1
THE "TRANSPARENT" TEXT

1
Fictions and Fabrications

When it comes to fiction I can write nothing but lies; and because I so believe, I simply cannot be serious about it.
Fubatatei Shimei, "Watakushi wa kaigi-ha da"

I think that truly creative writing involves ... discovering something in yourself, not dreaming up some formally satisfying story.
Yasuoka Shotaro, "Gendai ni okeru watakushi shosetsu"

There is an unmistakable mystique about the shishosetsu , begotten by the special status it enjoys as a "truthful" account of its author, which has traditionally discouraged anything more than a cursory analysis of it as a literary artifact. Writing about the shishosetsu is not unlike pursuing a desert oasis only to watch it recede and finally disappear altogether as you approach it, or like peeling an onion, skin after skin, in a vain attempt to get at the "core"; there is something inherently elusive about the entire project. How is one to analyze a form that critics have debated for well over half a century but for which they have failed to come up with a workable definition? How does one go about isolating a form that is commonly thought to differ significantly from the shosetsu , yet has no clearly identifiable linguistic marks? And finally, how is one to describe in English a form the very name of which comprises two seemingly neutral but in fact highly problematical and quite untranslatable elements, shi (watakushi ) and shosetsu ? Raising such awkward questions at the commencement of a study perhaps bodes ill for its outcome, but it seems wisest to alert readers at the outset to the illusory nature of this most "realistic" (as Japanese critics would have it) of narrative modes.

The shishosetsu appears all the more illusory in translation. "I-novel" is an unfortunate misnomer for a form that differs fundamentally from the novel and contains no "I" in the sense we are accustomed to thinking of the word. Shosetsu , a Chinese compound with its own long etymological and cultural history, is the word chosen by the early-Meiji students of western literature to render "novel" into Japanese. From the beginning, however, its much broader meaning and scope defied the neat definition of "novel" that the students, in all good faith, originally assigned to it. In the first place, shosetsu refers to a prose fiction of any length. The distinction made in English between the novel and the short story is nonexistent, although the Japanese do classify texts as "short," "medium," or "long."^[1] In the second place, shosetsu can refer to texts that westerners ordinarily do not think of as fiction. Essays, sketches, memoirs, and other discursive and reflective pieces, which we would normally subsume under the rubric of nonfiction, very often fall under the category of shosetsu in Japanese.

shishosetsu is a loose approximation of the German Ich-Roman , after which it appears to have been named. The Ich-Roman is a full-length novel in which the narrator gives a first-person account of his experiences. This genre is hardly unique to German literature (which has examples from Werther to The Tin Drum ) or to western literature as a whole; modern Japanese literature, too, has its examples—Wagahai wa neko de aru (I am a cat, 1905-6) and Botchan (1906), both by Natsume Soseki, being among the earliest and best. Japanese readers would never think of calling these texts shishosetsu , however, because the narrators bear little or no resemblance to their author. Soseki is no alley cat, after all; nor is he a reckless, ne'er-do-well teacher in Shikoku like Botchan (although he did teach there once). The shishosetsu , moreover, need not have a first-person narrator at all. A story narrated in the third person may still be considered a shishosetsu if the hero is clearly modeled after the author.^[2]

[1] Tanpen, chuhen , and chohen , respectively. These forms have no distinguishing structural characteristics or absolute word count. Hirano Ken, for example, calls Chikamatsu Shuko's Ko no ai no tame ni a tanpen shosetsu in NKBD 2:382 but calls the same work a chuhen shosetsu in Ito Sei et al., eds., Shincho Nihon bungaku shojiten , 765.

[2] For a discussion of the Ich-Roman and the shishosetsu , see, for example, Sato Koichi, "Ihi-roman." For a comparison of the shishosetsu with the French roman personnel , which is no more closely related, see Shirai Koji, "Roman perusoneru ni tsuite."

The nature of the shi/watakushi in shishosetsu , however, invites further examination. Watakushi can hardly be translated other than as "I," but it most assuredly does not mean the same thing. Indeed, the single first-person pronoun nearly all languages employ is no more useful to Japanese than the word "snow" is to Eskimo; it is much too general to serve any purpose. Watakushi is actually one of a half dozen or so first-person pronominals that any one person commonly uses to refer to oneself, depending on the occasion (e.g., public or private) and on the relative social position of the listener and/or the referent. A man might use one pronominal (e.g., ore ) to refer to himself when speaking to family members or to close friends, another (e.g., boku ) when speaking to his peers at work, and still another (e.g., watashi ) when speaking to his boss. (A woman is generally limited to the use of watashi or watakushi in the standard dialect.)^[3] Each first-person pronominal has distinct nuances that greatly affect a writer's style. We shall see in Part 3 how Chikamatsu Shuko, for example, uses the relatively formal watashi to lend an air of solemn, matter-of-fact dignity to his zany accounts of depraved love; how Shiga Naoya uses jibun to distance himself from his experience and at the same time draw the reader into it; and how through Kasai Zenzo's kaleidoscopic usage of first-person pronominals we "see" one man's sense of self disperse before our eyes.

This remarkable variety extends to the second-and third-person pronominals as well, which if anything are even more numerous. The incidence of so many pronominals in the language speaks eloquently for a very protean notion of self, one that depends for its existence more on the person or situation with whom or with

[3] Watakushi , the most public and formal of the first-person pronominals currently in use, is a curious choice of labels for a form that presumes to describe the author's "private" life and thoughts. The original meaning of watakushi , to be sure, is "private" or "personal" as opposed to "public" or "communal" (whence such words, still in use, as watakushigoto ["private matter"] and watakushigokoro ["private emotions"]), but the watakushi in watakushi shosetsu is clearly an example of the pronominal usage, as is illustrated by the term once used alternately with watakushi shosetsu : “watakushi wa" shosetsu . (Wa is the topic-indicating particle.) See the entry on watakushi shosetsu in NKBD 4:539. As an alternative, Kasai Zenzo elected to call his shishosetsu jiko ("self") shosetsu , although this term, too, is problematic.

which one is associated at a given moment than on one's own unilaterally initiated thoughts and acts. We can think of a true pronoun as a sign of separate and autonomous presence, marking art indelible boundary between self and other. The existence of only one first-person pronoun in western languages like English makes that presence all the more inviolable. The abundance of first-, second-, and third-person pronominals in Japanese, however, each used in accordance with the speaker's social relationship with a specific hearer and referent, tends to undermine that sense of separate, autonomous presence and blur the boundary between self and other.^[4] In short, self-expression in Japanese is forever a contingent activity, dependent on the relationship between speaker and hearer and, by extension, between writer and audience. Shi/watakushi shosetsu becomes in this way a metonymy for what is in Japanese a continually variable communicative act. The narrator cannot even utter the word watakushi or boku or ore until he has posited a specific relationship with the narratee. The shishosetsu , then, is as much "we-novel" as "I-novel."

Reality, Mediation, and the Fictional Contract

Creating a viably intimate narrative relationship is, however, fraught with contradictions and forces each writer to come to terms with the nature of writing. Insofar as the shishosetsu is a prose fiction, it can never be a completely personal communicative act; the narrator does not know the recipient of his discourse in the same way as, say, the writer of a letter. As Roland Barthes exuberantly claims: "Writing is the destruction of every voice, of every point of origin.... As soon as a fact is narrated no longer with a view to acting directly on reality but intransitively ... this disconnection occurs, the voice loses its origin, the author enters into his own death, writing begins."^[5]

[4] It is interesting to note in this connection that words like onore have been used historically as either a first-person or second-person pronominal, depending on context, while the modern reflexive pronominal jibun , depending on context, can signify the first, second, or third person.

[5] "The Death of the Author," in Image, Music, Text , 142. (Emphasis in original.)

And yet, we find, the shishosetsu is weaned most reluctantly from its author. Or, to put it another way, the shishosetsu resists becoming "writing" in Barthes's sense and flaunts its "personalism" by encouraging the reader to disregard its textual boundaries and view "real" world and "fictional" world as an unbroken continuum. The question of an author's "presence" in the text has assumed great importance in Japanese letters, as it alone is believed capable of establishing the text's "authenticity" (that is, autobiographical purity or nonfictionality), the highest value attached to writing. We shall discover that in Japan the notion of what is "real" or "authentic" is traditionally limited to personal observation and experience, with the result that fiction, insofar as it deviates from what "actually" happens, connotes a "fabrication" inapplicable to reality rather than a plausible, equally valid version of it. A story written in 1920 by Uno Koji (1891-1961), which relates one man's affair with a country geisha, plays on this notion. The narrator-hero interrupts his account and addresses his readers:

I have mentioned that I am a writer. When Hitogokoro , which I had completed during my previous stay in Shimosuwa ... was published last January, it created something of a sensation there, small town that it is, because I modeled the character "Yumiko" after Yumeko. Surely any intelligent reader has noticed a peculiar development in much recent Japanese fiction, namely the appearance of a nebulous figure known only as "I." One reads nothing about his looks, much less about his behavior or his profession. What does one read, you ask? Merely a string of impressionistic musings. Soon you realize that the "I" is none other than the author himself; indeed, this is almost always the case. The "I," therefore, is a writer, and the reader never seems to question this curious convention whereby the author uses first-person narrative to point to himself. Now, even though there is nothing wrong with having the narrator double as the hero and the hero pose as a writer, it is regrettable that readers have come to equate him automatically with the author and think of all the story's incidents as actually having taken place. At any rate, because Hitogokoro is written in the first person, people believed everything in it to be true and assumed that I had based my character "Yumiko" on Yumeko—which was all right by me, but I'm afraid it caused Yumeko no little consternation I hurriedly sent her a letter of apology.^[6]

[6] Amaki yo no hanashi , in Uno Koji zenshu 2:442-43. First published in Chuokoron (Sept. 1920). Translations from the Japanese are original to this study, unless otherwise indicated.

This often-quoted passage may be the first literary reference to what only later became known as the shishosetsu , and it well illustrates the kind of osmotic relationship that exists between author and narrator-hero—the one seemingly flowing into the other through a very permeable text. Uno knew that his audience would immediately recognize the by-then familiar formula of author = narrator = hero. A careful reading of Uno's discursive aside, however, suggests that it is not a critique. What sounds at first like an ironic disclaimer turns out to be an apt description of this very story. Indeed, it seems calculated to collapse the distinction between real world and fictional world by enticing readers into believing that the author has recorded events from his own life rather than invented them. The narrator laments the "regrettable" equation that readers ordinarily make between author and literary persona, yet implies that he too is "none other than the author himself" by listing in rapid succession the many properties that he the narrator and Uno the author have in common: the same profession ("I have mentioned that I am a writer"), the same piece of writing (Hitogokoro , which appeared in the same magazine as this story), and the same personal life ("people ... assumed that I had based my character 'Yumiko' on Yumeko—which was all right by me “).^[7]

Much of this story's impact, then, derives from the conceit, which Uno simultaneously ridicules and exploits, that the real-life author and his literary persona are one and the same. This, we might think, is impossible. Uno the author exists independent of the literary medium; Uno the persona, because of it.^[8] And in fact

[7] Elsewhere the narrator engages in a game of name-dropping, presumably to establish himself further in the readers' minds as a historical personage rather than as an imaginary character. He chats with another guest at the inn where he is staying who claims the acquaintance of Tanizaki Jun'ichiro, the famous author. "In that case, you must know Sato Haruo, too," he retorts, for he is aware, as no doubt are his readers, that Sato and Tanizaki were close friends. The guest quickly takes the hint and tells of his drinking acquaintance with Nagata Mikihiko, another of Tanizaki's literary friends. Ibid., 427.

[8] Distinctions between author, "implied" author, and narrator on the one hand and between reader, "implied" reader, and narratee on the other should be kept in mind when analyzing a highly reflexive form like the shishosetsu . See Seymour Chat-man, Story and Discourse , 147-51 and passim, for an account of the narrative process. See also Gérard Genette, Figures of Literary Discourse , esp. chap. 7 ("Frontiers of Narrative"), for an explication of narrative distinctions relevant to this study, as well as Narrative Discourse , by the same author. References in this study to shishosetsu writers, insofar as they are the "heroes" of their own texts (as suggested by William E Sibley in his The Shiga Hero ), will accordingly be to the literary personae as opposed to the extraliterary persons.

what will be argued here is that author and persona even in a shishosetsu are not fully interchangeable, because authorial "presence" must always be, in the final analysis, a product of representation. This is not to say that we can simply isolate persona and text from the author and his life, for they are all part of a literary and cultural field that determines in large measure how the reader regards—or indeed whether he has any interest at all in—such notions as "author," "intention," "text," and "fiction." To confuse persona with author, however, no matter how close the resemblance, is to confuse the telling with lived experience: the former is accessible to the reader in a way that the latter simply is not. Gerard Genette, in his study of narrative, reminds us, for example, that "even the references in Tristram Shandy to the situation of writing speak to the (fictive) act of Tristram and not the (real) one of Sterne; but in a more subtle and also more radical way, the narrator of Père Goriot 'is' not Balzac, even if here and there he expresses Balzac's opinions, for this author-narrator is someone who 'knows' the Vauquer boardinghouse, its landlady and its lodgers, whereas all Balzac himself does is imagine them; and in this sense, of course, the narrating situation of a fictional account is never reduced to its situation of writing."^[9]

The common practice in the Japanese literary establishment, or bundan , however, has been to reduce the narrating situation to its situation of writing. We shall see in Chapter 2 that the structure of the Japanese language provides some justification for this reduction, since the language's fundamental speaker orientation is most compatible—regardless of narrative person—with a limited point of view that is easily identified with the author's own.^[10] There is of course no dearth of examples in twentieth-century western literature that invite a similar practice (one recalls the writings of Gide, Miller, and Proust), but the emphasis on fictional autonomy has prevailed.^[11] The bundan , meanwhile, has historically believed lived

[9] Narrative Discourse , 214. (Emphasis in original.)

[10] Genette's strict distinction in Narrative Discourse , 185-89, between "point of view" (who sees?) and "voice" (who speaks?), therefore, is not entirely valid in the Japanese case. We will find that ambiguity of narrative voice is structurally built into the Japanese language. See the discussion in Chapter 2, below.

[11] Proust, for example, insists in Contre Sainte-Beuve that a literary text not be confounded with the life of its author and that in the text the writer creates a self distinct from his biographical one. See Stephen G. Kellman, The Self-begetting Novel , 17, 21.

experience to be more "authentic" than any literary version of it, and has read the shishosetsu in the confidently referential manner in which we read a biography or legal document, finding meaning in it only insofar as it sheds light on the author's private life. In this view, the shishosetsu narrator is less a storyteller than a spokesman who has eschewed fictionalization in favor of direct expression. Narrative is thus wholly subordinated to discourse.

A writer's notion of fiction is of course predicated on his concept of the "real." The Japanese view of history, the individual's place in society, the native literary tradition, and the language itself have all been instrumental in forming the concept of reality through which the early twentieth-century writer regarded his world and against which notions of fiction had to compete. As used here "fiction" suggests any plausible, alternative vision of "reality," rather than simply the "fantastic" or the "nonreal." John Fowles offers a credo for the fiction writer when he says, “We wish to create worlds as real as, but other than the world that is. “^[12] The great tradition of realism that flourished in the nineteenth century attests to the essential compatibility in western culture between the "imagined" worlds one reads about in novels and the "real" world of personal experience. It is less important that an event has actually occurred than that it could have occurred, given the particular situation. Fictional world and phenomenal world are both based on the premise, underwritten by the post-Enlightenment secularization of culture, that "reality" is as much mediated by the human mind as it is an entity unto itself; it is a function of the very manner in which it is apprehended and then represented.^[13] Fictive imagination is just as much a part of this reality as actual experience, because both are products of mediation and are in this sense created.

Fowles's fictional credo, however, does not square with the Japanese perception that the unmediated presentation of lived experi-

[12] The French Lieutenant's Woman , 81. (Emphasis in original; the quotation appears in chapter 13.)

[13] We should note the important distinction between the "gentile" or historical, and the "sacred" or original, that Vico makes in his philosophy of a human history. See the concluding chapter of Edward W. Said's Beginnings , esp. 347-53, for an excellent introduction to Vico's thought and a discussion of the evolution of an intellectual environment conducive to narrative creativity.

ence is possible when "freed" of fictional elements. Maruyama Masao, in a highly suggestive discourse, attributes the dearth of fictional "imagination" in modern Japanese literature to a traditional social organization that can accommodate only experience that is directly perceived. Maruyama distinguishes between two kinds of reality, "mediated" and "immediate," and argues that the western intellectual tradition stresses the former while the Japanese tradition stresses the latter. Modern society is founded on "mediated" reality, that is, "fictions" of social organization that transcend direct and tangible interpersonal relationships (such as the lord-vassal contract) and see people as individuals capable of independent action. Westerners, very much at home since the Renaissance with abstractions to govern their legal, philosophical, and social lives, have applied them naturally to the literary realm as well. The term "fiction," then, can have very positive connotations in the west, for it means "to fashion" or "to invent" as well as simply "to imagine" or "to pretend." Such a view of fiction derives from a fundamental faith in "mediated" reality, a faith so great that it rates the product of intellectual activity higher than perceptual or "immediate" reality. Fiction is natural in a society that believes individual action to be a determining force in social behavior and the natural order. "It is precisely because the reality does not appear directly," says Maruyama, "but as a 'mediated reality' depending on the positive participation of the human spirit, that we can call it fiction."^[14]

In Japan, however, man's intellectual and spiritual side is neither differentiated nor independent from perceptual nature. Like medieval European society, Maruyama argues, Japanese society sees interpersonal relationships as static and irrevocable. It naturally follows that the social environment, which modern society recognizes as man-made, assumes for the Japanese the same kind of reality ("immediate" rather than "mediated") as the natural world. Human intercourse is conceived of in the same dimensions as birth, death, or the passing of the seasons. "When people view their society as a fait accompli, “ Maruyama suggests, "they don't just automatically start thinking in terms of fictions."^[15] The result is that the Japanese, who feel uncomfortable about fiction, try

[14] "From Carnal Literature to Carnal Politics," 251.

[15] Ibid., 255.

to push it in the direction of firsthand, perceptual experiences. Whereas the western view of fiction is rooted in a fundamental faith in mediated reality, unmediated "fact," in the Japanese view, becomes itself a kind of faith. Maruyama concludes that the notion of fiction as the intellectual construct of an individual mind is rooted in modernity, for it could not exist without the individual imagination freed from a predetermined social order. Fiction becomes "real," that is, a viable entity, only when one sees oneself as independent of the natural environment and views intellectual constructs not as useless abstractions but as thoroughly relevant to daily life.

Although Maruyama's analysis reveals a socioevolutionary slant that virtually equates post-Restoration Japan with pre-Enlightenment Europe, it articulates persuasively a "modern" society's need to employ fictions and a "premodern" society's absence of such a need. In a culture that views "reality" only as immediate experience of the natural world, literature not surprisingly becomes a chronicling or transcribing of that experience rather than an imaginative reconstruction of it. And in a culture that views human relations as a predetermined part of the natural order, a fictional form such as the novel, which thrives on fluidity in human relations, is clearly out of place. Traditional Japan was such a culture; so, to a considerable degree, were Meiji and Taisho Japan. And if one accepts Nakane Chie and others' arguments that the transformation in Japan from a preindustrial to an industrial society was not accompanied by the kind of radical change in social organization that occurred in the west,^[16] then it comes as no surprise that the novel's emphasis on individual autonomy and on serendipitous relationships is not to be found in abundance in the early twentieth-century shosetsu .

Yet it was precisely the rise of individual autonomy and serendipitous relationships, Ian Watt suggests, that paved the way for the development of modern fiction in the west. Like Maruyama, Watt argues that modernity is the product of an individual-centered world in which one is "responsible for his own scale of moral and social values." Not only did the rise of individualism weaken communal and traditional relationships, it fostered the

[16] See Nakane Chie, Japanese Society , esp. 7-8.

"stress on the importance of personal relationships which is so characteristic both of modern society and of the novel—such relationships may be seen as offering the individual a more conscious and selective pattern of social life to replace the more diffuse, and as it were involuntary, social cohesions which individualism had undermined."^[17]

Beginning with Montaigne and Rousseau, the self in western literature has continually been celebrated and probed. The driving force in "classical" fiction (whether autobiographical or not), from Tom Jones and Emma to Wilhelm Meister and Rudin , has been socialization through individuation, achieved by rigorous personal scrutiny and dynamic interpersonal relationships and leading to the protagonist's confrontation and integration with society. A character's sense of "self" is brought about by a strong awareness of, and interaction with, the "other." Even such antisocial acts as, say, Werther's or Madame Bovary's suicides are depicted as having been committed in despair by individuals who had been thoroughly engaged in their societies.^[18] The Japanese, however, have been less convinced of (we might even say they have been uninterested in) the self's tangibility or value, even after their massive exposure to western civilization in the Meiji period. The driving force in much of Japanese fiction, therefore, has been what might be called "naturalization" rather than socialization, achieved by a studied withdrawal from society and often leading to the hero's identification with nature.

The often-discussed "search for selfhood" (jiko tankyu ) in modern literature actually harks back to the traditional "pursuit of the

[17] The Rise of the Novel , 177.

[18] The goal of self-determination in the west may have proved increasingly elusive since the late nineteenth century, but the individual has remained, nonetheless, the point of reference. Despite volumes of criticism in Japan and the west to the contrary, however, all emphasizing the impact of western thought, modern Japanese fiction springs from a tradition that did not take the individuated self as its point of departure and against which, therefore, it could not react in expressions of fragmentation and alienation so familiar to us in modernist literature. It is true, of course, that many heroes in western fiction, beginning perhaps with Crime and Punishment and continuing with such works as A Portrait of the Artist as a Young Man and The Stranger , have become less and less socialized and that alienation from one's society is a theme common to both twentieth-century western and Japanese fiction. The difference, however, is that alienation in western fiction still represents a deviation from the norm, whereas in Japanese fiction (such as the works of Shiga, discussed in Part 3) it can actually approach an ideal.

Way" (gudo ). Both are steeped in the Buddhist emphasis on liberation from the bonds of society and on yielding to life's inexorable cycle: the "four trials" of birth, aging, sickness, and death. The achievement of selfhood in the Japanese context, then, means in a very positive sense the loss of one's individuality. This is a recurrent theme in many shishosetsu but most conspicuously in those of Shiga Naoya, whose "Kinosaki nite" (At Kinosaki) and An'ya koro (A dark night's passing) are particularly powerful depictions of the authorial persona's communion with nature and of what the early twentieth-century philosopher Nishida Kitaro calls "emptying the self," achieved by the perceiving subject's identification with the object.^[19]

shishosetsu are saturated with the author's personal experiences, to be sure, but not with the aim of revealing a core of personality that is clearly defined in the process of socialization or self-analysis, as we might expect in western literature. It is ironic that the so-called "I-novel," treated by many Japanese critics as the showcase of the "modern self" (kindai jiga ), actually questions what it ostensibly champions.^[20] Indeed, the shishosetsu is less the vehicle for personal affirmation that critics frequently make it out to be than a theater where the hero acts out his ambivalence toward the self. The medieval scholar Nishida Masayoshi, one of the few to reexamine the shishosetsu from a non-Eurocentric perspective, sees little validity in claims for the shishosetsu as an expression of a burgeoning individuality appropriate to a "modernized" Japan; indeed, he is skeptical about the entire enterprise of reading into the shishosetsu the awakening of a western notion of selfhood.^[21] Takahashi Hideo, in his revisionist study, also urges that we revise our notions of the self in Japan. Rather than equate it with modern west-

[19] The Problem of Japanese Culture , 869.

[20] See Yasaki Dan, "Jiga no hatten ni okeru Nihonteki seikaku" (part of a longer work entitled Kindai jiga no Nihonteki keisei [1943]), for a sample of that thriving subgenre, the kindai jiga essay, which takes the "modern self" to be, in the words of Kenneth Strong, "the key to the vagaries of modern Japanese literature, or at least of modern Japanese fiction" ("Downgrading the 'Kindai Jiga,'" 407). (Emphasis in original.)

[21] "The code of the ‘watakushi ‘ in shishosetsu ," writes Nishida, "is the obliterating of self by confession. It would seem to have succeeded remarkably in enriching the writer's spiritual quest by aspiring ultimately to the kind of 'ego-shedding' [da-tsugasei ] that approaches the emancipating experience of the 'no-self' [muga ] of traditional Buddhism" (Watakushi shosetsu saihakken , 72).

ern society's individuated self, which has been largely weaned from nature and tradition, he concludes it is more properly conceived of as a premodern "element" of man that aspires to fusion with its surroundings. The shishosetsu's watakushi has brought new meaning to the "self," Takahashi argues, not by an assertion of individuality but by its reversion to selfless ness, as exhibited in the protagonist's return to the comforting embrace of nature, family, and tradition—to which we might add, in the case of those politically motivated writers in the 1930s who underwent ritual "conversion" (tenko ), the state.^[22]

In short, the diffuse, involuntary social cohesions of a strictly hierarchical society suggested by Maruyama were still very much intact in early twentieth-century Japan. The Japanese writer in the Taisho, let alone Tokugawa, period, never knew the range of selection and choice in human relations experienced by an individual in the England of Defoe, Richardson, and Fielding that Watt describes. Contact with western culture, to be sure, awakened writers to the idea of individualism. Yet while individualism in the west suggests a dynamic (as opposed to regulated or tradition-bound) relationship between the self and the other, it came in Japan, because of the traditional equation of spiritual autonomy with aloofness from society, to imply just the opposite: a withdrawal into the world of nature and private experience. This notion of individualism as a form of isolated self-contemplation may have been the only avenue to spiritual independence in a society that placed severe constraints on interpersonal relations. The typical Japanese writer, moreover, aware of the limits of his intellectual and social liberties, may have seen in his private life the only area he could exploit with confidence. In a society generally hostile to nonconformist behavior, he discovered in confession a literary form that matched his conception of self: something fulfilled by unilateral and almost instinctive expression rather than by integration in a matrix of human relationships. He saw confession, moreover, as an embodiment of his uniqueness, insofar as it could be authored only by one person, namely, himself. The shishosetsu was eminently suited to this view of individualism. Its protagonist becomes an "individual," as it

[22] Genso to shite no "watakushi. “ See the introductory and concluding essays, esp. pp. 23-24 and 288-89.

were, by virtue of his monopolized point of view and, not infrequently, by his occupation of an otherwise empty stage.

The Place of Fiction in Japanese Literature

As we have seen, the starting point in much Japanese "fiction" has not been the construction of a hypothetical situation but the observation of an actual one. This modus operandi is hardly unique to the shishosetsu ; precedents can be found in such classical forms as the zuihitsu (discursive essay), kana nikki (poetic diary), haibun (haiku and prose), and kikobun (travel sketches), all of which are literary descriptions of lived experience. Even the stories and settings in an innovative and "fantastic" form like joruri are generally rooted in actual events or previous literary accounts. Both "contemporary" (sewamono ) and "period" (jidaimono ) pieces are based on a large but finite number of "worlds" (sekai ), well-known historical events or literary depictions with fixed casts of characters.^[23] The accepted practice was to work with material already familiar to audiences rather than fashion entirely imaginary characters or situations. Playwrights worked within these established frames of reference, fleshing out the action and characterization as they saw fit.^[24]

The radical transformation of the old literary language into something approaching the colloquial, the experiences of modernization and urban growth, and the influx of western thought, all of

[23] The Sekai komoku (late eighteenth century) lists some 150 "worlds" along with their historical or literary source and names of characters. See Barbara E. Thornbury, Sukeroku's Double Identity , 22-23.

[24] James R. Brandon notes, "Precisely because a world was already significant in legend or history, or in the case of sewamono through public scandal, it was considered appropriate for the stage" (Kabuki: Five Classic Plays , 25). The same is largely true of other theatrical forms, including No and Bunraku. Kato Shuichi argues that there is little in Kabuki of "the universal, the paradoxical or the abstract" and that the playwrights' lines concern specific characters and situations but "say nothing about human feelings in general or about the human condition" in the manner of a Shakespeare. The reason for this, he concludes, is that Edo-period townsmen were interested in "specific subjects, not universal considerations." The somewhat negative cast of Kato's argument aside, this concern with specificity over universality would seem to apply equally well to post-Edo fiction. See A History of Japanese Literature , vol. 2, The Years of Isolation , 201.

which preceded the shishosetsu's emergence, insured that the gap between classical and modern Japanese literature would be unbreachable in certain respects; yet we should not overlook the modern writers' attempts to come to terms, in various and subtle ways, with the classical forms. It was, of course, never simply a matter of incorporating the classical vocabulary or aesthetic. For more than a millennium, the language served as a repository of richly connotative images and associations shared by the literate community.^[25] With the demise of old Japanese after Ozaki Koyo (1867-1903), however, the familiar vocabulary of pivot words (kakekotoba ), prefaces (jo ), epithets (makura kotoba ), association words (engo ), and poetic place-names (uta makura ), which for many centuries had provided a context for meaning and had made such ostensibly "personal" genres as the zuihitsu and kana nikki both highly conventional and more accessible to their audience, died out rather suddenly with no comparable system of literary conventions readily available to take its place. By the turn of this century, Japanese writers were looking to naturalism and other European literary movements for inspiration; but the western tradition, itself the culmination of many centuries of independent development, could not be readily assimilated. Writers now faced the task of formulating a new poetic vocabulary and repository of associations on which to draw. Unable to rely any longer on the "worlds" and associations of classical literature or in any coherent way on an alien literary tradition, they began exploring the possibility of using their own lives as "world." Once the writer established his persona as a legitimate subject of literary discourse, he was working, as far as he and his audience were concerned, with familiar material and could allude to it in subsequent works in the knowledge that readers would be conversant with it. This "world" gained further legitimacy as its author gained a name; and personal experience, as presented in the work, became part of the public literary domain. Just as Sei Shonagon (fl. early eleventh century) challenged her readers in The Pillow Book to pit themselves along with her against the tradition of literary allusions, the early shishosetsu writer chal-

[25] The extraordinary richness of this repository is demonstrated by Mark Morris in "Sei Shonagon's Poetic Categories," 29.

lenged his readers to gain a similar fluency in the newer and narrower "tradition" of his recorded life. He sought in this way to validate his experience as a kind of history.

The shishosetsu writer was preoccupied with such a legitimizing project in the first place in part because of fiction's subordinate relationship to history in traditional Japanese letters. We have observed that any author's view of fiction is predicated on his notion of reality, which is itself a mutable entity; concepts of just what constitutes reality vary with the age, culture, and artistic temperament, as E. H. Gombrich, among others, makes clear.^[26] The common understanding in the west has been that literary fictions are very much a part of "reality"; fiction, no less than history, aims at a credible representation of the world. Far from being opposed to history, it offers a competing and basically similar strategy of representation. (Henry Fielding, after all, calls his great novel about a foundling a "history.") Many novels could pass for histories, and vice versa. Hayden White goes so far as to say: "Viewed simply as verbal artifacts histories and novels are indistinguishable from one another. We cannot easily distinguish between them on formal grounds unless we approach them with specific preconceptions about the kinds of truths that each is supposed to deal in. But the aim of the writer of a novel must be the same as that of the writer of a history. Both wish to provide a verbal image of 'reality.'"^[27]

The difference between the two, then, is a very subtle one indeed. It certainly does not lie in the dubious notion that the historian "finds" his stories, whereas the fiction writer "invents" his. Such a notion obscures the extent not only to which "invention" plays a part in the historian's operations but also to which "discovery" can play a part in the fiction writer's operations. It is tempting to say that history deals with what actually happened (events already constituted) and hence with "truth," and fiction with what might have happened (events not already constituted) and hence with "untruth." Already constituted events, however, are by no means the historian's exclusive domain. They are a part of any "fictional" work as well. Likewise, events that exist "inside" the writ-

[26] See Art and Illusion , esp. 84-86, which discusses the triumph of an established artistic vocabulary over an unfamiliar "reality" or setting.

[27] Hayden White, "Fictions of Factual Representation," 122.

er's consciousness are not the exclusive domain of the fiction writer. In his Metahistory , Hayden White posits the fictive nature of historical reconstructions, which are generated necessarily by various contending modes (romance/metaphor, tragedy/metonymy, comedy/synecdoche, and satire/irony) of emplotment, and argues convincingly that no mode of conveyance of historical "fact" is neutral or definitive or inherently truer than any other. "Commitment to a particular form of knowledge," White concludes, "predetermines the kinds of generalizations one can make about the present world."^[28]

At the center of this view of historiography as an essentially poetic act lies the assumption that history, like fiction, is a form of narrative. Although White never specifically defines "narrative," it is clear that he uses the term to designate a verbal construct that is more complex than either a chronicle (a chronological, open-ended arrangement of data) or a story (a chronicle to which culminations and resolutions have been added) and utilizes a plot—that "provider of significance to mere chronicity" as Frank Kermode puts it.^[29] Unlike a simple story that is content with a chronological ordering (asking the question "What happened next?"), an emplotted narrative imposes an internal, nontemporal, overriding order on events (asking, "What is the point of it all?"). The aims of both novelist and historian are realized through this synchronic ordering process.

White takes both "chronicle" and "story" to be "primitive elements" in the historical (and by inference the fictional) account;^[30] but it is precisely within this "primitive" realm that much of Japanese history and fiction reside. Although narrative in the west, beginning with the Old Testament, has been informed by a dramatic poetic of emplotment that insists on a causal relationship between events,^[31] the traditional Japanese "narrative" (monogatari ) is informed by the highly conventionalized aesthetics of the native verse forms and the nondramatic historiography of Chinese chronicles.

In the absence of a highly representational mode, the influence

[28] Metahistory , 21.

[29] The Sense of an Ending , 56.

[30] Metahistory , 5.

[31] See, for example, Erich Auerbach's analysis of Genesis as an emplotted narrative in Mimesis , chap. 1.

of a more self-consciously presentational mode on Japanese fiction has been enormous. The latter mode has played no small role in western literature as well—as the continuing interest in tropes, for example, demonstrates—but it has not had the sweeping impact, especially on prose, that it has had on Japanese literature. Whereas prose in the west, whether "history" or "fiction," has been inextricably tied to the emplotted narrative, the Japanese have been more at home with literary forms that tend to undermine or circumvent the narrative flow. The presentational mode infuses not only waka, haikai , and other poetic forms but also such prose forms as the "fictional tale" (tsukuri monogatari ), "poem tale" (uta monogatari ), "story of the 'floating world"' (ukiyo zoshi ), "poetic diary" (kana nikki ), and discursive essay (zuihitsu ), to. name the most prominent examples. In all these forms, imagery, polysemy, and canonical allusions do not simply counterpoint the action but may actually interrupt or even overwhelm it.^[32]

Although Japanese prose had by the early twentieth century lost not only the rhythm but the poetic vocabulary of the old literary language, it was by no means on a comfortable footing with the emplotted narrative still so prominent in western fiction. The difference, of course, is the cause of much consternation for western readers of the shosetsu who even today talk of its "inaccessibility," bemoan the "lack" of an ending, and ask what is the "point" of the story they have just read. The shosetsu , however, needs no overall plot signifier to enable its discourse. Getting there, for readers of the shosetsu , is not just half the fun; it is all the fun. In fact, there is no "there" or "here"—that is to say, no privileged event indicated by the emplotting narrative mechanism; the shosetsu narrative goes everywhere and nowhere.

The Chinese annalistic model for the writing of history also had a considerable influence on Japanese letters, including modern shosetsu . Beginning with Nihongi and the other "Six National Histories" (Rikkokushi , 720-902) and continuing with the hybrid "fictional histories" (rekishi monogatari ) written in the late Heian (794-

[32] The presentational and serial nature of Genji and other Heian monogatari is brilliantly analyzed in Richard Hideki Okada, "Unbound Texts." Of particular interest is his discussion of the tenseless character of old Japanese and the ambiguity of narrative voice in monogatari , a feature that survives in the modern shosetsu . See Chapter 2.

1192) and Kamakura (1192-2333) periods, Japanese texts take after the Chinese chronicle in their essentially descriptive rather than investigative bent and likewise make no attempt at synthesis or at establishing a hierarchy of significance. What mattered was that an event had actually occurred; having: occurred, it was dutifully recorded. Even more important, the continental tradition guided Japanese thinking as to what constituted history and ultimately what constituted literature. The Chinese regarded literature as a public institution that had as its purpose the maintenance of the Confucian orthodoxy, a political and social system of bureaucratic government entrusted to a class of highly literate civil servants who owed their place in the administrative ranks to their fluency in the four major categories of literature: histories and biographies, the classics and their commentaries, philosophy, and belles lettres.^[33] The last category included poetry but no prose; fiction was not included in scholarly discourse until the influx of western literary theory in the twentieth century. Traditional commentators who wrote on prose fiction at all quite naturally criticized it in terms of history, the highest form of literature and the only body of prose narrative that, because of its unquestioned respectability, could serve as a standard of comparison.^[34] Even during the heyday of Chinese prose fiction in the Ming (1368-1644) and Ch'ing (1644-1911) dynasties, most authors wrote anonymously and lived out their laves in obscurity—the price to pay for indulging in such a minor art.^[35]

The content of "literature," then, for all intents and purposes, was poetry and nonfiction prose. The former served as a vehicle of expression and the latter as the moral guideline for the literati class. The highest forms of literature were public and utilitarian in orientation and meant to aid in the art of government. The Japanese civil service, under the control of either the aristocratic or military classes, never became as dominant as its Chinese model, and we see both literature and history take on a less public face. Rekishi monogatari , for example, are more likely to describe an outing of

[33] Liu Wu-chi, An Introduction to Chinese Literature , 5.

[34] C. T. Hsia, The Classic Chinese Novel , 14-15. Ssu-ma Ch'ien's Shih chi (Records of the Grand Historian ) was the most frequently invoked historical text against which classical Chinese prose fictions were measured.

[35] Liu, An Introduction to Chinese Literature 195, 228.

court ladies than a meeting of ministers. In the Edo period, however, the Tokugawa shogunate developed a deeply conservative political system that looked eagerly to the Chinese Neo-Confucian philosophy to legitimize its rule. Literature came to be regarded, officially at least, solely as a vehicle for propagating the dominant sociopolitical ethos. Indeed, literature was understood to be useful only insofar as it contributed to the Confucian worldview or "Way."^[36]

Literature that inspired the populace to socially correct behavior, then, clearly served a useful function in the Confucian scheme of things. In Tokugawa Japan, the writing of histories, biographies, and moral treatises was encouraged, while other forms were merely condoned. The poet Matsuo Basho's (1644-94) witty, self-deprecating assessment of haikai —"My art is like a fire in summer or a fan in winter; it serves people no purpose"^[37] —expresses, however ironically, the prevailing official bias against any forms of poetry except kanshi (poetry written in Chinese) and waka . The bias against prose fiction was even stronger and also reflected Chinese literary tastes. The word for prose fiction itself (shosetsu [hsiao-shuo in Chinese] originally meant "unofficial history" and referred to popular, loosely historical accounts written in the vernacular) is indicative of the low esteem in which it was held. Shosetsu might be translated literally as "small talk"; a shosetsuka was therefore someone who collected "street talk" and "roadside gossip" and committed them to writing.^[38] The Chinese expended great efforts preserving and handing down histories and the other categories of "refined" literature; the "vulgar" texts of prose fiction, however, were commonly ignored, lost, and even destroyed.^[39] The Han shu (History of the Former Han) lists in its catalogue of learning ten "sages" in descending order of importance—the Confucianist, Tao-

[36] The Confucian scholar Hayashi Razan (1583-1657), adviser to the first Tokugawa shogun, writes: "Where there is the Way [do] there are letters [bun ].... They are different manifestations of the same principle.... Letters branch out from the great trunk of the Way. Since the branches are few, they are healthy and firm" (quoted in Hiraoka Toshio, Nihon kindai bungaku no shuppatsu , 13). Razan's concept of the nature of literature echoes the words of Emperor Wen of the Wei (A.D. 186-226): "Literature: a vital force in the ordering of the state" (quoted in Burton Watson, Early Chinese Literature , from the epigraph opposite the title page).

[37] Quoted from "Kyoroku ribetsu no kotoba," in Sugiura Sei'ichiro et al., eds., Basho bunshu , 205.

[38] Liu, An Introduction to Chinese Literature , 141.

[39] Watson, Early Chinese Literature , 4.

ist, Legalist, diviner, logician, Moist, diplomatist, eclectic, farmer, and finally "small-talk writer"—but quickly adds that only the first nine are worth mentioning.^[40]

The notion of what constituted literature in Tokugawa Japan is examined with remarkable insight by Kitamura Tokoku (1868-94), a noted poet, essayist, and critic. Tokoku posits two distinct strains in Tokugawa letters: "refined" nonfiction literature that commented seriously on life and "vulgar" nonliterary fiction that buriesqued life.^[41] Into the former category went the histories, biographies, and other erudite texts that appealed to the ruling samurai class, in which fiction had no place; into the latter, the witty picaresques and other "popular" narratives that appealed to commoners, in which fiction figured large. Tokoku's distinction is noteworthy, for it shows that Tokugawa literature had no lack of fiction. On the contrary, a great deal was being written in the form of gesaku , a generic term for such forms as the kibyoshi (illustrated "yellow-covered books"), sharebon ("sophisticated books" about the gay quarters), kokkeibon ("humorous books"), ninjobon ("amatory books"), and yomihon (didactic "reading books"). Its very name (literally, "playful composition"), however, suggested its "frivolous" and therefore nonliterary character. Gesaku were often authored (usually anonymously) by samurai intellectuals, but their primary audience was the lower classes, mainly townspeople.

The idea that fiction and literature were entirely separate concerns—the former frivolous, the latter serious—continued into the Meiji period. It held true even for the man who made the first sustained attempt at writing in the colloquial idiom. Futabatei Shimei (1864-1909), author of Ukigumo (Drifting clouds, 1887-89), abandoned his effort at fiction writing for nearly two decades before trying again in 1906, only to abandon it for good a year later. Influ-

[40] Noguchi Takehiko, shosetsu no Nihongo , 10-11. That this low view of fiction has continued even into modern times is demonstrated by an influential essay by Chou Tso-jen, written in 1919, which argues for the superiority of the classical genres of prose and poetry over fiction and drama. See Hsia, The Classical Chinese Novel , 3-4.

[41] Nihon bungaku shi kotsu (also known by the subtitle of its only extant portion, Meiji bungaku kanken ), in Kitamura Tokoku shu , 124. My discussion of Tokoku is based largely on Hiraoka, Nihon kindai bungaku no shuppatsu , esp. the section entitled "Futatsu no bungaku," 7-24. See also H. D. Harootunian's stimulating article on Tokoku and Takayama Chogyu, "Between Politics and Culture."

enced as much by his samurai lineage and Tokugawa heritage as by his studies of western literature, Futabatei was very uncomfortable with the idea of a literary career, which he regarded as decidedly unprofessional. The contemporaneous success of the Ken'yusha, a school of writers that carried on the Edo gesaku tradition, no doubt increased his distaste for such a "frivolous" calling. In a late essay, he insists on the impossibility of writing fiction of any worth. "No matter how good one's technique," he argues, "one cannot write the truth. One may know the truth, but it inevitably becomes distorted when one speaks or writes of it.... When it comes to fiction [shosetsu ] I can write nothing but lies; and because I so believe, I simply cannot be serious about it."^[42] In the end, the only way he could be honest with himself was to quit the life of a writer and embark on a more respectable career in the bureaucracy. Futabatei continued his literary activities both as a translator of stories from the Russian (which won him his initial fame) and as a contributor of kansobun (short pieces that allow a writer to reflect on virtually any subject) to various journals, but he considered this part of his life an avocation.

By the early twentieth century, the original sense of shosetsu had been largely supplemented, if not superseded, by the western concept of the novel that Tsubouchi Shoyo (1859-1935), author, translator, critic, and student of English literature, first grafted onto the Chinese word when he published his famous interpretation of European literature (shosetsu shinzui ; The essence of the novel, 1885-86). And yet, the west's literary impact notwithstanding, the word shosetsu could not be stripped so quickly of its two-thousand-year-old connotations. As Noguchi Takehiko suggests, such connotations explain not only why Edo-period writers, with a mixture of shame and defiant pride, would use the derogatory term haishi shosetsu ("trivial history and small talk") to refer to their literary productions but also why the term sakka ("artist"; literally, "maker of things") is looked on even today with more favor than shosetsuka in identifying someone as a writer.^[43]

Early twentieth-century writers, then, could not help feeling ambivalent toward their literary heritage. To wear the label of sho-

[42] "Watakushi wa kaigi-ha da," in Futabatei Shimei zenshu 5:230-31.

[43] shosetsu no Nihongo , 12.

setsuka was to acknowledge the vulgarity of one's calling. The Confucian tradition had disenfranchised shosetsu from any of the legitimate categories of literature. Moreover, the new European model was still not so well established that it could counter the prejudice that shosetsu , unlike the classics or histories, did not tell the "truth," that is to say, a referential truth that pointed to some external reality outside the text and gained its significance from that reality. A few, like Noguchi, argue that it was the destiny of the shosetsu as fiction, aided by the positive western conception, to transcend the limits of referentiality imposed by the dominant culture and to present its own kind of truth, a reality that exists nowhere but in language.^[44] Pressures on the shosetsu and on its cultural context, however, worked both ways. To the same extent that its verbal energy as a fictive text tended to propel the modern shosetsu beyond referentiality, the modern shosetsuka , in hopes of gaining respect for his work, tried to contain it within the referential framework of traditionally accepted prose forms and thereby elevate it to the level of "true" ("refined" as opposed to "vulgar") literature. Thus, the legitimizing rationale for the shosetsu as fiction was soon countered by a movement to free the shosetsu of fiction altogether. Out of this movement came the shishosetsu .

The concern about telling "lies" is still very much alive today. Witness the following sentiments, delivered to a lecture audience in 1971, by the contemporary shishosetsu author Yasuoka Shotaro (1920-):

In a shishosetsu you write, somewhat tediously, about your own life and nothing else. A shishosetsu has ... no shape, no form, no style. It has none of these—and yet in writing about that ordinary, everyday life ... a person's unique individuality manages to surface. "Everyone has his quirks," as the saying goes ... but it's very difficult to discover those quirks for oneself.... Still there are those fleeting moments when you become detached and see yourself for what you really are ... just as clearly as if your eyeball were attached to the wall over there. I think that truly creative writing involves that kind of searching for and discovering something in yourself, not dreaming up some formally satisfying story. Anyway, that's how I got started. It took me about half a year to write my first story, a mere thirty pages long. I'm in no position to judge its worth, but writing has made me understand—how shall I put it?—the value of life. Or,

[44] Ibid., 20.

to put it another way, I feel that by writing I'm continually able to affirm my existence.^[45]

Yasuoka is speaking in an age of high democratic capitalism beyond the postwar era, at a time when the sociocultural constraints on "mediated" expression to which Maruyama Masao alluded might be presumed no longer operative. And yet there is no sign, if these words are any indication, that the privileging of lived experience has waned. Its appeal lies on one level in its patently narcissistic value as a kind of literary graffiti that affirms the scribbler's existence. Even more important for Yasuoka are its revelatory powers, its ability to plumb the "truth" about the writer and about life in a way that no "formally satisfying" fictional work can. A story is naturally formless if it follows life, naturally "unique" if it accurately describes its author, whose life is by definition one of a kind. For Yasuoka, then, true originality lies not in the possible but in the actual, not in imagining but in living. Any attempt to stray from lived experience is bound to result in mimicry and fabrication, which in this scheme are seen in opposition to originality. In a move that effectively challenges the western conception of the fictional enterprise, he tells of his efforts to "invent" an original, formally convincing prose fiction—the product, he claims, of a pure flight of imagination—only to discover that a virtually identical story already existed in the rakugo (comic monologue) repertoire.^[46] For Yasuoka, the lesson is clear: no such duplication is possible if a writer draws solely from his own life. Fabrication can only lead one away from the truth of lived experience.

There is still another kind of appeal, however, which Yasuoka only hints at, namely, the voyeurism inherent in reading what one knows to be a record of experience and not invention, a record not so valuable for what it says in itself as for whom it says it of. It may be impossible for the writer to see himself, as Yasuoka argues, except during a few fleeting moments of heightened awareness; but he is always in view of the reader, who is constant witness to the author's "quirks." The shishosetsu may have no shape or form, but it does have a peculiar "style," contrary to Yasuoka's assertion, which is none other than its reputed absence of style. The shishose-

[45] "Gendai ni okeru watakushi shosetsu," 34.

[46] Ibid., 33-34.

tsu's whole raison d'être rests on the powerful illusion of its textual transparency—its sincerity—which lets the reader view the author's experience "unmediated" by forms, shapes, structures, or other "trappings" of fiction. Again and again we shall encounter the belief, articulated here with such conviction by Yasuoka, that the shishosetsu lets the reader view through the window of its "transparent" language the private goings-on in the author's glass house in a way simply not possible through the mediated language of fiction.

One of our central concerns, of course, is just how "transparent" the language of the shishosetsu really is. We may not notice that the glass in this window on life is tinted or warped or even there until it is opened, rolled down, or otherwise exposed. How, if at all, does it differ from the language of fiction? And if it does not, then what makes readers and writers believe nonetheless in a difference, as if literary will took precedence over linguistic evidence? We shall grapple with these questions in the following two chapters.

2
Language and the Illusion of Presence

The essence of writing is to prevent any reply to the question: who is speaking?
Roland Barthes, S/Z

In most books, the I, or first person, is omitted; in this it will be retained.... We commonly do not remember that it is, after all, always the first person that is speaking. I should not talk so much about myself if there were anybody else whom I knew as well.
Henry David Thoreau, Walden

Why the preponderance of first-person narration in modern Japanese shosetsu ? We are led by the insights gained in Chapter 1 to conclude that it is due paradoxically to a lack of a sense of self—and to the concomitant lack of an isolable, autonomous "other." If this is so, however, the term “first-person narration" itself becomes problematic, because, as we noted earlier, pronominals in Japanese do not mark an indelible boundary between self and other. The shishosetsu turns out to be not so much a first-person as a single-consciousness narration (whether first or third person), and one, moreover, that makes no distinction between the narrated and the narrating subject.

To repeat, then, in terms that do more justice to the texts we shall analyze: why the preponderance of single-consciousness or subject-oriented narration in modern Japanese shosetsu ? In Chapter 1, we attributed the phenomenon in part to a worldview, rooted in the indigenous and continental cultures, that privileged "unmediated" over "mediated" reality and made the Japanese (as Maruyama Masao put it) try to push fiction in the direction of firsthand (although not particularly individuated) experiences. In later chapters, we shall examine the influence of contemporary intellectual, criti-

cal, and literary developments on early twentieth-century writers' narrative technique. In this chapter, we shall see what the Japanese language itself can tell us about the preference for subject-oriented narration.

Examples of such narration can be found in all periods of Japanese literature: waka , of course, and in such prose forms as the zuihitsu (Makura no soshi , Tsurezuregusa, Hojoki , etc.), kana nikki (Tosa nikki, Kagero nikki, Sarashina nikki, Towazugatari , etc.), and haibun (Oku no hosomichi, Shin hanatsumi, Ora ga haru , etc.). Even in an ostensibly "third-person" form like monogatari , the narrator's presence is too tenuous and particularized to be considered omniscient; indeed, the storyteller seems at times to merge with the characters.

In the classical western novel, from Fielding to Fowles, the narrator's omniscient perspective is validated solely by the novel's internal consistency; the narrator need not be situated vis-à-vis the characters in any concrete relationship, since s/he commands a suprahuman authority. Note, for example, how George Eliot's narrator in Adam Bede (1859) establishes complete control over her fictional world:

With a single drop of ink for a mirror, the Egyptian sorcerer undertakes to reveal to any chance comer far-reaching visions of the past. This is what I undertake to do for you, reader. With this drop of ink at the end of my pen, I will show you the roomy workshop of Mr. Jonathan Burge, carpenter and builder, in the village of Hayslope, as it appeared on the eighteenth of June, in the year of our Lord 1799.

The Japanese monogatari , on the other hand, features a linguistic relationship between narrator and characters that automatically situates the former within the world depicted. Even The Tale of Genji , whose narrator is perhaps the most ubiquitous presence in the classical canon, is not a truly omniscient narrative. The Genji text contains frequent asides about the inaccessibility of certain information concerning the characters, and an honorific language that places narrator, characters, and audience in a specific social and linguistic relationship. At the same time, the virtual absence in classical Japanese of pronouns, which would serve to identify characters plainly and distinguish them from the narrator, as well as the absence of clear diacritical and grammatical demarcations between the "framing" discourse and the "framed" story, radically

closes the distance between narrator and characters. Compared to the autonomous and highly authoritative narrator in Adam Bede , who reigns so confidently over her novelistic kingdom, the Genji narrator is indeed an amorphous presence.^[1]

In short, omniscient narration is not operative in classical Japanese literature. This is not to say that the language itself lacks the tools with which to create a truly omniscient narration, only that it would be very unnatural to use them to that end. A number of recent studies have drawn attention (with the intent of finding parallels in western languages) to the existence in Japanese of two distinct narrative styles: one that does and one that does not observe the speaker-hearer paradigm and epistemological restrictions of actual linguistic performance. However, none have made the strong case needed to be made for the preferred usage in Japanese of the former over the latter. An examination of this clear preference should shed some light on the nature of narration in the shishosetsu and Japanese prose fiction in general.

Representation Versus Transcription

In the early 1970s the linguist S.-Y. Kuroda posited two mutually exclusive narrative styles in Japanese, the "reportive" and the "nonreportive," which distinguish grammatically between the restricted knowledge of a speaker or narrator in actual linguistic performance and the more general kind of knowledge to which a fictional narrator is privy. A narrative is reportive if related by a single-consciousness narrator who addresses a specific audience as if engaged in an act of communication; it is nonreportive if related by a multi-consciousness narrator whose utterances are not a part of actual linguistic performance. The reportive style offers a single point of view (that of the narrator's own consciousness); the non-reportive style, any number of points of view. Since the former situates the narrator specifically vis-à-vis the characters and events in the story and the latter depersonalizes the narrator entirely, the two narrative styles cannot be used interchangeably.

[1] For a thorough study of the mechanics of narration in Heian texts, see Richard Hideki Okada, "Unbound Texts," esp. the discussion of Genji , 206-12.

Kuroda's main insight is that the reportive style's single-consciousness narrator^[2] can speak in a cognitively definitive sense only about what he "knows" experientially, which is to say, his own feelings and sensations, and must resort to a different morphology when observing or describing the feelings or sensations of any other person. In short, grammar mirrors epistemology in the reportive style.^[3] "The narrator in the reportive style," Kuroda observes, "however successfully he might otherwise transcend the world the story describes, can be pointed to by a mechanism of reference in grammar which exists independently of any assumption we might make concerning the ontological status of the narrator." He can never, in other words, be completely effaced from

[2] Not to be confused with the first-person narrator; for as we shall see below, the "third-person" narrator in the shishosetsu is no less a single-consciousness narrator who identifies psychologically and grammatically with the hero.

[3] "Where Epistemology, Style, and Grammar Meet." Kuroda argues the grammatical distinction by presenting two classes of sensation words, the first of which (e.g., adjectivals such as atsui ["hot"] or kanashii ["sad"]) the speaker/narrator can use to describe only himself, whom he knows experientially to be in a particular state, and the second of which (e.g., verbals such as atsugaru ["to act hot"] or kanashigaru ["to act sad"]) he uses to describe someone other than himself, whom he must judge through observation to be in a particular state. Thus, the speaker/narrator in the reportive style can make an epistemologically unequivocal statement (that is, one based on experiential knowledge and not merely on judgment) only about himself. To make epistemologically unequivocal statements about others requires the use of the nonreportive style.

The above example is actually one of several grammatical forms in the language that make the same epistemological distinction. Suffixes such as -rashii, -yo (na ), and -so (na ) indicate the narrator's judgment (on the basis of observation or hearsay) about attributes of people or things beyond the narrator's own experiential knowledge. (Such words as kawairashii ["cute"] and kawaiso [na ] ["pitiable"], in which the suffix has become fused with the root word to form an essentially new word, are excluded from this characterization.) Although it is true that a narrator in the reportive style can use -garu and similar "judgmental" suffixes to refer to himself, their use is limited to those instances, such as the state of drunkenness, when the narrator is unsure of his own actions and thoughts or when he wishes to objectify himself by distancing narrating and narrated consciousness, as in the sentence, "Omoshirogatte yatte mita ga, tsumaranakatta" (I gave it a try, making a show of interest, but found it a bore).

Although an impossibility in Japanese, a first-person narrator can make unequivocal statements about other characters' sensations and feelings in English. Here is an extreme example, in which Mike Hammer describes his gunning down of two opponents in Mickey Spillane's One Lonely Night : "They heard my scream and the awful roar of the gun and the slugs stuttering and whining and it was the last they heard . They went down as they tried to run and felt their legs going out from under them. I saw the general's head jerk and shudder before he slid to the floor, rolling over and over" (quoted in Umberto Eco, The Role of the Reader , 164; emphasis added).

the narrative. The depersonalized narrator in the nonreportive style, meanwhile, communicates "in a way which is essentially different from the paradigmatic linguistic performance. One might say that the secret of the writer's artistic creation lies partially here."^[4]

The interest for our discussion lies in Kuroda's claim that his "case study from Japanese" may be universal in its application. In her study of prose fiction narrative, Unspeakable Sentences , Ann Banfield takes a cue from Kuroda's argument to describe a form of depersonalized narrative (variously described as style indirect libre, erlebte Rede , and in her own terminology, "represented speech and thought") that presents the character's own point of view directly rather than through any mediating voice. Banfield argues that Kuroda's distinction between the reportive and nonreportive styles does exist in the grammar of English and the other European languages, revealing itself however not in a morphological distinction between sensation adjectives and verbs but rather in the grammar of reported speech. Banfield's study bears close scrutiny, then, because of its self-professed ramifications for narratives in all languages including Japanese. It describes an "unspeakable sentence," peculiar to narrative, which reconstructs or "represents," in a fictional "here and now," the speech and thought of characters in a prose fiction.

Banfield notes that the development of represented speech and thought parallels that of the novel between the seventeenth and the nineteenth century and that the form comes into its own with the emergence of such "stream-of-consciousness" novelists as Joyce, Woolf, Lawrence, and Proust. Contrary to direct and indirect speech, which reports a speaker's utterances with varying degrees of accuracy, represented speech and thought brings to light what a character may never have actually said or even consciously thought.^[5] Further, it is marked by shifted pronouns and modals that make possible the use of the past tense to describe thought and speech

[4] Kuroda, "Where Epistemology, Style, and Grammar Meet," 388.

[5] Thus, such direct- and indirect-speech forms as: (1) "She said, 'I am tired,'" and (2) "She said (that) she was tired," to quote from the example that begins Ban-field's book, become: "She was tired, she thought," in the represented speech/ thought paradigm, in which parentheticals such as "she thought" are a commonly used but optional phraseology. See Unspeakable Sentences , 23.

in the fictional present. This style has the advantage, Banfield argues, of being able to present the consciousness of a character directly without the intervention of a narrator. In ordinary linguistic performance, subjectivity is limited to the "I" of the speaker, whereas in narrative, and particularly in the mode of represented speech and thought, an entirely separate sense of subjectivity can be expressed. Indeed, rather than being "narrated" at all, consciousness in this style is represented without the intervention of any judging point of view. In Banfield's words, "No one speaks."^[6]

Since Ban field draws her conclusion from the premise that represented speech and thought is not modeled after actual linguistic performance, she is led to a radical counterposing of narration to discourse and valorizes the former in literary texts. "It is the dominance of the communicative function in speech," she argues, "which ... accounts for the absence of the features of narrative style in speech, and it is writing ... which frees linguistic performance from the tyranny of the communicative function."^[7] Represented speech and thought is essential to the fictional enterprise, she concludes. "The linguistic cotemporality of PAST and NOW and the conference of [the nonnarrating] SELF and the third person supply a language for representing what can only be imagined or surmised—the thought of the other. By separating SELF from SPEAKER, this style reveals the essential fictionality of any representation of consciousness. “^[8]

Like Kuroda, Banfield argues against the hypothesis put forth by Todorov and others that every text has a narrator and hence conforms to the communication model.^[9] Yet Todorov's model of literature (including, of course, the category of represented speech and thought) as an act of communication cannot be dismissed lightly. According to this model, the narrator, regardless of whether he actually "appears" on the scene, arranges the material and its se-

[6] Ibid., 97. Although certain literary forms, such as the epistolary novel and the skaz , do clearly address a second person, most narrations, Banfield contends, need no first person to intervene; where one does, "a narrator narrates, but addresses the story to no one" (ibid., 171).

[7] Ibid., 227.

[8] Ibid., 260. (Emphasis added.)

[9] See for example Tzvetan Todorov, The Poetics of Prose , 26-28, which asserts the complementary relationship of story and discourse, both of which, he argues, are aspects of "utterance" or communication.

quence, orders its telling and its perspective, chooses between dialogue and "objective" description, and otherwise makes his "presence" felt, however indirectly. Is it possible, we want to ask, for there to be an unnarrated consciousness? To whom, if not a narrator addressing an audience, do we attribute the parenthetical consciousness verbs such as "think," "reflect," "feel," and "wonder" that appear so commonly in represented speech and thought for the purpose of establishing a point of view? The narrator may not be "needed" to "report" every sentence to the reader, but we sense his presence whether he comments on, or simply makes us aware of, the reflection that takes place in a character's mind. Regardless of how "unspeakable" certain narrative sentences seem, none of them can be said to be unlistenable —which is why the argument for an author addressing his narrative to an audience of readers, in conformance to the speaker-hearer paradigm, remains so compelling.^[10]

More than in any specific problems with Banfield's discussion, our interest here lies in its applicability to the Japanese narrative, and in particular to the shishosetsu . Banfield's argument, which is not without merit in the English case, rests of course on Kuroda's analysis, which she credits for presenting evidence that Japanese also possesses a narrative sentence of represented consciousness. The evidence to support this claim, however, turns out on closer inspection to be surprisingly slim. Kuroda's entire article, cited earlier, quotes not a single passage from an actual Japanese text. (His examples appear to be reworkings of English sentences into Japanese.) A later article on narrative theory quotes only one—but how revealing an example it is! Kuroda sets out to prove that the grammar of the reflexive pronoun jibun differs in narration and in discourse (arguing that in the former mode it can indicate the point of view of a referent who is not the topic/subject) but demonstrates in the process the curiously reportive nature of his nonreportive style in the Japanese case.

[10] Gerard Genette addresses the issue succinctly with an example from the Odyssey : recounting Ulysses's slaughter of Penelope's suitors is just as much an action as the slaughter itself. "Without a narrating act, therefore, there is no statement [i.e., narrative text], and sometimes even no narrative content [i.e., story]" (Narrative Discourse , 26). Of these three senses of the term "narrative," Genette suggests that the narrative text of "discourse" is the only one directly available to textual analysis, but he never loses sight of the other two.

Let us examine the lone textual example on which Kuroda bases his analysis of jibun and, by extension, the nonreportive style.^[11] After admitting that such examples are not abundant in written Japanese, he then presents, in a footnote, the opening passage of Mon (1911), the third book in Natsume Soseki's famous trilogy, quoted in part below:

Sosuke wa sakki kara engawa e zabuton o mochidashite, hiatari no yosaso na tokoro e kiraku ni agura o kaite mita ga.... Sono sora ga jibun no nete-iru engawa no, kyukutsu na sunpo ni kuraberu to, hijo ni kodai de aru. Tama no nichiyobi ni ko shite yukkuri sora o miageru dake demo, daibu chigau na to omoinagara.

(Sosuke had brought a cushion on to the veranda and plopped himself on it, cross-legged, and was now basking in the midafternoon sun.... Viewed from the tiny veranda, [the sky] seemed extremely vast. It made quite a difference, he reflected, to be able on an occasional Sunday to gaze leisurely at the sky like this.)^[12]

Here, indeed, is a typical example of the reflexive pronominal jibun working to establish the point of view of the referent, in this case the hero Sosuke: the sky as seen through Sosuke's eyes. The reflexive pronoun is absent from the English translation, but the verb "seemed" in "[the sky] seemed extremely vast" leaves no doubt who the referent is. Kuroda suggests in the same footnote that the sentence "Sono sora ga jibun no nete-iru engawa no, kyukutsu na sunpo ni kuraberu to, hijo ni kodai de aru" (Viewed from the tiny veranda, [the sky] seemed extremely vast) could be made more "objective" or at least more "grammatically neutral"—closer, in other words, to an omniscient narration—by replacing jibun with the third-person pronominal kare and that it as well as the entire passage could be rewritten in the first person by replacing the words Sosuke and jibun with watashi . It is doubtful, however, that a Japanese writer would in fact wish to be "objective" by using kare

[11] Kuroda offers his own rather unidiomatic example of the shifting point of view made possible by a depersonalized narrator roaming from consciousness to consciousness—an unlikely occurrence in Japanese fiction. See "Reflections on the Foundations of Narrative Theory from a Linguistic Point of View," 119. The example also appears in "Where Epistemology, Style, and Grammar Meet."

[12] "Reflections on the Foundations of Narrative Theory," 122-23. Kuroda uses the English translation by Francis Mathy, p. 5.

in place of jibun ; for to do so would seem to introduce another point of view rather than point back to Sosuke. Neither would jibun be replaced by watashi in a first-person narration. The sentence, quite as it stands, is perfectly interchangeable with first-person narration and could be so read in another context. It is the absence of proper names and pronouns in a narrative, once a point of view has been established, that makes this possible. There is nothing in the sentence itself that marks it as third-person narration. Indeed, there is nothing in the entire opening passage that marks it as third-person narration, other than the single mention of the name Sosuke at the very beginning. Unlike the English "he," which of course refers to any male person previously mentioned or understood, kare has a far more circumscribed denotation in Meiji and Taisho literature: specifically, the protagonist, through whose eyes the narrator sees and through whose mind the narrator thinks. Indeed, what appears to be a pronoun, a placeholder for any subject of discourse, is in fact more correctly thought of as a proper name, because the use of kare is restricted to a single character.^[13] That is why a story's protagonist may be introduced from the very first as kare and only later, if at all, by a given name.^[14]

Nor is that all. In the very next sentence we find what looks like a typical example of represented speech and thought: "It made quite a difference, he reflected , to be able on an occasional Sunday to gaze leisurely at the sky like this." But is it really typical? The use of the past tense contemporally with the fictional present and the use of the parenthetical insert ("he reflected") identify the style

[13] This tendency can be observed in the contemporary colloquial language as well. Young women often identify their boyfriends exclusively as kare (or kareshi ), using names or titles to identify other men. Similarly, young men often identify their girlfriends simply as kanojo ("she").

[14] Very little has been written in Japanese on this subject. For a brief review of the use of kare in modern literature, see Yanabu Akira, “An'ya koro ni okeru 'kare.'" Yanabu notes that only the hero in Tayama Katai's Futon , for example, is designated as kare . (The hero's given name is not introduced until the second chapter.) Although kare does not refer exclusively to the hero in a work like Shiga's An'ya koro , the exceptions seem almost to prove the rule (see Chapter 8 below). This usage is observed by writers of both sexes. In Okamoto Kanoko's Boshi jojo [1937], for example, kanojo denotes only one character, the heroine, throughout the story, while various nonpronominal words are used to denote other female characters. See also Kumakura Chiyuki, "Nihongo no shukansei ni tsuite," for an important analysis of what he calls the Japanese language's "subjectivity" (i.e., speaker orientation), to which my discussion is indebted.

for us clearly enough in the English translation. A closer look at the original Japanese reveals a very different arrangement, however. "Daibu chigau na to omoinagara" (It made quite a difference, he reflected) is more literally rendered, "It makes quite a difference!—thus, as [I/someone] reflect[s/ed]." There is nothing to distinguish this expression of "represented" thought from direct utterance. (The particle to normally follows utterances/reflections of any kind.)^[15] It is context, then, not grammar, that determines whether the sentence in Mon translated as "It made quite a difference ..." is to be read as represented speech/thought rather than as direct utterance/ reflection. Indeed, the general sense one gets from reading the Japanese is that there is not the essential difference between these two forms that there is in western languages. The staple features of represented speech and thought, namely, the shifts in tense and pronoun usage, are nowhere to be found. The Japanese sentence, in short, is doing precisely what the English version cannot do: "representing" a character's speech or thought by imitating that character's very own words/thoughts.^[16] The "unspeakable" in English becomes the only way to talk—or write—in Japanese.

The Written Reportive Style

That such sentences are in fact grammatical is not so surprising in itself; we might expect certain expressions, rejected in one language, to be perfectly acceptable in another. What is significant for our discussion is the impact of their grammaticality on the Japanese narrative enterprise. The first effect of this rejection of the past tense for utterances in "represented speech/thought," combined

[15] It is deleted only after certain forms of indirect command. See Samuel E. Martin, A Reference Grammar of Japanese , 998.

[16] Two more examples of what are considered "unacceptable" expressions in English may help clarify what is meant here. Banfield asserts that consciousness verbs cannot introduce direct quotations and may be used only as parentheticals following an expression of represented speech and thought: thus, *"She noticed: 'He is getting fat'" and *"She remembered, 'Ah dear, it is Wednesday in Brook Street'" are unacceptable transformations, respectively, of "He was getting fat, she noticed" and "Ah dear, she remembered—it was Wednesday in Brook Street" (Unspeakable Sentences , 45). But the structure of "represented" speech and thought in Japanese (if one were to translate the above examples) is in fact almost identical to the unacceptable form in English, except that the communication/consciousness verb usually comes after the utterance/reflection in Japanese.

with the absence of pronouns or other semantic markers that distinguish types of narration, is the extraordinary sense of immediacy that obtains. The use of the past tense in English (even if it loses its "pastness," as Banfield argues) has the effect of creating a much sharper distinction between utterance and narration, between character consciousness and reader awareness, than exists in Japanese. Without a tense that insists on its own fictional autonomy or a pronoun usage that insists on the otherness, the externality, of the character(s) in the narrative, the Japanese narrative presents the speech and thought of a character in a way that not only posits an audience but also invites the reader's identification—indeed, assimilation—with the character's train of thought. And if this is true even for "fictional" narrations like Mon , it is all the more true for "autobiographical" (whether first- or third-person) narrations like the shishosetsu .

Second, and a consequence of the first, the reportive and non-reportive styles in Japanese, which are quite distinct in Kuroda's theoretical argument, merge in narrative praxis into a hybrid style that rejects depersonalized, multi-consciousness narration and approximates, but does not exactly duplicate, the speaker-hearer paradigm. Because of its special language, such a style—let us call it here the written reportive style—posits an audience without precisely identifying it. Locating the audience too specifically or denying it altogether would push the style toward either the reportive or the nonreportive pole. The written reportive style, with its restricted point of view, is the overwhelming preference of authors writing in a language that favors the transcription (in as literal. a way as possible in writing), over the representation, of speech and thought.

To return to Banfield's conclusion, we see that whereas "the linguistic cotemporality" in western languages "of PAST and NOW and the coreference of SELF and the third person supply a language for representing what can only be imagined or surmised—the thought of the other," the absence of those two features in Japanese makes representation (in Banfield's sense) a linguistic impossibility. The written reportive style would seem to undermine the Japanese writer's belief in the legitimacy of "invention"—that is, the right of (or simply the possibility for) individual imagination to supplant an

experiential, single-consciousness "reality" with a more compelling one (abstract and remote from immediate experience though it may at first appear) told from an omniscient perspective. Rather than acknowledge "the essential fictionality of any representation of consciousness, or any approximation of word to thought," Japanese yearns for the authenticity—however illusory—of an unmediated transcription of consciousness, in the language of an ever-present speaker.

For such a transcription to be truly convincing, however, it must be of the one consciousness that the narrator knows directly—namely, his own. Japanese prose fiction is therefore quite comfortable with first-person narration, since it does not aspire to a universal brokerage of consciousness afforded by omniscient narration or the nonreportive style. Yet this particularity of narrative voice, with which a Japanese reader identifies so closely and which the written reportive style exploits so expertly, works just as well with "third-person" as with first-person narration. In a "third-person" shishosetsu , the pronominal kare typically has only one referent: the hero himself. Since voice and point of view readily converge in the absence of grammatical and tense markings narrator and hero easily merge: not only do they appear as one in a shishosetsu , they even "speak" and "think" as one.

Perhaps it would be useful here to offer a new narrative paradigm consistent with the written reportive style: let us call it the recorder-witness paradigm, which rejects the wholly representative author-reader paradigm of the nonreportive style while approximating the communicative speaker-hearer paradigm of discourse. This paradigm is central to the shishosetsu , in which narrative voice and point of view become thoroughly intertwined. The grammar of the written reportive style reflects an epistemology that asserts that knowledge founded on personal experience is the only kind worth relating. A narrator situated within the confines of this epistemology is not aiming to reveal "the essential fictionality of any representation of consciousness," as is the case with Banfield's narrator of represented speech and thought, nor is he interested in the kind of "artistic creation" that Kuroda suggests is linked with the use of the nonreportive style. He is a happily nontranscendent figure, anchored in the narrative he both acts in and produces. And

whereas the omniscient narrator in a classical novel distances himself from the narrative and invites the reader (as Henry James does so explicitly and eloquently in his prefaces) to join him in witnessing his production unfold, the shishosetsu narrator "reports" on experience from an epistemological perspective that sees no further than the narrated subject.

The shishosetsu's recorder-witness paradigm thus pretends to a far more direct relationship than does the author-reader paradigm and can itself become a story's central or unifying element. Although it differs from the speaker-hearer paradigm in being a product of writing and not of speech, it presents, as the latter does, an intelligence no greater—or more distant—than that of any sentient being. Yet this intimacy between recorder and witness derives ironically from the rejection of a specific reader as the narrator's intended audience. Given a language in which a precise and subtle system of honorifics operates to situate speaker, hearer, and referent in a particularized and hierarchical relationship, the recorder had to discover a mode of expression that would radiate an equal sense of authenticity to any witness regardless of the actual social relationship that obtained.^[17] The recorder was in no position to maintain the fiction—and it could only be a fiction—of a universally valid relationship with his audience as superior or inferior or equal. This inherently ambivalent narrative situation stimulated the development of a style that was familiar yet circumscribed. Unable on the one hand to address a specific reader without potentially ruling out discourse with another, and on the other to address all readers as one without distancing them, the recorder essentially became his own audience, writing a "story" in the style of a sketch or memoir and inviting the witness—any reader—to examine it.

And yet although the shishosetsu writer's goal was to have the recorder-witness paradigm come as close as possible to the communicative act and thus do away with the mediative quality of written language, it was doomed to failure, as each writer was to dis-

[17] The great task of modern Japanese writers in general was to develop a colloquial style that would close the immense, centuries-old gap between the spoken and written languages and call less attention to itself as a rhetorical medium. Masao Miyoshi, in his already-classic treatise, argues persuasively, however, that the language of Japanese prose fiction never did develop into a truly colloquial idiom, because spoken Japanese precluded the kind of undifferentiated audience that the novel demands. See Accomplices of Silence , esp. xiii-xiv and 11-14.

cover in his own way. Chikamatsu Shuko began his first important series of writings by positing a specific, epistolary audience (his wife, addressed as "you" [omae ]) but eventually abandoned the project as unworkable. For Shiga Naoya, the only way a writer could maintain the pretense of "talking" naturally while writing was to talk to oneself; but doing so resulted in a style that was strangely aloof. Kasai Zenzo, meanwhile, discovered for himself, and revealed to us through "slips" in his colloquial narrative style, the nature of his writing as writing. He was forced to "talk" in unexpected ways and engage in a series of feints that only served to remind him of the textuality of his "discourse," the closer he approached a pure colloquial style.

All these writers worked to camouflage the fictionality of their writing, a fictionality produced not by the nonreportive style's "lie" of narrated consciousness but by the lie of authorial "presence" featured in the written reportive style. From this style has sprung the myth of "sincerity," in which the totally accessible author relates his experiences through the totally transparent text. Yet these writers would discover, in ways that will be elaborated in Part 3, the surprising opacity of their supposedly glass-clear medium. The written reportive style turned out not to be fiction-free: the very act of expressing themselves in writing, they realized, was in effect to don a mask, to supplant person with persona. Each effort to suppress the textuality of their enterprise and to valorize narrative immediacy was undermined by the very process of mediation in which they were engaged. As a result, many shishosetsu that seem at first glance to be neutral descriptions of experience, dependent for their legitimacy on an extratextual, referential world, take on a decidedly metafictional tone.

Such arguments for the shishosetsu's awareness of its own textuality, however, are not intended to equate it with "postmodernist" fiction: the latter thematizes overtly—and the former only covertly—the text's linguistic self-awareness and ontological independence from the referential world.^[18] Rather, they simply confirm

[18] See, for example, Linda Hutcheon, Narcissistic Narrative (esp. chaps. 1 and 6), which treats fiction as an autonomous universe, dependent on intratextual validity rather than extratextual "truth." See also Patricia Waugh, Metafiction , 21-61, for a discussion of the self-conscious nature of textual production, which is to be found in the shishosetsu (esp. in the writings of Kasai Zenzo) as well.

the assertion that "all great fiction, to a large extent, is a reflection on itself rather than a reflection of reality."^[19] The question of the shishosetsu's "greatness" aside, we shall find numerous examples in Part 3 of the form's reflection on the medium it is constrained to employ in the textual reproduction of lived experience.

While Taisho-period writers were forced by the very mechanics of their narrative art to reconsider the meaning of such concepts as sincerity and authorial presence in literature, critics were busily engaged in the mystification of these terms. That many of these critics were also writers does not undermine the argument that writers were in fact aware of the paradoxical nature of their project; it only suggests the degree to which that awareness had been suppressed to a subconscious level. Writers, as writers, reckoned with the language of the text; as critics, they drew back from the texts and embraced the mythology of sincerity and authorial presence. If we find that Japanese writers were more linguistically than critically attuned to the nature of fictionality and referentiality, it is because the shishosetsu had so thoroughly established its own legitimacy as a cultural (and not simply as a literary) artifact, weighed down with a large set of conventions to determine meaning. Both the shishosetsu's admirers and its detractors were slaves to these conventions, founded on a referential view of literature, and both addressed the issue of the shishosetsu's beneficial or deleterious effect on Japanese letters from essentially the same position. It is a position that we must now locate and explore.

[19] The assertion, by Raymond Federman, provides the epigraph to Hutcheon, Narcissistic Narrative .

3
Shishosetsu Criticism and the Myth of Sincerity

When one looks at the statue of Kannon adorning the Dream Hall in Horyuji Temple, the farthest thing from one's mind is the person who carved it. This is a special case, for the statue has taken on a completely separate existence from its producer. If I were able to produce literary text of equal caliber, I doubt that I would feel compelled to sign my name to it.
Shiga Naoya, Preface to the 1928 Gendai bungaku zenshu single-volume edition of his writings

If we forget that fictions are fictive we regress to myth.
Frank Kermode, The Sense of an Ending

Thus far we have seen how a literary tradition of equating "serious" writing with nonfiction and a language well suited to the reportive style's epistemology have nurtured the belief in an author's "unmediated" presence in the text, especially in a text so clearly autobiographical as the shishosetsu . This belief has in turn prompted the shishosetsu admirers and detractors alike to appeal to a biographical "pre-text" in establishing the proper context for evaluation. In the eyes of the detractors, shishosetsu writers were unimaginative gossips whose stories' very intelligibility depended on a prior familiarity with details of their private lives. In the eyes of admirers, they were a special breed of artists whose candid confessions inspired a confidence between author and reader that did away with the need for fiction. These seemingly contradictory but in fact interlocking views have informed the shishosetsu from its emergence in the critical consciousness at the end of Taisho down to the present day.

Three major periods of critical debate concerning shishosetsu —the first in the mid-1920s, the second in the mid-1930s, and the third in the 1950s—will provide the focus of our discussion here.^[1] We should note that even the earliest criticism appeared when the shishosetsu's popularity and influence had already reached their zenith; nearly all texts by the writers analyzed in Part 3, for example, date back to the 1910s and early 1920s. But there is no question that the values that eventually surfaced in the form of criticism were in place from the shishosetsu's very beginnings. An examination of those values should tell us a great deal about a cultural product that was packaged from its inception as a "purer" and "sincerer" prose form than ordinary fiction, and by that measure somehow more uniquely Japanese.

The First Period: "Pure" Shishosetsu Versus "True" Novels

In 1924 Nakamura Murao (1886-1949), author, publisher, and critic, inaugurated the first period of critical debate with an essay championing what he called the "true novel" (honkaku shosetsu ) against the shinkyo shosetsu (a variant term for watakushi shosetsu / shishosetsu ). The true novel's principal concern, Nakamura writes, is the realistic depiction of characters in society. Since the author reveals his philosophy only parenthetically, by his characters' speech and actions, the reader's understanding of such a text does not depend on a familiarity with the author's private life. The shinkyo shosetsu's principal concern, by contrast, is the direct communication of lived experience. Since the author has no interest in creating characters or dramatic situations and it is just as important to know who is writing as it is what is being written about, the reader's understanding greatly depends on his knowledge of the author. The descriptive shorthand used in the shinkyo shosetsu makes the form resemble the haiku in its economy, to be sure; but economy is no virtue in fiction, Nakamura insists, and he offers Anna Karenina as the consummate example of what he looks for in

[1] Some critics alter this periodization slightly. For example, Katsuyama Isao divides the same three decades into four periods: the first in the mid-1920s, the second in the mid-1930s, the third during the Pacific War, and the fourth right after the war. See "Shoki watakushi shosetsu ron" (1953), in Taisho , watakushi shosetsu kenkyu , 187 n. 1.

a novel. Japan has no hope of producing a Tolstoy, because writers have become too bogged down in their private lives and because the critics have become voyeurs whose preoccupation with autobiographical truth has undermined the incentive to create imaginative fictions. In an atmosphere where curiosity about private life reigns supreme over concern for literary excellence, any writer's attempt at authoring true fiction is derided as "vulgar" literature.^[2]

Although we have used watakushi shosetsu / shishosetsu exclusively in our discussion, we should note here that the term shinkyo ("mental state") shosetsu has also had wide currency down to the present day. The terms are often used interchangeably. Kume Masao (1891-1952), for example, argues that the shinkyo shosetsu is a "purer" form of watakushi shosetsu (see below), but either term would be appropriate in Nakamura's essay. Some critics have worked out elaborate schemes for distinguishing one from the other. The predominant postwar view is that the watakushi shosetsu is a literature of crisis, to which writers turn in desperation and at the risk of disrupting their private lives, while the shinkyo shosetsu is a literature of harmony and salvation, through which writers find peace with themselves and their surroundings. The latter is commonly conceived as an even more evocative, discursive, and nondramatic presentation of experience than the former.^[3] A writer is usually regarded as the author of one form or the other, depending on his commitment either to unmitigated self-exposé or to more controlled confession. Hirano Ken (1907-78) offers the best-known exposition of this critical perspective in his opposition of the presumably "unabashed" writing of Chikamatsu Shuko and Kasai Zenzo to the more "guarded" prose of Shiga Naoya. The rationale for distinguishing between two varieties of confession lies, not surprisingly, in the pre-text of private life. Writers like Shuko and Kasai, their personal lives a shambles, had nothing to lose by their candor, Hirano argues, while those like Shiga, their domestic lives more or less intact, had everything to lose by letting their "confessions" get out of hand.^[4] This distinction is not really useful, since

[2] "Honkaku shosetsu to shinkyo shosetsu to."

[3] Howard Hibbett reflects this critical view, which ascribes a more resonant and elevated tone to the shinkyo shosetsu , when he calls it a "contemplative" (as opposed to a "confessional") novel. See "The Portrait of the Artist in Japanese Fiction," 348.

[4] See Geijutsu to jisseikatsu , 25-45 and passim. Hirano's view is shared by Honda Shugo, who likewise makes an explicit contrast between the "destructive" literature of Shuko and Kasai and the "constructive" literature of Shiga. See “Shirakaba" ha no bungaku , 125. Hirano's strict opposition between these two types of literature has been questioned by later critics. See, for example, Katsuyama, "Taishoki ni okeru watakushi shosetsu no keifu" (1966) in his Taisho , watakushi shosetsu kenkyu , 248-53. Few critics, however, question their own ability to ascertain the accuracy of an author's confessions by an examination of the life.

it is impossible to determine objectively just how much an author "reveals" of himself in his work. It is valid, however, with regard to the author's literary persona. Thus, the Shiga hero's "confessions" predictably conform to the persona of a morally conscientious author-sage, for example, while the Kasai hero's conform to that of a forever-victimized, ne'er-do-well artist. We shall examine these personae in detail in Part 3.

But now to continue our discussion of early critics. Nakamura is representative of those who believed that the shishosetsu (by which is meant either watakushi shosetsu or shinkyo shosetsu ), although a faithful record of "real-life" events, was inevitably inferior to the true novel as art because it was not informed by an overall conception that transcended the author's personal experience. He was supported by an even severer critic of the shishosetsu , Ikuta Choko (1882-1936), who insisted that the world's truly great authors, from Shakespeare to Saikaku, best expressed themselves through imaginative creation rather than documentary accounts of their private lives.^[5]

Support for the shishosetsu came first from a practitioner, Kume Masao. In perhaps the most widely quoted essay on the subject, Kume confides that as an author he has always felt more comfortable writing shishosetsu than "fiction" and that as a reader he has always placed his greatest trust in other shishosetsu writers. He then presents an argument that epitomizes the Japanese writer's skepticism of the fictional contract.

I cannot believe that art in the true sense of the word is the "creation" of someone else's life.... I see it rather as the "recreation" of a life, of an experience, that actually took place. To be sure, a Balzac can write a voluminous narrative like The Human Comedy and portray his usurers and aristocratic ladies so vividly that they seem to come alive. Yet I cannot but regard such a narrative as a fabrication, nor can I place in it the trust I do in even the most off-handed remark the author might make about, say, the difficulties he encountered

[5] "Nichijo seikatsu o hencho suru akukeiko" (1924).

writing his stories. The world has known a handful of great authors—true geniuses (Tolstoy, Dostoevsky and in particular Flaubert)—who have been able to communicate something of themselves in their writing. The moment these authors express themselves through other characters, however, they distance themselves from their readers. Inevitably, embellishments and technical flourishes—convenient fictions, all—creep in. Their novels may be superior as entertainment, but they do not ring true to me. Once, during a lecture, I went so far as to say that Tolstoy's War and Peace , Dostoevsky's Crime and Punishment , and Flaubert's Madame Bovary were really no more than popular novels [tsuzoku shosetsu ]—first-rate examples of their kind, to be sure, but popular novels nonetheless. In the final analysis, they are mere fabrications—just so much entertainment.^[6]

For Kume, then, the shishosetsu is the purest of prose forms because it allows the author to express himself candidly without having to "fabricate" his thoughts in the guise of a novelistic character. Wary of the essential fictionality of represented consciousness, he yearns for a literature of "unmediated" experience told directly by the subject as opposed to one told through the "mediation" of an omniscient narrator. To a reader with Kume's sensibilities, Balzac's memoirs would have greater appeal than any of his novels (which, however brilliantly fashioned, were mere "fabrications"), because they brought the reader closer to Balzac the man. Kume appears to qualify his faith in unmediated experience when he later insists that the true shishosetsu is not "mere" autobiography or confession but "first and foremost a work of art." But art, in Kume's view, is a product of one's mental state. Only properly harmonized thoughts can generate an unerring depiction of self. The line Kume draws between art and nonart corresponds to the distinction he makes between the shinkyo shosetsu (a word that he claims to have coined) and the ordinary watakushi shosetsu .^[7] Kume argues that the shinkyo shosetsu is not simply a random account of personal experience (as he implies the watakushi shosetsu is) but a scrupulous self-portrait that can be drawn only when one has reached a suitably contemplative frame of mind. In Kume's view, then, art is a discipline, the

[6] "'Watakushi' shosetsu to 'shinkyo' shosetsu" (1925), 52-53.

[7] According to Kume, the word shinkyo was used in his haiku circle and referred to the poet's mental state during composition. (Some critics dispute his claim to be the first to use the term in connection with prose writing.) For a discussion of the origins of shinkyo shosetsu and watakushi shosetsu as critical terms, see Katsuyama Isao, "Shoki watakushi shosetsu ron," in Taisho , watakushi shosetsu kenkyu , 174-89.

shinkyo shosetsu a vehicle for meditation, and the writer his own audience.

Many critics, whether or not they subscribed to Kume's watakushi shosetsu / shinkyo shosetsu distinction, believed that the shishosetsu enabled the author, in the words of Uno Koji, to "plumb the depths of the self" in a way that the ordinary novel, fettered by its conventions of "fictionalization," could not.^[8] They also saw in the classical western poetics of plot, characterization, and dramatic scene a built-in artificiality that undermined the aims of sincerity, and argued that fiction was a crutch an author relied on only when he had exhausted his own life's experience or when he could not address his readers with complete candor. It followed that the best preparation for writing—itself no more than an unpretentious presentation of one's philosophy of life—was living. In this vein, Kikuchi Kan (1888-1948) insisted that a writer had to serve a rigorous apprenticeship to life before committing himself to art, which was a kind of "report card" of the life.^[9] He also insisted, in a remarkable statement linking art and morality, that individual character and literary style were inseparable:

It has long been the case that outstanding writers are also outstanding people.... Writing is a reflection of the entire person, and if the person's character is defective, so will be the reflection.... One improves one's writing by first improving one's character.... A writer will be successful as long as he describes his thoughts and feelings sincerely and faithfully—never mind how naive they may be. The worst possible thing is to imitate another's style and write about what one has never seen or felt.^[10]

[8] "'Watakushi shosetsu' shiken" (1925), 64. Uno singles out Kasai Zenzo as the form's most brilliant practitioner.

[9] See "Shosetsuka taran to suru seinen ni atou" (1922), in Kikuchi Kan zenshu 12:373-76. See also "Geijutsu to tenbun" (1920), in ibid., 29-32, and finally "Chikamatsu Shuko-shi no kinsaku" (1924), in ibid., 312, in which Kikuchi uses the term to describe Shuko's work Ko no ai no tame ni . Many critics continue to speak of the shishosetsu as the author's personal "report card." See, for example, Yamada Akio, "Watakushi shosetsu no mondai," 43. Both Kikuchi and Kume Masao eventually turned away 'from the shishosetsu : Kikuchi, because he believed that writing as a financially viable career depended on pleasing a mass audience (see, for example, his Han jijoden [1929-47], in Kikuchi Kan: tanpen sanju-san to Han jijoden , 490); and Kume, because he believed that one could write successfully only as an avocation, without the pressure of deadlines and editorial whims (see his "Junbungaku yogi setsu" [1935], 411-13). But their critical stance toward life and art remained unchanged.

[10] "Bun wa hito nari."

The assumption made by both sides in the debate, however—and here we see their common ground—was that fictionality was operative only in the author's "absence" from the text. A narrator-hero dearly modeled on the author was a sign to the shishosetsu's admirers and detractors alike that all experiences related in the text, being the author's own, belonged ipso facto outside the realm of fiction. There is no more telling evidence of their common ground than the comments of Nakamura Murao, the erstwhile shishosetsu critic. Ten years after he declared his preference for the "true" novel, he now confessed his partiality for "pure literature" (junbungaku , which, as many have noted, has been equated with the shishosetsu since the Taisho period)^[11] precisely in terms of the attitude toward fictionality we have discussed.

I regard the shishosetsu as the ultimate form of pure literature. It is surely the purest and the most candid of prose forms. In a conventional, objective novel [kyakkan shosetsu ], fiction [uso ] inevitably creeps in; it simply cannot be written otherwise. Tolstoy, Balzac, and Flaubert are revered as "gods" of the novel. All their work contains fiction.... The subjective passages, in which the author reveals himself directly, are much more real to me than the objective passages, in which fiction prevails. Anyone with a modicum of talent can write a novel. That is why a novel, no matter how deft or serious its fictionalizations, does not claim my respect or appeal to me as pure literature.^[12]

The notion of "pure literature" as the honest depiction of the author's own life, which might otherwise be "sullied" by incretions of fiction, was common throughout the bundan and continues to the present day.^[13] To claim, however, as we do here that Naka-

[11] See Katsuyama, "Shoki watakushi shosetsu ron," in Taisho , watakushi shosetsu kenkyu , 185. See also Oi Zetsu, "Watakushi shosetsu ron no seiritsu o me-gutte," 46-47, which links the equation specifically to Kume Masao's essay, discussed above.

[12] "Junbungaku to shite no watakushi shosetsu," 5. The prejudicial view of literature not based on the author's own life as being somehow less "authentic" is noted nearly three decades later by a well-known writer of detective fiction, Matsumoto Seicho, who insists that his own literary intentions are serious despite criticisms that the "fictionality" in his work makes it less "real." See Hirano Ken and Matsumoto Seicho, "Watakushi shosetsu to honkaku shosetsu."

[13] For example, Isogai Hideo, a contemporary critic, describes the author Makino Shin'ichi in the following manner: "Makino began his literary career writing stories in the shishosetsu mode; but his was not a temperament that could bear the strain of unvarnished confession, and he gave his works a comic and fictional veneer “ (NKBD 3: 225c; emphasis added).

mura's statement is a particularly cogent articulation of the myth of sincerity is to address squarely the myth's power and to acknowledge its evolution from a cultural to a precritical, "natural" artifact. The myth of "pure literature" became the shishosetsu touchstone; perhaps even more important, it became the touchstone for critical judgment of all Japanese prose fiction, whether shishosetsu or not. "Pure literature" was an original site, a virgin territory that could be recovered only if purged of "objective" passages having nothing to do with lived experience. The latter, according to Nakamura and the vast majority of critics until quite recently, could be made available only in writing like the shishosetsu . Such reasoning would not have made sense to the mass audience, which had no inkling of the real-life author who produced stories for their consumption, but it held sway in the bundan as long as writing was seen as a vestige of the life that preceded the text.

Akutagawa Ryunosuke (1892-1927), the famous short-story writer who was frequently criticized by the bundan for his forays into "nonpure" fiction, attempted to collapse the distinction between shishosetsu and true novel altogether by arguing that no art form was concerned with referential truth. Yet even he agreed that the shishosetsu was a "fictionless" writing.^[14] Tempted though we might be when reading a shishosetsu , however, to exclaim, along with Tanizaki Jun'ichiro (1886-1965), "Aha! The man is writing about himself!" we need not jump to the conclusion that we are "reading a nonfiction piece" and "immediately lose patience with it."^[15] If we do, we only confuse the fictive product (however "true-to-life") with the producer whom we do not know and whose connections with the text are severed once pen leaves page.^[16]

[14] "'Watakushi' shosetsu shaken" (1925), in Akutagawa Ryunosuke zenshu 8:39-46. Akutagawa's statement is important nonetheless, for he suggests that a work's basis in fantasy or fact has no bearing on its artistic value. He uses an example from painting to illustrate his point. The physical "reality" of the Mount Koya monastery's Red Fudo , with flames rising from the deity's back, is not at issue, he argues, but rather the painting's overall design. Likewise, the shishosetsu's apparently faithful representation of life does not by itself make it superior as art. "Honest" representation may be a moral imperative, he concludes, but it is not an artistic one.

[15] Jozetsu roku (1927), in Tanizaki Jun'ichiro zenshu 20:72.

[16] Tanizaki himself took advantage of just this sort of confusion and enticed naive readers to equate fictive and real-life authors. He made enough independently confirmable referential asides in his Shunkinsho (1933), for example, to have sent some gullible critics scurrying to an Osaka cemetery in search of the heroine's nonexistent grave.

shishosetsu authors may lie or they may tell the truth about themselves. But unless we are prepared to privilege the biographical pre-text over the written text to the extent that we deprive the latter of any autonomy, we gain nothing from having distinguished "truth" from "falsehood." The act of writing, however confessional, actually liberates the writer, however unwillingly, from lived experience. Therefore, a text whose protagonist has a clearly recognizable model is no less autonomous than a text that has none. If any one shishosetsu text does not stand up on its own, it is simply because it exists in a discursive field where it is radically contiguous to other texts (not to the life it presumes to "represent"). Either way a text is condemned to independence from its author.

Thus, the anonymously produced statue of the Kannon goddess adorning the Horyuji Dream Hall, which Shiga Naoya gazes at so fondly in the epigraph to this chapter, is not a "special case" after all. Shiga, even by signing the texts he writes, paradoxically confirmed his biographical absence from them. The seeming overlap of author, narrator, and hero notwithstanding, the author's "presence" in a shishosetsu is, finally, an artistic illusion like perspective or coloring or shading: it may add a dimension of "realism" to a text (the kind that Taisho-period readers were particularly enamored of, to be sure), but it does not bring the text any closer to "reality." Nor should it; only by transcending life—by transforming experience and perception into the written word—can literature be free to represent it. Even Shiga, the shishosetsu author par excellence, was apparently aware of this theorem of fictionality and its corollary of the illusion of authorial presence.

The Second Period: Writer and Society

When the debate over the shishosetsu intensified a second time in the mid-1930s, attention shifted from fiction's place in literature to the citizen's place in society—a legacy of the proletarian movement's concern for social engagement. The shishosetsu came under attack by various groups, from the modernist school known as the Shinkankaku-ha (Neo-perceptionists) to the proletarian writers, for its extreme introspection and lack of social consciousness. In

his essay "Junsui shosetsu ron" (1935), Yokomitsu Riichi (1898-1947), the Shinkankaku-ha's leading member, urges writers to offer a broader vision of society than can be had in the autobiographical junbungaku ("pure literature"), yet to avoid the frivolity of popular fiction (tsuzoku shosetsu ). He borrows a term from André Gide and champions what he calls the "pure novel" (junsui shosetsu ), which combines the former's moral seriousness with the latter's structural cohesion. Only the pure novel, Yokomitsu maintains, can bring together these two divergent strains, which he argues are carryovers from the "autobiographical" diary (nikki ) and "imaginative" narrative (monogatari ) traditions, respectively.^[17]

Yokomitsu's advocacy of a new literary synthesis leads him to challenge the notion of "realism" prevailing in the bundan . For junbungaku writers, Yokomitsu claims, realism simply means the faithful chronicling of personal experience. Following in the nikki tradition, the author-narrator focuses exclusively on a single character—himself. Because Japanese "realism" has no place for sustained plot and character development, the typical junbungaku shosetsu is fragmented and short-winded. Junbungaku writers, Yokomitsu argues further, have misjudged the serious artistic intentions of the great nineteenth-century western "realists" like Balzac or Tolstoy or Dostoevsky in whose works accident often plays a crucial role, because their narcissism has rendered them oblivious to anything outside their own immediate experience and consequently to the links between the particular and the universal. A great novelist skillfully orchestrates fortuitous occurrences in his characters' lives, he concludes, whereas the junbungaku writer, preoccupied with authenticity, seeks to eliminate them altogether.^[18]

Yokomitsu's answer to "pure literature's" limited narrative perspective and resultant particularity is a radically different, al-

[17] "Junsui shosetsu ron," 146—47. For additional commentary on this essay and its dubious debt to Gide, see Dennis Keene, Yokomitsu Riichi: Modernist , 181-83.

[18] "Junsui shosetsu ron," 144-45, 149-50. (One is reminded of Akutagawa's comment that "true-to-life" fiction contains far fewer coincidences than one probably encounters in one's own experience ["Shuju no kotoba," in Akutagawa Ryunosuke zenshu 7:448].) Yokomitsu would no doubt have agreed with Georg Lukács, who wrote a year later in his classic comparison of Tolstoy and Zola that without chance all narration was dead and abstract and that the secret of representation lay in its elevation of chance to the inevitable. See "Narrate or Describe?" (1936), in Writer and Critic and Other Essays , 112.

though vaguely defined, point of view he calls the "fourth person," capable of seeing both the inner self and the world at large; he holds up Stendhal's Charterhouse of Parma as a model for writers to emulate.^[19] "Fourth person" suggests that Yokomitsu realized the easy tendency of "third-person" narration to merge with first-person narration in Japanese and hoped to construct a narrative in which the distance between narrator and hero did not collapse. Yet this comprehensive narrative perspective, which directly challenged the epistemology that underlies Japanese realism and the pervasive mode of expression we have provisionally called the written reportive style, was never to gain much currency in the bundan . One reason, perhaps, is that Yokomitsu's proposed "synthesis" of two traditional forms (nikki and monogatari ) is of elements not as disparate as he makes them out to be. By contrasting the "autobiographical" nikki with the "imaginative" monogatari , Yokomitsu equates the latter with western fiction; yet we have already noted that both forms lack the teleological worldview that predominates in western narrative. Moreover, monogatari did not provide the precedent for omniscient-narrative overview that Yokomitsu was looking for. Japanese realism was not to be divorced so easily from the narrator-hero's perspective and voice.

Shortly after the appearance of Yokomitsu's essay, Kobayashi Hideo (1902-83) responded with "Watakushi shosetsu ron" (1935), a kind of postmortem on the shishosetsu , although he acknowledges in his famous conclusion: "The shishosetsu is dead, but have people vanquished the self? The shishosetsu will doubtlessly reappear in new forms, so long as Flaubert's celebrated formula, 'Madame Bovary—c'est moi,' lives on."^[20] He argues that the shishosetsu is rooted in Japanese naturalism and that Futon is its prototype, a view that has assumed the status of critical orthodoxy.^[21] He discounts any similarities between the shishosetsu and European personal fiction. Unlike the latter, which has its roots in the romantic movement inaugurated by Rousseau's Confessions and the subsequent flowering of self-awareness, the shishosetsu is a product of

[19] "Junsui shosetsu ron," 150, 154.

[20] "Watakushi shosetsu ron," 20l-2.

[21] Oi Zetsu, "Watakushi shosetsu ron no seiritsu o megutte," 43. Yoshida Sei'ichi offers a representative example of this view in "Watakushi shosetsu no mondai ni tsuite," 19.

the Japanese naturalist movement's faith in literature as nonfiction document. This faith, he notes, was crystallized in Kume Masao's 1925 essay that denounced several great nineteenth-century novels as "mere" fabrications. Kobayashi argues that Kume spoke for an entire generation of skeptics about the fictional enterprise.^[22]

Kobayashi's most penetrating remarks about the shishosetsu have to do with the relationship between self and society and how radically it differs in Japan and in the west. Kobayashi maintains that Japanese writers casually imported the idea of "self" from European literature without domesticating the social, intellectual, and scientific institutions on which it was based. The individual in modern European literature has always been situated in a broad social milieu. However much the romantics questioned the individual's place in society, they did not turn their backs on that society. Rousseau and his successors never sought refuge in depictions of private life for their own sakes; they confidently featured their self-portraits in sweeping canvases of society. Writers of personal fiction from Goethe to Gide were thoroughly assimilated into their communities and steeped in a tradition of individualism nurtured by what Kobayashi calls the "socialized self" (shakaika-shita watakushi ). Japanese writers, of course, lacked this tradition. And although fascinated with naturalism as a literary technique, they were incapable of truly appreciating, let alone transplanting, the positivistic philosophy from which it derived. "No writer, however great a genius he may be, can create singlehandedly a zeitgeist or social philosophy," writes Kobayashi; "... he can only articulate in his works a philosophy that already lives among the people."^[23] Unable to take root as a system of thought, then, the literary naturalism imported by Japanese writers yielded, inevitably, nothing more than technically brilliant depictions of the writers themselves living in studied isolation.

Kobayashi makes his point by contrasting the views of Flaubert and Shiga on the place of private life in art. Flaubert epitomizes for Kobayashi the European tradition of subordinating private life to artistic production. "The artist must conduct himself in such a way as to make later generations believe he never existed," he quotes

[22] "Watakushi shosetsu ron," 182.

[23] Ibid., 185.

Flaubert as saying. And further: "The only way to escape misery is to cloister oneself in art and completely disregard all else. I have no desire for wealth or love or passion. I have divorced myself irrevocably from my private life."^[24] This credo, argues Kobayashi, is antithetical to that of a shishosetsu writer like Shiga, whose desire to overcome crises in private life become both the motivation for and the motif in his writing. The notion that private life was off-limits to his creative career was no doubt behind Flaubert's yearning for anonymity. For Shiga, however, private life was the main arena of his art. Shiga recognized that a truly great work of art was, in the final analysis, independent of its author, as his remarks on the Kannon statue (which Kobayashi cites) make clear. Yet this recognition, Kobayashi suggests, did not translate into action. What are we to make of the long silence that followed those remarks? Shiga was able to gaze complacently at the statue at a time when he had already achieved personal contentment and had exhausted his store of creative material. Anonymity was no longer a choice; it was his destiny.^[25]

Shiga's case may have been extreme, but it was far from unique; sooner or later all shishosetsu writers faced this crisis of creativity. Having exhausted his private life as a creative resource, Shimazaki Toson (1872-1943), for example, turned to the writing of history, while Masamune Hakucho (1879-1962) turned to criticism. Still other writers turned to popular fiction, including Kume Masao, whose bold pronouncements a decade earlier had figured large in the first shishosetsu debates. Kume now argued (in "Junbungaku yogi setsu," 1935) that the shishosetsu could be pursued only as an avocation, that it was not worth sacrificing one's livelihood for, and that making such a sacrifice actually defeated the purpose of writing by destroying the elevated mental state one wished to write about. Kobayashi calls Kume's essay an example of wishful thinking by a writer unable to turn his own life into a work of art.^[26]

Kobayashi and Yokomitsu, then, were in general agreement over the baneful influence of junbungaku on Japanese letters. Both, however, were quick to pronounce the shishosetsu a failed form without

[24] Ibid., 186.

[25] Ibid., 186-87.

[26] Ibid., 187-88.

coming to grips with its continued popularity. Like most writers and critics of the time, Kobayashi and Yokomitsu were influenced by their readings of European, and especially French, literature, which enjoyed tremendous prestige. (Yokomitsu goes so far as to say that Japanese writers knew more about Europe than Asia and that their own tradition was French and Russian literature.)^[27] A decidedly evolutionary view informed their thinking; they were confident that Japanese letters would follow in the path of, and eventually merge with, European literature, the unquestioned standard. To be sure, Yokomitsu perceptively linked junbungaku to the native tradition when he noted its similarities to the poetic diary; and Kobayashi described Japanese naturalism (and by extension the shishosetsu ) as more "feudal" (i.e., premodern) than "bourgeois" (i.e., modern and western)—an idea that shocked his contemporary audience, which had thought of it as having made a dean break with the past.^[28] Yet neither they nor their contemporaries seriously questioned the belief that Japanese literature ought to have developed along the lines of European "antecedents" and that any deviation was in fact a "distortion." Their literary evolutionism and corollary expectations about the self's relation to society blinded them to their own shrewd insights into the nature of Japanese naturalism and the shishosetsu .

Kobayashi in particular interpreted the shishosetsu in terms of the western literary aesthetics he had so thoroughly assimilated. He accused Japanese writers of adopting the trappings of European literature and placing them in a barren context, and he accused Japanese society of depriving writers of the means to relate to a world larger than their own immediate surroundings. Yet although he remained discouragingly silent, except for the one parenthetical remark noted above, about their native lineage, one can nonetheless surmise what he imagined those roots to be. In an essay on the medieval poet-priest Saigyo (one of a series of essays he wrote during the Pacific War addressing various topics and figures in classical literature), we find a remark that might qualify as an answer.

[27] "Junsui shosetsu ron," 153.

[28] "Watakushi shosetsu ron," 191. To Honda Shugo and his compatriots, Kobayashi's characterization came like a "bolt from the blue." See "Kaisetsu," in Nakamura Mitsuo zenshu 7:620.

Saigyo introduced afresh the concept of man's loneliness into the world of poetry. This theme permeates his verse. It might be said that loneliness was a treasure with which Saigyo was born. I think it no exaggeration to call his way of life as a priest and a recluse a mere expedient that served to preserve this treasure.^[29]

It seems safe to say that Kobayashi saw the shishosetsu writer in terms of the time-honored aesthetic recluse who engages in solitary contemplation at a distant remove from society. If we substitute "prose" for "poetry" and "writer" for "priest" in the passage above, we have a fair picture of the shishosetsu writer's situation some eight centuries after Saigyo. Kobayashi suggests that the only modern writers who were in a position to reflect seriously on the nature of the self in society were Mori Ogai (1862-1922) and Natsume Soseki (1867-1916). Authors of unusual learning, they understood fiction's ability (and Japanese naturalism's inability) to universalize experience. Neither cast his lot with the shishosetsu . The other writers, lacking fully "socialized selves," were left to seek a kind of "self-purification" achieved by recording one's mental state.^[30]

Yet purification was surely more than an aesthetic exercise; it was for the shishosetsu writer the staking out of a position vis-à-vis society as well. The isolated writer quite naturally wrote about the writer in isolation. He received no aesthetic or social encouragement to write of anyone else. That such a literature would appeal to readers may be surprising; yet the shishosetsu did in fact command a steady although by no means vast audience, as we shall see in Chapter 6. This "distortion" of European models flourished in spite of the continual pronouncements against it. That is why critics of this period regarded the shishosetsu as such a mystery; like a bumblebee, it seemed to have no business getting off the ground, yet it managed to buzz effortlessly about the bundan and beyond despite the efforts of the best critics to shoot it down. Many decried its lack of content. Kobayashi was particularly concerned about its preoccupation with "style." It was not enough that junbungaku writers had turned their private lives into open books; their obses-

[29] "Mujo to iu koto," in Shintei Kobayashi Hideo zenshu 8:35.

[30] "Watakushi shosetsu ron," 187-88.

sion with writing seemed to trivialize the content of their stories even further. Japanese naturalists, who took it as an article of faith that art imitated nature, always had before them something to write about, he argued. However much they struggled with their material, they never thought to publicize that struggle. The shishosetsu writers who followed in their wake, however, did just that. Circumstances forced them to produce stories whose subject matter was none other than the writing of those stories, and the presentation of reality became a more important theme than reality itself.^[31]

And yet this concern with "style," which Kobayashi so abhorred, may well be the saving grace of the shishosetsu writer's art. The content of daily life becomes far less "trivial" when it is problematized by its container—the "prison-house of language," to borrow Fredric Jameson's term. Efforts at sincerity of presentation are undermined time and again by the writer's own awareness of the sheer artifice of presentation, of the frame that shapes the "content" of his life. If a literary form like the shishosetsu does not even pretend to talk about society and if the aesthetics of isolation are its guiding force, then it makes little sense to censure it for being a "distortion" of the western novel, with which it has only a tenuous relationship. It is one thing for a Japanese author to read widely in an alien literature, but quite another to write in a manner that actually ignores the indigenous tradition's deeply held epistemological and linguistic assumptions. It does not seem at all strange that modern Japanese prose fiction (and the shishosetsu in particular), with a centuries-old history antedating all contact with the west, should have taken on an entirely different character from that of the novel. As long as confessional "content" remained the issue, however, the critics of Kobayashi's generation naturally judged the shishosetsu as inferior to its purported model, the western narrative.

[31] Ibid., 197-98.

The Third Period: Fugitives and Masqueraders

Nakamura Mitsuo (1911-) wrote even more disparagingly of the shishosetsu than Yokomitsu and Kobayashi did. With "Watakushi shosetsu ni tsuite" (1935) he inaugurated a career that was to culminate in such devastating critiques of modern Japanese literature as Fuzoku shosetsu ron (1950) and Shiga Naoya ron (1954). As with Yokomitsu and Kobayashi, Nakamura's strong background in French literature deeply colored his judgments of the native literature and particularly of the author's relationship to society. After reiterating Kobayashi's assessment of the individual in the west as a "socialized self," however, Nakamura carried his analysis of junbungaku one step further. "Unlike western romantics," he writes, “shishosetsu writers had no awareness of any confrontation between society and the individual. Indeed, they lacked the concept of 'society' altogether. Society for these writers was only those people who had a direct impact on their sensibilities: family, friends, lovers, etc."^[32]

This is a shrewd observation. Rather than argue, as Kobayashi did, that the Japanese writer possessed no "socialized self," Nakamura redefined society on a scale commensurate with the writer's consciousness, which extended to his immediate acquaintances ("family, friends, lovers") and not to some larger, more abstract system of institutions and relationships. His comment that personal experience in early twentieth-century Japan was felt to be a perfectly adequate mediator between the writer and this micro-society, moreover, underscores Maruyama Masao's theory of traditional social organization that we reviewed in Chapter 1.

Nakamura summed up his argument fifteen years later in a sweeping critique of modern Japanese literature, Fuzoku shosetsu ron . This work differs markedly from the period's other major essays on shishosetsu in its relentlessly critical tone. Nakamura singles out 1906-7 as a watershed in modern Japanese literature, when a "duel" took place between Toson's Hakai (1906) and Katai's Futon (1907). The literary "revolution" sparked by the former's broad social awareness was nipped in the bud, he argues, by a second revo-

[32] Nakamura Mitsuo zenshu 7:134.

lution sparked by the latter's claustrophobic self-consciousness.^[33] Now the naturalists and their followers were exempted from modern literature's most difficult task, which was to write about oneself and one's life in universal terms. Futon's success changed the course not just of the naturalist movement but of all twentieth-century Japanese letters. "Most of the Taisho period's best works ... were written in a style imitative of Futon ," he writes. "This is most unfortunate when we consider its innumerable defects and shortcomings."^[34]

An obvious difficulty with Nakamura's analysis lies in his acknowledgment of good works, and even masterpieces, in a form he argues has no redeeming qualities. It is a paradox he never quite resolves, because his assumptions have inevitably generated a set of literary guidelines that modern Japanese letters simply do not meet. Even as he redefines the perimeters of "society" in accordance with the writer's marginalized existence, he cannot bring himself to abandon the social vision and narrative perspective offered by the classical western novel. Although he repeatedly notes the early twentieth-century Japanese writers' great attraction to western literature, it is finally Nakamura himself who succumbs most completely. By western standards, Japanese naturalism, not surprisingly, is inferior to the European model. Nakamura never questions the motives, which he doubtlessly considers reasonable, of Japanese writers who try to assimilate the western view of literature. Assimilating this particular aspect of western culture is just one more way of competing at all levels with a hegemonic power. Nakamura speaks to this idea very bluntly in an essay he wrote two years earlier.

Foreign influence has a characteristically great impact on the literature of second- or third-rate nations. It would not, however, result in an endless succession of schools or movements in a first-rate nation—that is, one with a first-rate literature. I think that it would be profitable to compare literary circles in Japan with those in such second- or perhaps third-rate nations as Romania or Poland. The latter are susceptible to any new literary trend that develops in Paris. I am sure that this is the case. In this regard, modern Japanese literature has little to brag about. Many Japanese, I know, believe that their

[33] Fuzoku shosetsu ron , 29-31. See also the discussion below, 113-14.

[34] Fuzoku shosetsu ron , 53.

own literature is not worth studying at all and that they would be better-off reading translated literature.^[35]

The political implications of this statement, written just three years after Japan's defeat at the hands of the Allied Powers, are unmistakable. A nation's literature is only as viable and as reputable as its global position, as can be seen in Nakamura's equation of a first-rate literature with a first-rate power. Nakamura is quite naturally at a loss how to assess a literature that appears to be subordinated to the hegemonic culture but is in fact anything but derivative.

Another difficulty with the analysis lies in its exaggerated assessment of Futon as the turning point in Japanese letters. Katai's narrative indeed helped spark the "revolution" in literary confession of which Nakamura speaks. But it provided, at most, the trigger, not the powder. When we note Futon's marked similarities to Katai's previous writings and its differences from the shishosetsu that followed (as we shall do in Chapter 5), we can only conclude that the work was not in fact a radical break with, but rather a stage in, the process of development in the Japanese narrative that took place over a significant period of time. If anything, Futon was as much a result as it was the cause of this narrative development, which we have already traced to the Edo-period literary tradition. This is not to deny the importance of Katai's particular achievement or to minimize the influence—perhaps it would be more correct to say the overwhelming presence—of a narrative form legitimized by (in the eyes of the Japanese) a superior culture. It is only to remind ourselves that these "revolutionary" developments actually grew out of the literary tradition.

Ito Sei (2905-69) thought carefully and productively about the shishosetsu as a species of narrative quite disparate from the novel. His shosetsu no hoho (1948) is probably the most comprehensive study of modern narrative written in Japan up to that time. Like Nakamura, he acknowledges the great prestige of western literature. In his preface, Ito provides readers with a list of "required readings," of which nearly half the titles are European.^[36] But his recognition of the indigenous narrative tradition is apparent as well

[35] "Kindai Nihon bungaku," in Nakamura Mitsuo zenshu 7:415-16.

[36] shosetsu no hoho , 9-10. Included, incidentally, in these "European" texts are two stories by Poe.

in his inclusion of texts from the premodern canon (from Genji monogatari to Koshoku ichidai onna ) and in his discussion of the modern Japanese narrative against the background of its premodern predecessors. Ito argues that it makes much better sense to talk about the stories of an author like Masamune Hakucho, for example, who writes about the literary company he keeps, in terms of Tsurezuregusa or Hojoki than in terms of western narrative.^[37]

Ito thus connects the shishosetsu with certain premodern forms and opposes them all to the western novel, arguing that comparison with the latter can only be counterproductive. He emphasizes, moreover, as no one did previously, the role that a small homogeneous audience played in the shishosetsu's development, spotlighting the peculiar literary economy known as the bundan . Regarded as social outcasts and usually living in poverty, bundan writers had little opportunity to experience the world outside their immediate surroundings. And so they wrote, inevitably, about themselves, their peers, their sordid affairs, their hand-to-mouth existence, and their struggles to meet publisher's deadlines, because this was the only life they knew. Since the audience consisted mostly of their own peers, such things as character depiction and plot were as unnecessary as introductions at a club. Poverty may have cramped the imagination, but the limited narrative scope also turned out to be a blessing, Ito argues: it eliminated the need for fictionalization and allowed writers to focus on what they considered to be the most pressing issue, which was not how to write but rather, as reclusive rebels unconcerned with social convention, how to live.

Ito's argument builds on the assumption that the self and the personal voice to which it gives rise are the primary concerns of modern literature, whether Japanese or western. The two literatures differ, Ito suggests, in the way the authors present their "selves." Paradoxically, it is the more modern society that places greater constraints on self-expression. European writers, unwilling to subject their private lives to public scrutiny, resort to fiction because it provides the facade they need to function in society. Japanese writers, on the other hand, excluded by their professions from respectable society, have nothing to fear from confession because they have no social position to lose in the attempt.^[38]

[37] Ibid., 15-16.

[38] Ibid; see esp. 55-56, 106-7.

"Masqueraders" (kamen shinshi , literally, "gentlemen in masquerade") and "fugitives" (tobo dorei , literally, "runaway slaves") are the picturesque terms Ito coined in a contemporaneous essay to describe these two groups of writers. In this essay, which summarizes many of the arguments in shosetsu no hoho , Ito poses several important questions. What is it about fiction (shosetsu ) that moves Japanese readers, and how does western fiction differ from it? Is fiction's essence in its structure or its philosophical content? And what is fiction's role in the two cultures? He argues that the shosetsu , as a "report" of the writer's life, moves the reader to the degree that it is able to depict that life unerringly. One cannot have one's "philosophy" and fictionalize it, too, Ito seems to be saying. Bundan writers may be social outcasts, but they are also part of a respected literary tradition that idolizes the writer—a Saigyo or a Kamo no Chomei or a Basho—who rejects society and the material world.^[39]

Whereas Nakamura measures the shosetsu against the standard of the European novel and argues that it comes up short, Ito argues that it is by no means inferior just because it moves away from fictional narrative and toward essay and autobiography. The shosetsu , he insists, is an ideal medium for intimate expression that would suffer from too much attention to structure. The only way to write successfully in such a medium is to live, as a morally free person in an otherwise restrictive society, with a mind to documenting one's life as faithfully as one can.

Ito's assessment of the shosetsu is a valid one. But his faith in the importance of content ("philosophy") over form ("structure") leads him inevitably into the trap we identified earlier as the myth of sincerity. "The Japanese have no use for masks," Ito states flatly. "Fiction is rubbish—good only for writers who would dress up in coattails for the evening. Fugitives need not stand on ceremony. They can dazzle their readers with the slightest handiwork. But they must take great care not to overdo it lest they be labeled phonies and ostracized by their peers."^[40]

[39] "Tobo dorei to kamen shinshi" (1948), in Ito Sei zenshu 16:286-91; see esp. 287-88.

[40] Ibid., 291. See also shosetsu no hoho , 71. In Literature and Sincerity , Henri Peyre provides a useful corrective to Ito's argument that the western writer is necessarily less sincere because he has determined (as in Stendahl's case, to take one of Peyre's examples) that "only in fiction could he reach truth" (190). Although this felt need for fiction may confirm Ito's characterization of western writers, it does not change the fact that the concept of sincerity and the struggle to give it form in literature has occupied countless writers in the last two centuries and given birth to what Peyre calls the "personal novel." See ibid., 161-202.

The Mirage of Authenticity

As Ito sees it, then, bundan writers in effect consigned fiction to the dustbin of literature and opted for a less structured, autobiographical medium. Yet we still want to know how such a move could make their writing sincere. In an essay in Writing Degree Zero on narrative presentation, Roland Barthes identifies the preterite and third person (to which we might add Banfield's represented speech and thought) as the signposts of fiction in the classical western narrative. Together they "are nothing but the fateful gesture with which the writer draws attention to the mask which he is wearing."^[41] There is no escaping the mask, because language itself points to it. "The third person, like the preterite ... supplies its consumers with the security born of a credible fabrication which is yet constantly held up as false."^[42]

Although Ito and others argue that the shosetsu's all-important philosophical "content" determines its "sincere," "structureless" form, Masao Miyoshi's commentary on the shosetsu in light of Barthes's analysis seems closer to the truth:

It is the reverse of the novel: rather than a "credible fabrication which is yet constantly held up as false," the shosetsu is an incredible fabrication that is nonetheless constantly held up as truthful. Art is hidden, while honesty and sincerity are displayed. Distance is removed, while immediacy is ostensive.... The shosetsu is thus an art that refuses to acknowledge art.^[43]

In short, narrative intimacy is itself an ideologically motivated form and in no way subordinated to content. Form is naturally suppressed in a culture that privileges sincerity over design, experience over word. It has not disappeared, however, only assumed

[41] "Writing and the Novel," 40.

[42] Ibid., 35. Although he does not elaborate, Barthes implies that first-person narration is just as much of a fabrication as third-person narration, when he notes the ability of the former to confer on the narrative the "spurious naturalness of taking the reader into its confidence (such is the guileful air of some stories by Gide)" (ibid).

[43] "Against the Native Grain," 233. See also Miyoshi's account of the western narrative form and the Japanese alternative, 231-33.

the guise of style. As a verbal construct, "sincerity" in the shosetsu , no less than design, is part of a rhetorical rather than a referential field.

This distinction is easily demonstrated by considering the criterion used to judge sincerity. We saw in Chapter 1 the Japanese writer's struggle to transcend his calling and achieve a moral legitimacy by offering his life (however derelict) as an example to his community of readers. The stakes of appearing sincere were therefore very high, and it is no wonder that a great many writers and a vast number of readers conceived of the shosetsu in Ito's terms: the choice was indeed between fiction and philosophy, between "how to write" and "how to live." It goes without saying that the writer who most successfully understood (dare one say "exploited"?) the myth of sincerity became the object of deification. Shiga Naoya's nickname, shosetsu no kamisama ("the god of the shosetsu "), is, then, no gratuitious label but the ultimate signifier of this mythical hierarchy. Criticism of Shiga (and by extension any less successful would-be recorder of "truth") necessarily centers on the touchstone of emotional honesty. In the words of Honda Shugo (1908-), for example, a principal defender of the myth:

The shosetsu of Shiga Naoya contain no false notes [uso , literally, "lies"].... There is not a single false note in the author's feelings toward his characters or toward nature. Instead of "no false notes," I could just as easily say "no idle phrases" or "no empty rhetoric." ... To put it more positively: each and every Shiga sentence harbors profound emotion and a powerful sense of authenticity.^[44]

Honda's statement is significant less as a critique of Shiga than as an implied ideal for Japanese literature, in which the authentic transcription of the author's feelings becomes the paramount aim of writing, and the correct identification of (perhaps even communion with) those feelings becomes the chief task of the reader/critic. But just what does this lack of deception really mean? Is the emotion in question felt at the time of writing or at the time of the incident about which the author writes? Honda argues in favor of the former; but to argue either way is in effect to acknowledge the mediation involved in any "authentic" transcription.

[44] “Shirakaba" ha no bungaku (1954), 155.

This leads to an even more awkward question. How do we know that Shiga or any author is not "lying"? That is, how exactly are we to judge the authenticity of the author's feelings? Clearly, we must engage in an act of faith and believe that the author is telling the truth. The critic's only recourse in determining the author's credibility, other than relying on documentation that is never fully verifiable, is to appeal to the author's style. Thus Honda's sonorous refrain: "No false notes." In that style, Honda suggests, is a leanness and a ruthlessness necessary and appropriate for honest self-scrutiny. In short, it is a style that says, "I mean business." This surely is what Honda means when he argues that Shiga's writing contains no idle phrases or empty rhetoric. But style itself is rhetoric, the literary equivalent of acting, the concealed art without which the honesty and sincerity could not be "revealed." It turns out, then, that form is indeed the key to content. For what is sincerity in literature but the donning of a verbally well-wrought mask, the masterful display of honest emotion? In the words of Benjamin Crémieux: "Is not the greatest artist also the greatest imposter , the most capable of giving form to whatever imaginary reality, of bestowing upon it, through expressing it, a soul and an appearance of truth which others than the creator will accept?"^[45] That a writer like Shiga really does sound more sincere than others, then, is a tribute not to his honesty but to his mastery of the rhetoric (the intimate voice, ellipses, allusions, etc.) of authenticity.

Honda and others argue that the test of any shishosetsu lies in the recognizability of its author. In fact, the author is "recognizable" only through a style made familiar from previous texts. The role played by sibling texts as a guide to reading an author, therefore, is a crucial one, and those who overlook its significance are all but forced to scavenge the author's private life in order to supplement a given text. (This matter will be addressed more substantively in Part 3 and especially in the chapter on Kasai.) Ito Sei clearly has an inkling of the problem when he writes:

[The shishosetsu ] was written with the expectation that the reader would know the hero's (that is to say, the author's) personal history without explanations of his circumstances and position. The hero in

[45] Quoted in Peyre, Literature and Sincerity , 336. (Emphasis added.)

such a text is therefore a virtual nonentity, so superficially is he described.... The reader had to glean what information he could about the hero's personality and circumstances from gossip about the author current in literary circles, or be familiar with the hero from previous texts that reported on the author's private life .^[46]

Ito, however, ignores the ramifications of his own insight: namely, that the reader need not rely on the finally inaccessible life. The oeuvre is enough. Regular readers of literary magazines, as we shall see in our discussion of the bundan in Chapter 6, would have no trouble piecing together the "lives" of the authors about which they read.

It is not just the writer, then, who must master the rhetoric and assume a role: the reader, too, must cast himself in a very specific relationship with the text. The shishosetsu's style demands it. We need not go as far as Walter J. Ong's claim, although it is probably a valid one, that the writer's audience is always a fiction to understand that the relationship between writer and audience is founded on rarely articulated but nonetheless concrete assumptions about the text and the world that must be accepted in order for communication to take place. Regardless of how seemingly intimate the narrative voice or how casual the allusions, the author does not, cannot, know his readers. He has to "make his readers up," as Ong puts it, "fictionalize them."^[47] Narrative intimacy—or narrative distancing, for that matter—is part of the fictionalization. Although it may surprise shishosetsu readers to learn that they are not the only audience intimately addressed or expected to possess background information, there is no question that Ong's observations concerning the Hemingway narrator in the opening passage of A Farewell to Arms are applicable to our study.

The reader—every reader—is being cast in the role of a close companion of the writer.... It is one reason why the writer is tightlipped. Description as such would bore a boon companion. What description there is comes in the guise of pointing, in verbal gestures, recalling humdrum, familiar details.... The reader here has a well-marked role assigned him. He is a companion-in-arms, somewhat later become a confidant. It is a flattering role.... Hemingway's exclusion of indefinite in favor of definite articles signals the

[46] shosetsu no hoho , 66. (Emphasis added.)

[47] "The Writer's Audience Is Always a Fiction," 11.

reader that he is from the first on familiar ground. He shares the author's familiarity with the subject matter. The reader must pretend he has known much of it before.^[48]

This argument recalls Ito Sei's on the bundan readership, with one important exception: Ong's emphasizes the pretense involved in the act of writing and reading. Sincerity has a new twist here: it is the product of style, not its generator.

The myth of sincerity develops to its logical conclusion in Hirano Ken's Geijutsu to jisseikatsu (1958), an exploration of the links between the modern Japanese writer's life and art. The discussion quickly gravitates toward the shishosetsu , which provides the subject matter for the work's major theoretical essay, "Watakushi shosetsu no niritsu haihan." We noted at the beginning of this chapter Hirano's characterization of the watakushi shosetsu as the literature of destruction and of the shinkyo shosetsu as the literature of harmony and salvation, a characterization generally supported by Hirano's contemporaries, including Ito Sei. Hirano's aim is to cast Japanese writers in two broad types: those who accommodate, and those who turn their backs on, the obligations of private life. Hirano suggests that the shishosetsu writer's art is by its very nature stimulated by crisis and stifled by tranquillity, anchored as it is to the vicissitudes of lived experience. Shinkyo shosetsu writers, therefore, whose twin goals of domestic harmony and candid expression harbor an insoluble contradiction, are faced sooner or later with the choice of betraying either their families or their profession. Accommodators that they are, they usually choose the latter course.^[49] No such contradiction exists for watakushi shosetsu writers, on the other hand, because they have already forfeited domestic tranquillity in order to chronicle lives that are often bent on destruction. Ostracized and poverty-stricken, their raison d'être lies entirely in their utter truthfulness as artists?

Hirano sees "sincerity" (and its companion ideals of "candor" and "truthfulness"), then, as the only positive, if ultimately undefinable, quality of writers who would mine their own lives for any material whatever the cost. The result of this reification is an in-

[48] Ibid., 13.

[49] Geijutsu to jisseikatsu , 29-37.

[50] Ibid., 36-37.

fiated and exalted signifier curiously drained of meaning, a capricious looking glass reflecting nothing but the authenticity of its own persuasion. The French poet Luc Estang's wry characterization is apt: "Sincerity is to itself its own mirror. In literature, we find only reflected sincerities."^[51] But Hirano, like Honda and Ito before him, is apparently satisfied with tautology, asking no more than that sincerity be sincere.

The myth of sincerity is founded, as we have seen, on the illusion of authorial presence. Japanese critics have expended enormous efforts buttressing the myth, perhaps because they have feared to ask whether the shishosetsu would still contain anything of value should the author in fact be inaccessible. To remove sincerity (and by extension authorial intention) from the text would be to deprive it, a record of otherwise "trivial" events, of its most important "content" and thus of any interest.

We have also seen that sincerity is not an ethical goal to which the artist aspires but a strategy of discourse, motivated by the desire of writers to legitimize or at least strengthen their position within the bundan and vis-à-vis society and governed by the shishosetsu form itself. Arguing that the author is "absent" from the work, therefore, far from denying the referential world to which shishosetsu writers constantly allude, actually helps us recognize a highly complex negotiation between art and life that a reified notion of sincerity fails to suggest. For, once we grasp sincerity as an ideology rather than as a vague, ultimately inaccessible emotion, we can understand how it controls the mode of literary production. Candor becomes a commodity that writers have no choice but to produce and critics no choice but to appraise.

The critical posture toward the shishosetsu during the period we have reviewed, from the mid-1920s through the 1950s, has remained essentially unchanged and continues to dominate present-day thinking. Critics of all .persuasions, including those most skeptical of the shishosetsu writer's art (from Akutagawa Ryunosuke to Nakamura Mitsuo), have unfailingly acknowledged the value of the shishosetsu as confession of a different order from other literatures and have thus actually contributed to the form's mystique. For that reason the shishosetsu , despite all attacks, has occupied a

[51] Invitation à la poésie . Quoted in Peyre, Literature and Sincerity , 328.

critical sanctuary in which a variety of powerfully nostalgic and ethnocentric emotions concerning the form's "purity," "Japaneseness," and philosophical "honesty" are heavily invested.

That some writers would rebel against this ideology was perhaps inevitable. What is surprising is how early in the shishosetsu's history the rebellion occurred; parody was common almost from the start. Yet the norm of sincerity was so pervasive that its deifiers generally failed to recognize the parody in other writers' texts. Members of the supposedly close-knit bundan would typically contrast their own productions, which they regarded as art, with the "bald" confessions of their peers. That they were able to question the absolute priority of experience over writing in their own work at least (as we shall see in Part 3), is justification enough for us to challenge the shishosetsu's claim to uniqueness in world literature (and the corollary claim of its unintelligibility to outsiders), even as we situate the form more precisely in its specific linguistic and cultural environment—ever mindful that particularity is no synonym for sacred exclusivity.