The Chinese Typewriter

by Thomas Mullaney

  34 Wyckoff, Seamans, and Benedict, The Remington Standard Typewriter, 33–34.

  35 Ibid., 16–17.

  36 Remington Notes 3, no. 10 (1915); Richards, The History and Development of Typewriters, 72. At the Columbia Exposition of 1893, Remington circulated postcards featuring Missipi, or Edna Eagle Feather, named the first Indian to learn shorthand and typewriting. Missipi was of the Osage Nation.

  37 The most celebrated index typewriter is the Mignon, first built in 1904 by the A.E.G. Company in Berlin, with a reported speed of 250 to 300 characters per minute. See Richards, The History and Development of Typewriters, 45.

  38 McFarland Papers, Bancroft Library, University of California, Berkeley, box 3, folder 14.

  39 “Typewriters to Orient: Remington Rand Sends Consignment of 500 in the Mongolian Language,” Wall Street Journal (April 26, 1930), 3.

  40 “Ce n’est donc pas sans fierté, que la Maison Olivetti contribue à leur marche en avant par son apport de machines à écrire.” “La Olivetti au Viet-Nam, au Cambodge et au Laos,” Rivista Olivetti 5 (November 1950): 70–72, 71.

  41 “Le Clavier Arabe,” Rivista Olivetti 2 (July 1948): 26–28, 26.

  42 For a fuller picture of the machines housed at this and other museums, please see the “Machines” section of the Bibliography.

  43 Samuel A. Harrison, “Oriental Type-Writer,” United States Patent no. 977448 (filed December 15, 1909; patented December 6, 1910).

  44 Ibid.

  45 Richard A. Spurgin, “Type Writer,” United States Patent no. 1055679 (filed August 11, 1911; patented March 11, 1913). Remington followed suit again in 1921, when Elbert S. Dodge filed a patent to adjust their machines for “Hebrew and similar languages.” “A specific object of the invention,” Dodge explained, “is to adapt the Remington typewriting machine with very slight constructional changes to reverse the usual direction of feeding movement.” Elbert S. Dodge, “Typewriting Machine,” United States Patent no. 1411238 (filed August 19, 1921; patented March 28, 1922).

  46 Selim S. Haddad, “Types for Type-Writers or Printing-Presses,” United States Patent no. 637109 (filed October 13, 1899; patented November 14, 1899).

  47 Ibid.

  48 Ibid.

  49 Ibid.

  50 A comparable idea is advanced by Vassaf Kadry, assignor to the Underwood Typewriter Company residing in Constantinople. See Vassaf Kadry, “Type Writing Machine,” United States Patent no. 1212880 (filed January 15, 1914; patented January 30, 1917). Assignor to Underwood Typewriter Company. Kadry identifies himself as a subject of the Sultan of Turkey.

  51 Baron Paul Tcherkassov and Robert Erwin Hill, “Type for Type Writing or Printing,” United States Patent no. 714621 (filed November 21, 1900; patented November 25, 1902).

  52 In 1910, Herbert H. Steele of Marcellus, New York, submitted a patent for an Arabic typewriter, assigned to the Monarch Typewriter Company. While bearing this general name—“Arabic Typewriter”—the focus of the patent is far more precise, focused on the carriage mechanism of the Monarch machine and its adjustment for the purposes of Arabic typewriting. “My invention has for its principal object,” Steele explained, “to produce an efficient carriage feed mechanism for use in a typewriting machine designed for writing the Arabic and similar languages wherein the paper carriage is required to have its step-by-step movement from left to right instead of in the other direction and wherein some characters require a greater extent of carriage movement than others.” H.H. Steele, “Arabic Typewriter,” United States Patent no. 1044285 (filed October 24, 1910; patented November 12, 1912). In 1917, Seyed Khalil (1891–1974), a self-described inventor and “freelancer” born in Kashan, Persia, filed for a patent assigned to the Underwood typewriter company and issued in January 1922. Khalil had immigrated to the United States in 1916, and shortly following his twenty-sixth birthday, offered up his vision of Arabic typewriting. He positioned his work somewhat in opposition to that of Tcherkassov and Hill. Their use of a nonsignifying graphic element served its function in reconciling Arabic writing to the typewriter, Khalil argued, but resulted in a method that was too slow and placed too onerous a demand on the typist. Moreover, Khalil argued, this approach to Arabic typewriting led to an erroneous belief “that certain letters of such languages must be distorted to be typewriting.” Each letter on Khalil’s machine had only two forms: a terminal form and a nonterminal form. Khalil saw broader application of his machine as well, looking beyond Arabic to Persian, Hindustani, and Turkish. See World War I Draft Registration Card (United States Selective Service System, World War I Selective Service System Draft Registration Cards, 1917–1918, National Archives and Records Administration, Washington, DC, M1509); World War Two Draft Registration Card (United States Selective Service System, Selective Service Registration Cards, World War II: Fourth Registration, National Archives and Records Administration Branch locations: National Archives and Records Administration Region Branches); Seyed Khalil, “Typewriting Machine,” United States Patent no. 1403329 (filed April 14, 1917; patented January 10, 1922); Fourteenth Census of the United States, 1920 (National Archives and Records Administration, Washington, DC, Records of the Bureau of the Census, record group 29, NARA microfilm publication T625).

  53 H.H. Steele, “Arabic Typewriter.”

  54 In 1917, John H. Barr and Arthur W. Smith submitted a patent application for an Arabic typewriter, in which the spirit of minimal modification was once again expressed. “A further object of our invention,” Barr and Smith wrote as assignors to Remington, “is to readily adapt an ordinary typewriting machine employed for writing English or other European languages to use as a so-called Arabic machine of the character specified above without modifying, or materially modifying, the structural features of such ordinary machines as they now exist.” What is more, the mechanical modifications they were proposing brought into reach not only Arabic but by extension Turkish, Persian, Urdu, Malayan, “and many others in addition to Arabic itself.” John Henry Barr was Associate Professor of Machine Design at Cornell University, part of their faculty of Mechanical Engineering and Mechanic Arts. See The Cornell University Register 1897–1898, 2nd ed. (Ithaca: University Press of Andrus and Church, 1897–1898), 18; John H. Barr and Arthur W. Smith, “Type-Writing Machine,” United States Patent no. 1250416 (filed August 4, 1917, patented December 18, 1917).

  55 Georg Wilhelm Friedrich Hegel, The Philosophy of History, trans. John Sibree (New York: Wiley Book Co., 1900 [1857]), 134.

  56 Edward W. Said, Orientalism (New York: Vintage Books, 1979); Rey Chow, “How (the) Inscrutable Chinese Led to Globalized Theory,” PMLA 116, no. 1 (2001): 69–74; John Peter Maher, “More on the History of the Comparative Methods: The Tradition of Darwinism in August Schleicher’s work,” Anthropological Linguistics 8 (1966): 1–12; Lydia H. Liu, The Clash of Empires: The Invention of China in Modern World Making (Cambridge, MA: Harvard University Press, 2004: 181–209). See also, for example, August Schleicher, “Darwinism Tested by the Science of Language,” trans. Max Müller, Nature 1, no. 10 (1870): 256–259.

  57 Samuel Wells Williams, “Draft of General Article on the Chinese Language,” n.d., Samuel Wells Williams family papers, YULMA, MS 547 location LSF, series II, box 13, 3.

  58 Samuel Wells Williams family papers, YULMA, MS 547 location LSF, series II, box 13, 3.

  59 Peter S. Du Ponceau, A Dissertation on the Nature and Character of the Chinese System of Writing, Transactions of the Historical and Literary Committee of the American Philosophical Society, vol. 2 (1838).

  60 “Du Ponceau on the Chinese System of Writing,” North American Review 48 (1848): 306.

  61 Ibid., 272–273.

  62 Henry Noel Humphrey, The Origin and Progress of the Art of Writing: A Connected Narrative of the Development of the Art, Its Primeval Phases in Egypt, China, Mexico, etc. (London: Ingram, Cooke, and Co., 1853). Creel cites the second edition, published in 1885. Herrlee Glessner Creel, “On the N
ature of Chinese Ideography,” T’oung Pao 32 (2nd series), no. 2/3 (1936): 85–161, 85.

  63 China As It Really Is (London: Eveleigh Nash, 1912), 154.

  64 Ibid., 160.

  65 W.A. Martin, The History of the Art of Writing (New York: Macmillan, 1920), 13. Cited in Creel, “On the Nature of Chinese Ideography,” 85–161, 85.

  66 Bernhard Karlgren, Philology and Ancient China (Cambridge, MA: Harvard University Press, 1926), 152.

  67 T.T. Waterman and W.H. Mitchell, Jr., “An Alphabet for China,” Mid-Pacific Magazine 43, no. 4 (April 1932): 353.

  68 Creel, “On the Nature of Chinese Ideography,” 85.

  69 Ibid., 86.

  70 Ibid., 160.

  71 Geoffrey Sampson, Writing Systems (Stanford: Stanford University Press, 1985).

  72 Jack Goody, The Interface between the Written and the Oral (Cambridge: Cambridge University Press, 1987), xvii–xviii.

  73 Ibid., 64.

  74 Jack Goody, “Technologies of the Intellect: Writing and the Written Word,” in The Power of the Written Tradition (Washington: Smithsonian Institution Press, 2000), 138.

  75 Havelock, Origins of Western Literacy, 18; Robert Logan, The Alphabet Effect (New York: William Morrow, 1986), 57.

  76 Goody, “Technologies of the Intellect,” 138.

  77 Goody, The Interface between the Written and the Oral, 37. Joseph Needham underwent his own transformation. In the second volume of Science and Civilisation, he voiced suspicions about the Chinese language, citing it as one potential answer to the question of why China had failed to achieve a scientific revolution on par with that of Western Europe. “At a later date,” he wrote, “we shall enquire how far the difference of linguistic structure between Chinese and the Indo-European languages had influence on the differences between Chinese and Western logical formations.” When that later date arrived, Needham announced his conclusion: the “inhibiting influence of the ideographic language has been grossly over-rated.” “It has proved possible to draw up large glossaries of definable technical terms used in ancient and medieval times for all kinds of things and ideas in science and its applications. At the present day, the language is no impediment to contemporary scientists. If the social and economic factors in Chinese society had permitted or facilitated the rise of modern science there as well as in Europe, then already 300 years ago the language would have been made suitable for scientific expression.” Joseph Needham, Science and Civilisation, vol. 2 (Cambridge: Cambridge University Press, 1956), 199; Joseph Needham, “Poverties and Triumphs of the Chinese Scientific Tradition,” in Scientific Change (Report of History of Science Symposium, Oxford, 1961), ed. A.C. Crombie (London: Heinemann, 1963).

  78 Alfred H. Bloom, “The Impact of Chinese Linguistic Structure on Cognitive Style,” Current Anthropology 20, no. 3 (1979): 585–601.

  79 Derk Bodde, Chinese Thought, Society, and Science: The Intellectual and Social Background of Science and Technology in Pre-Modern China (Honolulu: University of Hawai‘i Press, 1991), 95–96.

  80 William C. Hannas, Asia’s Orthographic Dilemma (Honolulu: University of Hawai‘i Press, 1996); William C. Hannas, The Writing on the Wall: How Asian Orthography Curbs Creativity (Philadelphia: University of Pennsylvania Press, 2003).

  81 William G. Boltz, “Logic, Language, and Grammar in Early China,” Journal of the American Oriental Society 120, no. 2 (April–June 2000): 218–229, 221.

  82 “Le macchine Olivetti scrivono in tutte le lingue,” Notizie Olivetti 55 (March 1958): 1–4.

  83 Allen Ginsberg, Howl and Other Poems (San Francisco: City Lights Publishers, 2001).

  2

  Puzzling Chinese

  Of all the languages in the known world, the most difficult to represent by movable type is, without controversy, the Chinese; having hitherto baffled the most skillful European typographers.

  —Marcellin Legrand, 1838

  That the mosquito-wing-thin pages of William Gamble’s notebook survived those four tedious years is a marvel. It would have accompanied the Irish-born American printer to China in 1858, where he was dispatched to oversee operations at the Presbyterian Mission Press in Ningbo, one hundred miles south of Shanghai. This was Gamble’s workbook for practicing Chinese characters.

  Each page of the journal was divided into a fifteen-by-fifteen grid, enabling Gamble (1830–1886) to practice just over two hundred Chinese characters before moving on to the next sheet. Within each cell, a smaller four-by-four matrix served as guidelines to help the printer work toward composing characters with structural balance and elegant dimension. Preserved now at the Library of Congress in Washington, DC, it is one of the more intimate parts of the Gamble collection.1

  Gamble did not use this particular journal to practice his penmanship, however. Instead, he used it as a kind of accounting ledger with which he and two Chinese assistants tallied data about the relative frequency of Chinese characters within an immense corpus of Chinese texts. For four painstaking years, Gamble and his Chinese colleagues—referred to by Gamble as “Mr. Tsiang” and “Mr. Cü”—examined roughly 1,300,000 Chinese characters in total, spread out over four thousand pages.2 Working line-by-line, they made a record of every one of these characters, counting every time each character appeared, and organizing the data into handwritten tables.

  If only we enjoyed a time lapse film of these three men’s labors, the obsessive quality of their venture would become clear: bodies gyrating in a mechanical anti-reading of such texts as the Great Learning and the Daodejing, dismembering and dissolving their substance into elemental units, and then ranking these elements into classes of frequency.3 The Zhuangzi as living text might confound and delight—Now I do not know whether I was then a man dreaming I was a butterfly, or whether I am now a butterfly, dreaming I am a man—but this was not Gamble’s principal concern. For Gamble, what mattered over the course of those four years was to determine the raw composition of the Zhuangzi—to determine, as it were, that the Zhuangzi was nothing more or less than a composite made up of approximately 8 percent zhi (之), 5 percent er (而), 5 percent bu (不), 4 percent ye (也), and only a small handful of “butterflies” (hu 蝴 and die 蝶).4

  Faced with the algorithmic obsession of William Gamble, a question that instinctively comes to mind might be: What did he discover? Which Chinese characters were the most frequent, and least? What might this mean for the question of Chinese technolinguistic modernity we are pursuing? Was this the answer? While this chapter addresses these important questions, our primary concern will be to dig beneath Gamble’s work and ask: How did he come to settle upon this puzzle in particular, and what were the consequences of this particular puzzling? How did he come to decide that, among all the possible questions one might raise about Chinese script, this was the one he would let dominate four years of his labor and his obsession? As we set out to examine the “puzzle” of Chinese writing, this is to say, the first questions we will ask are not about the solutions offered to this puzzle, but about the puzzles themselves. How and why did the Chinese script come to be viewed as a puzzle in need of a solution in the first place, let alone one so addictive that it could command someone’s concentration for four painstaking years? Moreover, why did Gamble choose to turn Chinese into this kind of puzzle—one whose solution was premised on counting and statistics?

  It would be tempting to assume that the Chinese language, owing to its alien immensity, simply demands this kind of obsessive research. The Chinese lexicon numbers in the tens of thousands, after all, and has undergone steady expansion over the course of history (figure 2.1). The Shuowen Jiezi, an early character dictionary compiled by Xu Shen (ca. 58–ca. 147) in the Eastern Han dynasty, included 9,353 characters and 1,163 variants, forming the foundation for character compilation work thereafter.5 The Da Song Chongxiu Guangyun, completed by Chen Pengnian in 1011, contained more than twice that number, with a total of over twenty-six thousand. In 1716, with the completion of the Kangxi Dic
tionary, the number of characters totaled some forty-seven thousand, escalating further in the twentieth century with three immense compilation projects—the Da Kan-Wa jiten, the Hanyu dazidian, and the Zhonghua zihai, tallying 49,964, 54,678, and 85,568 characters, respectively. It seems only natural then that practically anyone—be they printer, scholar, or student—would instinctively arrive at the same riddle: how can one possibly “contain” this overwhelming Chinese abundance, whether within the confines of human memory, a printer’s rack, a telegraph code, or a typewriter? The language itself seems to be all the evidence one needs as to its puzzling inscrutability.

  2.1 Expansion of Chinese lexicon over time

  Common sense fails us here, however, as it so often does. As this chapter will argue, no matter how natural or inevitable a particular “Chinese puzzle” might seem to us in retrospect, all “Chinese puzzles” are in fact historically constructed and variable. Phrased differently, before Chinese can be said to be “puzzling,” it must first be “puzzled” by a particular group of people, in a particular time, and within a particular technolinguistic framework. There has never been an inherent or a priori “Chinese puzzle”—one simply willed into existence by the intrinsic complexity or strangeness of Chinese writing—only specific puzzles, formulated within specific historical and technolinguistic contexts, some of which succeed, others of which do not. For puzzlings that succeed, they go on to be reinscribed and remembered as stable, natural, and a priori; while those that fail, fail because they could not for whatever reason attract would-be puzzle solvers into the ludic labor of their solution. These are the puzzlings that fail to become puzzles, and are quietly forgotten.

  The choice of words here—puzzling Chinese rather than Chinese puzzles—might seem peculiar at first, and yet is motivated by two core observations. First, while it might be tempting to think of Chinese script as inherently puzzling owing to its lack of an alphabetic writing system, history teaches us that China has for the greater part of its history been marked by a spectrum of civilizational achievement and catastrophe as awesome as what one finds among any of its alphabetic neighbors. At the midpoint of the last millennium, Ming dynasty China was one of the engines of the world economy, one of its largest population centers, and a sphere of unparalleled cultural, literary, and artistic production—all without an alphabet. To have visited the China of the sixteenth and seventeenth centuries would have been to find a society undergoing accelerated urbanization, a demographic explosion, and with a vibrant print culture, not to mention the rise of fantastically wealthy families who made their fortunes as part of a growing interregional trade, and through an empire-wide banking and financial system ingesting ever greater quantities of silver from New World mines in Potosí—all without an alphabet. Thinking ahead to 1911, six years before the great Russian Revolution, Chinese revolutionaries overthrew a 2,000-plus-year-old imperial system without an alphabet. From the 1940s onward, the largest Communist state in human history was formed, a Great Leap Forward launched to cataclysmic ends, a Cultural Revolution waged, a post-Mao Reform Era inaugurated, and a new economic superpower established—all without an alphabet. To imagine that Chinese is intrinsically puzzling—and thus that it has always been so—is pure fallacy. Second, even as we move into the nineteenth century, and into a time when the emergence of powerful alphabet-centered information technologies placed Chinese script at an objective disadvantage, even here the Chinese script never constituted a “puzzle” in any singular or stable way. To the contrary, the “puzzling” nature of Chinese was always in the eye of its beholders, and within particular technolinguistic contexts—never due simply to any inherent properties of the Chinese script itself.