57 Ibid., 11.
58 It bears emphasizing that, circa 1960, the development of predictive text tray beds had clearly reached a sufficient degree of commonality and sophistication that Wang and Lin were able to present not merely one technique, but five.
59 Wang Guihua and Lin Gensheng, Chinese Typing Technology, 8.
60 Chinese typewriter formerly employed at the United Nations (Geneva), Double Pigeon style, manufactured in 1972 by the Shanghai Calculator and Typewriter Factory (Shanghai jisuanji daziji chang), housed at the Musée de la Machine à Écrire, Lausanne, Switzerland; Chinese typewriter formerly employed at UNESCO (Paris), Double Pigeon style, manufactured in 1971 by the Shanghai Calculator and Typewriter Factory (Shanghai jisuanji daziji chang), housed at the Musée de la Machine à Écrire, Lausanne, Switzerland.
61 On “tinkering,” see Adele E. Clarke and Joan Fujimura, “What Tools? Which Jobs? Why Right?” in The Right Tools for the Job: At Work in Twentieth Century Life Sciences, ed. Adele E. Clarke and Joan Fujimira (Princeton: Princeton University Press, 1992), 7. For “reconfigurations,” see Lucy Suchman, Human-Machine Reconfigurations: Plans and Situated Actions (Cambridge: Cambridge University Press, 2006).
62 Chinese typists in the post-1949 period pushed their experiments in vernacular taxonomy to their very extremes, going so far as to rearrange punctuation marks and numerals into predictive clusters. In the case of punctuation, rather than grouping the comma, period, question mark, semicolon, etc., together, as on Chinese typewriter tray beds in the first half of the twentieth century, both the UNESCO and UN typists located them adjacent to those characters with which they either frequently or always appeared. The case of the question mark (?) is the most illustrative. Insofar as Chinese expresses interrogatives through the use of a limited set of particles, which themselves are normally followed by a question mark, the UNESCO typist decided to place these particles and the question mark together. Specifically, the question mark on the machine is flanked by ma (吗), the most common interrogative particle, used to transform a statement into a question without changing word order; ba (吧), used at the end of an utterance to indicate a “How about x?” suggestion; and ne (呢), used at the end of a sentence to indicate a rhetorical question, a suggestion, and certain other contextually specific meanings. As with the example of “Mao Zedong” earlier, moreover, the differences between the UNESCO and UN machines are equally revealing. Even as these typists shared certain taxonomic instincts, these instincts manifested in significantly different tray bed configurations. Song Mei Lee-Wong, “Coherence, Focus and Structure: The Role of Discourse Particle ne,” Pragmatics 11, no. 2 (2001): 139–153.
63 Geoffrey Bowker and Susan Leigh Star, Sorting Things Out: Classification and Its Consequences (Cambridge, MA: MIT Press, 1999). Note that a heat map is a data visualization wherein data values are represented as color values within a matrix.
64 Geoffrey C. Bowker, Memory Practices in the Sciences (Cambridge, MA: MIT Press, 2005).
65 Chinese Typewriter Tray Bed Comprehensive Character Arrangement Reference Table (Zhongwen daziji zipan zi zonghe pailie cankaobiao) [中文打字机字盘字综合排列参考表], appendix of Zhu Shirong [朱世荣], ed., Manual for Chinese Typists (Zhongwen daziyuan shouce) [中文打字员手册] (Chongqing: Chongqing chubanshe, 1988).
66 Natural-language tray beds became so popular at the time, in fact, that the method traveled all the way to the offices of Cambridge University. During my visit in July 2013, Charles Aylmer, head of the Chinese Department at the Cambridge University Library, kindly showed me the natural-language Chinese typewriter tray bed he himself had designed decades earlier.
Conclusion: Toward a History of Chinese Computing and the Age of Input
The forthcoming sequel to The Chinese Typewriter will be the first history in any language of Chinese computing, tracing this history from its inception in the immediate postwar period to its efflorescence in a burgeoning network of Chinese, Taiwanese, and Japanese computer scientists from the 1970s onward. The pathway of this history will take readers through subjects as diverse as machine translation, computer graphics, the rise of computer programming, the software revolution, the feminization of Chinese intellectual labor, and the growth of personal computing. At the center of the story will be a cast of eccentric personalities, drawn from the ranks at IBM, RCA, MIT, the CIA, the US Air Force, the US Army, the Pentagon, the RAND Corporation, the British telecommunications giant Cable and Wireless, Silicon Valley, the Graphic Arts Research Foundation, the Taiwanese military, the Soviet military, Japanese industrial circles, and the highest rungs of mainland Chinese intellectual, industrial, and military establishments.
As we will examine, continued experiments with mechanical Chinese typewriting became the doorway into new domains of information technology more properly called word processing and early computing. Using a variety of input devices—custom-made keyboards, QWERTY keyboards, pressure-sensitive surfaces, and even early pen tablets—inventors and companies from China, Taiwan, the United States, Western Europe, and the Soviet bloc engaged in ever more sophisticated explorations into Chinese technolinguistic modernity. With the dawn of Chinese computing, the already porous borders between common usage, combinatorialism, and surrogacy eroded completely, with the strategies and materialities of these once relatively distinct modes rushing together, swirling and eddying into new technolinguistic configurations. What is more, the border that once separated inscription from retrieval—the border that Lin Yutang first crossed in his work on MingKwai—disappeared entirely. Approaches to Chinese typewriting, typesetting, dictionary organization, telegraphy, and other domains and practices were brought into single devices in various combinations.
Most importantly, it will provide readers with the first-ever examination of input, a new technolinguistic condition that is quietly transforming the relationship between humans, machines, and language—not only in China but worldwide. Input constitutes the core technolinguistic condition of a new era in the history of Chinese, a condition that no longer adheres to the assumption that took hold more than a century ago as part of the collapsing imaginary of the alphabetic world: the all-pervasive fiction of Tap-Key and his monstrous Chinese keyboard. This fiction told the world that for every symbol there is a key, and by consequence, that alphabetic scripts were blessed with a kind of technolinguistic efficiency and immediacy that character-based Chinese writing could never hope to achieve.
If the appearance of Tap-Key marked the formation of this fiction, MingKwai and input marked its annihilation, rescuing Chinese script—and perhaps script itself—from the powerful myth that some writing systems are more immediate than others, and deeper still, that any script is immediate, ever. And so, if our story began in a time when the very notion of a “Chinese keyboard” was an oxymoron, our story now ventures into a time when the keyboard is ubiquitous in China, and yet when typing itself no longer exists—a time when QWERTY is everywhere in China, and yet when the keyboard as we know it is dead. For the keyboard and Chinese script alike, in order for everything to stay the same, everything needed to change.
The rise of input was by no means inevitable, nor is its history one easily celebrated. Input took shape as the result of a 150-year, sleep-deprived, torment-ridden history in which human beings functioning within the Chinese-character information environment were never permitted to drift off into the comfortable dream of immediacy like their counterparts in the alphabetic world. While the alphabetic world fell deeper and deeper into the comforting depths of Q-W-E-R-T-Y, A-Z-E-R-T-Y, and so forth, the world of Chinese characters was pierced day and night by near-constant alarm bells: Morse code, braille, typewriting, Linotype, Monotype, punched-card memory, text encoding, dot matrix, word processing, personal computing, optical character recognition, and the alphabetic order of the Olympic Parade of Nations, each expelling Chinese characters yet again from the bounds of the “universal.” Each expulsion, in turn, buoyed up to the surfa
ce of consciousness fundamental truths about the inherent arbitrariness of Chinese script—of all script—in a recapitulation of that most basic of principles within Semiotics 101: that there is no inherent, invariant, or natural relationship between the signifiers we use and those concepts we wish to signify. At the aggregate level, this longstanding condition of technolinguistic sleep deprivation translated into more than a century in which a transnational cast of engineers, linguists, telegraphers, education reformers, phone book compilers, library scientists, typists, and more were left with little choice but to exploit the vast spaces of possibility that open up at precisely the moment when one gives up on the myth of immediacy—when one accepts as one’s starting point a condition of nonidentity between keys and screens—and then begins to play.1
The rise of input is also not easy to grasp, particularly for those of us whose entire life has been conditioned by the what-you-type-is-what-you-get framework. As we venture into the next volume, three analogies help us distinguish between input and type: stenography, telecommunications, and MIDI. On a stenotype machine, such as those used within court stenography and elsewhere, only a fraction of the letters of the Latin alphabet is present. To produce letters that are absent from the machine—such as b, d, f, g, and many others—the typist must use the letters that are present to represent or stand in for them: the letter “f,” for example, is not found on the keyboard, but must be symbolized by simultaneously typing two of the letters that are: in this case, “t” and “p.” Meanwhile, the letter “b” is produced by simultaneously striking the “chord” composed of “p” and “w.” When the typist reads back over the transcript, and sees the letter “p” by itself, he or she knows that—in this case—“p” equals “p.” But when “p” appears alongside “w,” the stenotypist knows that—in this case—these letters are being used to represent another letter, rather than themselves. By reviewing these particular letter chords, he or she knows which letters are being indicated and can thereupon translate this coded primary transcript into a legible secondary transcript.
In one sense, computer-age China is a country of stenotypists. The symbols one desires to see on the page or screen are not present on the keyboard in any naked-eye sense; instead, everything that is typed upon the “primary transcript” is but a temporary and disposable set of instructions that subsequently need to be translated into a secondary, “plaintext” transcript in accordance with a set of protocols. In the context of input, it is this “secondary transcript”—the characters that appear on the page—that is endowed with preeminent value, while the original transcript—the keystrokes that are intercepted by the input method editor (IME)—is discarded immediately after the translation is complete, without ever being viewed by human eyes.
Telecommunication provides a second analogy. In China, all text input—even text input that takes place within ostensibly “non-transmitting” programs like Microsoft Word—is in fact a form of communication at a distance with oneself—or auto-telecommunication. Even though this human-computer interaction seems to be taking place entirely “locally,” confined to one person’s relationship with his or her machine, nevertheless the model at play is classically telecommunicative: the operator is not sending telegrams to another party, of course, nor signaling from ship to shore, but is sending out coded transmissions to the IME, which are then translated and retransmitted back to the operator in the form of Chinese “plaintext.” Taking place along the way is a process of retrieval: Chinese script is being summoned from elsewhere. In this way, Chinese input is a form of retrieval-composition, rather than the inscription-composition of typing.
The third and perhaps most evocative analogy is drawn from the domain of electronic music. With the advent of MIDI, or the Musical Instrument Digital Interface, entirely new modes of performing and composing musical scores became possible in the second half of the twentieth century. While patterned after and often indistinguishable from conventional instruments, MIDI pianos, guitars, drum pads, and woodwinds, among many others, are in fact instrument-agnostic controllers with which a performer can use the technosomatic format of one device (e.g., piano playing) to perform another (e.g., a violin). One can use a piano-shaped MIDI controller to play cello, a woodwind-shaped device to play the drums, a guitar-shaped MIDI device to play piano, and so on. The relation between a MIDI controller and its output is so plastic, in fact, that the controller need not resemble any conventional instrument at all. One could just as readily manipulate sound—any sound—through actuators embedded in the fabric of one’s clothing (as in choreographer Gerry Girouard’s Songs for the Body Electric), in a banana (as in the Sonic Banana project), or even the structure of an entire building (as in David Byrne’s Playing the Building installation piece mounted in Stockholm, New York, London, and Minneapolis).2
Whatever analogy we choose, one fact remains clear: in China, the QWERTY keyboard and typing as we know them are dead, and have been reborn in the service of something altogether different: input. In the Chinese context, the QWERTY keyboard was long ago transformed into a “smart” peripheral, in contemporary parlance, that derives speed, power, and accuracy in direct proportion to the increasing processing power, algorithmic sophistication, and memory richness of consumer-grade computers, tablets, smart phones, and more. Meanwhile, QWERTY in the alphabetic world has remained largely unchanged since the era of typewriting since, as we recall, “computer designers were happy not to reinvent text-based input and output.”
Chinese input is becoming more sophisticated with each passing year, moreover, with the steady advancement of predictive text, autocompletion, and most recently, a Wi-Fi-augmented input framework known as “cloud input” (yun shuru). In cloud input, IMEs use the cloud to provide smarter and smarter “suggestions” to users by comparing their QWERTY keystrokes against those inputted by Chinese computer users elsewhere in the network. Not unlike autosuggestions encountered in the Google search bar, Chinese cloud input has one all-important difference: this Wi-Fi-augmented process is not limited to the web, but is becoming a core part of all text input—even input that takes place in ostensibly “local” or “non-transmitting” programs. Whether one is searching Baidu or composing a Word document, one’s keystrokes are being intercepted by and ingested into third-party cloud servers, with Chinese character suggestions being retransmitted back. As Chinese input moves further in the direction of the cloud, then, there is increasingly no such thing as “entirely local” text, contained exclusively on one’s own computer or device.3
As we venture more deeply into the history and practice of Chinese input, however, it bears noting that our imaginations will continue to be stalked by the imagined Chinese monstrosity of Tap-Key. Long after the Chinese typewriter will have been displaced by Chinese computing and word processing, visions of immense, antimodern machines will continue to make regular appearances. “The Chinese typewriter is a longstanding joke in the West,” a 1973 piece for the UK-based Times will read, “where it is almost synonymous with the paradoxical or impossible.” Likening the machine’s tray bed to an immense, strange, lunar surface (when, in fact, it measures only eighteen inches by nine inches in “terrestrial” dimensions), it will describe the process of Chinese typing as an “operation similar to landing on the Moon.”4 Another British journalist, this one writing in 1978, will describe the Chinese typewriter as a “cumbersome form of miniature dive-bombing, on which a really proficient typist can achieve only ten characters a minute.”5 Another Chinese typewriter will appear at the Museum of the History of Science in Oxford in a special exhibit tellingly entitled “Eccentricity: Unexpected Objects and Irregular Behavior.”6 Still another will form part of the permanent exhibit in Sweden at the Tekniska Museet.7 Here in Stockholm, the machine will seem at first to occupy a privileged position, stationed in a display case at the very beginning of an exhibition dedicated to printing and other forms of writing technology. The accompanying description, however, will reveal an altogether different motive
:
Hieroglyphics are picture characters, pictograms. … Today’s traffic signs are a type of pictogram. The Chinese alphabet is also pictographic and contains tens of thousands of characters. … As opposed to pictographic script, alphabetic script has very few characters, or “letters.” It is much easier to develop printing techniques if you use an alphabetic writing system.8
Even The Simpsons will enter the fray in 2001. In his new job writing copy for a fortune cookie manufacturer, Homer Simpson will be shown extemporizing terse jewels of wisdom to his daughter, as she takes dictation on a Chinese typewriter. “You will invent a humorous toilet lid”; “You will find true love on Flag Day”; “Your store is being robbed, Apu.” He will pause for a moment to confirm that she is keeping up. “Are you getting all this, Lisa?” The frame will switch to Lisa, poised tentatively in front of yet another absurdly complex machine, pressing buttons cautiously and with great hesitation. In elongated syllables, she will respond: “I don’t knowwww.”9
The monstrous Chinese typewriter will not only persist, moreover—it will have reinforcements in the form of monstrous, imaginary Chinese computers. As exemplified in an online Q&A exchange from 1995, many of the same tropes of Chinese typewriting will soon migrate unconsciously from one arena of IT to another:
Dear Cecil, How on earth can the Chinese and Japanese use computers, given that their writing uses thousands of different characters? The keyboard must look like something off a Wurlitzer pipe organ.
Dear Nora, Nah, it looks pretty much like any keyboard, and using it is a piece of cake. All you have to do is adhere closely to the following six hundred steps. You might want to pack a lunch.10
The Chinese Typewriter Page 44
