicising,' or 'archaistic,' is widely found nowadays in two quite distinct uses, one referring to an Edo period literary writing style, 1916 (rpt. 1977-1978), V o l . I l l , pp. 37, 259 (Kamakura Period); V o l . V I I , pp. 344, 375 (Edo Period); N o m u r a Hachiro, Kamakura  Jidai Bungaku Shinron,  1922, Chapter 2,  ' G i k o B u n g a k u ' ; M i u r a Keizo, Sogo Nihon Bungaku Zenshi, 1924, Chapter 60, ' G i k o Bungaku' (Edo); Hisamatsu Sen'ichi, Nihon Bungaku Hyoronshi, 1936, p. 499 (Kamakura); Hisamatsu S e n ' i c h i , Yosetsu Nihon Bungakushi,  1952; rpt. 1965, p. 197, Chapter III,  Section 7, ' G i k o Monogatari' (Kamakura); Nihon  Koten Bungakushi,  1992, p. 83  (Kamakura), p.131 (Edo). 3  O f course, no pre-World War II texts mention Ariake in any capacity other than  as the name of a 'lost' monogatari, since it was only found after the war. 35  Chapter II - ' G i k o ' from Style to Genre gikobun,  and the other, giko monogatari, referring to certain monogatari of the Late  Heian and Kamakura period. Interestingly, neither term was used at the times of the literary phenomena to which they refer, though as we shall see, there is some evidence for awareness, at those times, of the categories to which 'gikobun' and 'giko monogatari' now refer. I intend to show that these two uses are related in an unexpected way, and one which throws some light on the nature of 20th century literary criticism of the subgenre to which Ariake may reasonably be assigned. The connotations of ' g i k o ' w i l l also be illuminating with regard to the discussion of the relationship between Ariake and its precursors, which I have begun to sketch out in the introductory chapter. Furthermore, there are certain ambiguities and contradictions in the later use of 'giko monogatari' which it shall also prove instructive to investigate. A very early appearance of the formulation, though not the term, ' g i k o ' may be identified in the Edo period itself, in the title of a work whose style parodied the Kojiki, the Gi-Kojikibun,  4  by 'Ashinaga Inagomaro,' a pseudonym of Yamada Hisatada roughly 5  6  between 1830 and 1844. Clearly, the logical segmentation of the compound does not 7  permit us to interpret this as an early use of 'giko' in the way it was to be used later, but ^ H I ^ ^ - b E , ^ . E d o period parodies of classical texts such as Ise  Monogatari  (parodied in the Nise Monogatari), and Genji Monogatari (parodied in Ryutei Tanehiko's Nise Murasaki Inaka Genji) are well known. Parodies of the Hyakunin Isshu have recently been examined by Joshua Mostow in Mostow 1997.  Ninon Koten Bungaku Daijiten (NKBD),  7  36  V o l . II, p. 124.  Chapter II - ' G i k o ' from Style to Genre the occurrence itself is worthy of note. Unlike the Edo pseudoclassical styles that were to become known as 'gikobun', the intention is parodic and humorous, blending ga and zoku with deliberate incongruity, rather than emulating the lofty Kojiki style to achieve 8  purity of wabun and an elegant tone. Nevertheless, the fact that such a work came into existence at all probably does signify an awareness of the 'gikobun' that by this time had been practiced by a large number of kokugaku  9  scholars, though generally choosing a  mid-Heian rather than Nara linguistic model. The fact that the title ended up serendipitously including the very character sequence that was to be coined as a compound later to refer to the object of its parody (that object being of course the classicizing kokugakusha rather than the Kojiki itself), is actually a rather delicious irony. The term 'gikobun,' though referring to the later, Edo period literary style, in fact came into use before 'giko monogatari,' the Late Heian and Kamakura period genre term. A widely available modern definition can be found in the Nihon Koten Bungaku  Daijiten™  which refers to it as a 'writing style,' buntai, from the mid-Edo period to the beginning of M e i j i , 'originated among kokugakusha and made an essential compositional exercise for the study of Japanese classics,' which was also promoted by Ogyu Sorai (1666-1728) and the Kobunji school. M i d - H e i a n (and sometimes earlier) vocabulary was to be used, and kanji were to be given their Japanese (kun) readings. Gikobun usage is identified in five  m,m.  s  m ^ , 'national learning, nativist studies,' E d o period scholarship focusing on  9  Japanese language, literature, religion and history, in contrasdistinction to Chinese, especially Confucian and Buddhist, studies. NKBD,  10  V o l II, pp. 124-125. 37  Chapter II - ' G i k o ' from Style to Genre contexts: 1) prefaces, postscripts and official records for which a miyabi"  tone was  required; 2) essays and lectures on kokugaku; 3) articles and travel diaries; 4) writings on current affairs (shozokubun); and translations and studies of the classics. It was also naturally used for text such as waka headnotes. Regarding the practicioners of gikobun, the Daijiten identifies three periods or major groupings. The first grouping consists of K a m o no M a b u c h i disciples, based in Edo. In works like Mabuchi's Niimanabi,  13  12  (1697-1769) and his  and Bun'iko  14  he valorized  the masuraoburi ('manly, heroic') qualities of ancient Japanese vocabulary (i.e. Kojiki and Man 'yoshu, cf. his own Goiko)  15  and the tawayameburi ('graceful, delicate, feminine')  qualities of mid-Heian vocabulary, and offered them as 'exemplary models,' or, in his usage, tataegoto or michiyukiburi.  This style was continued by his disciples Kato Chikage'  (1735-1808) and Murata H a r u m i (1746-1811), and their disciple Kato U m a k i 17  1777), and in turn by his disciple Ueda A k i n a r i (1733-1806)  i n K y o t o , often highly ornamented  ^^XS  U  i8  ,9  2 0  19  18  6  (1720-  (1734-1809), and by B a n K o k e i  2 0  w i t h stylistic devices such as  I JHyjft, 'courtliness,' especially associated with the Heian era aristocracy.  M^ym  i:fflW. {«irJ|. 38  Chapter II - ' G i k o ' from Style to Genre makurakotoba.  In fact the first object of this classical revival seems to have been waka  21  poetry, but the principles were soon applied to prose also. A second grouping is identified around Motoori Norinaga (1730-1801) (himself 22  a disciple of Mabuchi's), and his disciples. Norinaga's major contribution to the development of gikobun emerged from his detailed philological studies, which gave him a heightened sense of the varying lexical and grammatical features of different periods, as discussed in his Tama A r a r e  2 3  A s a result, he studiously avoided mingling vocabulary and grammar  from different periods, producing a more harmonious and less mannered gikobun. This style of his was continued by his pupils Fujii Takanao  24  (1764-1840, Sakigusa,  Shozoku  Bunrei), Nakajima Hirotani (1791-1864), and Kurosawa Okinamaro (1795-1859, Gagon 25  26  Yobunsho), apparently arousing considerable interest. The third and final gikobun grouping that is identified is that o f the gikobun novelists such as Katano Ariamaro (1703-1751, Shirozaru Monogatari) and G o i Ranshu (Zoku Ochikubo Monogatari). From Kamo no Mabuchi's school there are Takebe Ayatari (1719-1774, Nishiyama (Tsukuribune  Monogatari  Monogatari),  - 'Tale of the Western H i l l s ' ) , M u r a t a H a r u m i  Ueda Akinari (Harusame Monogatari - 'Tales of the Spring  £fc]fs3, ' p i l l o w word,' a word or phrase conventionally attached to a specific  21  following word or phrase, thus functioning as a standardized epithet. 2 2  *^Jt  ^.&btl.  2i  BJRtm.  26  39  Chapter II - ' G i k o ' from Style to Genre Rain,' Kuse Monogatari),  and Ishikawa Masamochi (1753-1830, Omi-ken  Monogatari,  Shimi no Sumika Monogatari, Miyako no Tenbun). Other literary histories and reference works  27  also usually add the name of Shimizu H a m a o m i  28  (1776-1824) to the list of  gikobun practitioners, particularly for his Sazanami Hitsuwa (1813). This rather lengthy listing should serve to show how widespread the practice of this style was, and how nearly coextensive it was with kokugaku itself. The Daijiten notes that, despite its enthusiasts, the style gradually died out and, after being labelled 'gikobun' in the Meiji period, it survived only as a scholastic exercise. W e shall see, however, that the M e i j i reflexes of 'gikobun' were in fact many and various. Although the precise style denoted by its narrowest definition - the imitation mid-Heian style promoted by Norinaga - is indeed rare, the pseudoclassical impulse certainly survived into the twentieth century in the guise of styles like bibun  29  and futsubun.  30  The Daijiten also draws attention to this style's artificiality by referring to it as a jinkoteki na bunsho, and it is interesting to investigate why this style rather than any other (for 27  E . g . Nihon Kokugo Daijiten, V o l . 5, p. 546; Yuseido Nihon Koten  Bungakushi,  p. 131.  2 9  3 0  H ^ t , 'bellelettristic prose,' an ornate, florid prose style. i a i l l f c , 'general style,' the Meiji Classical Standard after c. 1897 in newspapers,  magazines, textbooks, and government business. A 'blend of the most familiar idioms and grammatical features of kambun, wabun, wakankonkobun, p. 188.  40  and sorobun,' Twine 1991,  Chapter II - ' G i k o ' from Style to Genre example kanbun),  31  should be considered 'man-made.' This entry draws attention to the  fact that the Edo period terms used were inishieburi no fumi, gibun, and kobunji?  2  with  'gikobun' not in general use until the M e i j i period. O n further examination, one finds that in fact 'gikobun' is not in general use for this style until quite late in the M e i j i period. For example, the earliest 'modern' comprehensive history of Japanese literature,  33  the 1890 Nihon Bungakushi  Sanji (1865-1939) and Takatsu Kuwasaburo (1864-1921), kanaburf  5  principally uses the term  rather than wabun for prose in Japanese rather than Chinese, particularly for  Heian monogatari 31  34  by M i k a m i  36  and n i k k i .  37  Gabun™ is used quite widely, to refer both to the Heian  Tsuda Sokichi does include kanbun in his wide-ranging use of 'giko bungaku.'  His appears to be the broadest use of the term, and is perhaps not typical.  v^  3  *9 <d&&, six fern.  F o r a discussion of the emergence of 'modern' comprehensive literary histories  33  in Japan, see M i c h a e l Brownstein 1987. H ± # & , i^t&HJtP.  M  B * £ # 5 t  ^ £ £ .  3  5  3 6  M i k a m i & Takatsu 1890, V o l . 1. p. 218 ff.  "Describing Tosa Nikki, V o l . 1. p. 302. 3 8  H ^ t . Some later writers prefer the characters I j l X H i g u c h i Ichiyo has been  referred to as 'the last writer of Gabun,' in Chizuko Ueno, 'Vernacularism and the construction of gender in Modern Japanese Language,' in Proceedings  of the  Midwest  Association for Japanese Literary Studies Vol.3 Ga /Zoku Dynamics in Japanese Literature, (West Lafayette (Purdue Univ.) 1997), p. 8.  41  Chapter II - ' G i k o ' from Style to Genre monogatari and diaries in Japanese - the mid-Heian grammatical standard with a minimum 39  of Sino-Japanese vocabulary - and when referring to the kokugakusha (or wagakusha) authors' writing in that style. called a tsukuri-monogatari, early as the Imakagami  40  A later monogatari such as Torikaebaya is still merely  a term which does go back a long way, appearing at least as  'The Mirror of the Present,' c. 1170, whose last chapter has the  startlingly modern title, 'Tsukurimonogatari no Yukue' - 'The Future of the N o v e l . '  41  A  wide range of Heian and Kamakura period monogatari such asAsakura, Katano no Shosho Monogatari,  Nezame Monogatari,  Monogatari  are gathered together without any subgeneric distinction.  43  Ide Chujd Monogatari,  42  and Umetsubo  Shosho  44  However, by the time of the 1916 Bungaku ni Arawaretaru Waga Kokumin  Shiso  no Kenkyu by Tsuda Sokichi (1873-1961), the usage to denote the Edo style is firmly 45  entrenched enough to figure as a chapter title, Gikobungaku oyobi Kanbungaku.  46  In this  chapter the mid-Heian monogatari are called ko monogatari, contrasting with all later 39  U s e d regarding the styles of Genji Monogatari and Makura no Soshi, in V o l . 1,  pp. 318, 321. The Genji is even referred to as gabun no kyokubi naru mono - 'the most extremely beautiful gabun,' V o l . I, p. 233. 4 0  V o l . 2, pp. 344, 345, 347, 349, a propos of Motoori Norinaga.  41  Harper 1971, p. 51. Harper also gives a translation of this chapter, ibid., pp.  52-55.  #^imm. mm^mm.  42 43  ^Sanji & Takatsu 1890, V o l . 1, p. 232.  46  Iwanami edition, V o l . 6, Book I, Chapter 12, p. 76. 42  Chapter II - ' G i k o ' from Style to Genre giko monogatari, ko bungaku is contrasted with giko bungaku; and the Mikami/Takatsu usage of gabun seems to be firmly replaced by gikobun. This substitutes a perfectly clear opposition of 'original' and 'imitation' in place of the previous lexical dyads that opposed wabun with kanbun, and gabun with zokubun. A s we shall see, the first appearance of the term 'giko' seems to be in about 1896. What might account for this change in usage over a period of about twenty years? Attempting to trace the origins and changing usage of a term in literary history raises considerable problems. Given the rarity of indexes for the texts surveyed (which would permit a properly quantitative stylometric analysis), it seems that one is limited to two kinds of assertion, focusing on texts where the term might be expected to occur (in this case, texts which examine Edo and M e i j i period literary styles, or which treat late Heian and early Kamakura period monogatari). First, one can report whether the term is, in fact, used in the expected context. Second, one can try to establish whether it is the generally used term by the author in that context, and go on to note what kind of definition is functioning, and with what connotations. Tracing the shifts in usage, both by one author in his lifetime and among several authors over longer periods of time can reveal quite surprising fluctuations and redefinitions. Despite the fact that, as already noted, comprehensive literary histories of Japan do not appear until 1890, the number o f texts that need to be surveyed for this investigation is quite extensive. Not only is there a substantial number of sizeable literary histories from the late M e i j i and Taisho periods, but there is another area which needs to be surveyed when tracing the emergence of the term 'giko,' namely treatises on writing style. A s it happens, the question of writing style was one of the great cultural debates 43  Chapter II - ' G i k o ' from Style to Genre of the M e i j i period, and one that was not entirely resolved until after the Second W o r l d War, i f even then. This issue had a prominence in Japan in the last decades of the nineteenth century that is perhaps hard for most speakers and writers of Japanese and English to appreciate now. So central was the issue to the entire perception of Japan's 47  modernization project, that a leading enlightenment journal such as Meiroku devoted its entire first issue to the topic in 1 8 7 4  49  Zasshi  48  The debate over whether Japan should  adopt a standardized written style based on the spoken language or on a more grammatically conservative style came to be known as the genbun'itchi  50  - 'unification of written and  spoken language' - debate and was at its height in the last three decades of the nineteenth century. The topic has been extensively researched in Japanese, especially in the works of Yamamoto Masahide, 47  51  and the debate at the time was the subject of dozens, i f not  T h o u g h not so hard, of course, for contemporary speakers of Norwegian and  Greek to appreciate, for example, for whom the very issues that so exercised the Japanese in the late 1800s are still the subject of intense polemics between the advocates of rijksmal vs. bokmal, and demotiki vs. katharevousa, being in each case the modern colloquial and conservative literary forms respectively.  49  T h i s comprised N i s h i Amane's famous essay promoting the adoption of the  western alphabet, and Nishimura Shigeki's essay on ' W h y the reform of writing should depend on the level of enlightenment.' In Braisted 1976, pp. 3-16, and 16-20.  5 1  A n extensive bibliography, both of the works of Yamamoto and other, appears  in Nanette Twine, Language and the Modern State: The Reform 44  of Written  Japanese,  Chapter II - ' G i k o ' from Style to Genre hundreds, of publications, largely journal and newspaper articles. Soon after, more extensive histories of written styles appeared, inevitably drawing on the topics and terminology forged during the height of the genbun'itchi debates. Since discussions of style are the natural place to look for the emergence of a term that originally denoted a literary style, these genbun'itchi debates and style histories, which are to some extent prescriptive and descriptive approaches to the same subject, need to be added to the analysis of literary histories that started to be produced at the height of the genbun'itchi movement and are thus inevitably connected with it. A typical example from a literary history is in M i k a m i and Takatsu's Nihon Bungakushi points out that the style of Genji, although not 'pure genbun'itchf its dialogues, 'actually not terribly far removed from it.'  52  itself. It  was, in the speech of  This shows that at this time  even critical discourse on Heian era prose writing style could be conducted in the terminology of genbun 'itchi. The resulting corpus of texts to examine is naturally quite vast, and for the Routledge, 1991, which is also the most detailed overview of the topic currently available in English. For an examination of its strengths and weaknesses, see Richard Torrance's review in JJS 19 (1), 1993. Nanette Twine, now Nanette Gottlieb, has recently published a related text focusing on the years from 1900 until the 1990s, Kanji Politics:  Language  Policy and Japanese Script (Kegan Paul International, London, 1995), which has been reviewed by Sharalyn Orbaugh in the Journal of Japanese Studies 24, no. 1 (1998): 201-208. O n genbun'itchi  see also Kobayashi 1994.  junsui nam genbun 'itchi itaru ni wa arazarubeki mo, taiwa mondo no ku wa,  52  tadashi jissai wo saru koto, hanahada tokarazarubeshi. V o l . 1, pp. 271-272. 45  Chapter II - ' G i k o ' from Style to Genre purposes of this study it has only been possible to survey a part of the material, in the hope that it may be representatative. Certainly, the bulk of the M e i j i and Taisho comprehensive literary histories have been examined, together with the most important essays from the genbun'itchi debates and the related surveys of writing styles. It is hoped that this can give a reasonable picture of the developing discourse about style during the late M e i j i and early Taisho periods. The following table gives a chronological list of some Japanese literary histories of the M e i j i and Taisho eras that are significant in this context.  Meiji / Taisho Literary Histories 1890 M i k a m i / Takatsu  Nihon  1890 Haga Y a i c h i , Tachibana S.  Kokubungaku  1896  Koji Ruien  1898 Aston, W i l l i a m G .  History of Japanese Literature  1900 Haga Y a i c h i  Kokubungakushi  1900 Haga Y a i c h i  Kokugakushi  1905 Fujioka Sakutaro  Kokubungaku Zenshi: Heiancho hen  1906 Sassa Seisetsu et al.  Nihon Bungakushi  1906 Florenz, K a r l  Geschichte der Japanischen  1907 Haga Y a i c h i  Meiji  1908 Haga Y a i c h i  Kamakura Jidai no Bungaku  1909 Hiraide Kojiro, Fujioka S.  Kinko Shosetsu Kaidai  1911 Fujioka Sakutaro  Nihon Hyoronshi (1908-1910)  1913 Haga Y a i c h i  Kokubungakushi 46  Bungakushi Tokuhon  Jikko  Gairon  Jiten Litteratur  Bungaku  Gairon  Chapter II - ' G i k o ' from Style to Genre 1915 Fujioka Sakutaro  Kamakura Muromachi Jidai Bungakushi  ('06-09)  1916Tsuda Sokichi  Bungaku ni Arawaretaru Waga Kokumin  Shiso  1922 Nomura Hachiro  Kamakura Jidai Bungaku  1924 M i u r a K e i z o  Sogo Nihon Bungaku  Shinron  ZenshP  In a movement that parallels the debates and developments regarding prose style in several respects, poetic diction also became the object of intense controversy in the late 19th century and spawned another stylistic application of the term 'giko.' In 1882 Toyama Masakazu (1848-1900), Yatabe Ryokichi (1851-1899), and Inoue Tetsujiro (1855-1944) published the treatise Shintaishisho,  which advocated what was  essentially a 'genbun'itchi' approach to poetic diction, though several years before that term achieved general currency. Traditional alternating 5-7 syllable rhythms were to be maintained, but poetic diction was to be modernized, both in terms of vocabulary and grammar, substituting the contemporary written standard (not yet itself the equivalent of colloquial) for the classical diction that had prevailed hitherto - largely 10th century diction for the tanka, and modern vocabulary with classical grammar for the haiku. 5 3  A remaining literary history that has proved unobtainable is W a d a M a n k i c h i  and Nagai Hidenori's 1901 Kokubungaku 54  54  Soshi.  M a k o t o Ueda, Modern Japanese Poets and the Nature of Literature  (Stanford  University Press, 1983). A s he points out, 'The "new style" was not strikingly new, because its basic rhythm, the repetition of five- and seven-syllable lines, was that of traditional Japanese verse. Its vocabulary was pseudoclassical, too. Yet poems written in this style looked radically different from haiku and tanka, since they were longer, and more open,' p. 6. 47  Chapter II - ' G i k o ' from Style to Genre A s with the case of prose, the adoption of norms closer to the spoken language did not proceed at an even pace, the concept itself was controversial, and a strong anti-shintaishi reaction was felt by the late 1880s. Shintaishi itself split into variants that were either more colloqual or more classicizing in their diction. The latter emerged especially prominently in 1896, with the publication of the anthology Bibun Imbun Hana  Momiji,  55  'Blossoms and A u t u m n Leaves of Belletristic Prose and Rhymed Verse,' comprising ornate prose and classicizing shintaishi by Omachi Keigetsu  56  (1869-1925),  whose  contributions to literary and stylistic criticism we shall have cause to examine later; Shioi Uko  57  (1869-1913); and Takeshima Hagoromo  58  (1872-1967). A l l three were classmates  in the Japanese Literature course of Tokyo Imperial University, taught at various schools, and had been publishing poetry in the conservative shintaishi style in the magazine Teikoku Bungaku for several years prior to the success of Bibun Imbun Hana  Momiji.  They were known as the ' A k a m o n h a ' or 'Daigakuha' - the 'Tokyo Imperial University 59  School' of poetry. However, they also became known as the ' G i k o h a , '  60  owing to their  classicizing shintaishi diction. Most importantly, as we shall see, Omachi Keigetsu went on to become a widely published literary critic and historian, as well as poet and bibun prose stylist.  57  it#Mrl.  48  Chapter II - ' G i k o ' from Style to Genre Interestingly, both of the major European commentators on the contemporary Japanese literary scene commented favourably on the popularity of this poetic faction. W i l l i a m G . Aston in England (1899) ' regarded the ' G i k o h a ' as the Japanese poetry of 6  the future, as did K a r l Florenz in Germany (1906) , referring to them as the 'Teikoku62  Bungaku Schule.' Aston went so far as to claim that 'the day of Tanka and Haikai seems to have passed,'  63  and closed his History of Japanese Literature with an unprecedentedly  long (two page) translation of a poem by Shioi U k o from Hana Momiji, albeit qualifying it as 'the following specimen, which may be taken as characteristic of the vague and dreamy style of most recent Japanese poetry.'  64  Yet anthologies of Japanese poetry  completely ignore their works now, and Donald Keene, in his one mention of Hagoromo in his compendious Dawn to the West, (in reference to a comparison made by Takayama Chogyu of Hagoromo and Shimazaki Toson) refers to him as, 'the now forgotten poet.'  65  The second major challenge to the genbun'itchi movement, following the resurgence of conservative styles from the late 1880s to the mid 1890s  66  began just after the turn of  the century. M o z u m e Takami (1847-1928) in his essay 'Genbun'itchi no Fukano' in the Yomiuri Shinbun, December 1902, is the most salient instance of this, owing to the fact 6l  A s t o n , 1898; rpt. 1972, p.395.  62  Florenz, 1906; rpt. 1909, p. 624.  63  A s t o n , 1898, p. 396.  64  A s t o n , ibid.  65  Keene, Dawn to the West, V o l . 2, Poetry, Drama, Criticism, p. 215.  66  Nanette Twine, Language and the Modern State, Chapter 7. 49  Chapter II - ' G i k o ' from Style to Genre that having been 'instrumental in leading the way from the theory to the practice of colloquial style in 1886,'  67  with his essay ' G e n b u n ' i t c h i , ' he then recanted his former 68  support for the movement in favour of a futsubun with more modern expressions. In fact, similar public advocacies of futsubun can be found in essays like ' K o n g o no Buntai' by Omachi Keigetsu, which appeared in the magazine Taiyd, of which he was literary editor, in A p r i l 1901. W e have already had occasion to mention Keigetsu in his capacity as a ' G i k o h a ' poet, so one might indeed expect to find him favourably disposed to a more classicizing idiom. He does, in fact, take a conciliatory approach, promoting the coexistence of the two as complementary styles. It is notable that the entire essay on the future of literary styles makes no use of the 'conservative coded' terms, 'wabun' and 'gabun,' but is conducted entirely in terms of the futsubun / genbun'itchibun binarism. In his earlier essay ' K e i c h u A z e r i ' (1897), Keigetsu spoke of Keichu's works as 'wabun,' 69  (in fact the sixth chapter of that essay bears the title 'Keichu no Waka, Wabun'). B y the time of his 1914 essay 'Sakubun Jusoku,' he was speaking in terms of a similar binarism, 70  but this time between jitsuyobun,  'practical prose,' and bibun. Keigetsu does not define  'bibun' as necessarily conservative or archaizing i n style, but focuses instead on its poetic, affective nature: What is good in poetry is good in bibun. ... B i b u n is principally a style that appeals to the emotions .... In their youth, everyone is in a time of 67  T w i n e ( 1 9 9 1 ) , p . 205.  68  I n Meiji Bunka Zenshu, 20, Nihon Hyoronsha, 1967, pp. 129-139.  Meiji  69  Bungaku Zenshu, V o l . 41.  Keigetsu  10  Zenshu, V o l . 8, Section 2, 'Jitsuyobun to Bibun no Kubetsu'. 50  Chapter II - ' G i k o ' from Style to Genre abundant emotions, and bibun is welcome in the soil of youth.  71  This obliquely recalls the connection between the acceptable vocabulary and grammar of waka and of wabun. The wabun aspect of 'bibun' is clear from Keigetsu's own bibun, to which one volume of his Zenshu is devoted, and which, like the futsubun in which his essays are written, uses a wide range of Classical forms.  72  The term 'gikobun' was certainly in use by this time. But how, i f ever, does Keigetsu use it? Certainly not for the style of post-Genji monogatari. In the entire Heian and Kamakura chapters of his 'Nihon Bunshoshi' (1907)  73  it makes no appearance, and  the monogatari style is referred to as 'wabun' for the late Heian period. monogatari are, in fact, scarcely discussed, in favour of wakankonkobun  15  74  Kamakura  and kanbun  texts. ' G i k o b u n ' is used however, for wabun-style prose in the Edo period, though often alternating with 'wabun' and usually with positive evaluation.  76  O n the other hand, in  'Bungaku no Shurui' (1903), where 'gikobun' is used alone, it is in a context with I b i d . p. 242.  7l  72  F o r a fine example of 'bibun' see the epigraph to this chaper, w h i c h also  provides as favourable a characterization of 'gikobun' as one could wish for. "Characterized as 'the first comprehensive style history,' in Kokugogaku  Kenkyu  Jiten, p. 40. Nihon  14  7 5  Bunshoshi, p. 1664.  f PyJttJItffX, a style of written Japanese with free use of Sino-Japanese vocabulary  and some Classical Chinese grammatical constructions, lexically intermediate between wabun and kanbun, but grammatically closer to wabun than kanbun. 76  pp. 1711, 1721, 1722, 1726, 1746, 1747, 1748, 1753.  51  Chapter II - ' G i k o ' from Style to Genre definitely negative connotations, but not necessarily because of the style per se: Norinaga was indeed a great classical scholar and philologist, but since his waka and gikobun are rather unskilful, though he is undeniably a man of letters, they should be omitted right away from the broad sweep of Japanese literary history.  77  This sparing use of 'gikobun' seems characteristic of late M e i j i literary historians, who, while not sticking exclusively with the 'wabun' and 'gabun' that the first modern literary historians of the 1890s employed, nevertheless use these more positive terms on a more regular basis to refer to the Edo period classicising style. Omachi Keigetsu is certainly a strong candidate for playing a role in the introduction of the term ' g i k o ' into literary critical discourse, since he was associated with its original application to the classicizing shintaishi school of which he was a member, however, as already hinted at, the first use of this term in this context in the texts surveyed actually predates Keigetsu's 1903 usage. The writer in question is in fact Haga Yaichi (1867-1927) who featured prominently as the author of numerous works in the list given above of relevant literary history related texts from the M e i j i and Taisho periods. He is one of the giants of late M e i j i and Taisho literary history and criticism, not to mention wider ranging lingustic, historical and cultural studies. It is in his Kokugakushi  Gairon of 1900 that we find an early, perhaps the  earliest, use of 'gikobun' in its later widespread usage for the Edo period kokugakusha's classicizing style. In fact Haga Y a i c h i shares with M i k a m i / Takatsu the distinction of authoring one of the first two comprehensive 'modern' Japanese literary histories, his Kokubungaku 77  Tokuhon which appeared in A p r i l 1890, whereas M i k a m i and Takatsu's  p . 177. 52  Chapter II - ' G i k o ' from Style to Genre appeared in the fall of that year. However, Haga Y a i c h i ' s expository text in his 1890 publication is limited to a 42-page introduction in what is largely an anthology, containing 402 pages of cited material. M i k a m i and Takatsu's two volumes, although containing 78  ample illustrative excerpts, fill most of their 451 and 540 pages with historical exposition and analysis. The 1890 Tokuhon,  after brief mention of Kamakura period 'wabun,' presents  wabun as one of six kinds of Japanese prose current in the Edo period and notes the use of 'gabun' in this context in relation to the wabun writings of Norinaga. A t this point 'gikobun' is still apparently absent from literary historical discourse, though this section speaks of K e i c h u and K i g i n ' s imitation (naraite) of mid-Heian style and the artifice (sakui) of their sentences. In the 1900 Kokugakushi 79  Gairon there is a distinct impression  that the use of 'gikobun' is indeed a recent one, since the phrase in question, referring to Mabuchi, reads 'iwayuru gikobun wo kaita no desu' initially, then with more assurance on the same page 'gikobun wo tsukutta M a b u c h i , '  80  which seems to credit M a b u c h i with  inventing gikobun. Elsewhere on the same page and later in the text  81  'gabun' is used for  the same kind of writing, and Kurosawa Okinamaro is even referred to as a bibunka. The associations do not seem to be pejorative, as the conclusion stresses that 'writers of 78  M i c h a e l Brownstein, ' F r o m Kokugaku  to Kokubungaku  the M e i j i Period,' HJAS 1987 (47): 435- 460, pp. 443, 444. Kokubungaku  19  Kokugaku  80  81  Tokuhon, p. 202.  Gairon, p. 213.  p . 220.  53  : Canon-formation in  Chapter II - ' G i k o ' from Style to Genre gikobun must always be treated with our respect.'  82  A s such a major figure, one can  assume that Haga Y a i c h i ' s choice of terms could be quite influential. However, in his Kokubungakushi  Jikko of the same year he makes no use of 'gikobun,' using 'gabuntai'  for Kamakura period kanabun, for the E d o period.  84  83  and using 'wabun' and 'gabun' largely interchangeably  N o r does the term come to displace 'wabun' and 'gabun' in his  writing over the next few years. In his 1908 Kamakura Jidai no Bungaku  there is the  important development of the application of 'gikobun' to Kamakura period writing, though 'gabun' is still the more frequent term. of the rekishi monogatari  96  fourteenth century), Meiji period  88  87  85  'Gikobun' is in fact used to characterise the style  the Mizukagami  (late twelfth century) and Masukagami  and it is pointed out that 'gikobun' writing continued on into the  There is still only one use in his Kokubungakushi  Gairon of 1913, and that  as an explicit alternate to 'gabun' - gabun wa sunawachi gikobun, kokugakusha.  89  (mid-  for the writings of  Elsewhere in the text 'gabun' prevails for both the Kamakura and E d o  period styles. The next year in his 'Kokutei Tokuhon no Bunsho ni tsuite,' 'gikobun 82  p. 224.  83  p. 260.  84  pp. 299, 300, 303. 'chukotai no bun' on p. 306.  85  ' g a b u n ' on pp. 486 (perhaps a consciously new usage is denoted by 'iwayuru  gabun'), 487, 489 (Izayoi Nikki). M$Lffi)!m, 'historical tales,' works in the tsurkuri-monogatari written style but  &6  with subject matter based on historical figures and events. 87  p.489.  88  Ibid.  Kokubungakushi  89  Gairon, p.392. 54  Chapter II - ' G i k o ' from Style to Genre sunawachi gabun' is defined as one of the things that futsubun is not,  90  and there is no  other use of 'gikobun' - only 'gabun' - nor is there any use of either in his short but related article 'Shorai ni Hyojun Buntai ni tsuite' of 1918. Overall then, although Haga Yaichi may possibly have originated the term 'gikobun' for its Edo and perhaps also its Kamakura uses, it remained interchangeable with 'gabun' for h i m and it cannot be considered his preferred term in these contexts. A s we shall see below, for the earliest widespread use of 'gikobun' in both of these contexts it seems that Tsuda Sokichi's 1916 Bungaku ni Arawaretaru Waga Kokumin  Shiso no Kenkyu  is the  key text. Fujioka Sakutaro in his 1905 Kokubungaku fact that he discusses in detail Sagoromo,  91  no Nezame  93  and Torikaebaya,  94  Zenshi - Heiancho Hen, despite the  Hamamatsu Chunagon Monogatari,  92  Yowa  and frequently alludes to their derivative or imitative  qualities, never categorizes them as 'giko.' Similarly, in his posthumously published Kamakura Muromachi Jidai Bungaku Shi (1935, edited 1915, from lectures 1906-1909),  9 1  V o l . I l l , pp. 153 ff. 'narai' p. 153„164; 'mogi' p. 163.  9 2  V o l . I l l , pp. 166 ff. 'moho', 'Genji ni gi suru' p. 174; 'zensaku o mo shite,'  'Genji no mosha', p. 175. 9 3  V o l . 4, pp. 63 ff. 'Genji ni uru tokoro ooshi' p. 77; 'mata Sagoromo ni uru  tokoro mo aru ga gotoshi'; 'mata Genji nado no komonogatari ni genwaku shi, kore ni naraitaru mono naru ga, tsutomete sono mogi no ato o kakusan to shite' p. 78. 9 4  V o l . 4, pp. 79 ff. 'genji moho no ato rekireki tari', p. 84.  55  Chapter II - ' G i k o ' from Style to Genre although there is an entire chapter on 'shosetsu,' no Miya Monogatari,  Iwashimizu,  95  including Kaze ni Tsurenaki,  Koke no Koromo, and Sumiyoshi  their penchant for imitation is stressed,  96  Matsura  Monogatari,  and  with regard to the question of style the only  reference is to 'kobuntai.' In Fujioka's essay 'Nihon Hyoron S h i ' (also posthumously 97  published in 1911) he does use the term 'giko' in the context of early Kamakura works in the Genji style, such as the Okugisho and Shinchushd, but only to state that this awkward imitative writing was different from the 'gikobun' that resulted from the kokugaku movement of the Edo period, and the latter are also referred to in the same paragraph as 98  'gabun.' H a d he managed to extend his Kokubungaku  Zenshi through the E d o period  before his early death we might have seen just how widespread or nuanced his use of ' g i k o ' was, but on the present sampling the use seems sparing and omits Kamakura ' P e r i o d II: Kamakura E r a ' , Chapter X I V , pp. 204 ff. Individual monogatari are  95  summarized and critiqued in Chapters X V - X V I I . Note that in this text the  Shinkokinshu  is treated separately as a period unto itself, 'Period I: Shinkokinshu E r a ' , lest it be contaminated by the concept of the Kamakura period. The problem stems from the widespread characterization of the Kamakura period and its associated monogatari as lacking in imagination (or 'self-confidence' even, as some critics would have it). H o w then to account for the quality of the Shinkokinshu,  a Kamakura period work that is  widely regarded as the best and freshest imperial anthology since the Kokinshu, by these same critics? 96  pp. 225, 226, 232, 236.  97  p . 225. Nihon  9Si  Hyoron Shi, in Meiji Bungaku Zenshu V o l . 44, p. 371. 56  Chapter II - ' G i k o ' from Style to Genre period application. One possible source for spreading the use of the term 'gikobun' may well be its appearance in the novella Seinen,"  ' Y o u t h ' by M o r i O g a i  100  (1862-1922), which was  published serially between M a r c h 1910 and August 1911 in the journal Subaru. In an early episode the young protagonist Jun'ichi visits the seasoned man of letters Oishi and tentatively suggests his ambitions of becoming a poet or novelist. Oishi implies that poetic ability is largely innate and one cannot study how to write poems, but presumably with novel writing in mind he says, O f course i f you want to imitate the style of ancient literature (gikobun de kako to iu ni wa), you may have to practice their sentences and such, but I myself can't do it. M y own work may have numerous inappropriate words in it itself, but I pay no attention to that. After all, it's the brain, the mind, that's important.  101  The implication here seems to be that the time for preoccupation with the superficialities of style is now past, together with the genbun'itchi debate (which seemed to have been largely resolved by 1905), and content and the imagination are much more valorized. Earlier in this chapter it was pointed out that the use of 'gikobun' to denote Edo or M e i j i period use of the Heian 'pure Japanese' style (also known as 'wabun'), while  , 0 0  «^.  Seinen,  101  in Gendai Nihon Bungaku Zenshu, V o l . 13, Mori Ogai (Part 2), p. 160.  Translated by Shoichi Ono and Sanford Goldstein in Thomas Rimer, ed., Youth and Other Stories - Mori Ogai, University of Hawaii Press, 1994, p. 393. 57  Chapter II - ' G i k o ' from Style to Genre seemingly unknown at the beginning of the M e i j i period, was extensively used by the time of Tsuda Sokichi's 1916 Bungaku ni Arawaretaru Waga Kokumin  Shiso no  Kenkyu.  It was being used for the Edo and Meiji wabun styles, passim in Tsuda's chapters on Edo period kokugakusha wabun writings,  102  and he even uses the term 'giko shosetsu'  103  for  the long prose fiction in imitation of Heian models, written by the fascinating figure of Arakida Reijo  104  (1732-1806), who, having produced a collated edition of the Utsubo  Monogatari at the age of thirty-six, turned to both long and short fiction as well as nikki, in the gikobun style. Evidently she was a Kamakura period woman writer of monogatari born after her time, but ironically it may have been this usage of 'giko shosetsu' that suggested applying the term to the late Heian and Kamakura period genre that are now known by that name. A s it happens, 'giko' was now being routinely applied to Kamakura period works as well as Edo and M e i j i period works in the wabun style. A n early use of this broadened definition of 'gikobun' is found in the 1906 Nihon Bungaku Jiten (perhaps the earliest comprehensive Japanese literary dictionary) by Sassa Seisetsu, Yamanouchi Soko, and Ueda Kazutoshi, where it is defined as simply the use of the style of Heian monogatari, nikki, and zuihitsu in later ages (kdsei).  W5  The same dictionary defines 'wabun' as the  style of Heian court monogatari, diaries, and zuihitsu, or texts written in imitation of 102  Tsuda 1916; rpt. 1977, V o l . 7, Chapter 11, p. 344, and Chapter 12, p. 375, both  subtitled'Giko Bungaku.' 103  Ibid. p. 401.  105  Sassa 1906, p. 52. 58  Chapter II - ' G i k o ' from Style to Genre them.  106  Thus this is a period of extremely broad use of the term, since books on style  regularly referred to gikobun as one of the contemporary style options and literary critics used it to characterize contemporary authors' styles. For example, the naturalist novelist Tokuda Shusei (1871-1943) in his 1914 Meiji Shosetsu Bunsho Hensen Shi refers to Ogai's 'pseudoclassical tone' - ' g i k o c h o ' 'gikoteki na d o r y o k u . '  107  and Rohan's 'pseudoclassical endeavours' -  108  Tsuda is also using 'giko' freely to refer to Kamakura period wabun writing in this work. Here finally, amidst terms like 'gikoteki shosetsu,'  109  there emerges the term  'giko monogatari': 'What one should regard as a second aristocratic literature are the giko monogatari.'  110  The list which follows the term indicates that it is the post-Mumyo-zoshi  monogatari that are so designated: Iwashimizu, Hydbukyo,  Sumiyoshi,  and Matsura no Miya,  Kaze ni Tsurenaki, Koke no  etc. Where Tsuda refers to Sagoromo, Hamamatsu,  Koromo,  Torikaebaya,  although there is much reference to 'imitation,' they are not  designated by the term 'giko monogatari.' The term is used fairly freely in this chapter, yet, as previously remarked, it was not found in Haga Y a i c h i ' s 1908 Kamakura Jidai no Bungaku, nor does it appear in Hiraide Kojiro and Fujioka Sakutaro's 1909 Kinko Shosetsu Kaidai, or Fujioka's 1911 Nihon Hyoronshi, or his posthumously published 1915 Kamakura 106  Ibid. p. 240.  107  Tokuda 1914, p. 117.  108  Ibid. p. 164.  109  Tsuda 1916, V o l . 2, p.37.  Kizoku  n0  bungaku no dai-ni to shite kangaeneba naranu no wa giko  de aru. Ibid. p. 49. 59  monogatari  Chapter II - ' G i k o ' from Style to Genre Muromachi  Jidai Bungakushi,  nor in Haga Y a i c h i ' s intervening 1913  Kokubungakushi  Gairon. In all of these, the style of the monogatari or 'shosetsu' may be referred to as 'gikobun,' but the term 'giko monogatari' either has not been coined or is not in general use. This seems to have changed notably by the time of Tsuda's 1916 work. B y the 1920s Nomura Hachiro's 1922 Kamakura Jidai Bungaku  Shinron has a  substantial section under the title ' G i k o Bungaku,' as Tsuda did, encompassing chapters on the Mizu Kagami, Izayoi Nikki and other diaries, in addition to 'Sumiyoshi Monogatari oyobi sono hoka no shosetsu,' under which may be found the usual culprits, from Matsura no Miya to Kaze ni Tsurenaki. The Taisho histories are rounded out with M i u r a K o z o ' s 1924 Sogo Nihon Bungaku  Zenshi, which reserves the term 'giko bungaku' for the  wabun writings of the kokugakusha. Showa and Heisei literary, histories are naturally too numerous to mention exhaustively. However, examination of a substantial sample of two dozen of them, which it is hoped w i l l be representative (see chronological list appended to this chapter), reveals that there are broad fluctuations in the use of 'gikobun,' from its widest sense, which extends to all post-Heian wabun, to a narrow use that restricts it to the Edo (and sometimes Meiji) wabun. Interestingly, there is no sense that the broader use is more common in earlier works. Works consulted from the 30s through the 50s generally used the narrower, Edo-specific definition, whereas there were examples from the 70s, 80s and 90s that encompass both the Kamakura and Edo meanings. The 1965 Bunsho Hyogen Jiten is, however, adamant that Edo period wabun should not be called 'gabun' but only 'gikobun,' which is not to be used for works like Tsurezuregusa. When it comes to the term 'giko monogatari' there is much more agreement in 60  Chapter II - ' G i k o ' from Style to Genre these texts to restrict it to Kamakura period monogatari. The main area of variance is on whether to include the pie-Mumyd-zdshi  monogatari Ariake, Asaji ga Tsuyu and Matsura  no Miya in this category. O f particular interest is Konishi Jin'ichi's argument for categorizing giko monogatari. According to Konishi, ' 11  both writers of fiction and their audiences realized that no one would ever write a better narrative than that found in the superlative Genji Monogatari (The Tale of Genji). Authors reacted in two ways. One method, the next best thing to rivalling the Genji, was to create a similar monogatari through the use of related subject matter and style. The second kind of response was to experiment with unusual subjects and so create something fresh, something different from Genji. Ariake no Wakare (Parting at Dawn) is a twelfth-century work belonging to the former category. The latter category comprises three twelfth-century  narratives,  Torikaebaya Monogatari (The Changelings), Matsura no Miya Monogatari  (The  Tale of the Matsura Shrine), and " M u s h i Mezuru H i m e g i m i " ("The L a d y who Loved Insects"). The former approach set the standard for fictional monogatari from the thirteenth century on. K o n i s h i thus offers a two-tier taxonomy of post-Genji  monogatari with a two-part  subcategorization of monogatari in the second tier. The two tier taxonomy seems to owe much to the contemporary judgements of the Mumyo-zoshi, canonization such as the Fuyoshu, pronouncements in the Mumyo-zoshi,  and other indexes o f  to judge by the space allocation and critical and the number of poems from the relevant  ' " K o n i s h i V o l . 3, pp. 284ff. (Japanese), pp. 284 ff. (English). 61  Chapter II - ' G i k o ' from Style to Genre monogatari in the Fuydshu, as discussed in the preceding chapter of this study. However, the heterogeneity of the later monogatari often makes classification by subject matter problematic. The Mumyd-zdshi  judges works largely by their similarity to  Genji, and indeed on the basis of subject matter Torikaebaya has much of its plot in common with Ariake (especially with the first book of Ariake). It is thus strange to place Ariake and Torikaebaya in different categories on the basis of subject matter. A s maintained elsewhere in this study, there are extensive connections between Ariake and Genji in terms of theme, plot, narrative structure, and style. But the same can also be said of Torikaebaya, and to almost the same degree. Equally, it must be claimed that Ariake deals with 'unusual subjects' at least to the extent that Torikaebaya does, since in most regards the subject matter of Torikaebaya is a subset of that of Ariake. Neither arguments based on chronology or subject matter lend themselves to sharp distinctions for classifying giko monogatari, and given the rather small fraction of the corpus that survives, probably under 10%, all generalizations at the level of genre must be extremely hazardous. There seem to be some grounds for hypothesizing that the term 'giko monogatari' came about more as a response to language than to subject matter. If the relevant background to the emergence of the term ' g i k o ' is indeed the genbun'itchi movement, then in that context monogatari were perhaps conceived of as becoming 'giko' to the extent that their language gradually distanced itself from that of their writers and readers. It is worth recalling that M i k a m i and Takatsu, as cited above, felt that the language of the Genji was almost, but not quite 'genbun'itchi.' In this sense, pursuing a sharp division between bona fide monogatari and giko monogatari really is to pursue a chimera. Nevertheless, tracing the emergence of the term itself does throw 62  Chapter II - ' G i k o ' from Style to Genre interesting light on a period of literary history suprisingly far removed from the texts in question. The term is still very much in use in the present vocabulary of Japanese literary history. The most recent literary history which I have been able to consult, Kubota Jun's 1997 Nihon Bungaku  Shi, has, very gratifyingly, a section under the heading ' G i k o  Monogatari.'" E v e n more gratifyingly, the paradigmatic giko monogatari described in 2  rather more detail, as opposed to others that are merely listed, are Matsura no Miya and 'once thought to be lost, Ariake no  Wakare.'  u3  The most incontrovertible aspect of the term 'giko monogatari' is that it foregrounds the genealogical relationship between any text i n that category and a given text or texts that are the originals that it imitates. To that extent, it is a well-chosen designation, because one's understanding and enjoyment of the giko monogatari is immeasurably deepened by knowledge of the models from which it might be maintained the ' g i k o monogatari' are deliberately imperfect copies. The interest is in the juxtaposition of similarity and difference, the interplay of what is expected in a Genji-style, monogatari and what is unexpected. One of the most characteristic aspects of the style of narration of monogatari like the Genji is the subtle interplay between various levels of narration, including a discrete narrator's voice, relatively impersonal narration, character's speech, and internal monologue. Ariake no Wakare's  treatment of this traditional aspect of  monogatari narration w i l l be explored in the next chapter.  " K u b o t a 1997, p. 167. 2  113  Ibid.,pp. 168-169. 63  Chapter Three Telling the Tale: Narrative Structure and Narrator Although almost nothing can be said with assurance about the author of Ariake no Wakare, by a curious irony it might be claimed that the narrator of the story is one of the more developed personalities among the unnamed narrators of monogatari i n the Genji tradition. In another unusual development, the shape of the narrative structure itself, and in particular the question of whether the monogatari is in fact complete, relates in a most intriguing way to the identity of the narrator. What I wish to propose, in fact, is that the ending of the story, peremptory though it might seem, is a vital clue to who is in fact telling the tale. Whether this narrator is, in fact, coextensive with the author to any degree is a question that anyone mindful of the duplicitous author of the Tosa Diary w i l l know to approach with extreme caution. However, first one should examine the narrative structure of the tale as a whole, which w i l l bring us back to some basic aspects of the manuscript itself.  The Structure of Ariake no Wakare The Manuscript and the Editions The unique manuscript of Ariake no Wakare,  in the Tenri University Library  consists of three books of very uneven length, with Book I twice the length of B o o k III, and B o o k II two thirds the size of Book I. Since this is a unique manuscript, there is no 1  'More precisely: B o o k I: 14,985 words, Book II 10,520 words, B o o k III, 7,207 words. W o r d definitions correspond to headwords in the Obunsha Kogo Jiten, 8th edition,  64  Chapter III: Narrative Structure and Narrator sure way of knowing whether this uneven distribution of text reflects an original division, or was arbitrarily imposed by the writing material available to the Edo period scribe, or is attributable to any other cause. One may, however, examine the contents to see whether the plot structure relates in any way to this division into books. The most cursory glance reveals that the breakpoints in the text's division into books do not constitute very logical divisions on the basis of the narrative structure. Book I does end with the abdication of the Emperor and Empress, and a round of court promotions, but a much more logical break point would have been the gap of twenty years or so i n the narrative that occurs in the middle of Section 83 in the middle of this seventeenth and final chapter of Book I. This break is marked by the comment "The months and years sped by fleetingly," and is followed by a detailed presentation of the 2  various ranks everyone has attained after this span of years has passed, especially the members of the younger generation, whose stories w i l l largely dominate Books II and III. The breakpoint between Books II and III is even less logical from this perspective of narrative structure, since it occurs at the height of the sequence of two dramatic spirit possession scenes, one of which closes Book II (Chapter X I V ) , and the other which opens B o o k III. Modern editions of the text have imposed further divisions onto it, in cases where they offer anything more than the bare text itself, in addition to relocating a 'misplaced 1994. Unbound honorific auxiliary verbs, e.g. tamahu, are counted as separate words. 2  6 t s u k i 1979, p. 222.  65  Chapter III: Narrative Structure and Narrator quire,' an emendation by Otsuki, which has not been challenged since it was proposed. 3  4  The first division not found in the manuscript is in Otsuki's 1969 edition of the text, which subdivides the three books into 17, 14, and 8 chapters respectively. These chapters are narrative structural units that correspond to the major episodes of the narrative structure, at a level that recalls the scope of the chapter units of the Tale of Genji, though not the length of those chapters, since the Genji chapters are generally rather longer. Typical chapters w i l l cover a unit such as a night of surreptitious spying by the invisible Ariake (Chapters III, I V , and V of Book I correspond to his nocturnal visits to the Sanjo L a d y ' s , L o r d Nakatsukasa's, and the Sadaisho's residences respectively), or a public event like the N e w Year's and Plum Blossom Banquets in Book I, Chapter VIII; or the F u l l M o o n Banquet and subsequent night with the Emperor in Book I, Chapter X I ; or the 40th Birthday Celebrations for the Former Emperor in Book III, Chapter V I ; or the later Sadaijin's trip to Yokawa and Ohara that constitutes Chapter X I I of Book II.  Is there any reason to believe that this kind of chapter structure may reflect an original division of the text? Whether or not Murasaki Shikibu divided her monogatari into the chapter divisions that we have now, it is clear that by the time when Ariake  was  probably written in the late 12th century, the Genji was being read in chapter divisions, and had been for a long time. Sugawara Takasue no Musume writes in her Sarashina Diary that she was presented "with fifty-odd chapters of The Tale of Genji in a special sakkan, §f ft!.  3  "Otsuki 1979, p. 492. Originally proposed in his article in Biburia 34 (1966) Oct.  66  Chapter III: Narrative Structure and Narrator case," and that is certainly close enough to the current count of 54 chapters for us to 5  surmise that the division of that text in about 1020 was similar i f not identical to the one we have now. The Mumyozoshi  and Teika's Meigetsuki  also indicate that the Genji  was being read in these chapter divisions at the time when Ariake  was written and first  read. Therefore this kind of chapter division for a monogatari text was well established at the time. If that was the case then we might have expected the author and contemporary scribes to have made this chapter division themselves at the time, i f that was how the text was being read. So the absence of chapter divisions may be a reflection of authorial choice. G i v e n the reading ease and general usefulness of chapter divisions, it seems unlikely that once established they would have been discarded at some point in the textual history.  6  A smaller level of textual division is found in Otsuki 1979, where the three books are subdivided into much smaller sections: 86, 64, and 41 respectively. These are smaller logical units of narrative structure, often corresponding to 'scenes' that have more or less unity of space, time, and action. They also often correspond to text segments that begin 5  Morris, p. 46.  6  O n the other hand, Teika does talk of writing in the chapter titles to the Genji  once the copying as a whole had been finished (perhaps as some kind of concluding ceremony). If this was a widespread practice, then i f one instance of the copying of the text was broken off before it was completed, then the incomplete ms. (and it can be argued that the Ariake ms. is incomplete, see below) would lack titles - but not necessarily divisions, since space would have been left to fill them in.  67  Chapter III: Narrative Structure and Narrator with the phrase 'saru ha,' which often signals the beginning of a new episode. However, 7  of the 40 occurrences of 'saru ha' in the entire text (of which 3 can be discounted since they occur in speech rather than narrative), only 12 signal section beginnings as established by Otsuki, so this linguistic marker has rather a limited validity. A similarly intermittent 8  indicator of the end of logical units is the narrator's trailing away with an apology that her memory fails her and she cannot remember any more (as at the end of III.vi.27), but such remarks also occur in the middle of episodes, and then she gets her 'second w i n d ' and continues with the narration. So overall, there are few indications of narrative subdivision, either in the physical arrangement of the text in books, or in the use o f explicit linguistic markers.  Narrative Structure The major logical division of the text, which w i l l be apparent to any reader familiar with the Tale of Genji, is the division of the narrative focus into two sections centering on the lives of two generations of family members. In the Genji  this break  occurs after Chapter 41, Chapter 42 beginning with the ominous words, "the shining Genji was dead," and it soon becomes apparent that a substantial amount of time has passed and the dramatis personae  have substantially changed. A s already remarked, a  similar division takes place in B o o k I, Section 83, with the major difference that the major characters in the monogatari up to this point, which I shall refer to as Part I, slip into the background for what follows, Part II, emerging however, at crucial junctures. In 7  See Otsuki's footnote to the first occurence in the text, I.ii.3, p. 38, note 2.  8  1.3,1.11,1.16,1.21,1.43; 11.15,11.55,11.61,11.62; III. 11,111.35,  68  in.40.  Chapter III: Narrative Structure and Narrator terms of the logical structure of the narrative, this break is very similar in both texts, but its location differs. In the case of the Genji, the break occurs three quarters of the way through the text, whereas in Ariake  it occurs slightly before the halfway point of the  complete text. Nevertheless, it is the major division of the text as a whole in both monogatari.  Within each part of Ariake there is a natural division into major scenes, as outlined above, roughly corresponding to Otsuki's 1969 chapter divisions. However, as in the case of Genji, major scenes are often preceded or followed by miscellaneous smaller scale connecting or ancillary scenes, as in the Genji chapters, or even two major episodes are merged together in one chapter, as is the case in Genji with the illness and death of A o i being followed by the consummation episode of Genji's relationship with Murasaki in chapter nine, ' A o i ' , and in Ariake with the two major banquets and music-related phenomena in B o o k I, Chapter VIII.  The thematic structure of Ariake  can also be viewed as a double structure, with  Part I focusing on the fate of the gender-crossing Ariake, and Part II focusing on the fates of the two characters he acknowledged as his children, now the new Sadaijin and Empress, but who are in fact the children of his uncle and cousin who were respectively the Sadaisho and Sanmi Chujo in Book I. This splitting of character focus from one protagonist in Part I, to two in Part II, one of whom predominates, also resembles the shift to Kaoru and N i o u in the U j i chapters of the Tale of Genji. Spatially, there is also a structural correspondence here. The first part of the Genji focuses on life at the Capital - other than in the Suma and Akashi 69  chapters - then the  Chapter III: Narrative Structure and Narrator action moves to U j i at a certain point in the second part, where the plot of the second part really gets under way with Kaoru's involvement with the U j i sisters. In Ariake, the young Sadaijin encounters the Sanjo Lady and her daughter, later Shijo-no-Ue, at A w a z u , and much of B o o k II is taken up with his visits there (Chapters VIII-XI), and the Sadaijin's visit to the priest at Y o k a w a in Chapter X I I . B o o k III does however firmly bring the action back to the Capital, whereas in Genji the action oscillates back and forth between the Capital and U j i .  Whereas these thematic threads serve at one level to reinforce the separation of the narrative into these two parts, they are themselves significantly linked by the generational theme. In addition, there are other narrative threads that may seem to have concluded themselves i n Part I, but which then resurface in the second part, thereby providing a strong sense of unity to the monogatari as a whole. Typical of this procedure is the re-emergence of significant background characters from Part I, such as L o r d Nakatsukasa's Wife, the Sanjo Lady, and the Sokyoden Lady, all of whom had faded out in Part I, and whom the reader could have been forgiven for assuming would not appear again. This is not a procedure that is notable in the Genji, where very few characters indeed from the first part reappear i n the second, though secondary characters do quite often 'go to ground' for long periods in the first part, only to resurface at crucial points.  The final effect of this structuring is to produce a highly wrought, complex narrative that balances the separation and the interrelatedness of the two generations, much as Genji does.  70  Chapter III: Narrative Structure and Narrator  Completeness A significant question relating to the narrative structure of the monogatari is', in the case of Ariake,  as it is with Genji, the question of whether the narrative is complete  (or mostly complete), or whether there is a substantial part either lost, or intended but never written. There have been no theories proposed that postulate missing chapters, as there have been in the case of Genji (giving rise to the substantial subgenre of the Genji supplement). O n the other hand, the final sentence of Ariake 9  is considered to end with a  lacuna, as some believe is the case with the end of the final chapter of Genji, ' Y u m e no Ukihashi.' H o w indeterminate, then, is the text as it now stands? T w o important kinds of resolution have, to my mind, brought the major themes of the text very close to complete closure: (i) the deaths of a large number of major characters, without the emergence of new ones to take their place, and (ii) the resolution of significant secrets from Part I, which had largely resulted from the circumstances engendered by Ariake's gender-crossing. Regarding the deaths of principal characters, Book III is almost characterized as a book of deaths, in the same way that B o o k I might be characterized as the book of pregnancies and births. The only death occurring other than in Book III is the death of the Sokyoden N u n in B o o k II, Chapter X I I . A s auspicious as her death is presented as being, the other deaths begin soon after. First there is the death of Lady Nakatsukasa (III.i.3) after her bitter spirit possession of Oigimi (Il.xiv) and Shijo-no-Ue (IILi). This is followed by the death of the Naidaijin (IILiv), who was Sanmi Chujo in Part I. Soon after, Tai-no-Ue 9  Cf. Aileen Gatten 1982. 71  Chapter III: Narrative Structure and Narrator dies (III.v.21). The Genji L a y Priest dies in III.vii.29. Then the Daijodaijin, who had taken religious orders (Ariake's father, the Sadaijin in Part I), dies in III.viii.38. Soon after, the Kanpaku (the Sadaisho in Part I), dies, in III.viii.40. In the final section, III.viii.41, Jijii herself is on her deathbed. The cumulative effect of all these deaths is to produce a very strong sense of closure indeed. Almost all of the older generation has passed away. Even Ariake herself has already been invited to leave this world, by the heavenly maidens in III.vi.27, and she opts to stay, though who knows for how long.  The resolution of the great secrets, which were associated with Ariake's life as the Udaisho in Part I, are now in great part revealed to the key characters who need to know about them. In the Genji, the paternity secrets of Emperor Reizei and Kaoru do not remain secrets from them for the greater part of the narrative, unlike the secret paternities of the Sadaijin and Empress in Books II and III of Ariake, which constitute elements that provide a strong degree of narrative tension right up to the end of the text. This is especially heightened in the penultimate chapter of the Ariake, III.vii. Here, the Sadaijin is vigorously pursuing the Sokyoden Consort of the Crown Prince (not to be confused with the Sokyoden Lady) who, unbeknown to him, is really his half-sister. She rebuffs him out of love for the Crown Prince, not because she knows of their relatedness. That the resolution of the paternity secrets is to be a major part of the 'business' of B o o k III, along with the extinguishing of most of the older generation, first becomes clear early in the book with the focus on the relationship between the Naidaijin and the Empress. F o l l o w i n g up on a dream in III.ii.4, the Naidaijin spies on the Empress in III.ii.6, and guesses that she is really his daughter, not the Udaisho A r i a k e ' s , by his  72  Chapter III: Narrative Structure and Narrator stepsister Tai-no-Ue. He becomes anguished and i l l with possession by an evil spirit after she bears her son in IILiii. He finally manages to reveal his relationship to her in I l l . i v , after the details are confirmed by Jiju, and promptly dies (Ill.iv. 19). This theme of paternity secrets is resumed with the Sadaijin's alarming pursuit of the Sokyoden Consort in I l l . v i i , and again, at the very end of the monogatari in III.viii, where Jiju on her deathbed has finally decided she must reveal the truth to the Sadaijin. He has just arrived for this fateful conversation when the tale breaks off, maddeningly to the modern reader, but perhaps quite properly for the Kamakura period reader, mindful of the Genji with its discreetly elided deaths of Fujitsubo and Genji, and the original affair with the Rokujo Lady, always pointed to as characteristic elisions of that text. The final 10  poem, which hovers detachedly from the final, incomplete sentence, could, in fact, easily be taken as the Sadaijin's response to the terrible news: 'What I have to say to you is quite extraordinary and somehow ^Labelled ^If,  shohitsu, in the schema of Genji narrative structural devices of  the late E d o period scholar Hagiwara Hiromichi (1813-1863), in his Genji  Monogatari  Hyoshaku. For a presentation of this system, see Noguchi Takehiko, 'The Substratum Constituting Monogatari: Prose Structure and Narrative in the Genji Monogatari,'  in  Miner 1985, p. 130 ff. The genealogy of this term and others used by Hagiwara is thought to derive from M a o Tsung-kang's 1679 Tu San-kuo chih Fa (How to Read the Romance of the Three Kingdoms), via Takizawa Bakin's Haishi Nana Hosoku (^^."b&M'J Seven Rules of Stories).  73  Chapter III: Narrative Structure and Narrator terrifies me. Although you have managed to reach this point in your life,  [...] [The Sadaijin could only respond,] omohi-iri no  ominously they  tera no kane-gane  toll, the bells of the temple,  kiku kara ni  and since hearing them,  kono kure bakari  this evening is by far  kanashiki wa nashi  the greatest sadness of a l l  12  W i t h this revelation accomplished, both of the characters who had been led to believe that they were descended from the Udaisho Ariake w i l l have been disabused of this notion. The revelations provide a somewhat punitive conclusion to the  irogonomi  activities of the Naidaijin and Sadaijin, though the Kanpaku never knows of his relationship to the Sadaijin before dying. Taken together, these deaths and revelations provide as strong a sense of closure to Ariake no Wakare  as is reasonable to expect from post-Genji  court tales. If there is  indeed any text missing it seems unlikely to be very much. There remains, of course, the unresolved question of the ever mysterious, clearly semi-divine Ariake herself. But i f we are never shown the demise of Genji, how gauche then to expect a return to the heavens a la Taketori Monogatari for Ariake. After Genji, such complete closure, and such theatrical  12  B o o k III, Chapter viii, Section 41.  74  Chapter III: Narrative Structure and Narrator manifestations of the unrealistic supernatural, would be avoided by those monogatari that could be considered truly 'imitative.'  Levels of Narration in Ariake no Wakare Numerous commentators and critics of the Genji, since at least K a m o no Mabuchi in his Genji Monogatari Shinshaku,  13  have identified the various structural devices that  are considered to be one of the characteristics of the monogatari.  14  In addition to these  structural principles, another element that is regarded as highly characteristic of monogatari is the presence of multiple levels of narration. A s early as the 1528 Sairyusho  15  Sanjonishi K i n ' e d a  16  (1487-1563), such narrative categories as 'plain narrative' (ji no  bun), 'narrator's comment' (soshiji), 'authorial intrusion' (sakusha nokotoba), (kaiwabun),  of  and 'character's inner thought' (shinnaigo) were recognized,  17  'dialogue' and it has  even been claimed that the presence of the narrator is the 'fundamental, defining characteristic of monogatari literature.'  18  Certainly, all these characteristic levels of monogatari narration are to be found in Ariake, from the most 'interior' shinnaigo, through kaiwa, ji no bun, and soshiji to the most exterior sakusha no kotoba, and their interplay in various scenes creates the very  "W&ffiMffilffl,  begun c.1740, completed c.1758.  14  Noguchi Takehiko in Miner 1985, p.131.  1  6  17  Ibid.,p. 138.  ,8  Ibid., p. 137.  H ^ f f i ^ .  75  Chapter III: Narrative Structure and Narrator 'narrative polyphony' of voices that Noguchi identifies as the sound of monogatari.  19  However, in the case of Ariake, the levels of the narrator and author (potentially, but not necessarily, the same) are of special interest, both in terms of identifying the narrative standpoint of the text, and also particularly because this is an anonymous text, and any information that has implications for authorship is also of great interest.  The Narratorial Presence The presence of a narrator is signalled in the text by a variety of devices, including the following: the use of respect language that situates the narrator hierarchically vis-a-vis the characters in the text and vis-a-vis an implied listener or audience; directly expressed doubts or rhetorical questioning about characters and events in the text; implications of the 'suffix of recollection,' ki; and explicit self-references that vary from apologies for 20  memory lapses, to feelings of inadequacy when faced with the task of capturing indescribable experiences in writing. The cumulative effect of these devices is very gradually to build up an increasingly detailed presentation of a narrator with a distinct social rank, personality, and ultimately perhaps, a role in the tale that is being told. Rhetorical questions and judgemental comments are distributed fairly evenly throughout the text, but narratorial self-reference becomes increasingly explicit and frequent in Book Three, which, coupled with certain developments in the denouement of the plot, seems to invite the reader to identify the narrator with one 19  Ibid.,p. 141, 148.  kaisdnojoddshi,  20  MM(D^W)M. 76  Chapter III: Narrative Structure and Narrator of the characters whose actions are being described. This level of narrative presence is a relatively frequent characteristic of monogatari based on court life, and the level of presence can vary from the anonymous but distinct voice (or voices) in Genji Monogatari, ' 2  court ladies of the Mumyozoshi.  to the named narrators of the Okagami and the  This kind of narration seems more prevalent than the  Princeton Companion allows, with its reference to "indirection of narratorship such as is usual in a monogatari or history." Named narrators may be rare, but a strong sense of 22  narrator's presence is not.  Respect Language The range of textual devices available for indicating respect in classical Japanese, as in modern Japanese, includes honorific prefixes; honorific, polite, or humble lexical items (either replacing neutral items, or in the case of verbs, suffixed to them as auxiliaries, or replacing them); and honorific inflections such as the honorific passive and honorific causative. Typical examples are the prefixes on-, mi-, and go-; the fully lexical verbs obosu, and notamau; and the fully lexical or auxiliary verbs tamahu, owasu, and  kikoyu.  A l l of these types of respect language are used extensively in Ariake. Since the great majority of the dramatis personae are of the higher nobility, more specifically the kandachime class ('Senior Nobles), the honorific prefixes and replacement 23  21  C f . Amanda Stinchecum, " W h o tells the Tale? Ukifune: A Study in Narrative  V o i c e , " Monumenta Nipponica 1980 (35): 375-403. ^Princeton  Companion to Classical Japanese Literature, q.v. 'Okagami'  ^Ranks I through III, together with the saisho, sanmi no chujo, and sangi of Rank 77  Chapter III: Narrative Structure and Narrator and auxiliary honorific verbs are used consistently in reference to them. In the case of Emperors and Empresses, both regnant and abdicated, honorific passives and causatives are also used with great frequency (and occasionally with the higher ranking non-imperial family characters). The principal characters treated with the kinds of respect language referred to above are the following:  First Rank: Daijodaijin,  Kanpaku  Second Rank: Sadaijin, Udaijin, Naidaijin Third Rank: Sadaisho, Udaisho, Gon-Dainagon,  Gon-Chunagon  Fourth Rank: Saisho, Sanmi no Chujd, Nakatsukasa-kyo,  To no Chujd  Others: Sanjo no Onna, Sadaijin I no Kita no Kata, Sadaisho I no Kita no Kata, 24  Tai no Ue, Nakanokimi, Kannotono, Oigimi, Onnagimi I, Nakatsukasa no Kita no Kata, Sokyoden no Onna, Sokyoden Nyogo, Sen'ydden Nyogo, Tajima no Suke} Uemon no Kami 5  26  I V , who had kandachime status, rather than mere tenjdbito status to which the remainder of Ranks I V and V were restricted. 24  T h e women in this list are all the wives or daughters of high ranking nobles  whose rank presumably accounts for the keigo used regarding their women. 25  Assistant Provincial Governors of Great or Superior Provinces (suke) were  normally Sixth or Seventh Rank posts respectively. However, the character in question has a concurrent appointment that is illegible in the original ms. (but thought by O g i to be Sanmi-no-Chujo). 26  See Otsuki, 1969, Book I, Chapter I, Endnote 7.  Commander of the Gate Guards of the Right is normally a Fifth Rank post, but  the son of the Genji Prince (Udaijin I) held the post concurrently with that of Udaisho, a 78  Chapter III: Narrative Structure and Narrator Where the functioning of this respect language becomes especially interesting for inquiry into the rank of the narrator, is in the treatment of the middle and lower ranking aristocracy. The cutoff point for respect language indicates the level at which characters are at the same status as the narrator. Characters treated without honorifics. are the following:  Fifth Rank: Kurodo no Shosho, Ben no Shosho, Shonagon II Sixth Rank: Shikibu no Jo Others: Chunagon no Kimi, Chujd, Shonagon I, Jiju / Jiju no Naishi,  Shosho,  Shikibu, Ben no Naishi, Shosho no Naishi, Ben Thus the implication that the narrator is probably of Fifth Rank is consistently felt throughout the text.  27  It is especially strongly indicated by the fact that she uses no  honorifics when referring to male Fifth Rank characters. In the case of Fourth Rank males, honorifics are consistently used, but females with corresponding titles are not treated with honorifics (e.g. Chujo, Chunagon no K i m i ) , indicating that whereas a male bearer of a title is a true holder of that post, females bear those titles by association (i.e. a Third Rank post. 2 7  A similar procedure is used to characterize the narrator in the Genji. 'Because  honorifics are used in this work only of characters of a particular court rank and above, we know that the narrator must be below that rank, but high enough to be serving at court; about the same status as Murasaki Shikibu, in fact,' (Bowring 1988, p. 59). The narrator of the Genji omits honorifics for the Sama no Kami and Shikibu no Jo, Fifth and Sixth rank posts, who participate in the famous 'rainy night discussion' of the 'Hahakigi' chapter. 79  Chapter III: Narrative Structure and Narrator male relative is the true holder of that post) so it is not a precise indicator of rank, though among female bearers of such titles, a corresponding hierarchic structure seems to have prevailed, so that a female Chunagon would outrank a female Shonagon.  Rhetorical Questions and Comments One of the most pervasive reminders of the narrator's presence is the frequency with which comments are made drawing attention to the act of narration. These comments broadly fall into three categories, which I distinguish from outright references to the person of the narrator, which I shall address separately afterwards.  i) Ineffability / Redundancy Topos First among these categories are remarks on the ineffability of something the narrator is attempting to describe, or the fact that a point is so obvious that it goes without saying. This is such a widespread conventional comment in monogatari that it seems appropriate to call it the 'ineffability / redundancy topos.' The comment is supremely uninteresting and formulaic in itself, but it is interesting to see just how widespread it is in the text, and how regularly it reminds the reader or listener of the narrator's presence. The following are the comments of this type occurring in the first half of B o o k I, together with some corresponding examples from the Genji. L i . 1: 34 we ni kaku tomo nude mo oyobu-majiku ya to medetaki 'even i f someone were to paint this scene the brush would not be adequate to the task, it was so exquisite. Genji, 'Kiritsubo,' 1.9: we ni kakeru Yau-ki-hi no katachi ha, imijiki we-shi to ihedomo, hude kagiri arikereba, ito nihohi sukunashi. Tai-eki no hu-you, Bi-yau no yanagi mo, geni, kayohitarishi katachi wo, kara-meitaru yosohi ha uruhashiu koso arikeme There are limits to the powers of the most gifted artist. The Chinese lady in the paintings did not have the luster of life. Yang Kuei-fei was said to have resembled 80  Chapter III: Narrative Structure and Narrator the lotus of the Sublime Pond, the willows of the Timeless H a l l . N o doubt she was very beautiful in her Chinese finery.' (S. p. 12) I.ii.4: 40 miru me no medetasa ha sara nimo ihazu 'It absolutely goes without saying that [Ariake] was splendid to look at.' I.iv.14: 62: ihu-tomo-naku magirahashitaru kehahi mo, T need hardly add, [the Sanjo Lady] made as if she had recovered her composure.' I.vi.23: 84 onna no mi-sama iheba sara nari 'the young girl, I need hardly say, looked quite resplendent.' [Tai arriving at Ariake's in moonlight] I.vii.26: 64 mashite ihan kata zo naki ya '[Tai's mother's exasperation] was all the more indescribable' I.vii: 28: 100 ihi-shiranu sa-hou wo tsukusaruru. 'they really put themselves out performing the ceremonies in a way that beggars description.' Genji, Yugao, 4.8: kono hodo no koto kudakudashikereba, rei no morashitsu. details are tiresome, and I shall not go into them.' (Seidensticker p. 65)  ' B u t the  Ill.iv. 18 on-gokoro-madohi, sara-ni ihan-kata-nashi. 'Her Majesty's grief I am completely unable to describe. [Empress II, hearing of the death of her father, the Naidaijin] III.vi.26 yagate kaheri-noborinuru kaze no nihohi, oho-dono no uchi ni michite, kasumi watareru sora no keshiki, sara-ni ihu ha oroka nari.  nagori  'Then straight away the palace was suffused with perfumed breezes as the maidens ascended back to the heavens leaving only a mist lingering in the sky. But it would be pointless for me to describe it in any more detail.' The first example given here is especially interesting, as it invokes another topos, that of the 'picture simile,' drawing attention early on in the text to the fact that the narrator, in relaying what she has heard and overheard, is akin to a painter 'picturing' a scene after viewing it. This concern with the gaze of narrator and characters w i l l be examined further in Chapter Five.  28  For an exploration of this topos, especially in regard to the Genji, see Mostow 1997. 81  Chapter III: Narrative Structure and Narrator  ii) Rhetorical Questions / Inferences The second category of comments that draw attention to the process of narration are rhetorical questions, marked by particles or particle combinations such as ka ha, ni ya, ka, or ya, and inferences about why particular things that are being described have happened, usually marked by conjectural inflections such as -kemu, -mu, -ramu or -beshi. Once again, this kind of comment draws attention to the narrator's reasoning, and occurs frequently in the text. However, great care must be taken to distinguish between the reflections of the narrator and the reflections of characters that are being described, the latter often being signaled by niya to.  This category of narratorial comment is especially interesting, since it not only draws attention to the act of narration, but also presents the narrator as a conscious interpreter of events witnessed, further developing the role of mediator between the reader / listener and the events being related. It has also been remarked that the use of emphatic particles, particularly namu, also foregrounds the presence of a narrator.  29  Such emphatic particles are ubiquitous in  the text, but they give a slightly less salient impression of the narrator's presence than these rhetorical questions, which shall be focused on here. Classifying these comments separately as questions and inferences, while having a fairly clear grammatical basis (the presence or absence of interrogative particles) achieves 29  C f . Stinchecum 1980a, p. 382, drawing on Sakura Atsuyoshi's discussion of  namu'mBunsho  toHydgen,  1975, pp. 24-41. 82  Chapter III: Narrative Structure and Narrator little, since from a semantic perspective they both offer similar kinds of narratorial speculation. O n the other hand, one might usefully subcategorize these comments into those in which the narrator speculates with some confidence about actions and motivations, often emphasizing important aspects of the plot, and those comments in which the narrator infers and sympathizes with a character's feelings, or expresses incomprehension over them. It should be added that the narration often aligns itself strongly with the point of view of one of the characters (especially Ariake or Jiju), though without the quotative particles or use of keigo that would make this identification explicit. Such passages have been kept in the voice of an empathetic narrator rather than transferred to the associated character.  The following are typical comments that emphasize plot elements or character motivation, also offering scope for some irony: I.ii.6: 44 kaku omohi no hoka naru on-sama ni mi-nashi-kikoe-tamahiteshi naru-beshi. 'That must have been why the child was brought up to look rather differently from what one would expect.' [Why Ariake was raised as a boy] 1.111.7 kokoro shiranu hito niya ... 'One of the women (I think she didn't understand the situation) was saying,' I.iv. 12: 58 naho susumu kata ya hukakaran. 'was it perhaps that [Sanmi Chujo's] violent passion was intensifying for someone else? [An inference that Ariake was probably making] I.iv. 13: 58 koto-ni, hito-ke sukunaki tokoro-gara ha anadzurahashiki mya 'Since there was practically no-one around, I suppose there was no-one to look askance at [Ariake's] behaviour, so he must have felt emboldened,' The following are typical comments in which the narrator sympathizes with character's feelings or expresses incomprehension: 1.111.8 itodo ihan kata nakaru-beshi. 'what could [Tai-no-Ue] have said in reply [to the 83  Chapter III: Narrative Structure and Narrator Sadaisho] i.iv.9: hitasura motetsuketsuru on-mi nareba. nani-ka tsutsumashikaran. 'since he handled his role with confidence, I wonder why he was so nervous.' I.iv. 13: 58 ikaga odorokarezaran 'how could [the Lady] not be astonished [at the invisible Ariake's words] I.iv. 14: 60 yauyau hito no on-sama no natsukashisa ha omohi-shiru niya T suppose she gradually realized how deeply appealing [Ariake] was' I.vi.20: 78 sarubeki tsuide tsukuri-ide, kikoe-shirasenikeru nochi, mashite nani-gokochi ka-ha shi-tamahan 'Jiju found a suitable opportunity and explained it a l l to [Tai-no-Ue]. I really wonder how the poor girl must have felt after that! I.vi.23: 86 uki hurusato kokoro ni tomaraneba, 'ika de mo, toku ide nan' to omohu niya '[Tai] felt no attachment to her ghastly home, so I am sure she must really must have felt she wanted to leave absolutely as soon as possible' I.vi.24: 88 kagirinaki on-sama nareba, kokoro-gurushiku ahare ni oboshi-itatsukubeshi 'seeing that [Tai] was such a nice girl, [Ariake's mother] must have felt terribly upset and sorry for her.' I.vii.28: 100 tare ni ka-ha notamahan  'so would [Tai] have told anyone?'  A s this sampling from part of B o o k I shows, the presence of the narrator is signalled quite regularly by these intrusions of the narrator's own conjectures.  iii) Narrator's Judgements The third category of comment is that of outright narrator's judgements on the events and characters being described. While not as frequent as the previously detailed kinds of narratorial comments, these in fact go rather further in characterizing the narrator's point of view, so that not only is the reader conscious of the presence of a narrator, but also becomes increasingly aware of the narrator's personality. Once again, this kind of intervention is found in the Genji, though, as numerous scholars have pointed out, this  84  Chapter III: Narrative Structure and Narrator kind of personalized narrator's presence is distinctly commoner in the earlier part of the Genji than in the U j i chapters. Typical comments in Ariake tend to be about the character 30  of individuals featured in the narrative, especially the protagonist of the first part of the text, who corresponds in a number of ways to Genji in the Genji. I.ii.6: 46 kami no on-shirube ya, geni imiiikariken 'That divine revelation must have been a wonderful thing indeed' I.iii.8: 50 geni omohu koto nakaru-beki mi ni, adikinaki on-motenashi nari. 'In fact, for someone who should have had nothing to worry about, [Ariake's] behaviour was terribly anguished.' I.vi. 19: 74 to, oboshi-utomaruru zo ainaki. '[Ariake] thought contemptuously, which I find really rather regrettable.' I.vi.21: 80 sukoshi omohi-yarinaki waka-bito nite '[Jiju] being a rather callow young woman' I.vii.27: 98 amah mote-hanare, kedohoki on-arisama wo zo, 'atara, hito no hushi' to omohubekanmeru.  kuchi-woshiki  'It seems to me that he was much too aloof and distant and it appears that people must have thought what a pity he had this one feature that was open to criticism.' I.xii.54 taga on-tame mo adikinaki ya. ' H o w terrible that all this happened because of Ariake!' I.xv.70 hashidika nam niyou-go niya to itohoshiki. 'Weren't such thoughts really below a Junior Imperial Consort? It was actually quite pitiful' Narrator's judgements of this kind are in fact found throughout the text.  'Suffix of Recollection' Somewhat more problematic than the type of narrator's interventions mentioned 30  C f . Stinchecum 1980a, p. 376. 85  Chapter III: Narrative Structure and Narrator so far is the so-called 'suffix of recollection' or kaiso no joddshi. This past tense suffix is associated in some Nara and Heian texts with the personal recollection of the speaker,  31  and for a Kamakura period example, its appearance in the first part of Kamo no Chomei's Hqjoki, is considered to be indicative of his personal experience of the various disasters described. Without going into a protracted investigation of when and whether one can distinguish the 'personal' past usage from a simple past or 'hearsay' past, it may suffice to say that although this suffix is found in passages where it is clear that the narrator can not have been present, such as the scene where Ariake is accosted by the Emperor in B o o k I, Chapter X I , nevertheless, it is found, perhaps with greater frequency, in contexts where the narrator may well have been present. Similarly, the suffix 'meru,' while not primarily assumed to indicate personal observation, may indicate highly personal conjecture, which ultimately foregrounds the narrator's presence just as much. Some examples that raise the possibility of this kind of 32  interpretation follow. I.v. 16 konata ha, naka-tsu-kasa-no-miya  sumase- tatematsuri -tamahu to kikishi zo kashi.  'So this was the place where he had heard the Udaijin let Prince Nakatsukasa stay' [But there is no honorific - perhaps T [the narrator] have heard'] I.vii.27: 94 u-dai-shiyau ha, yo no hito ni sahe amah nam made atsukahare-tamahisM, nagori naku sadamari-hate-tamahinum on-keshiki ha, tare bakari naran to, omohiyari kagirinaku omohubekameredo, sono hito to sashite kikoenu wo, naho motekashidzukare-tamahu haenaku zo omohubekameru. 'Major Captain of the Right Ariake had apparently behaved rather too reservedly with people at Court [perhaps, 'we noticed']. N o w there were indications that he 31  F o r a general description see Iwanami Kogo Jiten, p. 1440, or M c C u l l o u g h ,  Bungo Manual, p. 14. 32  C f . Iwanami Kogo Jiten, p. 1435, or McCullough, Bungo Manual, p. 20. 86  Chapter III: Narrative Structure and Narrator had settled down with someone who had no lingering attachments, and they just speculated endlessly about who she might be since nobody could identify her. Moreover, it seemed they were disappointed at how she was carefully being looked after and kept from sight.' hito-me ni ha, uchi nado ni mo tatematsuri-tamahazu, hedate-arite motenashitamahishi urami ni koto yosete '[The Sadaisho] had not presented [his WifeJ's daughter at Court as a lady-in-waiting to the Emperor, which everyone was aware of [Perhaps 'we all were aware o f ] It w i l l become apparent, in fact, from evidence presented in the next section, that there are good grounds for believing that these suffixes can indeed be interpreted as indications of the narrator's recollection of personal experience rather than the mere relaying of hearsay.  Direct Self-Reference / Apologies Unlike the Tale of Genji, where the narrator becomes less overt towards the end of the tale, especially in the U j i chapters, in Ariake no Wakare the narrator's presence becomes more marked in the later part of the text, especially in Book III. This presence is signaled here by direct self-reference, usually slighting, self-deprecating references that constitute apologies to the reader for the narrator's manifest shortcomings. The typical context is that of a particularly spectacular and elegant court function, which defeats the narrator's ability to render it into words for the reader, but which calls forth more personalized comment than mere ineffability of the scene. There are five key passages of this nature in Book III: III.Hi.8 ohoyake-zama ni-te, sama kahari, itsukashiki koto wo tsukusaretaru, sara-ni kakitodomemu kata nashi. anagachi-ni mono-oboenu onna nareba, mi dani wakazu. [The post-natal ceremonies] were performed publicly and differently from usual. They were absolutely as majestic as could be, and there really is no way that I can capture it in writing. I really am a dreadfully forgetful woman, since I don't even understand what I see myself. 87  Chapter III: Narrative Structure and Narrator IILiii. 12 rei no sono hi no sa-hou, ika-de-ka  mi-todomemu.  A s is so often the case with me, I really have to wonder how on earth I could capture the sight of the Fiftieth Day ceremonies. III.vi.24 ihan-kata-naku omoshiroki ni, shi-gaku nani-ka no hodo, rei no mi-gurushiute morashitsu.  monomanebi  Words cannot express how charming it all was as the musicians were all thoroughly rehearsed. In fact, as I often find, it really taxes me quite cruelly to capture such things in writing to let you know just what it was like. ko-mono nanigashi to tatage-atsumetaru nado ha, rei no hakanaki onna no kakitsudzuku-beki ni mo arazu. Baskets of fruit had been put together and placed here and there in the garden as offerings, but I can hardly be expected to be able to write a complete description of the whole scene in detail, typically inconsequential woman that I am. IILvi.27 kaherase-tamahu mo oboezu.  gi-shiki, hito no notamahi-aherishi koto-domo, rei no hito-tsu  Regarding the Former Emperor's ceremonial return to his residence and everyone else's comments on the situation, they were as usual, and I cannot remember a single thing about them. These asides really do give a remarkably strong sense of the narrator's presence, as well as furnishing us with some very welcome information about her identity. Especially important are the two references to herself as a woman (IILiii.8 and IILvi.24). The narrator of the text is explicitly a woman then, though this does not necessarily tell us anything about the gender of the author. In addition, IILiii.8 with its use of 'mi-wakazu' indicates that she really is talking about personal experience in this instance, not her difficulty in recalling what someone else has told her. This may increase the likelihood that the ' - k i ' suffix actually is being used for personal recollection elsewhere i n the monogatari, and strengthens our perception that the narrator really was present at many of the events described. It is also noteworthy that there is a distinct intensification of the  88  Chapter III: Narrative Structure and Narrator narrator's presence and identity during this last book. A s we shall see, this is probably no coincidence. Is this narrating presence necessarily to be considered a single entity? Both Richard Bowring  33  and L y n n M i y a k e  34  make the case for perceiving multiple narrators in the  Genji, based on the different kinds of narrator-foregrounding devices described above. B o w r i n g postulates two overt narrators and two covert narrators, the two overt narrators corresponding to the explicit intrusions or soshiji, and the rhetorical questions or conjectural suffixes, and the two covert narrators corresponding to the merged discourse of narrator and character, as in the presentation of the dying K a s h i w a g i ' s thoughts, and to the 'equivocal narration' that seamlessly shifts perspective between several characters. M i y a k e notes that 'although the narrators appear one at a time, there are many voices, for it would have been logically impossible for any one narrator to have been i n all places at all times. Yet because the narrators appear so unobtrusively, a reader may not always be aware of their plurality.' Yet both of these arguments seem to require a 35  very restrictive definition of unitary narration. It does not seem unreasonable to assume that a single narrator may shift between differing modes of narratorial salience, and between directly observed, overheard, and hearsay evidence, without forfeiting their subjectivity.  To this observation, one may add another line of enquiry. The argument for a 33  B o w r i n g 1988, pp. 58-67.  34  M i y a k e 1993, pp. 78-87.  35  M i y a k e , p. 78.  89  Chapter III: Narrative Structure and Narrator unitary narrator might be plausibly reinforced by aspects of the text that do in themselves suggest a specific identity for the narrator. A s it happens, in Ariake no Wakare there are a number of clues to just such an identity for a unitary narrator.  Conclusion: The Identity of the Narrator Constituting another important difference from the Tale of Genji, and indeed from most other court monogatari, all of the preceding information about the narrator leads to a plausible identification of the narrator with one of the characters in the tale itself. The intersection of characteristics indicating that the narrator is a woman of about Fifth Rank, who personally observed much of the sequence of events described in the story, and is still living to tell the tale, leads to a strong suspicion that the tale is being told by none other than Jiju herself. Since Jiju was originally an attendant on Tai-no-Ue, and then passed into the service of the Former Empress as Jiju-no-Naishi in IILiv. 14, she was ideally placed to know all of the secrets regarding Tai-no-Ue's children's secret paternity, as well as the history of Udaisho Ariake / the Former Empress. In fact a number of scenes are explicitly narrated from Jiju's point of view, such as the Sadaisho's molestation of T a i - n o - U e , T a i - n o - U e ' s abduction by Ariake, which she observes approvingly, 36  37  36  and A r i a k e ' s  I . v i . 2 1 . F o r example: 'Just then, Jiju managed to make out the absolutely  unmistakable form of someone behind the screen, even though it was very dark there. She recognized the Sadaisho's Wife and thought...' 37  I.vi.24: ' Y e t she had certainly seen some things herself in that house and had  got only too used to them, so she was certainly not at a loss.' W i t h these thoughts and 90  Chapter III: Narrative Structure and Narrator unconsummated intimacy with Tai-no-Ue. That the last prose words of the tale should 38  be hers seems no coincidence, and in fact the final poem could also be Jiju's lamentation at what terrible news she had to break to the Sadaijin (HI.viii.41).  39  Moreover, the text shows considerable sympathy for a number of attendants at others, Jiju's point of view closes this particular scene. 38  I.ix.37. 'The sight of the two of them lying side by side as he poured out his  heart to her was quite remarkable, yet the lady-in-waiting Jiju who was observing all this at some distance must have thought something terrible was happening.' 39  It is especially interesting to note that Wilfrid Whitehouse, in his discussion of  the narrator and author of Ochikubo Monogatari in the appendix to his translation, also admits that ' I f the reader forms the opinion that the writer portrays the character of A k o g i , the L a d y Ochikubo's attendant, with more autobiographical feeling than the rest of the characters and that a hint to the same effect is given by the closing words of the story [ ' A n d she who was formerly called A k o g i became Naishi no Suke. A n d it is said that the Naishi no Suke lived to the age of two hundred years' (Whitehouse and Yanagisawa 1935, p. 234)], his feeling w i l l agree with my first impression. M a n y details of style and expression certainly do point to a man as author, yet still I feel reluctant to give up my first impression that a woman of the same station in life as A k o g i was the author, even though the whole weight of the opinions of all the experienced commentators is against this conjecture.' (Whitehouse and Yanagisawa 1935, p. 239.) More recently the gender of the author of Ochikubo is thought to be uncertain, but it has been argued that there were two and the second author (of Books III and I V ) was probably a woman (Fujii Sadakazu 1989, p. 409). 91  Chapter III: Narrative Structure and Narrator about her rank, not just Jiju. There are several sympathetic references to the attendants of the Udaisho Ariake, who feels sorry for their loss of patronage occasioned by his death (I.xv.69). Their corresponding elation on being taken into the service of the Junior Consort (I.xiv.66), or young Sadaijin (I.xv.69), is also stressed. The key role played by Jiju is given great prominence, especially in the scenes with the young Tai-no-Ue (I.iii.vi), the dying Naidaijin (Ill.iv), and the Sadaijin at the end of the story (111.41), and the narrator seems to share with the female Shonagon her celebration at the great good fortune of the Empress Ariake, to whose trials and tribulations as Udaisho, Shonagon also seems privy (I.xvi.74). M a c k Horton has stressed the important role that ladies-in-waiting play in the Tale of Genji, 'the fictional world of the shining prince and the structure of the narrative evoking that world would utterly collapse without them.'  40  Not only are they significant  players in the plot, but they frequently act as narrators within the tale itself, as when U k o n tells of Y u g a o ' s life after Yfigao herself has died enigmatically. Developing the 'oral performance theory of monogatari' of Tamagami Takuya, Horton links this role to 41  the regular duty of ladies-in-waiting to read aloud to their mistresses, and notes how their function as oblique purveyors of narrative information corresponds to the ficelle  roles of  characters in the novels of Henry James, a character like U k o n being 'characterized just 40  Horton 1993, p. 95. '  41  Tamagami Takuya, "Genji monogatari no dokusha - monogatari ondokuron,"  in Tamagami 1966: 247-265, and "Onna no tame n i , onna ga kaita, onna no sekai no monogatari," ibid., 432-440.  92  Chapter III: Narrative Structure and Narrator enough to give her plausibility without detracting from the centricity of the main character.'  42  Just like her counterparts in the Genji, Jiju 'serve[s] the rhetorical purposes of the author by moving the plot forward,' thanks to her 'willingness to look with favor on male suitors and to encourage their diffident mistresses to accept them,' just as Jiju persuades Tai-no-Ue to elope with Ariake at the end of Book I, Chapter v i .  43  Tamagami's model  posits that 'events in the story are witnessed by various ladies-in-waiting, who then years later tell their stories to another lady-in-waiting while adding their own occasional comments. That second woman, the scribe and editor, records those accounts with additional personal commentary. That written account is then read aloud by yet another lady-in-waiting to an audience, with further personal remarks.'  44  However, in the case of Ariake no Wakare, we have a lady-in-waiting, Jiju, whose life spans all the key episodes in the text from Ariake's young adulthood on. She is by far the most important of the likely informants for the story, relying on others only for details of Ariake's childhood and the younger Sadaijin's romantic peregrinations, since she serves both Tai-no-Ue and Ariake as Empress and Former Empress. Lastly, i f the narrator is in fact Jiju, what could be more appropriate than to have 42  Horton, pp. 97-99.  43  Ibid, p. 101.  93  Chapter III: Narrative Structure and Narrator the narrative break off when she is on her very deathbed,  45  suggesting that we have the  story of the great events she witnessed in her life, up to the point where she was no longer able to write them down? After all, as narrations of one's life, autobiographies must, by definition, always be incomplete. They cannot help but lack the narration of life's closure, which is death. The final impression that one takes away from the text is that it really is, in a variety of senses, Jiju's story.  Horton also notes (p. 105) that such nyobo are kept peripheral, and they 'appear and disappear according to the dictates of the characterization of the principals and the emplotment of their stories, not for the intrinsic interest of their own situations.' Especially interestingly, ' o f the eighty or so nyobo named in the text, we are told of the death of only one.'  94  Chapter Four The Language of Ariake no Wakare: 'Giko' - from Genre back to Style Chapter T w o was concerned with the genealogy of the term 'giko' and the process by which it designated first a literary style, and then a literary subgenre, moving from a primarily linguistic designation to one with largely content-related connotations. This chapter seeks to return to the original focus of the term, to answer the related question, ' H o w may the language of Ariake no Wakare be characterized?,' in particular addressing to what extent, and in what sense, the language is indeed 'giko' or not. G i v e n the vast body of rules that define a language in toto, the scope of this chapter must necessarily be highly selective in terms of the linguistic features on which it shall focus. Since the presentation of Ariake's language is intended to situate its usage in the context of comparable preceding and roughly contemporary texts, the scope of this part of the study is strongly limited by the kinds of pre-existing linguistic studies of Heian and Kamakura monogatari. W i t h this in mind, the focus of this chapter w i l l be 1  limited to two linguistic aspects of the text, one lexical and one primarily grammatical, (or perhaps, more properly, sociolinguistic, involving both lexicon and grammar). The aspects under examination w i l l be 'subjective adjective' usage, and respect language or keigo. Early literary historical references to the texts now treated as 'giko monogatari' tended to dwell on the extent to which their language conformed to Heian era monogatari usage as established by Genji Monogatari, setting up a linguistic relationship between later monogatari and these texts that is quite comparable to the relationship between the 'See the bibliography for principal representatives of such studies. 95  Chapter IV: The Language of Ariake no Wakare Kokinshu and the subsequent imperial anthologies of waka poetry. Edo period kokugakusha were the first to codify this kind of linguistic relationship between the Genji and post-Genji monogatari, and it is the most characteristic feature associated with giko monogatari in 2  M e i j i and Taisho literary histories, when style was at least as salient a literary issue as 3  content was.  "''  A s was traced in Chapter T w o , the early Showa period saw a foregrounding of content issues and the 'literary imagination' in its widest sense, over questions of style, and the resulting association of 'giko monogatari' with imitation rather than authenticity has continued to overshadow the term up until the present. In what follows, I intend to 4  show that the level of 'conservatism' in linguistic usage in Ariake no Wakare is indeed quite considerable, measured in the two areas under consideration. In fact, from what evidence is available for other giko monogatari, there seems a strong indication that Ariake is more conservative than most. O n the other hand, some pronounced idiosyncrasies also emerge, especially in the lexicon. 2  M o t o o r i Norinaga, in his essay 'Style"Bunsho,' dating from about 1757, he  calls the language of the Genji the only 'true' Japanese (cited in Harper 1971, p. 103). 3  E . g . Haga Y a i c h i 1908, p. 484, which speaks of style being a matter of 'heritage'  seshu in the Kamakura period; Tsuda Sokichi 1916, p. 49, charecterizes the style of giko monogatari as gikobun, drawn directly from the Genji. 4  E . g . Konishi 1985, V o l . I l l p. 284: 'One method, the next best thing to rivalling  the Genji, was to create a similar monogatari through the use of related subject matter and narrative style'; Kato 1979, seems to make no remarks on style or language i n his discussion of giko monogatari p. 191 ff. 96  Chapter IV: The Language of Ariake no Wakare  The Lexicon of Ariake no Wakare H o w to compare so vast a topic as the lexicon of a substantial prose work, especially when the most natural comparison is with a text as substantial as the Tale of Genjil The physical extent of a text is no fixed indicator of the scope of its lexicon, but in the case of the Genji, the size of the text does indeed offer a premonition of the nature of the lexicon contained therein. The lexical richness of the Genji is in fact quite comparable with the fabled breadth of vocabulary to be found in the plays of Shakespeare, only i n this case the assertion holds on the basis of a single text rather than an oeuvre. B y one count, the 207,808 'running w o r d s ' ('tokens,' words counted according 5  6  to their number of occurrences) of the Genji represent 11,423 different 'headwords' ('types,' word entries in a dictionary, that are 'lemmatized,' i.e. different inflections are subsumed under one heading). These are the necessary figures that are required to make the fundamental measure of vocabulary richness, the 'type/token ratio,' which is frequently used to characterize the lexicon of a work or author, or to chart the fluctuations i n richness over time or between works. However, there are two problems with most 7  published information on the lexicon of the Genji. O n the one hand the information tends to be incomplete - most Genji dictionaries, i f they give citations in addition to definitions, 5  M i y a j i m a Tatsuo, Koten Taisho Goi-hyo (Kasama Shoin, 1971), pp. 334-335.  T h e Japanese term is nobe-go-su. 7  For a fuller explanation and discussion, see Anthony Kenny, T h e Measurability  of Literary Phenomena,' in Kenny, The Computation  of Style (New Y o r k : Pergamon  Press, 1982). See also T. H . Howard-Hill, Literary Concordances: A Complete for the Preparation of Manual and Computer Concordances (Pergamon, 1979). 97  Handbook  Chapter I V : The Language of Ariake no Wakare either give only a small number of exemplary citations, or cite all the page numbers where occurrences may be found, rather than the absolute number of occurrences. Since there are frequently more than one occurrence of a word on a given page, no useful statistic can be derived from counting these citations. O n the other hand, where exhaustive counts are available, as in the case of the Miyajima's Koten Taisho Goi-hyd, these are often restricted to 'content' words, and very high frequency words such as particles or auxiliaries are omitted, though these are often highly sensitive indicators of style. Similarly, one needs detailed information about the definition of ' w o r d ' being used when it comes to categories like auxiliary verbs. Without such information, even complete counts given for 'verbs,' as in Miyajima, must be used with great caution. What methods and materials are available for investigating the lexicon then, and what aspects of it are susceptible to being investigated and compared? In addition to the published dictionaries and Miyajima's statistical study, a public-domain text of the Genji is available, and although a text without word boundaries does not readily submit to the automated compilation of wordlists and concordances, manual searches and counts using a search function facilitate the investigation of words that are not of extremely high frequency. For Ariake, Otsuki's 1969 edition contains a lexicon of non-function words, indicating all pages on which they occur, so for lower frequency words accurate counts can be obtained by checking manually for multiple occurrences on the pages listed. For the Genji there are also various Japanese studies of particular aspects of the lexicon, and one statistical study in English, of subjective adjectives, Andrew A r m o u r ' s 1985 article, 'Analyzing an Author's Idiolect: Murasaki Shikibu.' 8  In Poetica 21, no.2 (1985): 164-180. 98  8  Chapter I V : The Language of Ariake no Wakare When it comes to characterizing the lexical usage of an author or text, there are two obvious approaches using contrasting kinds of words as indicators of idiosyncratic usage. The first kind are content words, in the kinds of usage that may appear salient even just on attentive reading, the 'favourite' words, or words used with a characteristically special meaning, or regularly i n specific contexts. Sei Shonagon's wokashi, and the Genji's aware come readily to mind. More precise statistical investigation can indicate exactly to what degree a given word is favoured in one author or work, and how precisely it compares with another author or work, thereby confirming or nuancing one's informal impressions. The second kind of words, function words, are more difficult to perceive in an author's relative preference, since they are so ubiquitous. Nevertheless, characteristic preferences can constitute a veritable 'fingerprint' of a text or author, and one can use lexico-statistical information either to reveal a pattern one would never have suspected, or to confirm or refute a hypothesis based on informal observation. Examples of both 9  kinds of study can now be supplied for Ariake no Wakare. Regarding content words, Armour's article mentioned above provides a convenient starting point for a comparative investigation of a subcategory of the lexicon of Ariake T^or an account of the identification and use of one such 'fingeprint,' contrasting usage of 'while' and 'whilst,' subsequently used for authorship ascription, see the account of the investigation of several anonymous Federalist papers, in Robert L . Oakman, Computer Methods for Literary Research (Univ. of Georgia Press, 1984). For a more extended account, see Ivor S. Francis, ' A n Exposition of the Statistical Approach to the Federalist dispute,' in Jacob Leed, The Computer and Literary Style (Kent State Univ. Press, 1966). 99  Chapter I V : The Language of Ariake no Wakare and Genji. His focus of interest is in examining Genji chapter groupings by style, specifically examining the distribution of a 'basket' of 50 high frequency subjective adjectives which might be considered more independent of subject matter than some other words, yet more characteristic of the author's style. B y 'subjective adjective' he understands words or locutions like aware naru ('touching'), wokashi ('charming'), ayashi ('untoward'), yoshi ('fine'), itohoshi ('pitiful'), to give the five most frequent in the Genji. Even a simple ranking like this is not without interest, since it adds valuable detail to the generalization that the characteristic concern of the Genji is aware, but that of Makura no Soshi is wokashi.  10  W e can see here that since wokashi is the second most popular subjective  adjective in the Genji, for a meaningful comparison more information is needed than just the first-place holder alone. Investigating Makura no Soshi, using Miyajima, we find that aware, although in fifth place in Makura, has a relatively low number of occurrences (to switch from the rank of a word's frequency to the absolute number of occurrences:)  Word  Genji Rank Genji Count Makura Rank Makura Count  aware naru  1  814  5  87  wokashi  2  643  1  446  ayashi  3  508  6  74  yoshi  4  489  2  181  itohoshi  5  359  13  27  Thus the usage can be much more interestingly contextualized by the addition of a small amount of further statistical information. W e can now say that Sei Shonagon's 10  M o r r i s 1967, note 1157, citing an article by Tanaka Shigetaro in  (November 1964): 141. 100  Kokubungaku,  Chapter IV: The Language of Ariake no Wakare relatively infrequent use of aware distinguishes her writing from that of Murasaki Shikibu's more than her favouring of wokashi does, since wokashi is relatively frequent in the Genji too. This also cautions against generalizing that the aware I wokashi divide captures a different focus of the two authors in a simplistic pathos vs. charm sense. It is perhaps revealing that Sei Shonagon's lower usage of aware is not indicative of a significant avoidance of negatively toned adjectives. If we fill in the missing third and fourth places for subjective adjectives in Makura, we find medetashi (140 occurrences, only 13th place in Genji with 212) but nikushi (123 occurrences, 48th place in Genji with 92). This immediately confirms (and quantifies) one's intuition that Sei Shonagon, although more drawn to the wokashi and yoshi, is also much exercised by 'hateful things.' Even this tiny sampling of words might lead one to observe that, as close as Murasaki Shikibu and Sei Shonagon were in time period, social rank, and the social milieu about which they wrote, their preferred subjective adjectives are significantly different, which raises an interesting question of whether differences of temperament or of genre are more at work here, a question which must be addressed elsewhere. What would a similar comparison have to tell us about Ariake no Wakare compared with Genji Monogatari!  Ariake was written nearly two hundred years after Genji, but i n  the same genre, in fact in a subgenre generally regarded as highly imitative. Furthermore, the social milieu described, and many plot elements and dramatis personae are notably similar. W i l l lexical statistics merely confirm and quantify the high degree of imitativeness, or might they point out some idiosyncrasies that disrupt the assumption that giko monogatari have a lexicon and usage that is also giko? After a l l , as shown in Chapter T w o , the earliest uses of the term seem to foreground language more than content and genre. 101  Chapter I V : The Language of Ariake no Wakare Taking the same notion of the 50 most frequent subjective adjectives, their frequency and distribution in Ariake were examined, to get some kind of measurement of 1) the variation between books, 2) the similarity of their order of frequency in Ariake vis-a-vis the Genji, and 3) the 'proximity' of Ariake to Genji in this regard, by comparison with a work like the Pillow Book. The findings are then tabulated (see Tables 1-2). Given the disparity of size of these works (Ariake being roughly a quarter of the size of Genji), it is more informative to focus on the order of preference of adjectives, rather than the brute number of occurrences. So what is compared is the relative rank of the adjectives rather than their absolute number of occurrences, though that is also given, and is what the rank derives from. Thus, i f the Genji-Ariake  'Rank Difference' for a  given adjective is zero, then, relative to the other adjectives, the two works favour that adjective to an identical degree. A large positive number indicates that the adjective in question is much more favoured in Ariake, a large negative number indicates that it is used much less. If the 'Rank Difference' column is mostly small numbers, then the overall usage of these adjectives is quite similar. If there are lots of large numbers, then the usage is notably different. If one takes the sum of the absolute values of each Rank Difference (i.e. converting all the negatives to positive) one gets a figure that is some kind of overall index of difference (regardless of whether each individual case was a positive difference or negative difference). B y looking at this total at the bottom of the Rank Difference columns, one can thereby assert, for example, that, characterized by these adjectives, the second book of Ariake is more like Genji than the other two, but the degree of difference is still not that great. 102  Chapter IV: The Language of Ariake no Wakare In fact, overall, the Ariake usage conforms to Genji usage to a quite surprising degree. The ten most frequent adjectives in the Genji are very close to the ten most frequent adjectives in Ariake, with aware taking pride of place again. Only three adjectives from the Genji 'top eight' are absent, most notably itohoshi being displaced (and by a long way) by medzurashi. This is, in fact, quite in keeping with the greater prominence of the supernatural in Ariake. Even taking the twenty-five most frequent of these adjectives in Ariake, only nine are displaced from their ranking in the Genji by more than eight places. So not only are the same words used, they are actually used with a very similar degree of preference. Nevertheless, there are some fascinating exceptions in the handful of adjectives with really large rank differences. Attention has already been drawn to medzurashi. If one focuses on those with a rank displacement for more than 20 places, (one finds only four: tsurenashi, osoroshi, yukashi, and nikushi), of all the top 50 subjective adjectives in Ariake, by far the greatest displacement from its Genji ranking is found in the case of tsurenashi. A s this adjective features prominently in Ariake's  'leitmotiv' poem, M i b u no  Tadamine's ariake no tsurenaku mieshi wakare yori akatsuki bakari uki mono ha nashi, which is invoked regularly in the text, this quirk of intertextuality readily accounts for many, but not all of these occurrences, which outnumber the number of apparent allusions to the poem. What this draws our attention to, is the fact that not only are larger segments of the leitmotiv poem frequently quoted, as is generally remarked on in commentaries, but other aspects of its poetic diction are more subtly invoked, resulting in a much greater salience for this adjective than one finds in Genji. The implication is that these adjectives with a large rank displacement offer some kind of indication of where Ariake's stylistic 103  Chapter I V : The Language of Ariake no Wakare individuality lies. Examination of the distribution in the three books of Ariake (see Tables 3-6), also shows that the subjective adjective preference is relatively consistent across the different books. If one switches now to function words, even taking a much smaller set of words, given the exigencies of the statistics available for the Genji, one finds a similar collection of evidence, with overall notable conformity to Genji usage disrupted in a few cases by what appears to be the individual character of Ariake's  style as constituted in lexicon.  G i v e n the constraints imposed by grammar, and their relative independence from the vagaries of subject matter, it is to be expected that usage of function words, especially particles, will show much less variation between texts and authors than nouns and adjectives do. Nevertheless, there is scope for stylistic preference to be marked, especially in combinations, and where two or more function words are largely synonymous. Cases in point would be exclamatory particles such as zo, koso, shi, namu, and intensifying adverbs such as ito, itaku, imijiku, and itodo. Statistics for Genji particles were not available in Miyajima, but a number of intensifying adjectives and interjections were available, and a tabulation of their rankings for twenty of these is given. O f these, no less than thirteen can be seen to occupy exactly the same rank of frequency of usage, or are only displaced by one place, with ito being the favourite intensifier by far in both texts. None of the others are displaced by more than five places. The question is, does this ordering represent a fairly fixed relative usage for these words over a wide variety of texts, so that the similarity evidenced here reflects not the closeness of Ariake's  style to Genji, but a more or less prevalent 'classical' usage of these words?  A s with the case of the subjective adjectives, one wishes lexical statistics were available 104  Chapter I V : The Language of Ariake no Wakare for a wider range of monogatari, but comparison with the Pillow Book is instructive. Examining the corresponding data for that text, it is immediately apparent that more words are displaced in rank order relative to the Genji, and the displacements relative to the Genji are also often larger than Ariake's are. Especially notable is itodo, down seven places from its ranking in Genji." When the amounts of displacement are totalled, one finds that Ariake, despite its greater removal in time from the Genjiby comparison with the Pillow Book, is nevertheless closer to the Genji in its relative preferences for these words than the Pillow Book is. The total 'displacement distance' between Genji and Ariake is only 30 places, whereas between Genji and the Pillow Book it is 48 (see Table 7). Returning now to the subjective adjectives, i f one adds the evidence from the Pillow Book and examines the kind and degree of differences in usage, a number of interesting points emerge. Expanding coverage from the five most frequent subjective adjectives in Genji to a set of fifty, one finds that by comparison with the relationship between Genji and Ariake there are more large displacements and some are very large indeed. O f the twenty-five most frequent of these adjectives in Ariake only four had really large rank displacements greater than twenty places different from those adjectives' places in the Genji. This increases to six for the Pillow Book, and the size of the greatest displacement is also disparate: 26 places for tsurenashi in Ariake relative to Genji usage, but 44 places for nikushi in the Pillow Book relative to Genji usage. When the placements " A spot check for occurrences of the collocation ito itaku revealed 15 in Ariake but only 3 in the much larger Genji, suggesting another idiosyncracy of Ariake's perhaps of later monogatari in general). 105  (or  Chapter I V : The Language of Ariake no Wakare are totalled, the results are even more striking than they were for the interjections and intensifying adverbs: the total displacement distance between Genji and Ariake is 474 places, but between Genji and Makura no Soshi it is 563 places. In both lexical subcategories then, Genji and Ariake are much closer to each other than either are to Makura no Soshi. These findings are extremely preliminary, but they seem to suggest a fruitful line of enquiry into the relationship of the lexicon of Ariake no Wakare with that of the Genji Monogatari. One would especially like to know how Ariake relates to other giko monogatari in this regard. It would be very interesting to be able to rank the language of a number of giko monogatari in relation to each other and in relation to the Genji. In particular, since there are such strong thematic similarities between Ariake and Torikaebaya, the two would make an exceptionally interesting comparative lexicostatistical study. It is to be hoped that, with the advent of a greater number of computer-readable texts, accurate lexical counts distribution studies for all parts of the lexicon, including function words, w i l l become available in the near future.  12  12  Another possibly fruitful line of enquiry might be to compare sentence openings.  Otsuki points out that saru ha is frequently used in Ariake (12 times out of the 28 occurrences, by my count). Genji has only four, but they are all sentence initial.  106  Chapter I V : The Language of Ariake no Wakare  Respect Language in Ariake no Wakare Given that the dramatis personae of a court tale like Ariake no Wakare are, very largely, the highest echelons of the aristocracy, respect language plays an extremely important role in characterizing the language of the text. A s with other aspects of the language of the text, it is interesting to situate it with regard to other similar texts, both of its own era and before, to identify possible relationships, or idiosyncrasies. G i v e n the issues surrounding the relationship between Ariake and the giko monogatari genre, it is important to try to establish the extent to which its language may be considered 'giko,' or imitative, in the same way that its themes and various stylistic features may. In this regard, a particularly apropos study of keigo is Negoro Tsukasa's 1991 Genji Monogatari  no Keigoho,  13  which takes into account much previous scholarship.  The seventh chapter, ' K a m a k u r a Jidai no Bungo ni okeru ' t a m a h u , " '  4  provides a  characterization of the use of keigo in Kamakura period manuscripts of the Genji, and in giko monogatari, against which findings from Ariake may be measured. Earlier in the book, Negoro provides a chart of keigo usage in the Heian period 15  that divides both honorific and humble keigo into three principal levels. The most intense 13  M e i j i Shoin.  14  Negoro 1991, p. 97. The chart is based on Tamagami Takuya's analyses in  Genji Monogatari Nyumon(\913), to Kansho  which derives from his "Heian jidai no keigo," Kaishaku  (1956) January, "Heian jidai no keigo," Kokubungaku  (keigo special issue,  January 1960), and "Genji monogatari keigoho" (Jidai-betsu Sakuhin-betsu Kaishaku Bunpo, Kaishaku to Kansho (January 1973). 15  Chapter Four, pp. 59-60. 107  Chapter IV: The Language of Ariake no Wakare of these, which he calls saiko keigo, might be termed 'Imperial Keigo' since it is restricted to verbs whose subject must be a high ranking imperial family member: emperor, empress, crown prince, or former emperor. Keigo verbs at this level include, tatematsuru (in the honorific uses of 'wear, ride,') mawiru (in the honorific uses of 'drink, eat,') notamahasu, tamahasu, ohashimasu, ohoseraru, oboshimesaru, etc. The level of intermediate intensity, niju keigo 'second rank respect language,' includes, obosu, -se-tamahu, sasetamahu, etc. and the level of least intensity, saitei keigo, includes, go-ran-zu, obosu, notamahu, ohotonogomoru, ohosu, ohasu, mesu, tsukahasu, -tamahu, -ru, -raru, etc. Summarizing honorific causative usage, Negoro notes that the causative jodoshi  only  functions honorifically when followed by -tamahu; that an honorific passive form is less intensely honorific than -tamahu; that -tamahu itself when followed by the passive or causative forms is more intensely honorific; and that passive forms preceding -tamahu are not functioning as honorifics. It is against the background of this specific system for Heian era usage that Negoro presents his data on Kamakura period usage, albeit with certain general reservations about the chart's rigidity. Noting that a modern reader might consider some uses of keigo in Genji, such as 'ohashi-tsukitari,' less honorific than expected because of the absence of -tamahu, he stresses that this is a post-Heian perception. In mid-Heian texts, -tamahu is  108  Chapter I V : The Language of Ariake no Wakare not added to a form such as 'ohashi-tsuki.' However, in Kamakura period giko monogatari, nikki, and zuihitsu, it is. For example, precisely 'ohashi-tsuki-tamahu' is found in Shinobine Monogatari!  6  keigo,'  This is widely regarded as an instance of giko monogatari 'excessive  of which the formula 'ohashi-X-tamahu' is the most characteristic example.  17  Similarly, in Ama no Karu Mo one finds 'tsukahashi-tamahu,' and in  Sayogoromo  one finds 'oboshi-tamahu,' which are not mid-Heian usage, as was stressed by kokugakusha like Motoori Norinaga and Fujii Kosho.  18  This rule applies not only to tsukahashi- and  oboshi-, but also to ohashi, as in the first example. Because the first element is already honorific, -tamahu is unnecessary. (Suffixing -tamahu to -se/-sase and -re/-rare is a different case). T o obtain a sense of the frequency of these forms, in the whole of Ama no Karu Mo there are three cases of 'tsukahashi-tamahu' and one of 'oboshimeshi-nageki-tamahu,' in all of Sayogoromo  there are two instances of 'oboshi-tamahu,' twenty-four of the type  'oboshi-nageki-tamahu,' and one each of 'oboshimeshi-tachi-tamahu' and meshi-idetamahu.'  19  If one turns one's attention to Genji, using the Aobydshi-bon  text edited by Teika  and his household between 1224 and 1225, there are 21 instances of the type 'oboshi20  tsudzuke-tamahu' and one case of 'notamahase-shirase-tamahu.' 16  This is not a particularly  Negoro, p. 98.  kato keigo.  xl  18  Norinaga, Tama-arare, 'tsukahasu, tsukahasaru;' Fujii, Shosoku Bunrei, 'oboshi.'  Negoro, p. 99. 19  Negoro, p. 100.  20  B o w r i n g 1988, p. 84. 109  Chapter I V : The Language of Ariake no Wakare high incidence, given that the Genji is considerably longer than both of the giko monogatari mentioned. There are also absolutely no occurrences of 'tsukahashi-tamahu' or 'oboshitamahu.' 1 find that the occurrence of 'notamahase-shirase-tamahu' may also be a special 21  case, since it is in a passage of direct speech where Kaoru is asking Nakanokimi to convey his feeling to Ukifune. Thus the two honorifics may relate to this more complex configuration.  22  A t this point Negoro turns to a comparison of the Aobydshi-bon  honorifics with  those of the roughly contemporaneous Kawachi-bon edited by Minamoto Mitsuyuki (11631244) and his son Chikayuki (both governors of Kawachi) completed in 1255. Focusing on keigo verbs showing respect for their subject, he provides extensive documentation of the discrepancies between the two texts, which may be summarized as follows:  AB-Bon Honorific  K-Bon Honorific  Count  none  -tamahu  277 (40 in speech)  -tamahu  none  197 (30 in speech)  none  obosu, ohasu 125 (11 in speech) notamahu, tatematsuru  obosu, ohasu etc.  none  21  60 (1 in speech)  I n the standardized N K B T e d i t i o n I do, however, find ten instances of 'oboshi-  notamahasu,' and one each of 'oboshi-okite-notamahasu,' 'notamahase-obosu,'  and  'notamahase-nasu.' 'Oboshi-nasu' is highly frequent, with a dozen occurrences through the Akashi chapter alone. The question of 'notamahasu' compounds is not addressed by Negoro. Perhaps this is because the other examples are of redoubling lower level keigo verbs, whereas 'notamahasu' is imperial keigo. There is also one case of 'oboshi-tsuzuketamahu,' E-awase, II, p. 361. There do not seem to be any others of this pattern. Azumaya,NKBT  22  V I , p. 48. 110  Chapter IV: The Language of Ariake no Wakare Negoro's interpretation of this widespread variation is that, whereas the mid-Heian honorific habits functioned according to living rules, in the Kamakura period there appears to be something of a generalized equivalency for many of these keigo forms.  23  I would  add that the relative proximity of the two manuscripts in time, and the fact that all forms are found in both texts, indicates that the variation seems purely random, rather than attributable to different systems prevailing in the two texts. The data thus far concern keigo forms at the lightest level of honorific intensity alternating with neutral forms. Following this, data that may cross between the different intensity levels is presented: ABB-tamahu A B B lexicalized honorific  KB lexicalized honorific KB-tamahu  94 (15 in speech) 24  83 (15 in speech)  In mid-Heian texts, -tamahu was a light honorific, obosu and go-ranzu were stronger, and tatematsuru (the lexical, honorific use, as opposed to the auxiliary, humble use) was at the strongest level. Yet between these two manuscripts they appear i n virtually free variation. Kamakura period users were losing these distinctions, as is clear from other Kamakura bungo. Kenko uses 'oboshi-idete' where Heian usage would prescribe 'omohiide-tamahite.'  25  It is to this loss of distinction between the levels that the apparently  'excessive keigo' of the Kamakura period may be attributed. The usage of -tamahu 23  Negoro, p. 106.  24  o b o s u , omohosu, obosaru, ohasu, ohashimasu, notamahu,  kikoshimesu, go-ran-zu, tatematsuru, tsukahasu. Tsurezuregusa,  25  Dan 104, towards the end. Ill  notamahasu,  Chapter I V : The Language of Ariake no Wakare seemed easiest, so it is the most generalized.  26  Where does Ariake no Wakare fit into this picture? To what extent does it conform to this stereotype of Kamakura period keigo usage, which ironically defines giko monogatari keigo usage as unsuccessfully imitative of Heian usage? First, concerning the combinations that are completely absent from Genji: ohashi-Y-tamahu,  tsukahasi-tamahu,  oboshi-tamahu  21  Despite the fact that all three of these are found in the giko monogatari cited by Negoro, there are no instances of tsukahashi-tamahu or oboshi-tamahu at all in Ariake. There are two instances of ohashi-V-tamahu, and one of oboshimeshi-V-tamahu. In this regard, 28  29  Ariake would seem to fall between Genji and the giko monogatari with respect to this kind of keigo. Regarding the double honorifics of the type 'oboshi-V-tamahu' which are rarer in Genji than in the giko monogatari, we find only three examples in Ariake  30  actually  proportionally much lower than the incidence in Genji (24), given that Ariake is roughly a quarter of the size of Genji. 26  Negoro, pp. 111-113.  27  T o these may be added the pattern 'oboshimeshi-(V)-tamahu,' which I also find  completely absent from the standard edition, and Negoro finds in Sayogoromo.. 28  'ohashi-tsuki-tamaheru,' B o o k Il.iv. 13: 258, and  'ohashi-somesase-tamahu,'  II.xi.41: 324. 29  'oboshimeshi-okitesase-tamahubeshi,' I l l . i i i . 12: 406.  30  'oboshi-isogi-tamahu,' I.vii.29: 102, 'oboshi-todome-tamahishi,' I.xiv.67: 187,  and 'oboshi-atsukahi-tamahite,' II.vi.23: 282. 112  Chapter IV: The Language of Ariake no Wakare To this might be added data concerning two other types of double honorific not mentioned by Negoro, and which are rare in Genji, namely the patterns 'meshi-V-tamahu' and 'ohose-tamahu.' in Ariake  32  31  There are six instances of the 'meshi-V-tamahu' in Genji and four  seven instances of 'ohose-tamahu' in Genji, and two in Ariake, a roughly  comparable proportion again, though data for other giko monogatari is not available. Similar proportions obtain between 'oboshi-notamahasu' (10 in Genji) and 'oboshinotamahu' (2 in Ariake), although precisely equivalent forms are not found. A n anomaly 33  is that 'oboshi-nasu' which, as mentioned above, is very frequent in Genji, has only one instance in Ariake. Nevertheless, all of the double honorific forms, which are not considered standard keigo in mid-Heian Japanese, are similarly absent or rare in Ariake also. In terms of 'excessive keigo,' an examination of all seven occurrences of the verb 'notamahasu' in B o o k I of Ariake shows that all uses have the Emperor as the subject, and it is never 'excessively' applied to non-imperial family nobles (See appendix for citation of all instances of 'notamahasu' in Ariake). In Book II, of the twenty occurrences, four uses do not have imperial family members as their subject (all the rest have the 3l  There is also one instance of 'ohose-tsukahashite,' Suma, II, p. 168, and one of  'meshi-ohoshite' in Yugiri, I V , p. 428. There are no 'ohosu' double honorifics after this point. Genji,  32  'Tamakazura,' III p. 109; 'Takegawa,' V p. 82; 'Hashihime,' V p. 126;  ' Y a d o r i g i , ' V p. 436, 465; 'Kagero,' V I p.232. Ariake I.vii.27: 96, I l . i . l : 228, II.i.2: 230, II.v.19: 272. 33  There are no double honorifics with 'notamahasu' in Ariake, and none of the  type 'oboshi-notamahu' in Genji. 113  Chapter I V : The Language of Ariake no Wakare Former Emperor or Empress). In two of these cases, much lower ranked persons are speaking about utterances of persons far above them in rank (a Gon-Chunagon speaks of the Sadaijin, the Sokyoden N u n speaks of the late Udaisho Ariake, and the lady-in-waiting B e n speaks of the Naidaijin). In the remaining one the narrator is telling of the Naidaijin trying to speak to the Sanjo Lady. This may be an attempt to capture the obsequious speech of the lower ranked when addressing those further above them, or this may be a case of Kamakura period keigo slippage, with keigo becoming relative rather than absolute. In the seven such occurrences in Book III, there is one other case of a lower ranked character, Shikibu no Jo reporting the Naidaijin's dying words to the young Empress. However, since the Naidaijin is secretly the Empress's father, the usage may be proper, for one who is in on the secret - and Shikibu-no-Jo is the very one who has been the go-between for the correspondence in which the Naidaijin reveals that he is her father. The remaining instances in this book refer to the Daijodaijin, who has been granted 'Equality with the Three Empresses,' jusangu,  34  as father of the Former Empress,  so again, this may explain the usage. Overall then, Ariake's keigo usage is clearly closer to the Genji standard than the other giko monogatari for which data are available. O n the other hand, for the most part, where Ariake keigo usage begins to depart from that of Genji, it is, as expected, in the direction of the other giko monogatari, with double honorifics of the type that reinforce Negoro's contention that the giko monogatari show that a strict awareness of the distinctions between the different levels of honorific language was being lost. Nevertheless, in its broad lines, especially in the distinction between 'Imperial keigo' and the less intense levels, the basis of the Genji keigo system  114  Chapter I V : The Language of Ariake no Wakare still remains.  35  The conclusions suggested by both the lexico-statistical and the keigo linguistic evidence are very similar. A s expected, Ariake's language, in these respects, is strikingly close to that of the Genji, closer even than some other giko monogatari, and a text like the Pillow Book which is contemporary with the Genji. Nevertheless, particularly in the realm of the lexicon, there are a small number of usages that are all the more salient for their smallness of number, and it is here that we might locate the individuality of Ariake's language. Without comparable study of other giko monogatari it cannot be ascertained whether these features, in addition to distinguishing Ariake from the Genji, also distinguish Ariake from other giko monogatari. However, the keigo evidence suggests that Ariake may have some linguistic distinctness from them. A s more quantitative linguistic research is performed on these texts we should gradually be able to locate Ariake more precisely in the discursive space of monogatari.  35  Further investigation of the functioning of keigo in Ariake could be expanded to  consider the frequency and usage of the different levels of humble keigo.  115  Chapter I V : The Language of Ariake no Wakare  C N VO C N  00  MS  CN  CN  VO  !ON  c  CO  CO  CN  CO  ON C N  CN  CO  ! vo  co:  ON  CO  C N CO C N  CN  i CN!  CN  LU  wgi s  in vo  00  00  00  in  CN  ON  t--  C N CO O N ;  CN  : CN C N  00!  O N CO  CO CO  CN  00  CO  co! 00 00! O N 00! V O ! O N I T )  VO ON  CO  00  CO  00  ON CN  VO  CO  O:  CN  CO! CN! CN!  CO  ON O N ! CN  o CN  CO  00  VO  ON  C N CO  00  VO  ON  !VO  CN! CO! CN! CN! CN  CN  CD  00!  \A  1  ^1  AW. V  -3  V <&)  ^  !  A  7i!  A  A  A  Ai  JrO! 0  CM  *f\  -gv  :  ^  !  A A'* A£> N / ! K M 0!  *r\ CM CO CM CM  116  CM  m  CM  A  AJ  CM  Chapter IV: The Language of Ariake no Wakare  co  CN  ON V O 1 T-H  Tt  1  Tt  T—1 IT- 00  T-H  T-H  1  1  i—i 00 CN CN  i  i  r-~ T-H  r-H CO  i—1  1  i  o  o  i  r  CN T t CO i—l T-i CN 1  co! CN V O CN i—i i  Tt  T—1  T-H 1  CNi  VO:  ri  i—1 1  T"H 1  in i  i—i i  in COi  voi T-H o CN CN i i  o  Tf  r- CO T-H 1  o  CN CN  i  00 CO CO CN o m CO oo CO i i CNi T—I 1  CN CN  i  VO  CN  i  o  T-H  CN i  1  oo  o Tt Tt  1—1  T-H  T"H  CN  Tf  CO  ON  l  i  m >ni l~H i—1 CN  Ti-  m i  Tl- O N T—1  CN  I  — l—l  T-H  1—1  o  CO CM T—1  VO  Tt  Tt  LL  CN ON  LU  VO  CO  H  O  T—1  T-H  VO  CO  VO  T—1  i—i  m CN CN  00 o ON  ON  T-H  T-H  T t CO CN: CN T t  ON  T-H  VO  00  ON  ON  VO  O  i-H  ON  T"H V O  T—1  CO CN CN i-H T — 1 .T—1 T ~ < i—1  CO  Tt  Tt  00  'in in  CO CN CN T t T t CN T t CO  Tt  T-H  Tt T—<  00 Q  00  T-H  T-H  O  o  o  i—I i—i  VO  CN  T-H  VO  ON  1—1  o  1—1  # o  ON  O  O  ON ON T"H  VO T-H  ON  in vo  CN r o CN CN  i-H  ON  CN CO CO CO O N  ON ON  T-H  i-H  T-H  T-H  i-H  Tt Tt  m  VO  ON  T-H  T-H  ON  Tt'  o  VO  ON  O  00 CO  Tt  in  ON  ON  T-H  r-  in  o  #  T-H  i—I V O O N CN o O N ooi CO T t CN CN: T t T t CN CO Tt- m  00  VO  Tt  Tt  o  T-H  CO CN  CN CN CO  T-H  i—I  CO  Tt  T-H  o  # m TCOt o T-H  m  T}- T l - CN CN  T-H  r-- 00 Tt  ON  m  •V  A  A A A  A  A A  A A  «  -¥ '>A  T-H  CO CO CN  T - H CN CO T t T - H CN CO T t 00 O N VO 00 O N VO CN CN CN CO CO CO CO CO COj CO co CO: CO T t T t T t T t T t Tj- T l - T f  </ A A AM  •O! : mi  A  /• ^>  A  Tt  A  HA  A  "R  vj xjUr\jSNJ>uDi xj oo  VO  T-H  A  <  o  1 - 1  CN V O O N CO T t CO T t T t  T—1  Tt  O CN 00 CO vol vo CN CO; CO T t CO CO c o : CN  00  00 o  T—1  00 CO  O) O  CM CM CO CO  ^> ^ i S N  CM CO *t CO CO co  A  & If)  co  <r> N co co  A  £6 00  0)  co  CO  o  T-  •9*. CM CO  t  117  *!• <*  in  (O  00  T> Tf  "5T  o in  1 -  CM  in  in  m  Chapter IV: The Language of Ariake no Wakare  C:  00  00 O N 00 O N CN o CO 00 VO O 00 00 vo 00 00 CN T " H CN CN VO CN in  H  CO  T—1  3i  oi  u  00 CN o CN T - H  ON  T-H  ON T-H  CO VO CN  o  T-H  T-H  T-H  T-H  CN  ON ON  ON  T-H  T-H  T-H  00  T-H  T-H  CO CO m CO CO r-H  T-H  o  00  in  m  m  00 o  CN  o T-H  CO  ini >n oi T-H  T-H  CN  CO 00 00 O N CN O 00 vo O N m T - H CN T - H T - H CO  r » vo VOi CN T - H T—1 coi T - H CO  Tt T-H  o  m  CO CN CN  C: KSi  8  CN CN  c  CO CN  Ui  CN U0  S3  ON  T-H  CO  ON  T-H  T-H  f - CO  T-H T-H  mri  IB\  CO  »i  CO  T-H  CN  T-H  S  00 CO CN CO CN CN  C:  ON  >  Si i—i  •T—l  o  00  00 CO  CO  si  T—1  c  3! Oi  Uj 1—1  s  iff  CD nr; B\  «n VO  T-H  T-H  00 VO  ON  VO 00 *ni  CN m  in  ON  T-H  T-H  CN CN  in  ON;  CN CO T - H CN  o T H  O N CN VO CN CO CO CO  vo i n o  T-H  T-H  in  VO CN  CN  00  VO  vo  00  T-H  CN O O CN 00 CO T-H CO CO CN CN  CN CO m  CN CN  T-H T-H  m  CO  00  T-H  T-H  T-H  cs  O  CN  m CO m T-H  VO m  T-H  T-H  T-H  T-H  o  00 CN  ON  T-H  T-H  T-H  T-H  r t i vo CN; T - H  VO  T-H T-H  T-H  00 VO CN T - H  ;.—  00 i  T-H T-H  CN CN  T—I  T-H  o  o  ON  CN  ON  00  T-H  T-H  T-H  00 o CO T - H m ; T - H CN T - H CN CN i  i  ON  ON  H  ON T-H  esi T—I  c oH  > • I-H  o u  ON  T-H  Ci  ••5* <  ON  T-H  T—1  3!  5? c«J!: LU  CO  T-H  o! T~H  B o  CO  C!  I—I  o in  o  3 O:  o  Cfl  ON  r - VO m i n m CO CO T - H o CN CN O N 00 CO o 00 VO VO m CO CO CO CO CN CN CN CN CN CN CN CN CN CN CN  ON  ON  00 r » i vo  T-H  T-H  T-H  T-H ; T-H  3 O  U i  T-H  m  VO  CN CO  00  ON  c  o  T-H  CN CO  T-H  T-H  T-H  T-H  m vo r - 00 T-H  T-H  T-H  T-H  T-H  ON T-H  o  T ^ H CN CO i ^ CN CN CN CN i cN  Pi  &>  o H  1  CD! •M  o:  to  <Di *'  T3;  <;  H  A  A  V  f5 OO  in  o  *!  A  T-  f 5 «0 V xj +6 xj CM co  #  A  A  6 XJ  1-  118  10  CO  co  0)  o  \A icM  V  0 XJ xj  A  AJ  &i  CM in CO N CM CM CM CM CN CM CM CM T-  Chapter I V : The Language of Ariake no Wakare  T-H o VO VO o O N W1 VO T - H CN CO ON T - H H O N C N ON VO T-H 00 o ON O VO ON Os o o C N ON O N ON T-H ON CO CO ON ON T t o O N T - H T-H T—H T-H T - H T-H T-H T-H T-H T-H T-H T-H CO  00  CN  8 o S3  CO ON 00 o C N T t T-H CO T t  T-H T t  T—1  CO  CN CN T t  CO  oo 00  Tt  Tt  T-H CO  O N T-H  o  wo  o  T-H VO C N CO T-H T t  CN T t  C N CO VO o \ C N CO C N  T-H T t  C N VO ON VO T t WO ON 00 r CO T t T t T-H CO T t C N C N T t  C N C N C N C N T-H C N  ON VO CO T t C N C N CO CO  CO CO C N CO T t  r- rco CN  CO T t  Tt  o  CO C N  ON C N WO T t CN CN  Tt  o  T-H  Tt  o  o 00 wo T t  VO O N C N o  T-H  o  wo t~- wo  CO  CO  wo  CO C N C N  T-H T-H T-H  T-H  00 O  CN CN  T-H T-H C N  00 T t wo Tt  Tt  CO T t  Tt  Tt  T-H  Tt  rH  T-H  C N CO T t  C N C N T-H C N O  VO O C N T t ON CO T-H wo VO 00 O 00 CO CO CO CO CO CO C N C N CO CO CO T t T t  VO o wo T t ON 00 CO O CO wo  T-H C N  T H  CN CN  T-H  o  o  VO  Tt  Tt  o  o  Tt  Tt  o  T-H  o  CO O N  T-H  O  O  o  O  ON  00  Tt  Tt  T-H  o  W0 T t T t T t  a o  VO  00 VO T t 00  T—<  T-H  wo  Tt  W0 T t CO WO CO CO VO T t  T-H T-H T-H T-H  LU  >  • i-H +-> CJ  T-H T t VO CO CO T-H C N C N  ON Tt C N ON T t VO C N CO T-H o C N CO C N C N CO CO CO CO T t CO T t T t  VO wo CO C N C N  >  T-H T-H  T-H T-H T-H o o o ON ON ON T-H T-H T-H T-H T-H r H T-H T-H T-H  00  o  00 CO W0 VO CO CO T t T t T t T t  Tt Tt  r - VO  Tt  Tt  Tt  CN CN  T-H C N  co  Tt Tt  wo VO r - 00  Tt  wo  • •-1 <4-»  o  wo VO r - 00 OS o CN CN  CN  T-H C N CO T t wo VO C N C N CO CO CO CO CO CO CO  00 ON o  CO CO T t  Tt  Tt  Tt  Tt  Tt  Tt  Tt  ON  Tt  ca  o wo  CO  & /•  O  m  O H cd H  *!  <y  4i V U  Xj  cn o CM CM co co  -¥ •0  ''A  xj  Xj  -ft T-  co  XJ  ^>  S>  !«  A  &  AJ  •H £ 6  CM CO T> IT) <£> oo 0) o CO co CO CO CO CO CO CO  &  A  A  T-  T>  119  A  H  f5  *PN  A  *U  CM CO t  <T "3-  -tt  in  ID rt  V  A  -ft  CO  «*  AJ  \J  &  XJ  ^»  to  -ft  A  >»  o in  T-  CM co  Tf  w  m  m  in  Chapter I V : The Language of Ariake no Wakare  T-H 1  CN i  o  in  ! r - CN co i  T-H  CO 1  co  o  O  o  o  to  ON  CO  T-H  CN  VO o  i  T-H  CN  o  CN  i o 00 o  VO CN!  O  VO CN 00  o  T-H  i  T-H  VO m  i  m  i  CN T-H  ;co  T H I  1  mr:  2>  o  I Tf 1  ON  i  i  T—1  VO  CNi  o  o  'EJ  T-H  T-H  CO  00 T-H  T-H  a  m  CN CO  c  ICN1  CN CO i n i CO  i ^°  00  T-H  1  ON  CO O CO i n CO CN  O  T-H  CN CO  T-H  T-H  T-H  T-H  CO  00  CN o ! i n CO VO t-~ CN T-H i CN CN i  m VO  i 00  ON  ^  T-H  T—1 1  T—1  T-H  T-H  m  f-  VO v o CN  T-H  T—<  T-H  T-H  CO  T-H  CN  ON;  o  T - H CN CN CN CN  CN  o  Mi &  C« CO  >  T-H  00  CO  CN CN  io  ON  CN  T-H  00  T-H  T-H  CN  00 VO  CO CO CO CO T - H  CO  o T-H  CO CN CN  VSl  • »-H  (4\  o  o  T-H  i  T-H  CN  1  T-H  CN CN  o  o  o ! VO 1  T-H  CN VO  T-H  ;ON  T-H  1  T-H  CN;  1  CD  > 4-1  CO iff CO  o  m o H  CO:  $r  -4-*:  (J  A  Oil  H  <! CN  to  A  A  V i V V  •53 00  o  T-  N  "53  A  Xj  to  co  1  £6  <*  to  CD  oo  X o  T—  CM CN  120 ^  A  6  "53  xj CN  — , \A  0 J  -ft  CM co CM CN  V  K> A  0 xj •53 AJ r XJ ^> 10 CD CM CM CM  CM  Chapter IV: The Language of Ariake no Wakare  8  T-H  r--  T-H 1  T—1  o  1  CN c o ! O N  T-H  T-H  T-H  1  VO  CO i n T-H  T-H  T-H  r-  CO CO 1  1  CN  m  1  T-H  CN v o  T-H  T-H  T-H  T-H 1  CO O N  0  1  00 :  m CN CN CO CO i T - H1 1  CO T-H  T-H  T-H  H  in T-H  O T-H  1  VO  T-H 1  m  ON  ON  1  1  T-H  T-H  CN i n CN  T-H  1  1  00 1  CN  r  CO T-H 1  CN O N i  CN CN CN CN 1  1  T-H  1  00  o a £3  r- i ri T—1  m  i  00  T-H T-H  T-H  CO T H  T-H  0  m T-H  1  00  0  CN  1  CN CN CN T-H  1  T H 1  T-H  r--  i>  CO  T-H  vo 1  T-H 1  1  ;  1  co  m T-H  1  T-H 1  CN CO CN CN i n 1 m 1  0.  B o CA  CD > •  O  z  • n vo 00 O N i n 1 V O t > 00 O N 0 T - H CN CO CN CN CN CO CO CO CO CO CO CO CO CO CO CN  i ^  00 I  »  T-H  n-  CN CN  CO O N T-H  00 CO  O  CN V O O N v o in CO 'St T - H c o  ON  0 -3-  T-H  I  CN 1 CO  i n VO •<fr  i  00  H  CN CN  1 v o O N CN i ^ CO  i  O  •*  00  ON  O <n  0  m r~- m CO CO CN CN  1-H  o CO O N V O 1 CN T - H 1  CD  00  T-H T-H  T-H  1  0  CN  CO  CN  T-H  1  CN v o  O  1  T-H  I  1  T-H  T-H 1  1  'in * -  1  1  t  r--  CO  T—.  1  1  CO CN i n T-H CN CN t 1 1  < > •  T-H  <*->  o  CD H ? 5»  CO  o m O  H  CD F-H  :  *! :  _j -ft A  *!-ft  — r j KT» V xj KJ x J A £  ''A  £ N  J )  a>  o  T-  CM  co  CM  CM  CO  co  CO  CO  CO  10  CD  CO  CO  \A  r» co  A —^  -R  co  0)  O  T—  CM  CO  CO  1-  «»  «*  121  CO  •R  A  AT  £6  (  •R  * J  H  1-  LO  CD  h-  "0  A  tS •R co *r  *R AJ  TON  A  H  •ft  +6  A J  XJ  co  xj &  •ft •ft A  •ft  NO  •R  -ft  xj  £ N  \A  SN  •ft  •R ±>  CO  CCS  Tol  CM  •ft o>  0  T-  CM  CO  "3-  to  w  10  to  LO  Chapter IV: The Language of Ariake no Wakare  Oi  Oi  Tt  Oi  o  Oi  Tt  Oi  CN  Oi  co  O CO  O i CO  wo  CN:  CO H i  CNi CNi T t i W0i  O  VO WOi wo  CO CNi O i  coi CO: coi r - i  H  CO woi CN  CN  CN  oi Ci C; o CSi OH!  o <+H  O Oi Ci <ui Ci «5i  O  COi T t Tt  U:  Q!  Ci  o oi C! (Di i<Di  Ui  M  "HH u- AH  Tt:  woi co Tti 00 o r - wo: Tt CNi r-: oo vo r-Hi T t WOi Tt CNi r H WOi Tt Tti H i CO r - i co H i CN  C 3 Oi  m  CO  CN  Tti  WO VOi CO oo:  oi  H i Tt:  O  Tti Tti  CN:  o  O i coi  00  Tt  r- o wo o CN  O i CN  CO WOi OOi O N  VO! VO ONi CO  WOi CO VOi  ON!  vo: C N  CC!  OH!  t"—: O r - i vo oo: oo C N : vo; vo: T t : wo: co COi O N CNi CN Oy O N VOi woi O N ; ooi coi wo C N : VO r - i oo VO: VO CN: T t : coi C N : C O ; C N  Ci 3! Oi  LU  VO CO CN  wo: T t  CNi r - i ooi n i  ONi  co  Tti  00  wo  cdi •*-»!  T-H!  Ci 3i Oi  o  ;  00 Tt 00 W0 C N  ON  TH i ON ON! r -  T t i TH  ¥ U  O 00 r-  o  Tti  ci  m?i  CN  H i H i H WOi t—! ON r - i r - i vo Tti co; C N  00i wo  voi  r-i o  H i  r H ; CN  ON  wo  00i CNi O  CO Tt  {  CN  ci  ffi  CO: T t  W0: VO  r—: ooi o\i o i  H i CN  CO T t i W0 vo  r - i OOi  ON  Oi  CN:  «Ji  c  _Oi •4-ji  x>i Oi >i TJ!  <\ l-  o; oi  CO  tJi  "C:  ~^:  " •  o c CM  jOi ll  ! J)  & hi S  40 CO  lO  CO  CM  \J A ^•6  A A j 1-ft  AJ  A  M  AJ AJ  A  /* i /•  H  A"  o A  ;  nr  E-  A 2>\  03  co  o CM  122  ci  CM  CM CM  CM  in CM  Chapter Five Gender in Ariake no Wakare: Gender, Genre, and Genealogy One aspect of Ariake no Wakare that must immediately strike most modern readers is the gathering together of several gender-related issues that have provoked intense interest in literary critical circles since the 1980s. Ariake gives an explicit critique of male behaviour in the context of courtship and marriage, together with an exploration of female-to-male crossdressing; the male gaze; incestuous sexual abuse; both male and female same-sex and same-gender love (drawing attention to that distinction by the very juxtaposition); spirit possession in the context of desire, pregnancy and marriage, together with other instances of the conspicuously gendered supernatural; and the gendered significance of genealogy. The mere co-occurrence of these topics in one text is fascinating enough in itself. The fact that these topics are most intricately woven together in the plot, together with the interpretations that it suggests, make this a remarkable text indeed. However, given this textual embarass de richesse, its treatment in the context of this study, which constitutes the first book length study of this text in any language, poses an acute problem. G i v e n the extensive specialized bibliographies of several of these topics, thorough treatment might well call for a substantial monograph on each one. O n the other hand, focusing entirely on one or two is hard to justify when one is largely introducing a text for the first time, quite apart from the fact that the impressively tight integration of these topics in this text makes it especially difficult to disentangle one or two strands and single them out for in-depth treatment at the expense of the others. The inevitable consequence of this chapter w i l l therefore be to give only a tantalizing and 123  Chapter V : Gender in Ariake no Wakare preliminary presentation of these issues and their handling in the text - an indication of some avenues that future research might find fruitful, rather than anything approaching a comprehensive treatment.  From 'Castigation of Irogonomi' to a Critique of Gender Relations In the last decade, a large number of critical readings in English of Heian women's writing have valorized what is perceived as a strong element of 'writing against patriarchy' in those texts, focusing naturally on the Heian women's diaries, especially the Kagero Diary, and on the Genji. These more recent interpretations are in relatively sharp contrast 1  to the traditional Japanese focus, traceable at least back to the Mumydzoshi, its most explicit early articulation in Motoori Norinaga's Shibun Genji Monogatari  Ydryo  but finding 2  (1763) and  Tama no Ogush? (1793-1796), which foregrounded mono no aware  and specifically its effect on both male and female experience.  4  'For the diaries, see especially Arntzen, 1997, M i y a k e 1989, 1990, and 1996, M o s t o w 1990 and 1991, Sarra 1996 (but see also Mostow 1992, which argues that Kagero Nikki was written 'on commission' for the author's husband, Fujiwara Kane'ie); for the Genji, see especially Bargen 1986, 1988, and 1997, Field 1987, Shirane 1987, B o w r i n g 1988, and Okada 1991. For later monogatari there is little more than M a r r a 1991, and Sarra 1990.  ''This dominant interpretation in Japanese criticism is echoed in works like Ivan M o r r i s ' World of the Shining Prince (Oxford University Press, 1964), passim. Similarly,  124  Chapter V : Gender in Ariake no Wakare Some earlier recent criticism had seen the Kagero Diary as an individual complaint of one woman against her husband's treatment of her, and the Genji as partly a presentation 5  of the difficulties of the female condition faced with the activities of the irogonomi  or  highly uxorious lover. M u c h of this more recent criticism posits a broader critique of Heian gender relations as a whole, especially as regards the courtship and marriage customs of the court aristocracy. It takes as its cue the undeniable fact that both diaries and monogatari add to their criticisms of specific males more generalized observations about the nature of men. Rather than address the vast issues of how tenable such a position is for the numerous texts in both of these genres, the focus here shall be on the extent to which Ariake no Wakare might support such a reading.  Material that most strongly supports the 'critique of patriarchy' reading can be found primarily at the level of explicit generalizing remarks in the text, directed at men, and in specific aspects of the plot itself. Indeed, for many late 20th century readers it would be hard to avoid at least an awareness of such a theme on the basis of the plot unfolding in just the first six chapters, even without the parenthetical remarks of characters and narrator about the behaviour of men. The leitmotiv poem by M i b u no Tadamine, which sets the tone of the opening in E a r l M i n e r ' s 1969 article 'Some Thematic and Structural Features of the Genji Monogatari,'Monumenta  Nipponica  24 (1-2), p. 11, he finds Only 'one matter in which  Genji fails particularly. His relations with Fujitsubo are, during his father's life, undutiful and impious. But it is not really clear what attitude we are meant to take to this behavior.' 5  Cf. the introduction to Seidensticker 1964. 125  Chapter V : Gender in Ariake no Wakare chapter has, from the earliest commentaries, been open to two interpretations.  6  Both  interpretations see the poem as a lover's complaint at the unfeeling moon that shines wanly on his departure at dawn, but one interpretation sees the departure as following a successful tryst, the other interpretations sees it as following an unsuccessful tryst. Given the conventions regarding the gendering of physical movement, and the protocols of Heian courtship, the speaker is presumed to be the man. Taking the 'unsuccessful tryst' interpretation, then the insensitivity of the moon is naturally associated with the coldness of the lady who has proved unresponsive. Apparently it was the 'successful tryst' interpretation that was favoured by Teika (and Kensho), although almost all medieval commentaries favour the 'unsuccessful tryst' interpretation.  7  In the opening scene, Ariake no Wakare recontextualizes the poem quite strikingly (and perhaps originally). The traditional interpretation of Tadamine's poem which is closest to this initial Ariake scenario is the 'unsuccessful tryst' one, since the context is one of 'not meeting.' However, instead of the man being the distressed party spurned by the woman, here it is the woman who laments the man's insensitivity. The man is characterized as 'unboundedly conceited,' and of 'extremely sensual disposition.' H e 8  seeks out numerous women who gratify h i m by their devotion - 'it was his nature to enjoy having people completely in love with h i m , ' and he leaves them waiting for long For the two interpretations, see Kubota Jun, Hyakunin Isshu Hikkei (Gakutosha,  6  1982), p. 59, and Mostow 1996, p. 232. Mostow 1996, p. 232.  7  fr€ V ft < &t>fctfS Qtct^V  s  ...  [ V ^ D < 7]&K&\*LT.  126  I. i.2.  Chapter V : Gender in Ariake no Wakare periods while they are 'surely being driven into a frenzy of anxiety.' The 9  irogonomi  stereotype hereby introduced is then exemplified by the recklessly aggressive erotic adventures of the first Sadaisho and Sanmi Chujo, the father and son philanderers whose activities largely dominate the next six chapters, and even beyond. Throughout, there is a consistently critical point of view of this conduct expressed both by the narrator and by the protagonist, the Udaisho Ariake, to whom 'even the most reckless conduct would have been permitted,' but who was 'extremely phlegmatic and restrained.'  10  Typical comments by Ariake are the following: " H o w hateful! The hearts of men are truly wretched." i i i . 7 . " "There is nothing quite so unreliable as a man." vi.19.  12  (On the Sadaisho's philandering with his stepdaughter) Ariake thought typical love affairs were really pointless things he wanted no part of. [...] "it really seems to be women's wretched lot to be deceived by their lovers." I V . 11. (On Sanmi Chujo's treatment of 13  the Sanjo Lady)  ^<Lo  I.i.2.  K< f C & T L<5#>.I.ii.3.  $>t£5fcX.  ll  5%  ft V)o  127  I.iii.8.  Chapter V : Gender in Ariake no Wakare Other characters also pass comments in a similar vein: "So now it's that Chujo," [Jiju] thought, " H o w awful! It's so shameful that it's him of all people." X . 4 1 .  14  (On his pursuit of Tai-no-Ue. The  other attendants are also clearly outraged by his violation of her). Most of the plot in the first seven chapters, in fact, bears out such an anti-patriarchal reading, since it largely proceeds as a series of tableaux of abused women viewed by the invisible protagonist, who is moved by their plight to intervene. In succession, the reader is presented with the 'waiting lady' of the prologue; the Sadaisho abusing his stepdaughter, and her mother's anguish over it; his son Sanmi Chujo's heartless seduction and abandonment of the Sanjo Lady and Lady Nakatsukasa; and his pursuit and violation of his stepsister Tai-no-Ue. Not only do the first seven chapters, in and of themselves, suggest this anti-patriarchal reading, but this reading tends to be further emphasized when these plot elements are put in the  context  o f prior H e i a n literary history. These chapters  abound  with  recontextualizations and inversions of traditional monogatari elements and in many cases the different usage serves to heighten this theme. The opening tableau of the matsu onna is an absolutely ubiquitous topos in much of the earlier Japanese prose and poetry, but the link into Chapter II, with its comment that such irogonomi are frequently of the highest rank and can expect to be permitted the 14  r£o,  *Ht?>j b&JbK,  128  rfc&v^D,  £bX£*)i>£t£  Chapter V : Gender in Ariake no Wakare most reckless conduct,' already suggests an uncommonly explicit tone of social critique. 5  Almost immediately, the character who then becomes the protagonist of the narrative is presented as an anti-irogonomi, in contrast to the majority of preceding monogatari save Torikaebaya (though the U j i chapters of Genji, as opposed to the Genji chapters, provide an early monogatari exploration of this theme, and Ochikubo is less concerned with irogonomi), and in marked contrast to his young contemporaries. The next chapter presents an inversion of two further gender-specific and interestingly related topoi: invisibility and kaimami.  In previous texts, both topoi  characterize irogonomi activity, specifically the discovery of and access to desirable women. In fact, invisibility of a male protagonist makes possible a kind of 'super-kaimami,' wherein the male gaze is unconstrained by any intervening object whatsoever. This is clearly the tenor of the surviving setsuwa that exemplify this topos and what we can know about the monogatari Kakuremino.  16  But no sooner is this familiar topos introduced  in Chapter III, than it is used to present an unequivocally pejorative instance of irogonomi conduct - the Sadaisho's molestation of his step-daughter - rather than to celebrate irogonomi access to a desirable female. Not only does the protagonist, the Udaisho Ariake, critically observe another male's irogonomi activity in this Chapter and the next (inversion of kaimami), but he then uses his invisibility to actively counter irogonomi l5  I . i i . 3 : ' S u c h situations are not necessarily all cases involving men of humble  station. W h e n one inquires about the men concerned, are they not descendants of such and such and emperor, or some chancellor or other?' 16  C f . Sei Shonagon's remarks in Makura no SoshiDan  in the Fuyoshu, and the plot of Konjaku IV.24. 129  100, the surviving poems  Chapter V : Gender in Ariake no Wakare activity in Chapters I V and V I , where he respectively comforts the seduced and abandoned Sanjo L a d y and abducts the Sadaisho's stepdaughter to protect her from his further advances. This last episode inverts another widespread topos, the abduction of the vulnerable and nubile girl. In a striking reversal of the famous abduction scenes in Ochikubo and Genji,  17  the abduction serves to prevent rather than facilitate sexual access, though the  'marital' purpose is still maintained, in a public sense. This, though, has also been presented as an 'inverted' marriage of two women, since the female gender of the protagonist has been revealed at the end of Chapter II. To return to the step-parent / step-child episode, in its configuration in Ariake it represents two further inversions. First, as an inversion of the Genji episode with Genji and Fujitsubo (also resulting in the birth of a child whose real paternity constitutes an important secret), the gender of step-parent and step-child is inverted, though the male source of sexual aggression remains constant. But the very fact of abuse by a step-parent thereby constitutes another inversion or refraction, that of the traditional and very widespread mamako ijime, 'tormented stepchild' motif, for which the step-mother is usually responsible. This is also found in Ochikubo, the locus classicus for the motif among courtly monogatari, and in Genji, where it is an important part of the initial plot and tone setting - Genji's relations with his step-mother Kokiden. It need hardly be added that this is also an inversion of the step-child Genji's desire for his step-mother Fujitsubo in the Genji. True, 17  F o r Ochikubo, see B o o k II (Whitehouse 1935, pp. 96 ff.); Genji Chapter 5,  'Wakamurasaki.' 130  Chapter V : Gender in Ariake no Wakare Genji's step-fatherly interest in his adopted daughter Tamakazura has strongly romantic overtones, but their relations do not advance beyond that point, and in fact soon retreat from it. This sustained practice of topos inversion presents us with difficulties of interpretation comparable to those of Torikaebaya and of the story Mushi Mezuru (Her Ladyship  Himegimi  who Likes Insects) in the Tsutsumi Chunagon Monogatari collection.  Critical debate is divided over whether the gender and plot inversions in these texts are intended as comic 'reversals of expectation' or seriously subversive challenges to the gender and discursive order.  18  In the case of Ariake, however, there seems less room for  uncertainty. Unlike Mushi Mezuru Himegimi, there are no clues indicating parody, such as the laughter of interpretive figures like the himegimi's  ladies-in-waiting. U n l i k e in  Torikaebaya, the role reversal is not a self-contained plot, but the premise and first half of a plot that takes on some of the darkening quality of the U j i chapters of the Genji, which are not generally considered comic to any degree.  Spirit Possession as a Critique of Gender Relations A further important piece of evidence to assess, regarding the case for a gendercritique interpretation of the text, is afforded by the instance of spirit possession. Postwar responses to the spirit possession scenes in the Genji have included valorizations of spirit possession as a celebration of female power, as in Enchi F u m i k o ' s Gen/i-inspired 1958 novel, Onnamen ( ' M a s k s ' )  19  and, more recently, interpretations of the spirit possession  18  C f . Marra 1991, pp. 63 ff.  19  E n c h i Fumiko, Onna-men (Kodansha, 1958); Fumiko Enchi, Masks, trans. Juliet 131  Chapter V : Gender in Ariake no Wakare scenes as outright critiques, i f not castigations, of patriarchy and of Genji himself as its principal agent, especially in the work of Doris Bargen.  20  Significant spirit possession scenes also appear in a number of monogatari other than the Genji, but the most notable are probably those in Ariake and Waga Mi ni Tadoru Himegimi. After a brief survey of pre- Genji spirit possession, Bargen uses 12th and 13th century pictorial data as evidence for her thesis (though, significantly, not the Me-nashi Kyo), so it is perhaps surprising that neither she nor any others who refer to this interpretation have brought these later monogatari into the debate, but the bulk of her study is based on close reading of the Genji possession scenes. A s one might anticipate, the contribution of these other literary treatments of the theme is extremely apropos. Juxtaposing Ariake with this theory permits us not only to explore possible interpretations of Ariake, but also to test the applicability of the theory to texts other than the Genji. Bargen's theory of Genji spirit possession as an 'oblique aggressive strategy' used by women to express grievances against men in societies where women are typically constrained from giving voice to such complaints, draws its inspiration from anthropological studies of spirit possession in various societies, but especially Melanesian and North African ones in this century.  21  In such societies, spirit possession permits the socially  Winters Carpenter (New York: Random House (Vintage), 1983). 20  Bargen 1997, summing up her work of the preceding decade or more.  2l  C f . I. M . Lewis, Ecstatic Religion: a Study of Shamanism and Spirit  Possession  (Penguin, 1971, 2nd ed. London and N e w York: Routledge, 1989). The second edition includes consideration of Doris Bargen's work.  132  Chapter V : Gender in Ariake no Wakare marginalized to express complaints that would normally be repressed, thus functioning as a socially sanctioned kind of 'therapy' for the possessed, and something of a 'safety-valve' for the society, since the status quo remains intact following the spirit possession episode. In the anthropological interpretation, the emphasis is on the spirit possession as a performance by the possessed to vent pent-up feelings, and as a bargaining chip. To this Doris Bargen adds a more specifically feminist reading for the cases in the Tale of Genji, that posits the spirit possession as a stratagem of 'gender solidarity' which responds to the needs of both possessor and possessed in venting feelings and shaming the source of their discomfiture, the man who has offended them both. In the Tale of Genji, this corresponds to Genji himself. Thus Lady Rokujo is seen as working together with characters like Yugao, A o i , and Murasaki, to j o i n with them in giving voice to frustrations at Genji's actions. The traditional view has seen spirit possession as a hostile act of uncontrollable jealousy and anger on the part of the possessor, aimed as much at harming the possessed female rival as at publicly shaming the man concerned by public denunciation. Thus Bargen's reinterpretation of the meaning of such scenes is a radical one. Her interpretation 22  means that these possessions are as much about the resentments of Yugao, A o i , and Murasaki as they are about Rokujo's jealousy. A n obvious difficulty with this interpretation 22  Bargen considers this the view of all traditional Genji criticism, summed up i n  the most recent detailed study of the topic in Japanese, Fujimoto Katsuyoshi's Genji Monogatari  no Mono no Ke: Bungaku  to Kiroku no Hazama (Kasama Shoin, 1994),  (Bargen p. 26 and n.116, p.285). 133  Chapter V : Gender in Ariake no Wakare is that although the possessed women may indeed have cause to regard Genji with dissatisfaction, they pay a heavy price indeed for the privilege of discomfiting him, all three dying shortly after their possession experiences, i f not as a direct consequence of them.  23  In addition to the central configuration trio of the possessor, the possessed, and the male object of the possession performance, Bargen also gives interesting consideration to the role of the exorcists, who in her interpretation are not merely ancillary figures without significance in the gendered critique, but on the contrary are also objects of the critique of patriarchy. Drawing on critical remarks in the Murasaki Shikibu Nikki and Sei Shonagon's Makura no Soshi, she identifies a current of criticism of male exorcists (as opposed to the mostly female mediums) in Heian women's writing, and feels that ' i n the 24  Genji, too, exorcists make a mostly pathetic spectacle of themselves.'  25  It w i l l also be  interesting to see whether this vein of gender critique is also identifiable in Ariake  no  Wakare. There are four instances of spirit possession in Ariake no Wakare, affecting the young Empress in Book II, Chapter xii; Oigimi and Shijo-no-Ue in Book II, Chapter xiv, 23  Compare Tania M o d l e s k i ' s study of mass culture texts in English for female  readers, Loving with a Vengeance: Mass-Produced Fantasies for Women (Hamden, C T : Shoestring Press, 1982), p. 12, ' A t the end of a majority of popular narratives the woman is disfigured, dead, or at the very least, domesticated.' 24  Bargen pp. 13-17.  25  Bargen, p. 17.  134  Chapter V : Gender in Ariake no Wakare and B o o k III, Chapter i ; and the Naidaijin (formerly Sanmi Chujo) in B o o k III, Chapter iv. O f these possessions, the first and last are by unidentified spirits, so the question of the disposition of the possessing spirit is more open to speculation (though there is a very plausible suspect). However, the ones in between are clearly ascribed to Lady Nakatsukasa, and are described quite spectacularly. In all cases our view of monogatari spirit possession is significantly broadened. The first spirit possession in Ariake no Wakare, rather than constituting a major episode in itself, provides instead a frame around another episode: the young Sadaijin's fortuitous encounter with the Sokyoden N u n while on a fruitless mission to summon a highly reputed exorcist, the reclusive former Bishop of Yokawa. The possession itself is initially presented as merely the conjecture of the narrator, relating it to the Empress's indisposition - T suppose it was because of spirit possession.' The Empress is pregnant, 26  a time when a woman is particularly prone to spirit possession, but there could be no motivation, or indeed literary precedent, for the Emperor to be the intended object of an 'oblique aggressive strategy,' since the narrative has provided no motive for any woman to wish revenge on him. However, in characteristically oblique narrative fashion, the reader is then reminded of the secret paternity of the Empress. Though publicly acknowledged as the daughter of the late Udaisho Ariake, the Empress was in fact fathered on Tai-no-Ue by the former Sanmi Chujo, now Naidaijin. A s he is an irogonomi character, there are certainly candidates for the role of vengeful former mistress.  However, at this point the narrative shifts to the Sadaijin's fruitless trek across the 26  II.xii.45, #P %  (D(D\-fK^>. 135  Chapter V : Gender in Ariake no Wakare snowy wastes of midwinter Ohara M o o r to summon the Bishop of Yokawa. Fortunately, the Sadaijin and his retinue seek shelter from the blizzard on their return, at a hermitage which turns out to belong to the former Sokyoden Lady, now a nun, who had enjoyed an amorous exchange with the Sadaijin's putative father, the late Udaisho A r i a k e .  27  The  Sadaijin is entranced to hear reminiscences of the father he scarcely knew, and hurries back to the Capital to report to the Former Empress, his aunt, who is, of course, none other than the former Udaisho Ariake, now living under a new identity and gender, concordant with her actual sex. Touched to be reminded of an old admirer, the Former Empress expresses her affection by sending an intimate poem and 'recalled the way she used to hold the brush long ago and wrote in exquisitely cursive 'grass-style."  28  'The way she used to hold the  brush long ago' seems to denote a male writing style, an act of epistolary cross-dressing that has remarkable consequences. Overwhelmed by the miracle of a letter from