Digital Humanities Workshop - Session 2

Impact of the Digital on Japanese Studies

Session 2 - Abstracts

Raja Adal, University of Pittsburgh

The Epigraphy of Business Documents

This presentation is part of a larger project that is concerned with the way in which writing mediates human relations, not only in terms of the semantic content of the text but in terms of its visual form. More specifically, it seeks to understand how the material aspect of writing mediated business networks in Japan from 1889 to 1940. To do so, it makes a digital analysis of more than ten-thousand letters in the archive of one of the largest Japanese companies of this period, the Mi’ike Mining Company. These documents were written with a panoply of instruments, from the brush to the pen to the typewriter, featured a variety of stamps, letterheads, and inks, and were duplicated using a range of reproduction technologies from the hectograph to the mimeograph, carbon paper, and printing press. Together, these technologies of inscription gave texture to this human network.

Amy Catalinac, New York University

On the Politics of Text

For most of the postwar period, politics in Japan revolved around individual Liberal Democratic Party politicians and their respective abilities –cultivated through personal campaign organizations and intra-party factions – to deliver pork to groups of voters in their electoral districts. Japan’s 1994 electoral reform changed all this.  Nowadays, politics looks much like it does in other parliamentary democracies: politicians contest elections as representatives of parties and promise policies their party will implement if it wins office. I apply cutting-edge unsupervised learning tools for quantitative text analysis to an original, digitized collection of 7,497 Japanese-language election manifestos produced by the universe of candidates competing in the eight House of Representatives elections between 1986 and 2009 to document when and how this profound shift occurred. Specifically, probabilistic topic models demonstrate that politicians’ electoral strategies shifted from pork to policy and scaling models show that parties became more ideologically cohesive after the reform.

Molly Des Jardin, University of Pennsylvania

On the Language of Empire in Taiyo Magazine (1895-1925)

How does the empire talk about itself? In imperial Japan, borders were in flux and a discourse of naichi (home islands) and gaichi (colonies) was established over time, making this question particularly compelling. Taiyō, a popular general-interest magazine, covers 1895-1925, and I have converted NINJAL's XML Taiyō corpus (tokenized by sentence) to word-tokenized UTF-8 text in order to discover what the publication might tell us about imperial language. I plan to begin with exploratory analysis using topic modeling as a starting point. One challenge of this corpus is its size: it consists of 1.2 GB of text and thus "breaks" nearly all available off-the-shelf tools when analyzed in its entirety (including Voyant Server and Topic Modeling Tool). Due to this, I want to think about productive ways to break up the text into meaningful chunks for analysis, as well as the analysis itself.

Mark Ravina, Emory University

Political Discourse in Early Meiji Japan

How did early Meiji-era political activists voice their criticism of the state? How did their language change as they adopted new concepts of sovereignty and legitimacy? My project addresses these questions through an analysis of kenpakusho, “memorials” or petitions. Between 1867 and 1889, the Meiji government received nearly 4000 petitions on an astonishing array of subjects: treaty revision, conscription, education, tax reform, postal regulations, religion, and marriage ceremonies. Petitioners included daimyo, samurai, priests, monks, and commoners. These thousands of petitions constitute a small “big data” corpus: small enough for analysis on a microcomputer, but too large for an exacting close reading of all the texts. My preliminary DH analysis of the petitions reveals how Meiji political discourse, especially that of the Freedom and Popular Rights Movement (Jiyū minken undō), built on older political discourse. I have mapped, for example, how the term jiyū took on its modern meaning of “freedom,” building on older Confucian critiques of despotism. Using techniques ranging from simple comparison of word frequency to topic modeling, I have begun to model the emergence of a new language that legitimized political contestation and dissent.