The Chinese Typewriter

by Thomas Mullaney

  China’s Second Vernacularization Campaign

  The origins of this diffuse, decentralized, and grassroots movement are impossible to identify with certainty, and yet available evidence helps us trace out its contours with some confidence. The earliest and clearest example is found in the activities, not of a typist, but of a typesetter named Zhang Jiying, who became known to readers of the People’s Daily in 1951 in an article entitled “Kaifeng Typesetter Zhang Jiying Diligently Improves Typesetting Method, Establishes New Record of 3,000-plus Characters per Hour.”14 Having worked as a typesetter for over a decade, first in the city of Zhengzhou and later in Kaifeng, Zhang was trained on both the older “24-tray character rack” and the newer “18-tray character rack,” and posted very respectable typesetting speeds (ranging from 1,200 to 2,200 characters per hour over the course of his career thus far).15 Only a few short months after the formation of the PRC, however, he experienced a reported surge of inspiration and began to engage in a sweeping, experimental reorganization of his character rack—an experiment that culminated in his 1951 feat.

  What caught Zhang’s attention in particular was his colleagues’ practice of pairing characters together on the character rack that they employed frequently in the course of daily work. Three characters in particular had been clustered together, in clear violation of radical-stroke organization: xin (新), hua (華), and she (社), which together form the name “New China Press” or Xinhuashe. “I thought, if one were to put a group of related characters together,” Zhang would later explain, “it would definitely be good for setting.”16

  Zhang set out to apply this principle across his entire character rack. His character rack would soon feature over 280 two-character compounds, eight three-character sequences, and even seven four-character sequences, a style of organization he termed lianchuan—meaning “series” or “chain.”17 He included terms and names such as “revolution” (ge­ming), “American imperialist” (Meidi), “liberation army” (jiefangjun), “agriculture” (nongye), and many others employed in Communist parlance. These were cliché, in the original French meaning of the term, signifying a printer’s “stereotype block” upon which is etched a commonly used phrase, rather than just a letter.18 Derived from the past participle of the verb clicher, or to click, the term was connected to the sound made by a printing sort being set in place. Over time, cliché drifted semantically to its present-day meaning of a “trite, worn-out expression.”

  Zhang extended this organizational scheme to practically the entire character rack, not merely a small, specialized region thereof as his colleagues had done. Zhang’s clichés were not consistent, moreover, but transformed depending upon the properties of the text under production and the overarching political environment.19 If “materials on the workers’ movement” constituted the operative theme of one period, Zhang explained, he prepared such compounds as “production” (shengchan), “experience” (jingyan), “labor” (laodong), and “record” (jilu). At other times, the demands of a specific propaganda campaign might dominate media attention, prompting Zhang to rearrange his character rack anew, prioritizing terms and phrases such as “Resist America, Aid Korea” (kang Mei yuan Chao) (the Korean War–era mass mobilization campaign).20 In this way, Zhang set about transforming his body and his character rack into Chinese Communist Party (CCP) rhetoric incarnate, not in the sense of parroting certain key terms, but in the sense that his fingers, hands, wrists, elbows, eyes, peripheral vision, joints, movements, anticipatory reflexes—every part of his body—would be intimately attuned and maximally sensitized to the distinct cadences of CCP rhetoric.

  Zhang pushed his new character arrangement system further, surpassing his own record: a record 4,778 characters set in one hour, or nearly eighty characters per minute, captured on film on July 29, 1952, by the South Central Film Team of the Central Film Production Company.21 In the meantime, Party-state authorities saw in Zhang’s accomplishment an opportunity to celebrate the kind of proletarian parable so vital at the time: the model worker who had used personal initiative and free time to push his industry beyond what others had imagined possible, overturning “tradition” broadly writ, and in doing so demonstrating to the masses both the possibilities of ingenuity and the impermissibility of complacency.22 It was Zhang’s departure from conventional practice—his heteropraxy—that helped him serve orthodoxy far more effectively than rote, conventional practice—or orthopraxy—ever could. “One person setting three or four thousand characters [an hour] doesn’t amount to much,” Zhang reflected. “But if everyone were to set three or four thousand characters, now that’d be something.”23 In quick succession, the Party extended an invitation to the typesetter to take part in the May Day celebration of 1952, helped him to co-author a book explaining his method in greater depth, sponsored his tour of publishing houses across the country, admitted him to the Party, and encouraged others to study and perhaps emulate his method (figure 7.4).24

  7.4 Zhang Jiying

  The case of Zhang Jiying was soon followed by others. In 1952, Commercial Press in Shanghai undertook character tray reform, implementing the same lianchuanzi system and witnessing a clear increase in speed as a result.25 The Jinggangshan Newspaper Printing House implemented the lianchuanzi arrangement as well, as part of their shift away from the industry standard twenty-four-part character tray to the so-called “‘eight’-character-style” character rack (ba zi shi). And just as the city of Kaifeng had its model typesetter in Zhang Jiying, Jinggangshan could take pride in their own: local typesetter Wang Xinshun, who set an all-province hourly record on April 10, 1958, with 3,840 characters set. Wang broke his own record later that same year, setting 4,100.26

  By 1958, lianchuanzi had become widespread enough to be spotlighted in a manual on page layout and typesetting published by the journalism research institute at People’s University.27 Here the method was referred to as the “Connected Language Tray Bed” (lianyu zipan) or the “Connected Character Tray Bed” (lianchuanzi zipan), and was explained as follows:

  As much as possible, you want to place compounds together that are used together and are related—for example, by placing the four characters “jie” (解), “jue” (决), “wen” (问), “ti” (题) in four adjacent columns.28 Another example would be putting characters together like “jian” (建), “she” (设), “zu” (祖), “guo” (国), and “ti” (提), “gao” (高), “chan” (产), “liang” (量), which would make setting up these characters a lot more convenient.29 We could even consider using radiating style (fangshexing) or chainlink style (liansuoxing) to set up a common usage tray bed, for example “ren” (人), “min” (民), “gong” (公), “she” (社) —> “hui” (会), “zhu” (主), “yi” (义).30 Radiating it out like this would resemble the word game “thimble linking” (dingzhen xuxian).31

  A similar technique, the manual explained, was to pair linked words and place them in thematic zones. In one zone, the typesetter might arrange clusters such as “American imperialism” (Meidi), “to invade” (qinlüe), and “to destroy” (pohuai), calling this the “negative connotation terms tray” (bianyi zipan).32 Another region could then be designated the “Social Structure Terms Tray” (shehui zuzhi mingcheng de pan), featuring, as the manual explained, “‘Socialism,’ ‘cooperative,’ ‘Chairman Mao,’ etc., etc.” (“shehui zhuyi,” “hezuoshe,” “Mao zhuxi” dengdeng).33

  To find “Chairman Mao” and “socialism” tucked inside quotation marks, followed by the term dengdeng (“so on and so forth” or “et cetera”), is deeply revealing, alerting us to the metacognitive distance that was central to this organizational method. In order to make use of natural-language arrangements, one could not be an unwitting parroter of routine political expressions, spouting Mao-era phraseology unreflectively. To the contrary, one had to be acutely aware of their roteness and regularity—to think of them precisely as cliché, in order to be able to do an effective job of preparing one’s character rack or tray bed for maximum efficien
cy. The success of Zhang’s method was a function of maximum sensitization to and anticipation of the distinct cadences of CCP rhetoric, not just with obvious cases such as “Mao Zedong” and “cadre” (ganbu), but also politically charged “neutral” terms such as “education” (jiaoyu), “to exist” (cunzai), “according to” (genju), and others. It was Zhang’s ability to establish a deeply private, individualistic, even esoteric relationship with the public, authoritative, and standardized language of the era that increased his capacity to serve political orthodoxy. Radical individualism was, in this context, completely compatible with and conducive to state power.

  Chinese Typewriting and “Mass Science”

  As illustrated in the case of “New China Press,” Zhang Jiying was not the first to conceptualize a vernacular taxonomic approach to Chinese typesetting. He was, however, perhaps the first to pursue this possibility to its logical extremes, extending the lianchuanzi system across the entirety of the character rack, rather than a limited region thereof. The same distinction held true for Chinese typewriting, in which—as we recall from chapters 3 and 4—character slugs in the Chinese typewriter tray bed were not fixed, but rather removable and replaceable. Although the application of natural-language clusters achieved unprecedented stature and levels of experimentation beginning in the 1950s, in fact there were already limited “predictive” elements to the Chinese typewriters we examined from the Republican period. Two in particular merit consideration. First, we recall there existed one region on Republican-era typewriter tray beds reserved for “special usage characters” (teyong wenzi), in which characters were not organized according to radical classes or stroke count. Instead, within this narrow strip on the tray bed, measuring four columns wide and thirty-four rows deep, characters were clustered so as to form common two-character compounds and multicharacter sequences such as “the Republic of China” (Zhonghua minguo).34 In particular, typists dedicated this small region to the characters that formed the names of Chinese provinces. For example, the characters meng (蒙) and gu (古) were located horizontally adjacent to one another, forming the name Menggu or “Mongolia.” This arrangement became more complex and interesting in the case of characters that appeared in more than one toponym, such as jiang (江), hu (湖), nan (南), xi (西), dong (東), and shan (山) (figure 7.5). To the left of jiang, for example, was the character zhe (浙), forming the name Zhejiang; to its upper-left corner, the character su (蘇), forming Jiangsu; to its lower left was long (龍), beneath which was located a third character hei (黑)—together forming the province name Heilongjiang.35

  7.5 Sample from “Special Character Region” of Huawen typewriter (pre-1928); from Chinese Typewriter Character Arrangement Table (Huawen daziji wenzi pailie biao) [華文打字機文字排列表], character table included with Teaching Materials for the Chinese Typewriter (Huawen dazi jiangyi) [華文打字講義], n.p., n.d. (produced pre-1928, circa 1917)

  The second example comes in a letter dated August 26, 1928, from newspaper reporter, editor, and language reformer Chen Guangyao to Wang Yunwu, editor in chief at Commercial Press and inventor of the “Four-Corner” retrieval system we saw in the previous chapter. Knowing of the company’s Chinese typewriter division, Chen suggested to Wang the possibility of reorganizing the company’s Chinese typewriter tray bed so as to maximize the adjacency of naturally co-occurring Chinese characters. By way of example, Chen cited the two- and more-character terms “then” (ranze 然則), “China” (Zhongguo 中國), “to invent/invention” (faming 發明), “Three People’s Principles” (Sanmin zhuyi 三民主義), “world” (shijie 世界), and even longer idiomatic expressions like “to be unexpected” (yiliao zhiwai 意料之外).36 When ordering the characters in this way, he explained, a compression of only a few characters would make possible many more compounds. Even a simple four-character sequence 發明達光 made it possible to write such common terms as “to invent/invention” (faming 發明), “to develop” (fada 發達), and “bright” (guangming 光明), all with much greater speed and ease than when using radical-stroke arrangement. Moreover, common particles could be organized more scientifically by placing them, not in accordance with dictionary order, but in proximity to those characters alongside which they normally appeared.

  For all his encouragement, however, Chen Guangyao’s experimental proposal to Commercial Press was never adopted by the company, and never took hold in other areas of Chinese typewriting. Indeed, within the rich archives on Republican-era typewriting, there is not a single further reference to the application of his system, whether in theory or in practice. To the contrary, all evidence suggests that, while elements of natural-language arrangement continued to be used on Republican-era machines, they remained entirely localized to the small “special usage” region of the typewriter.

  The early PRC period marked a different era altogether. In the wake of the Zhang Jiying mini-phenomenon, natural-language organization was quickly appropriated from typesetting and applied to the field of Chinese typewriting. In November 1953, readers of the People’s Daily encountered Shen Yunfen, a young woman who had joined the People’s Liberation Army two years prior, at the age of seventeen.37 A native of Shanghai, she was appointed to the North China Military Region Headquarters to serve as a typist in October 1951 during the Korean War. By her own account, her performance was slow at the outset, which led her to experience despondence—even to the point of losing weight from the anxieties involved. Under the guidance of a colleague, Shen reported increasing her speed to 2,113 characters per hour—a respectable improvement, but one that left her dissatisfied.

  Shen decided to pursue the “New Typing Method” (Xin dazi caozuofa), a method she had learned of recently, attributed to one Wang Jialong.38 Applying the same principle of adjacency as Zhang Jiying’s “serial” method, Wang and soon others exploited the shape of the typewriter tray bed to extend Zhang’s linear, one-dimensional organization into a two-dimensional, x-y matrix. The operative principle of the New Typing Method was termed “radiating compounds” (fuci fangshe tuan), explained as follows: by “selecting one character as the core and then radiating outward from it,” the typist could populate the three to eight spaces around each character with as many related characters as possible.39 Owing to this multidimensionality, the typist could push beyond the left-right sequencing of related characters on the compositor’s character rack, and begin to experiment with both vertical and diagonal arrangements. It also made possible the stringing together of these mini-regions into widening associative networks.

  In moving away from radical-stroke organization, a space of practically infinite possibility was opened up. Shen Yunfen’s speeds steadily increased, reaching 3,012 characters per hour and then leading her to a first-place finish in the North China Military Region Typing Competition in 1953 with a record-setting 3,337 characters in one hour. On January 25, 1953, the young Shen was granted the title of “First-Class Hero” (yideng gongcheng) and “second-level model worker” (erji mofan). In September 1955, she was received by Mao Zedong at the National Conference of Youth Activists in Socialist Construction.40

  An essential question emerges out of the stark contrast between Chen Guangyao’s failure in the Republican period and the vibrant experimentation of the 1950s: why was it only during the Communist period that this already known technolinguistic technique—“new,” “China,” “press”—became the object of intense focus and exploration? How did the logic of xinhuashe, a region accounting for only a fraction of the overall character rack, become the organizational principle for the entire character rack? Likewise, how did the logic governing a thin strip on Republican-era tray beds, a “special” region that accounted for not even 6 percent of the total lexical space, eventually conquer the surrounding 94 percent? One potential answer that can be disqualified immediately is any suggestion that the Republican period was in some sense less innovative, experimental, or anti-traditional with regard to language reform overall. To the contrary, as we saw
in the preceding chapter, the late Qing and Republican periods witnessed a practically uninterrupted exploration of alternate character organization and retrieval systems, and scores of reformers heaped criticism upon the organization of the Kangxi Dictionary and its attendant radical-stroke system. Why then was it not until the 1950s that predictive text strategies began to proliferate within typesetting offices and on Chinese typewriter tray beds? How do we account for the sudden rise of natural-language experimentation in the Maoist period?

  To understand the emergence of natural-language experimentation in the early Maoist period, we must consider three key political changes that took place following the Communist revolution of 1949: the Chinese Communist endorsement and celebration of what have been termed “popular” or “mass” knowledges; an increasingly routinized if not “predictable” Chinese Communist rhetoric; and unprecedented time pressure experienced by typists during the immediate postrevolutionary period.41 Just as Communist authorities made and endorsed radical pushes for nonelite participation in fields as diverse as paleontology, medicine, and seismology, so too did they endorse calls for a sweeping, bottom-up reorganization of the Chinese language—a mass taxonomy that would depart from earlier modes of categorizing Chinese characters, and better reflect the way in which “common people” organized their linguistic universes.42 If it was possible to proletarianize medicine and the physical sciences, and to “challenge the notion that science was the province of elites,” why not the systems by which script was organized?43 The second transformation involved the development of a political discourse of unprecedented routinization—a “systemization of ideological categories and language” that came to influence, if not define, entire domains of textual production in the early PRC period.44 This was true not merely for overtly Chinese Communist keywords such as “struggle” (douzheng) and “proletariat” (wuchan jieji), but also for an ever-expanding repertoire of what Franz Schurmann has aptly termed “seemingly conventional words with special significance”—terms such as “opinion” (yijian) and “discussion” (taolun).45 A third factor was the unprecedented employment of Chinese typists in the conduct of government affairs, a topic broached at the beginning of this chapter. By the mid-to-late 1950s, Chinese typists and typewriters could be found throughout China, serving an increasingly regular role in the everyday functioning of government, as well as in the unrelenting series of Mao-era mobilization campaigns—campaigns for which typists were tasked with producing economic reports and low-run mimeographed materials. During the Great Leap Forward, in particular, this increasing pressure on typists developed into a metanarrative all its own, with typists beginning to adopt the conventional language of “quotas” and “output.” Typists and work units competed with one another to outstrip target goals and to increase the efficiency of their “production”—that is, the total number of typewritten characters produced per month and per year.