Oceanic Exchanges: Transnational Textual Migration And Viral Culture

Chair: Marc Priewe
The dramatic expansion of newspapers during the nineteenth century created a global culture of abundant, rapidly circulating information. For scholars of nineteenth-century periodicals and intellectual history, the digitization of newspaper archives and the ever-growing array of tools for accessing and assessing them provide a fruitful platform of new evidence to re-evaluate how readers around the world perceived each other and to obtain fresh insights on the global networks through which news and concepts traveled. Through the identification of patterns and trends, text mining is particularly suited for this task. In this panel, we want to focus on the array of methods within text mining (e.g., text categorization, text clustering, concept/entity extraction, production of granular taxonomies, sentiment analysis, document summarization, entity relation modeling, word vector models) that are used in current research to identify and model patterns of information flow as well as to trace the migration of concepts across different communities over space and time.
The panel brings together the efforts of scholars from different disciplines within the humanities (e.g., cultural history, comparative literature, historical linguistics) and computer scientists to explore and share how text mining methods and tools are currently used to answer research questions such as:
Which stories spread between nations and how quickly?
How did the migration of texts facilitate the circulation of knowledge, ideas and concepts and how were these ideas transformed once they migrated from one side of the Atlantic to the other?
How did geopolitical realities (e.g., economic integration, migration) influence the directionality of these transnational exchanges?
How did reporting within ethnic communities differ from reporting in surrounding host countries?
The panel will begin with a five-minute introduction to the
Oceanic Exchanges project and then present five ten-minute test cases exploring sets of data in a variety of languages (i.e., English, Italian, Spanish, French, German, Finnish, Swedish and Dutch) and employing different text mining methods. The panel’s main objective is not to prove one method “right” and another “wrong” but to explore the dominant strategies and methodological approaches that unveil the thematic and textual complexities between historical newspapers archives. Throughout the five papers of this panel, we offer multi-layered text mining approaches that, encompassing several methods, varying sets of data, and different discourse scenarios in a range of languages, are structurally coherent, methodologically solid, and comparatively rich. The panel will appeal to historians, linguists,  media and communication scholars, literary scholars, and computer scientists interested in textual and conceptual changes, continuities and replacements, but also to those focusing on representation of actors and events in public discourses.

Oceanic Exchanges between Italy and the United States: How Italian Americans became White

(Lorella Viola, Jaap Verheul)

Presenter: Lorella Viola, Jaap Verheul
Between 1880 and 1930, it is estimated that more than 22 million people from all over the world migrated to the United States, 4 million of whom were Italian. As immigrant communities grew, the immigrant press boomed accordingly. As far as the Italian language is concerned, between 1880 and 1920, there were between 150 and 264 Italian language newspapers published in the United States. Diasporic newspapers became an instrument for community building and helped immigrants cope with life in the New World, including easing their transition into American society as well as serving as powerful tools of language retention and national identity preservation. At the same time, acting as advocates for the rights of the respective immigrant communities, they performed social control by drawing attention to what was acceptable within the Italian immigrant community but also within the dominant norms and values of the American society.
This paper explores the role of printed media in constructing the concept of Italian identity, the so called
italianità, between the end of the nineteenth century and the beginning of the twentieth century in Italy and the United States. The overarching aim is to investigate the dynamics of knowledge transfer between the two Italian communities: Italians of the newly formed Italy and Italian diasporic groups in the United States. Using a mixed-methods approach that pairs Digital Humanities technologies such as text mining and semantic modelling with the discourse-historical approach pioneered by Ruth Wodak (2001), this study compares Italian newspapers published in Italy from 1867 to 1900 with Italian language newspapers published in the United States from 1880 to 1920. The results will show that the concept of
being Italian, after originating in Italy as a synonym for national identity, travelled to the United States, where it was reshaped by the Italian language press into a means for fighting against marginality and vindicating whiteness and social inclusion. In this way, in Italy
italianità embedded the ideals of the
Risorgimento and conveyed connotations of patriotism and nationalism. In the United States, on the contrary, Italian ethnic newspapers became powerful tools through which the Italian immigrant community could negotiate social integration in the host country.
Italianità became a way to uplift the Italian “race” which was simultaneously distancing itself from African-Americans.

Tracing the Traffic of Cholera in Nineteenth-Century Newspaper Repositories
(Jana Keck, Moritz Knabben, Steffen Koch)
Presenter: Jana Keck, Moritz Knabben
In this presentation we introduce a new corpus exploration tool which offers access to datasets of historical newspapers from North America, Europe, and Australasia (1840-1914), including at least six languages. With this tool, we are able to search for news articles containing specific keywords. However, since our corpus covers over one hundred million articles, the set of documents containing even specific keywords might be extensive. We address this problem by using visualizations to summarize the results and to find ways of deriving meaning from a vast body of texts. Our project is framed around a historical case study examining the traffic of cholera, which was one of the deadliest and most feared diseases. The nineteenth-century press played a crucial role in helping to construct the public image of cholera and other diseases. Contemporary medical scientists and practitioners, especially in the United States, wrote mostly negatively about the role of newspapers in insufficiently informing the public about possible causes, transmission roots and routes, and treatment methods. In this paper we will show how “cholera” was covered by the press in different languages and in different countries. To this end, we add functionality to a quantitative search method by applying further visualization techniques. This allows both a “distant reading” of a large newspaper corpus, while retaining some of its contextual meaning by filtering and displaying the set of found articles. Questions we ask include, but are not limited to: Can we identify a narrative of disease propagation circulating in the press that was absent, or even suppressed and disagreed with, in the official conduits of medicine? Did the increasing discrediting of miasma theory and of cholera itself cause a similar decline in the news coverage? We highlight the potential of this search tool to show how newspaper articles de-centered medical knowledge from being located only within the medical field to illustrate instead that the acceptance of these medical developments was not only a scientific, but moreover a social and cultural  effort. By placing the non-scientific newspaper articles into conversation with the social and medical developments, a more complete picture of the modernization and professionalization of medicine and (scientific) journalism is possible.

The Origins of Fake News: Lajos Kossuth, Political Celebrity, and Dis/Information in the Nineteenth-Century Press
(Paul Fyfe, Jana Keck, Mila Oiva, Jamie Parker)
Presenter: Mila Oiva
In December of 1851, the Hungarian revolutionary Lajos Kossuth landed in New York City at the outset of a publicity campaign to secure American support for Hungarian independence. His trip along the eastern seaboard, ultimately arriving in Washington, DC, was exhaustively covered by not only the American press but by an increasingly connected network of international newspapers. In many ways, that network established Kossuth as an international celebrity, stirring the crowds that flocked to his speeches, inspiring far-flung editors to adapt Kossuth’s messages, and inviting Hungarian political operatives and others to discredit Kossuth and manipulate the news network itself. This paper argues that Kossuth’s journey—and the international news coverage and censorship it inspired—showcases the mid-century operations of a networked mass media system. Connected by domestic telegraph wires, railways, steamships, exchange networks, and extensive reprinting practices, the international newspaper network by the mid-nineteenth century was increasingly functioning
as a system. The Kossuth case shows how its properties as a system were exploited for various ends, including not only Kossuth’s political aspirations but a complex set of goals that shift with language and location. The manipulation of the international news network shows the stirrings of dissemination, disinformation, and censorship that now shape political discourse via globally-connected social media platforms.

Our sources span the United States, Britain, Germany, Austria, Finland, and Russia, comprising thousands of articles published in multiple languages following Kossuth’s arrival in New York City in December 1851. We identify news reprints manually and with the help of computational linguistic tools, including the use of rare word tokens and Named Entity Recognition. This selected corpus is then used to create visualizations of news reprints and their dissemination. The patterns and disparities of those reprints, in turn, allow us to evaluate the accuracy of these reports, to identify “fake” news, and to explore how and why it spread. Much like contemporary social media, international media in the nineteenth century allowed for disinformation practices based on the political goals of national stakeholders. Then as now, these practices exploit the paradoxes of any international media system: its seeming connectedness and yet its enduring distances. As Lajos Kossuth carried his message personally across the Atlantic, stories about his trip circulated and spun behind him. The international news network at once consolidated Kossuth as a global celebrity and fractured his message for various regional political ends.

“Remember the
Maine”: Newspapers in the Spanish-American War

Ernesto Priani Saisó, Jamie Parker, Marc Priewe, Isabel Galina Russell, Miriam Peña Pentel, Rocio Castellanos, Laura Martinez Dominguez, Laura Lopez, Ximena Gutierrez-Vasques, Adan Lerma.
Presenter: Ernesto Priani Saisó
On February 15, 1898, an explosion destroyed the US-American battleship
Maine stationed at Havana’s port and the following morning the news hit the front pages of many cities in the United States, Mexico, Spain, the United Kingdom, Germany, Austria, and Finland. The explosion happened amidst already existing diplomatic tensions that would result in the Spanish-American War, which in turn marked the emergence of the United States as a new imperial actor in international politics. The objective of this paper is to show how the information about the explosion of the
Maine was disseminated throughout selected European countries (Spain, UK, France, Germany, Austria and Finland), the United States, and Mexico by analyzing news reuse in the press using several digital tools as well as text mining and Natural Language Processing (NLP) techniques. In addition, we employ Automatic Author Profiling and Vector Space models to characterize the news semantic content applied to a corpus constituted from several National Newspaper Libraries. This allows us to compare historical news coverage of events in ways that were hitherto impossible. In Mexico, for instance, the news diffusion occurred through newspapers from the Spanish, French and American communities living in the country. Internationally, the news dispatch system depended on telegraph cables connecting the US and Europe and an alternative one between Spain, Cuba, Mexico, and Central America, inaugurated two years before the outbreak of war in April 1898. Our analysis shows further that the location source of the news was initially Havana but then moves to Washington, New York, and Key West in the United States. Quickly the American voice was the dominant one on the event; meanwhile, there was a debate in virtually all national presses about the explosion, whether it was an accident or provoked and what the political implications might be of either cause of the explosion.

News Flows in 1904: The Media Coverage of Nikolay Bobrikov's Assassination
(Otto Latva, Asko Nivala, Mila Oiva, Hannu Salmi, Marja Jalava)
Presenter: Hannu Salmi
General Governor of the Grand Duchy of Finland Nikolai Bobrikov was assassinated in Helsinki in 1904. While the first shot was heard in Helsinki at noon on 16 June, the news was published in Mexico City, thanks to the eight-hour time difference backwards, the next morning, on 17 June. The shooting of Bobrikov offers an illuminating case for the study of European and global news circulation at the beginning of the twentieth century. The efficient international telegraph network and press agencies forwarded information that was estimated to be of interest for the news market quite quickly. The system of news delivery emphasized big centers, but the Bobrikov case shows that news from the remotest corners of Europe could also get a wide circulation. The rhizomatic communication network had its centers, but it was not centralized.
This presentation studies the tempo and the ways in which news about Bobrikov was disseminated in the network of newspapers. It concentrates particularly on the viral expansion of the news during the first week after the assassination. What kinds of temporal rhythms did it have? What narrative elements did it include? What kinds of spatial ramifications did the news have? What is the diagram of news flow behind the spread of this particular case? The presentation also focuses on the thematic differences in the discourse on Bobrikov’s murder in different cultural settings. The presentation utilizes digitized newspaper collections from Finnish, French, German, Dutch, US, Australian, Austrian, Swedish, and Danish archives. In addition, we have used microfilm and physical newspaper collections of Russian newspapers at the Finnish National Library. The study takes advantage of the transnational news dataset and follows the most frequent paths of news flow, and visualizes them using tools like Palladio.