This study uses cutting edge computational algorithmic tools to identify and analyze the emergence and structure of a Hebrew periodical network from the second half of the nineteenth century. Our corpus consists of five Hebrew periodicals published 1874-1893: HaMagid, HaMelitz, HaTzfira, HaLebanon, and Habazeleth, which provides a rounded perspective of the major Hebrew periodicals of the time, representing different geographical regions and sects.
Computational tools are essential in analyzing networks because of the need for a distant overview rather than a close reading of the text. In the case of the Hebrew networks, the need for computational analysis is greater due to the facts that these articles were often reused for various reasons and their authors remained anonymous. As a result, links often remained hidden to the readers of the text. To analyze these dynamic networks, in which everything—news, issues, agents, and institutions—was in motion, and reveal often-concealed textual and social relations, textual reuse analysis and stylometric methods are used. The reuse analysis is done through a commercial algorithm developed to detect academic plagiarism in the modern Hebrew language. The stylometric analysis uses an existing open-source code to identify clusters of similar writing styles.
We implement and customize network analysis methods to the unique circumstances of Jewish journalistic networks, specifically regarding the geographical spread of readers and authors spanning state and continental borders. In this environment, the definitions of "local" and "us"—from a national, cultural, and social perspective—were complicated, delicate, and elusive. Furthermore, offices of Hebrew periodicals often migrated from one state to another due to licensing issues and personal circumstances of the editors. This in itself contributed to the forming of long-distance personal and institutional relationships. This Hebrew journalistic network constituted a radical model of communal, rather than individual, authorship.
This research identifies two primary types of networks: first, the flow of news along with their "viral" contents, and second, the human agents central to the limited Hebrew "Republic of Letters," of which the Hebrew press was a significant component. Stylometric analysis allows us to cluster authors and literary circles according to similarity in writing style and argumentation. This will identify significant agents and groups within the journalistic Hebrew network.
This study will be the first large-scale computational network analysis of historical Hebrew periodicals, and thus a methodological and historical contribution to periodical and journalism scholarship.