The Chinese Typewriter

by Thomas Mullaney

  In this chapter, we will examine three remarkably different “puzzlings” of Chinese, each of which emerged during the nineteenth century. We will call these three puzzlings common usage, combinatorialism, and surrogacy. As exemplified in the work of William Gamble above, the first mode—common usage—was premised upon an assumption about Chinese that is so widespread and taken for granted that it might strike us as unnecessary even to articulate: namely, that the elemental unit of Chinese writing is the “character,” and that there are tens of thousands of these characters in Chinese. Setting forth from this point of departure, the common usage approach to Chinese technolinguistic modernity was one geared toward reducing the Chinese lexicon to its most essential units, which in turn required the kind of painstaking statistical work undertaken by Gamble and his associates. The technolinguistic objective of this common usage approach was that of building inscription technologies that contained only the most frequently used characters within the overall language. The common usage approach to Chinese technolinguistic modernity will be the focus of the first part of this chapter.

  Precisely at the same time that Gamble was feverishly counting up each of the language’s tens of thousands of characters, however, others were problematizing Chinese in a dramatically different way, based upon far different assumptions about the script. The second mode—combinatorialism—was premised upon reimagining Chinese writing as a form of quasi-alphabetic script in which one could decompose Chinese characters into a set of modular shapes that the operator would then use to build, or “spell,” characters piece by piece. Chinese characters could not be “spelled” in the sense of phonetic letters, yet perhaps they could be “spelled” using a set of recurring, modular shapes. In this approach, what was called for was not a reduction of the Chinese lexicon, but a critical reimagining of the very essence of Chinese writing itself, transposing the idea of “letters” and “spelling” atop Chinese writing and reimagining the structural components of Chinese characters—often referred to as “radicals”—as the equivalent of letters in the Latin alphabet. In forming the puzzle this way, combinatorialism thus denaturalized one of the most seemingly self-evident and “obvious” characteristics of Chinese—namely, the idea that Chinese was made up of an overwhelming number of elemental units called “characters.” If the Chinese character was not the irreducible atom of Chinese writing, as maintained in the common usage approach—if there existed something still more fundamental and repetitive into which characters could be further divided—then the real puzzle of Chinese technolinguistic modernity was waiting to be solved here at this level. The method that grew out of this puzzle would come to be known as “divisible type,” and will be the focus of the second part of this chapter.6

  Still others puzzled and problematized Chinese in yet another way—a third mode we will refer to as surrogacy. In contrast to both common usage and combinatorialism, this third mode was premised neither on counting and ranking Chinese characters nor on cutting them up into pieces, but on symbolic systems that could be used to stand in for, or refer to, Chinese characters—particularly in the arena of the emerging technology of telegraphy. In this mode, Chinese characters remained the fundamental unit of Chinese, but were not to be trafficked in directly. Rather, they were to be sequestered to separate spaces—whether a character code book, a database or, most abstractly, to the realm of human memory and the “mind’s eye”—and thereafter “retrieved” according to specific protocols. In this way of puzzling Chinese, the primary task of the puzzle solver was neither the statistical conquest of Chinese lexical abundance, as in common usage, nor the dismemberment of Chinese characters into their component elements, as in combinatorialism, but the development of ever more efficient techniques of reference, address, databasing, search, and retrieval. Surrogacy will occupy our attention in the final part of this chapter.

  Instead of treating the “Chinese puzzle” as a taken-for-granted starting point for our history, then, this chapter excavates the underlying and often invisible assumptions that made each of these distinct types of puzzles imaginable, meaningful, solvable, and worth puzzling over in the first place. As we will see in subsequent chapters, moreover, it will be these logics and assumptions that persist and continue to shape the pursuit of Chinese technolinguistic modernity long after Gamble and his contemporaries have been forgotten. Each of these three logics will reappear, that is, in the age of typewriting to come.

  Settling the Nomadic Typesetter: Movable Type, Common Usage, and the Desire to Surround Language

  Returning to the question raised at the outset of this chapter—what propelled William Gamble in his algorithmic, statistical labors—we find important clues from the printer’s own accounts. As Gamble explained in the preface to his 1861 work, Two Lists of Selected Characters Containing All in the Bible and Twenty Seven Other Books, “not only have the type themselves taken up much room,” referring to Chinese fonts, “but the compositor in going from case to case for each type has unavoidably consumed so much time, as thereby to render composition both expensive and tedious.”7 Decades before linguists, engineers, and language reformers would attempt to resolve the “incompatibility” they saw between Chinese and typewriting, then, Gamble sought to resolve a still earlier “incompatibility”: between the Chinese script and movable type.

  During the mid-nineteenth century, it became commonplace for foreign printers to criticize Chinese for the challenges it presented to movable type printing—a curious criticism when we consider that the technology of movable type printing had been invented in China four centuries prior to Johannes Gutenberg in Mainz. In the eleventh century, Bi Sheng (990–1051) developed earthenware movable type, which was arranged into an iron frame lined with paste to secure the fonts during the printing process. In the fourteenth century, records by Wang Zhen (fl. 1290–1333) cite the usage of wooden movable type, with bronze and other metals being used in the process by the late fifteenth century.8 While Chinese printing was dominated by xylographic or woodblock printing methods throughout this period, nevertheless the technical prowess of the system continued to be refined. An illustrative case in this regard comes from the early Qing period. In 1773, work began on a massive compilation project authorized by the Qianlong emperor (r. 1736–1795). Its objective was the collection, publication, and circulation of 126 rare Chinese works selected from the larger Four Treasuries (Siku quanshu) collection.9 In his first memorial to the emperor, supervisor of the Imperial Printing Office (Wuying dian) Jin Jian requested authorization to cut a total of 150,000 movable wooden type, and further stipulated the need to cut between 10 and 100 duplicates of approximately 6,000 of the most commonly used characters within the font.10 In contrast to xylography and the hand carving of master copies, movable type printing required operators to be acutely aware of the total number of distinct characters needed to print a given text and, relatedly, the number of duplicate sorts that would be needed for each of these characters—that is to say, their relative frequency.

  Jin Jian partitioned the Chinese lexicon, and with it the printing office itself, into two broad categories in accordance with the energetic output of the typesetter’s body—with the lexical categories of high and lower frequency characters being translated into somatic categories of near and far, vicinity and distance. “There are some rare characters that are seldom used,” Jin wrote, “and for which few type will have been prepared. Arrange small separate cabinets for them according to the twelve divisions mentioned above, and place them on top of each type case from where they may be seen at a glance.” It was only then, after this primary partition, that Jin organized his characters secondarily according to the “radical-stroke” system, a taxonomic system dating back to the late Ming dynasty (1368–1644) and then canonized in the early-eighteenth-century Qing dynasty (1644–1911) reference, the Kangxi Dictionary.11 In this dictionary, named in honor of the Kangxi emperor (r. 1661–1722) who ordered its compilation, upward of 40,000 characters are organized into 214
categories based on the primary character component (or radical [bushou]) out of which each character is composed. Characters such as ta (他 “he”) and zuo (作 “to do/make”) are categorized along with others built using the “person” radical (亻), while characters such as hong (洪 “flood”) and hu (湖 “lake”) are assigned to the group of characters listed under the “water” radical (氵). A second-order taxonomic rule places each of these 214 categories in a sequence based on the number of strokes in the component radical. For example, insofar as the person radical is composed using two brushstrokes and the water radical using three strokes, the characters ta and zuo above precede hong and hu in the dictionary.12 Within each of these 214 classes, a third-level taxonomic rule further organizes characters according to the number of residual strokes required to complete the character, thereby placing the five-stroke character ta prior to the seven-stroke character zuo.13 Well into the twentieth century, this was the dominant method of organizing Chinese-language dictionaries in China.

  When Jin Jian and his typesetters walked from case to case, and between cases and the printing press, then, they were quite literally walking through a physical model of the Chinese language itself.14 Rather than surrounding the Chinese script, the way their counterparts in the Western world surrounded the alphabets that lay before them in the type case, it was the Chinese script that surrounded them. This distinct spatial quality of Chinese movable type vexed Gamble and a number of his contemporaries, whose increasing criticism of Chinese movable type was premised on what we might term an almost imperceptible “drift” in the operative definition of “movable type” itself. While aware that this printing technology had been invented in China and for Chinese characters, Gamble and others nevertheless expressed their belief that a time still awaited when Chinese would finally be brought “within the compass of European typography,” as one contemporary phrased it.15 “European typography” in this formulation carried with it a particular meaning, namely the idea of Chinese movable type as a lesser form thereof, lesser precisely because of the nomadic, peripatetic practices of Jin Jian and others—this process of walking through the Chinese script. For Gamble, true movable type was clearly tied to an ideal of sedentary mastery—the way in which a Western typesetter could stand fixed before his type rack and surround script. Operating beneath all these assumptions was a shifting definition of the term “movable type”: no longer limited to its strict technical meaning, signifying a technique and technology of printing in which “type” were cut and cast as modular, “movable” pieces (a criterion Jin Jian’s method undoubtedly fulfilled), a new paradigm had emerged in which “movable type” was only truly movable to the extent that the typesetter was stationary (something the Chinese typesetter was most certainly not).16

  The first of our three “puzzlings” of Chinese took shape in this context—one in which algorithmic reading, frequency analysis, and ever-increasing Chinese corpora would quickly become the mainstays of a new epistemic framework. A foundational moment for this new “Chinese puzzle” came in 1810, when the English-speaking world was introduced for the first time to the Great Qing Legal Code, as translated by George Staunton. Practically from the moment of its publication, scholars and statesmen alike crowded about the window it opened up into the once mysterious legal logic of the Celestial Empire—a “landmark in Western knowledge of not just Chinese law but also Chinese civilization in general,” as historian Li Chen has argued.17 Called Britain’s first sinologist by some, Staunton enjoyed a personal and professional pedigree that established him as an ideal translator and editor for such a work. The son of diplomat and Orientalist George Leonard Staunton, and page boy to George Macartney during the 1793 embassy to emperor Qianlong’s court, Staunton commenced his training in the Chinese language at a precocious age, later going on to serve as a member of British Parliament and as a senior official in the East India Company.

  Staunton’s translation of the Great Qing Legal Code was foundational for the formation of modern sinology in a second, less apparent way. Upon its publication, Orientalists, printers, educators, and publishers became fascinated by a realization that Staunton had stumbled upon in the course of his translation work—specifically, a number that Staunton arrived at when he tallied up how many distinct Chinese characters he encountered along the way. For all its size and complexity, Staunton revealed, the Great Qing Legal Code was composed of roughly two thousand different Chinese characters in all—but a small fraction of the approximately forty-seven thousand found in the Kangxi Dictionary, the leading Chinese dictionary of the day. This remarkable number—2,000, more or less—implanted a thrilling possibility in the imagination of many: if a mere five percent of all Chinese characters were needed to print and read a legal text of such importance, what might also be true of the Chinese-language canon more broadly? What might this mean for foreign printers and students of the Chinese language, who had long believed it impossible to conquer its “tens of thousands” of characters? Far more than providing one potential solution to the “Chinese puzzle,” Staunton’s work offered up something even more powerful: a means by which to transform Chinese into a puzzle, or in other words, the puzzle itself.

  Staunton’s “discovery” of Chinese common usage sent a shock wave through the transnational community of sinologists that could be felt for decades. For foreign printers of Chinese, his observations raised the possibility of dramatically scaled-down Chinese fonts—fonts containing perhaps five thousand sorts, rather than fifty thousand—that would nevertheless be sufficient for their needs. As suggested in the preliminary findings of Staunton, the objective of surrounding the Chinese language—and thus sedentarizing the nomadic Chinese typesetter—might be achieved by erecting fences to divide up the rolling terrain of the Chinese language. This puzzling of Chinese script was thus primarily one of boundary formation: determining through rigorous and painstaking analysis which characters in the Chinese lexicon were truly “essential” and which were frivolous or incidental. For foreign learners of Chinese, the notion that their energies could be focused on mastering a core subset of the language—rather than dispersing their energies across “useless” characters—was particularly attractive in an era when no formalized pedagogical system was in place for acquiring the Chinese language. Printers and students alike quickly took up the task of establishing boundaries between essential and inessential characters, dissolving Chinese texts in the acid bath of reason with the goal of scientifically determining where to concentrate their mental energies and financial resources. The age of Chinese “distant reading” had begun.

  In quick succession, scholars began to extend Staunton’s initial observations far past the Qing legal code. Bengal-based missionary Joshua Marshman announced to his readers that “all the works of Confucius contain scarcely three thousand characters.”18 The epic novel Romance of the Three Kingdoms was composed of a mere 3,342 distinct characters, the Missionary Herald reported. William Gamble, the missionary printer we met at the start of this chapter, determined more precisely that the Four Books of Confucianism required only 2,328 unique characters to print; the Five Classics only 2,426. In the aggregate, the Thirteen Classics—embodying The Book of Poetry (Shijing), The Book of Documents (Shujing), The Rites of Zhou (Zhouli), The Ceremonies and Rites (Yili), The Record of Rites (Liji), The Book of Changes (Yijing), The Commentary of Zuo (Zuo zhuan), The Commentary of Gongyang (Gongyang zhuan), The Commentary of Guliang (Guliang zhuan), The Analects (Lunyu), Luxuriant and Refined Words (Erya), The Classic of Filial Piety (Xiaojing), and The Mencius (Mengzi)—contained only 6,544 characters.19 With each passing year, it seemed, foreigners were steadily solving the puzzle of common usage Chinese.

  In puzzling Chinese so, however, the “common usage” framework gave rise to a profound and arguably irresolvable tension—a politics that would haunt this approach to Chinese technolinguistic modernity, perhaps forever. In contrast to our analogy of nomadic and sedentary, plains and fences, the terrain of Chinese
language itself was one that had always morphed and changed, and always would. As historians know well, the Chinese lexicon changed and expanded dramatically during the closing decades of the nineteenth century and the opening decades of the twentieth, with thousands of new Chinese-language terms pouring in from neighboring Japan, and still more developed by Chinese translators of foreign-language texts. In the context of common usage Chinese movable type, neologisms and other lexical changes exerted an influence unknown in Western, alphabetic movable type. If we imagine typesetters in the Western world having, at certain points in history, to confront neologisms such as “hegemony,” “colonialism,” and others, such terms posed no steeper challenge for the alphabetic typesetter than to set the three comparably common words “my,” “he,” and “gone,” or “is,” “on,” “oil,” and “calm,” respectively. No matter the novelty or obscurity of a word in German, English, French, or Italian, nevertheless its composition in alphabetic movable type required the same letters as its more common counterparts—save perhaps in the printing of foreign loanwords that required special letters not present in one’s script. In the Chinese case, by contrast, “common usage” was a zero-sum game, in the sense that the inclusion of any new character necessitated the exclusion of another—or phrased differently, it required constantly reestablishing the boundary between “frequent” and “infrequent.”