6
The Bundan: Readers:, Writers, Critics

I really get the feeling that your reading ... is based on a knowledge of, and a sympathy for, the author's private life.
Chiba Kameo, in a roundtable discussion

I simply could not force myself to become more intimate with society merely in order to write stories that any young girl would then be able to understand.
Tayama Katai, in a roundtable discussion

The shishosetsu so dominated the Taisho literary world that the phrase "Taisho literature" (Taisho bungaku ) now connotes its heyday. This phenomenon was due in no small measure to the rise of the bundan . The existence of a literary subculture in which writers came to associate with one another more as social acquaintances than as artists and which encouraged gossip about one's peers, by word of mouth and ultimately in print, contributed immensely to the critical consciousness of the shishosetsu as being uniquely true to life and therefore the only shosetsu form of any importance in Taisho letters. Although it would be wrong to isolate the bundan as the shishosetsu's sole formative element when we have already linked several significant literary, linguistic, and intellectual factors to its inception, there is no doubt that it played a crucial role in legitimizing critical focus on the writer's life as much as on his writings. In such a climate, the writer freely assumed readers' familiarity with—and curiosity about—the details of his personal life, and publishers actively solicited stories that exploited this curiosity.

Although the bundan in its broadest sense includes any person or coterie active in the literary world, it is used here in its more restricted sense to include only those writers, critics, and publishers associated with what is commonly called junbungaku , or "pure"

(i.e., confessional or autobiographical) literature, as opposed to taishu bungaku , or "popular" literature. The distinction between "pure" and "popular" is almost certainly a carryover from the Edo-period dichotomy, noted in Chapter 1, between the nonfictional literature of edification and the fictional "nonliterature" of entertainment, bolstered by the naturalist school's stress on lived experience as the only legitimate source on which the author could draw. It was also intimately related, surely, to the values of a select as opposed to a mass audience.

The bundan , in the narrow usage considered here, was a product of three crucial trends that came together during the Taisho period: the literary journal's emergence as the "pure literature" writer's single most important medium, the popularity of shorter works, and a camaraderie among writers nurtured by physical proximity, social alienation, and contemporary journalistic demands. These trends were so pervasive and intimately related that they became self-perpetuating and often transcended their status as mere professional realities, with which an author had to cope in the process of writing a story, to become a central theme in a story. (We shall see this theme explored to its fullest extent in our discussion of Kasai Zenzo.) And since literary purity was measured in terms of how closely the author adhered to the details of his personal life, a writer had a good deal of motivation to model his characters after his acquaintances, for example, or to cite his previous stories (as Uno does in Amaki yo no hanashi , quoted in Chapter 1), or even to refer self-consciously to the very manuscript which he was then writing. These factors in turn provided the bundan audience, composed largely of writers and would-be writers, with a powerful incentive to read and criticize stories entirely on a referential level.

At the heart of the bundan's raison d'être is the junbungaku writer's elitist consciousness, born of common education (most writers went to universities, typically Tokyo or Waseda or Gakushuin, before launching their literary careers), geography (Tokyo, the hub of cultural activity, was the home or adopted home of virtually every junbungaku writer), and aloofness from a society that generally took a dim view of the writer's profession. Traditional social prejudice against the writer of "nonliterary" fiction (gesaku ) worked, as we have seen, against the early twentieth-century shosetsu writer as well. The authors of "pure literature" believed that their writing

was not "fiction" in the sense that popular literature was, since they, too, were imbued with this prejudice. But society as a whole (i.e., non-bundan society) was less inclined to distinguish one kind of writer from another. In an achievement-oriented world that paid special reverence to university graduates, who were expected to pursue socially useful careers, writers, with few exceptions, were second-class citizens. Mori Ogai earned respect as a man of letters largely because he had already distinguished himself in the medical profession. And Masamune Hakucho suggests that Natsume Soseki's reputation as a scholar helped sell his novels. (Soseki had been a lecturer at Tokyo Imperial University before joining the staff of the Tokyo newspaper Asahi in 1907 as editor of the newspaper's literary page.)^[1] Hakucho and his naturalist colleagues, meanwhile, met with little commercial success, in part because their many stories set in the provinces did not appeal to an urban audience but also because of the traditional view of the "fiction" writer as an outsider or outcast.

Just what constitutes an "outsider" is of course in itself problematic, but it depends at the very least on the existence of an "insider." Kawakami Tetsutaro defines insiders as those established intellectuals who maintain the orthodox system of thought; he suggests that modern Japan has had no orthodox, traditional system (such as Neo-Confucianism in Tokugawa Japan) to which an intellectual could ally himself, and thus no true insiders. The most likely candidate for an "orthodox" system of thought in post-Restoration Japan, Kawakami argues, was the utilitarian philosophy, with its emphasis on success and service to the state. By this standard, at least, the man of letters—whose very act of writing, when not clearly utilitarian in intent, was considered antisocial in Meiji and Taisho Japan—was very much an "outsider."^[2] This common view of the writer manifested itself even at the level of daily life. One writer reported to his friends in astonishment that he was willingly let a house even after he had defiantly announced his calling to the landlord—rare treatment then for one of his kind because of his low social and economic status.^[3]

[1] Shizenshugi seisui ki , in Masamune Hakucho zenshu 11:308-9.

[2] Nihon no autosaida , 11-12 and 206-7.

[3] The writer was Hirotsu Kazuo. See Okubo Fusao, "Bundan ni tsuite," in Bunshi to bundan , 243. In the same book, Okubo notes the literati's growing social prestige and affluence since the war ("Bundan no sengo," 163-234).

The junbungaku writer's career revolved, at least at the beginning, around his coterie and its publication, known as a dojin (or donin ) zasshi , which catered to a small and homogeneous audience. Unlike contemporary coterie magazines, which often have a nationwide membership, the Taisho magazines were very exclusive and their memberships defined by mutual acquaintance and common purpose, a fact that resulted both in fast friendships and bitter infighting.^[4] Kasai Zenzo, for example, published his earlier stories in one such coterie magazine, Kiseki , which he and his Waseda University colleagues put out monthly from September 1912 to May 1913. Kiseki was but one of several dozen dojin zasshi appearing at the time. Although most were quite short-lived, they made up the Taisho bundan's lifeblood. Shiga Naoya established his career by his many contributions to a dojin zasshi called Shirakaba , which was one of the most successful. The coterie magazines, edited and distributed by the contributors themselves, provided an ideal vehicle for the aspiring young author who lacked a name but possessed the boundless energy needed to publish a magazine on a shoestring budget. Coterie members were plagued by a lack of funds, and the attrition rate was high; a typical magazine lasted perhaps a dozen issues. Kiseki (1912-13), for example, had 9 issues; the first four and by far the most significant series of Shinshicho (1907-17) had anywhere from 6 to 11; Ningen (1919-22) had 24. There were exceptions to this law of evanescence, the most notable being Shirakaba . A total of 160 issues appeared from April 1910 to August 1923. Its longevity was due in large measure to its strong financial backing and to the contributing members' relative affluence.^[5]

Printings of most dojin zasshi were in the very few hundreds, or less. Only a smattering of copies were sold at bookstores on consignment; the majority were distributed gratis to important literary figures in hopes of catching their attention.^[6] The Kiseki group

[4] See Yamamoto Kenkichi, "Dojin zasshi hyo," 177-78, for a general description.

[5] Chikamatsu Shuko was the only writer discussed in Part 3 who never belonged to a coterie. He began his career as a critic, writing in newspapers like the Yomiuri , before starting to write fiction.

[6] Ibuse Masuji, in a conversation held on 1 March 1985, recalled that the strategy used to catch the influential critic's eye in his early dojin-zasshi days (late Taisho-early Showa) was to send the magazines by registered mail. See also Ibuse, "Dojin zasshi no koro," 222-24.

printed two hundred copies of their first issue and placed one hundred of these in bookstores like Kinokuniya on consignment sale. The twenty-eight copies that actually sold, however, are said to have been bought in twos and threes by the dojin zasshi members themselves in an effort to save appearances.^[7] Again, the exception is Shirakaba , whose monthly circulation numbered in the thousands and peaked at well over ten thousand around 1920. Its broad appeal as a magazine of the arts helped insure the popularity of its major contributors: Arishima Takeo (1878-1923), Mushanokoji Saneatsu (1885-1976), and Shiga Naoya.^[8]

By the time a coterie magazine had folded, however, some of its contributors would have made enough of a name for themselves to be invited to write for the more prestigious literary and commercial magazines, which included Shinshosetsu (1889-1950, with interruptions), Chuokoron (1899-present), Shincho (1904-present), Kaiho (1919-33), and Kaizo (1919-55). Fully half of Kasai's fiction (roughly forty stories), for example, appeared in these five magazines; the rest is scattered in some thirty other periodicals. Perhaps a third of Shuko's nearly two hundred stories appeared in these same magazines, and a good many more in Bunsho sekai (1906-20), one of the naturalist movement's strongholds; Waseda bungaku (1891-present, with interruptions), one of the great university-supported magazines along with Keio University's Mita bungaku (1910-present, with interruptions); and Bungei shunju (1925-present), a well-known general-interest magazine, which began as a literary magazine. Shiga emerged as a major literary figure with the publication of Otsu Junkichi in Chuokoron in 1912, and he serialized his crowning achievement, An'ya koro (1921-37), in the pages of Kaizo .

Most readers of dojin zasshi and of all but the largest literary magazines were writers and would-be writers themselves, nearly all of whom lived in or near Tokyo. Precise figures are hard to come

[7] Kono Toshiro, “Kiseki kaisetsu," 13.

[8] See the entries in NKBD , vol. 5, on the various magazines in question. The information on Shirakaba circulations is gleaned from a personal correspondence (12 Feb. 1985) from Miyazaka Eiichi, the magazine's last editor.

by, but the circulation of a purely literary magazine like Shincho in early to mid-Taisho is thought to have ranged from three to five thousand. Magazines like Kaizo and Kaiho , which also catered to literary audiences, had circulations of around thirty to forty thousand in 1920. Only Chuokoron , far and away the largest general-interest magazine, could boast a circulation of over a hundred thousand at this time.^[9]

This high concentration of writers in Tokyo bred familiarity on a social as well as a professional level. The major literary magazines, moreover, served as clearing houses of information and gossip about writers. They also served as forums: published interviews with writers and roundtable discussions between writers and critics provided an inexhaustible supply of literary grist. Writers, therefore, sharing not merely the pages of the same magazines but also frequently participating in the same roundtable discussions, were in constant contact with one another and held few secrets. Rare was the bundan critic, therefore, who failed to note the correspondences between a story and what he knew of its author. In the following panel-discussion excerpt, we see how the discussion of a short story ("Isan," 1924) quickly gravitates toward its author, Kasai Zenzo.

[9] The figures are based on the following sources: Hirano Ken and Matsumoto Seicho, "Watakushi shosetsu to honkaku shosetsu," 133; Domeki Kyosaburo, Shinchosha hachiju nen shoshi , 14; Yokoyama Haruichi, Kaizo mokuji soran somokuji , 11; and Maeda Ai, Kindai dokusha no seiritsu , 173. See also Yokozeki Aizo, Watakushi no zakki cho , esp. 5-9, for further information on circulations and on the bundan in general.

Nakamura Murao's claim that he read "Isan" as an autonomous text of fiction in effect goes unsubstantiated. Like nearly every critic of the age, he reads a text in such a way that knowledge of the author's private life becomes essential to its understanding. It follows, then, that Nakamura would say of Kasai, in another discussion: "What makes [Kasai's] stories interesting is the writer himself. They may not be so interesting to readers who don't know him, though."^[11] Here again, private life is seen as the Ur-text of pure literature.

Although literary magazines reached their zenith in the mid-Taisho period, daily newspapers like the Tokyo Asahi and the Yomiuri were also an important medium for prose writers through the end of Meiji and early Taisho. One of their greatest contributions was serial publication. Many of the naturalist writers' longer works first appeared in the major dailies: Tayama Katai's Sei (Yomiuri , 1908), Tsuma (Nihon , 1908-9), and En (Mainichi shinpo , 1910); Shimazaki Toson's Haru (Tokyo Asahi , 1908), Ie , part 1 (Yomiuri , 1910), and Shinsei (Asahi , 1918-19); Tokuda Shusei's Kabi (Asahi , 1911) and Arakure (Yomiuri , 1915). (The Asahi also published all of Natsume Soseki's fiction from mid-1907 until his death in 1916.)

It is difficult to understand why newspapers would have an interest in publishing some of these authors (who, however important to literary history, had only a small following) unless we realize that the newspapers themselves did not enjoy anything close to the mass audience that they do today. Although all the major dailies now boast circulations in the many millions, they served a far more elite clientele until mid-Taisho or later, as Table 1 shows.

Papers like the Asahi, Tokyo nichi nichi, Jiji shinpo , and the Yomiuri appealed to a highly literate audience and devoted far greater space to literary matters than any of the major papers do today. Let us

[10] Kasai Zenzo zenshu bekkan , 410. Originally published in Shincho , Mar. 1924.

[11] Ibid., 408.

examine the one that devoted the most space to literature: the Yomiuri .^[12] In early Taisho it, like most other papers, was only eight pages long. Page 1 of our randomly chosen 4 May 1914 issue is dominated by two serials, plus the daily editorial. Political and international news is relegated to page 2; local and cultural news, to page 7. Page 3 includes a travel essay and an installment of translated literature. Two critical essays appear on page 4 and two more on page 5, along with a long poem in the "new style." An installment of still another shosetsu appears on page 6, and of a historical tale told in the old kodan style on page 8. All this—plus the usual assortment of haiku, tanka, senryu, and reviews. In addition to publishing fiction and criticism by professional authors, the Yomiuri held competitions in various categories for aspiring writers. The paper advertised its first short-story competition during May 1914, and the prize-winning story appeared in the 8 June issue, complete

[12] See Ikegami Kenji, “Yomiuri shinbun to Tayama Katai," for an analysis of the importance of the Yomiuri as a literary organ in mid- to late Meiji.

with the judges' critiques in a format anticipating the Akutagawa and Naoki prizes begun by Bungei shunju in 1935.

“Yomiuri bundan" (page 4), a regular feature, always contained one or two essays, reviews, or travel pieces by well-known authors. Newspaper photos are still rare in this period (averaging only about one per page), yet one often finds a commemorative photograph of a gathering of writers at some restaurant or publishing house. But the most remarkable item on this page is a daily column called “Yomiuri sho" (Yomiuri notes), which kept readers up to date on recent and future publications as well as the activities of the nation's major writers. On 3 May 1914, for instance, we learn that a volume of essays by Shimamura Hogetsu has gone through the final proofs and now awaits publication by Shinchosha. On the sixth, we learn that a two-hundred-page manuscript by Iwano Homei will appear at long last in the June issue of Chuokoron . On the nineteenth, we learn that Shincho's May issue has been taken off the market by authorities, the victim of official censorship. On the twentieth, we learn that Hogetsu's above-mentioned volume of essays was published the previous day. Throughout the last week of May the column lists articles and stories (including Homei's) that have appeared in the major literary magazines' latest issues.^[13]

The activities of Chikamatsu Shuko, already an established writer by this time, are faithfully reported. We learn that he will publish a short story entitled "Haru no yukue" in the June issue of Bunsho sekai (5 May). He leaves Nara, his first stop on a provincial tour, on the fourth for Dogo Spa in Shikoku (8 May) and later returns to his home in Bizen (Okayama) for a brief visit (18 May). On the nineteenth he moves to Kyoto and lodges at a boarding house (22 May). The column dutifully notes his address. Finally, on 5 June, Shuko boards the 9:00 A.M. express for Tokyo (7 June). Shuko himself reports on his travels in a number of articles appearing on the same literary page. In "Ryojin" (Dusty road, 18 May), he describes his journey from Tokyo to Nara, and in "Bunraku-za yori"

[13] Thanks to Professor Sasaki Yasuaki of Ibaragi University for introducing this fascinating column. A similar but less detailed column appears in Jiji shinpo ("Bungei shosoku") and also in the major literary magazines like Shinshosetsu , Shincho , Waseda bungaku , and Bunsho sekai , although the latter could not compete with the dailies for timeliness and comprehensiveness.

(From the Bunraku Theater, 8 June), he writes further about his sojourn in the Kansai area. Read together, “Yomiuri sho" and the articles provide thorough coverage of Shuko's movements. The latter, in the popular travelogue style, rewards the curiosity that the former has piqued.^[14]

Since coverage of Shuko represents only a fraction of the total, it is easy to imagine from this sampling the tremendous reader interest in even the most peripheral literary matters. The writer lived the life of a celebrity, although on a much smaller and more intimate scale than the word implies today, and virtually any activity was considered "news." That his "personal" life (e.g., travels, boardinghouse addresses) was just as newsworthy as his "professional" life (e.g., writing, publication plans) reminds us once again of the general disinclination to distinguish between the two. In his readers' eyes (and no doubt in his own), the writer comported himself necessarily as writer in everything he did. Being an author was by definition a twenty-four-hour-a-day occupation—not simply a livelihood, but a way of life.

Ironically, it was when the newspapers began commanding truly sizable audiences (circulations of most major dailies had reached the hundreds of thousands by the end of Taisho) that their literary significance declined. As circulations increased, literary editors looked for authors with broader appeal, forcing junbungaku writers to turn to the literary magazine as their principal medium. "Forcing" may be the wrong word, given the bundan writer's elitist consciousness. There are stories of bundan literati who in some cases did not wish to appear even in the same magazine with a writer who was thought to have made concessions to "popular" tastes. Akutagawa Ryunosuke, for example, himself a victim throughout his career of critical attacks against his tendency to "invent" stories rather than tell the "truth" about his own life, once refused to have a story published in the same issue of Chuokoron with Muramatsu

[14] A passage in Shuko's Giwaku zokuhen (1913) gives us an idea of the monitoring function of this and similar columns. The story's narrator-hero is surprised to learn from a maid at the Nikko inn where he is lodging that his former wife, whom he has not seen in years, knew that he was in Nikko and had inquired about him when she arrived. "How did she know that I was here?" he muses. "She must have read about me in the newspapers. She obviously checks up on me every chance she gets" (CSS , 59).

Shofu (1889-1961), a very popular (and very lowbrow, in Akutagawa’s opinion) writer of historical fiction.^[15]

One of the victims of this popularization of newspaper shosetsu was Tayama Katai. A pioneer along with Shimazaki Toson of a radically new kind of serialized fiction that seemed devoid of any concession to reader interest, Katai saw his own popularity decline as the literary audience of the expanding dailies grew more diffuse. His rationalization of his position is worthy of any bundan writer in its disdain for the nonliterati:

I had believed that if I only knew more about society as a whole, I could write stories with greater popular appeal. I tried, and I failed ... and I ultimately abandoned the attempt. I simply could not force myself to become more intimate with society merely in order to write stories that any young girl would then be able to understand. I feared that I would end up sacrificing my art in the attempt. And that frightened me.^[16]

Beginning in late Taisho, Katai published an increasing number of his serialized shosetsu in the less high-powered provincial dailies.

Thus, the literary magazine became the Taisho junbungaku writer's mainstay. Fees were low, however, and even frequent publication in the more influential magazines did not provide a sufficient income, although it did enhance a writer's reputation. (Nakamura Murao writes in 1925 that a literary magazine like Shincho , with a circulation of barely ten thousand, could afford to pay a writer only seven or eight yen per page, while women's and entertainment magazines, with circulations in the hundreds of thousands, paid writers twenty to thirty yen per page.)^[17]Kaizo editor Yokozeki Aizo recalls that "Kura no naka" (1919), for example, earned its author Uno Koji instant recognition and numerous solicitations from literary magazines but no appreciable change in living arrangements. Uno shared his quarters with another writer and usually had only enough money to buy a one- or two-day supply of rice at a time.^[18] No early Taisho-period writer—not even the likes of Soseki or

[15] NKBD 4:476d.

[16] Tayama Katai et al., "Shincho gappyokai," 59.

[17] "Bungei zasshi no koto," in Bundan zuihitsu , 112.

[18] Watakushi no zakki cho , 8-9.

Toson—could afford his own home, aside from certain independently wealthy "Mita School" and "Shirakaba School" writers, who inherited theirs. (Tayama Katai, a rare exception, managed to build a small house in the suburbs in late Meiji.) Writing, even full-time writing, was not a money-making proposition. Since most writers had little to live on, they commonly took advances, and it was the policy of most literary magazines to offer them freely in order to encourage—or prod—writers into producing manuscripts.^[19]

Nor could most writers count on book royalties to supplement their income by any substantial amount. A first printing often numbered in the mere hundreds of copies, and a famous author did well to sell at the very most one or two thousand copies in this period. Even Soseki's books sold only a few thousand copies in their initial years of printing.^[20]

The best way to grasp the scale of the bookselling business in late-Meiji and Taisho Japan is to review the admittedly scanty data on sales of that period's most popular works. Hakai (1906), which catapulted Shimazaki Toson (then known only as a poet) into the ranks of leading fiction writers, met with extraordinary success. The first private printing of fifteen hundred copies sold out almost immediately, and within a year ten thousand copies had been sold. This not-enormous figure was virtually unprecedented for its time. We have already noted the sensation that Tayama Katai's Futon created. When it first appeared in book form (in a collection of Katai's stories entitled Katai shu , 1908), it went through seven printings in six months, which seems remarkable until one realizes that each printing after the first most likely numbered only in the hundreds.^[21]

The data on Natsume Soseki's books are of considerable interest, because they bring home to us the limits of success that one could expect as a writer in the early to mid-Taisho period. Soseki was far

[19] Ibid. See also Nakamura Murao, Meiji Taisho no bungakusha , 47-48 and passim, for useful information on the bundan .

[20] Nakamura Mitsuo, "Taisho bungaku no seikaku," in Nakamura Mitsuo zenshu 7:503.

[21] The figures on printings are gleaned from Senurea Shigeki, Hon no hyaku hen shi , 123-28. Hototogisu (1898-99), the most widely read book of the Meiji period, sold just nine thousand copies in its first year but eventually sold a half-million copies over a twenty-five-year period. Konjiki yasha (1897-1903) also sold extremely well, but Senuma offers no concrete figures. See ibid., 77-84.

and away the most popular writer of his time; yet his books hardly made him a wealthy man. Matsuoka Yuzuru (1891-1969), Soseki's son-in-law and an author in his own right, compiled figures on most of Soseki's books and surmises that sales of all books during Soseki's own lifetime (from the publication of Wagahai wa neko de aru in 1905 to his death in 1916) totaled perhaps one hundred thousand copies, or roughly ten thousand a year, and that sales of his four most popular works—Neko, Botchan (1906), Kusamakura (1906), and Gubijinso (1908)—accounted for fully half the total.^[22] The first printings of books for which reliable figures are available generally numbered only two to three thousand, and later printings only a few hundred. Sales did not pick up substantially until the appearance of popular editions late in the author's life; then they rose dramatically with his death.

Soseki could command higher royalties than most: 15 percent for the first printing (after the first 130 copies), 20 percent for the second through fifth printings and 30 percent thereafter. These terms, gleaned from a contract drawn up by Shun'yodo, Soseki's principal publisher, apply specifically to Uzurakago (1906), Soseki's second through fifth printings, and 30 percent thereafter. These terms, gleaned from a contract drawn up by Shun'yodo, Soseki's apparently good enough to arouse the jealousy of other writers. Nagata Mikihiko (1887-1964) recalls that Chikamatsu Shuko became infuriated when he learned that Soseki's share reached 31 percent. "I don't care how famous he is," he complained to Nagata, "that's highway robbery!"^[24]

And yet, although the royalties attest to Soseki's popularity, they did not bring him great wealth. His annual salary from the Asahi was three thousand yen, but his contract with the newspaper forbade him publication in other periodicals. Royalties of perhaps another two thousand yen on book sales brought his income to a total of around five thousand yen.^[25] This was by no means a large sum, compared to what could be made in other professions.

[22] Soseki no inzei cho , 11. See the tables on pp. 6-17.

[23] Ibid., 24-25. Matsuoka reports that Shun'yodo eased the terms in later years, reducing the number of first-printing copies and thereby increasing the chances for further printings and higher royalty shares.

[24] Bungo no sugao , 148-49.

[25] Matsuoka, Soseki no inzei cho , 26.

It was not until the beginning of Showa and the advent of the inexpensive enpon ("one-yen book") that junbungaku writers gained some degree of affluence. The appearance of these cheap anthologies, each featuring one or more writers and costing as much as a taxi ride (entaku ), greatly increased sales, readership, and writers' revenues. Kaizosha (the publisher of Kaizo ) launched the enpon phenomenon with Gendai Nihon bungaku zenshu (1926-29), a thirty-eight-volume collection of modern Japanese literature that sold 380,000 copies, an astronomical figure for the times. Twenty-five more volumes were soon added to the set. Shun'yodo, a rival publishing house, immediately followed suit with the fifty-volume Meiji Taisho bungaku zenshu (1927-32), which sold 100,000 copies. Shinchosha extended the enpon frontier to world literature, and its thirty-eight-volume Sekai bungaku zenshu (1927-30) sold a half-million copies. In all, some two hundred collections were published in the early years of Showa alone. Enpon -based revenues may have been the single most important factor in dissipating the elitist, guild consciousness of the bundan writer and in integrating him into commercial society. Having become a much more broadly based and businesslike institution, bundan no longer had quite the same connotations in Showa Japan.^[26]

Despite the growing number of books, the shishosetsu author wrote few full-length stories, and his books were usually collections of previously published short stories. Although writers in the west have generally made their reputation on book-length fiction, the serious writer's career in Japan, especially in Taisho Japan, depended absolutely on publication of short fiction in the major literary magazines.^[27] The typical junbungaku writer was a short-story writer: Kasai, of course, and Shuko (although the latter's stories were frequently of novella length), as well as more famous contemporaries like Shiga and Akutagawa Ryunosuke. Even a long-winded writer like Tokuda Shusei had earned a solid reputation as a short-story writer by mid-Taisho. "In our age," one critic laments

[26] Suzuki Haruo, "Enpon to bungaku zenshu." The sales figures are derived from Senuma, Hon no hyaku nen shi , 171-79; and Yokozeki, Watakushi no zakki cho , 6.

[27] See Nakamura Mitsuo, "Bungaku zasshi to zasshi bungaku," in Nakamura Mitsuo zenshu 7:366-67. Masamune Hakucho states flatly that he could not have written fiction had the magazine format not been available to him. "Watakushi shosetsu no miryoku," in Masamune Hakucho zenshu 7:378.

in a roundtable discussion (Shincho , March 1925), "most bundan- oriented criticism, since it appears in the monthly literary magazines, passes over full-length stories in complete silence."^[28] Kasai Zenzo's condescending description in Funosha (1919) of long fiction well illustrates the bundan's belief that extended prose was by definition flaccid, unwieldy, and inartistic:

I'm going to do my darnedest to be more productive. And I'm going to try my hand at long fiction, too. There is nothing wrong with this homely, horse-faced creature. It is no freak. Its features are all in the right place; they simply aren't arranged in the most becoming manner. And if people let you get away with writing it, I don't see why I too shouldn't give it a try.^[29]

Long stories or short, the editors of the major literary magazines did not sit idly and wait for manuscripts to trickle in. They actively solicited material from writers, with the result that they devoted more of their energies to enforcing deadlines than to providing stylistic guidelines.^[30] In addition to offering advances on fees, editors often sent writers to a hot spring or mountain retreat in the hope that an out-of-the-way setting would be more conducive to creativity than the bustling city. This strategy frequently backfired, however, especially when tried on the likes of Kasai Zenzo. In an absorbing account of numerous Taisho-period authors, Yokozeki Aizo tells how he sent Kasai to a spa in the mountains of Shinshu in May 1919 to write a story (Funosha ) for his magazine. Kasai, however, proceeded to spend not only his advance but also his entire manuscript fee and more on drink and geisha, forcing the inn to hold him hostage until Kume Masao arrived with additional funds in the form of a second advance from another publishing house. Both writers then proceeded to spend the "ransom" money

[28] The critic is Kano Sakujiro (1885-1941), and the discussion can be found in KZZ, bekkan : 432.

[29] KZZ , 1:331.

[30] Yokozeki Aizo, Kaizo's first editor, recalls that the magazine could never get enough from certain writers and was constantly pressing them for manuscripts of any sort to satisfy a devoted readership. The most sought-after junbungaku writers in the magazine's early days (which is to say the early 1920s) were Akutagawa Ryunosuke, Mushanokoji Saneatsu, and two of the writers under study here, Shiga Naoya and Kasai Zenzo. (This information was gleaned from a conversation held on 20 February 1985 with the Kasai scholar Omori Sumio, who interviewed Yokozeki before his death.)

on more drink and geisha and had to be bailed out by yet another messenger. Kasai did not return to Tokyo until September. The example may be extreme, but it was apparently not unusual.^[31] One of Kasai's stories ("Furo," 1921) describes a similar situation in which the writer travels to a seaside resort on borrowed money only to spend the entire sum and more before his creative juices begin to flow.

When an author had become fairly well established, it mattered little what he wrote; and thanks to the short-fiction format, he could satisfy a schedule-conscious editor with an open-ended story or a string of impressionistic musings, the full implications of which might be accessible only to those readers who, like members of a secret society, were already quite familiar with the author's previous work. Soma Taizo (1885-1952), for example, a member of the Kiseki group along with Kasai Zenzo, Hirotsu Kazuo (1891-1968), and Tanizaki Seiji (1890-1971), could write an "open letter" to a penniless, feckless friend who (he claimed) constantly grubbed money from him, knowing that his readers would recognize Kasai as the story's model, although the content is not so specific as to be obvious to the uninitiated.^[32] Pressed by editors and the ever-present deadline, the Taisho-period junbungaku writer turned regularly to his own life for material, as it seemed the most readily accessible, with the knowledge that such material would pique his readers' interest—and elicit even more solicitations (and deadlines) from his editors. This journalistic pressure frequently induced him to make "confessions" he might not otherwise have penned and sometimes to commit bizarre acts that might not otherwise have occurred to him, in his desperate search for new material.

Once an author had fashioned a particular persona, however, he could alter it only at the risk of appearing to act out of character and perhaps offending or even losing his audience. In order to make his writing "true to life," then, he had to continue acting out the role dictated by this literary self-image. Tail began wagging

[31] Omoide no sakkatachi , 83-87.

[32] Omori Sumio, "Kazai Zenzo," 156. The story is "Rinjin" (1919), first published in Bunsho kurabu . It is rather easy to tell that the model is Kasai in such stories as "Ashizumo" (1929) and "Shichigatsu niju-ni nichi no yoru" (1932) by Kamura Isota (1897-1933), although the character is identified only by initials.

dog: persona now on occasion shaped the person in real life. Some writers succumbed to the illusion that they could solve their personal difficulties on the pages of their shishosetsu , and not a few came to rely on the form for its practical function as an open forum. But here always lay the seeds of destruction. Shishosetsu writers often discovered to their dismay that the solutions they had worked out on paper came back to haunt them in real life.

Few writers could bear such self-abuse; most turned to other projects, which gave them more narrative freedom or at the very least did not jeopardize their spiritual and even physical well-being. Some diversified in order to survive: Uno Koji, Kikuchi Kan, and Kume Masao, for example, ventured early and successfully into the realm of popular fiction. Yet there were others, Shimazaki Toson and Dazai Osamu (1909-48) among the most prominent, whose failures to keep their lives even one step ahead of their writing resulted in domestic tragedy and in death. These writers succumbed to the temptation inherent in a form bound intimately with private life: that of closing the temporal gap between experience and writing.

None of the three writers under study here went to quite such extremes, but neither were they particularly successful at diversifying their writing. Chikamatsu Shuko dwelt incessantly on a very few episodes in his life and found only late in his career a diversion in historical fiction. Shiga Naoya, who did not depend on the income from his writing for a living, simply stopped writing altogether for several long stretches in his career. And Kasai Zenzo, who was by far the most dependent, both financially and psychologically, on the form, perhaps luckily had his career cut short by consumption. All three nevertheless succeeded in coming to grips with the singular treachery of writing about oneself and at their best made their lives speak with extraordinary power.

The following chapters examine how these three writers coped with the contemporary journalistic realities we have discussed, in their struggle to transform experience into art. That struggle was rendered all the more complex by a growing awareness of their medium's treacherous nature. Their success depended largely on how adeptly they extricated themselves from the quicksand of referentiality, how shrewdly they exploited the myth of sincerity, and finally how well they learned the lesson of textuality: that the pro-

duction of literary art (including shishosetsu ), its grounding in "real life" notwithstanding, entails a mediative process that challenges the one bastion of "truth" that seems inviolable—lived experience. Such realizations rarely came easily and often came under duress. We shall look for flickers—sometimes flashes—of awareness in the seams, the folds, the fissures in their texts.