Perceptions of the Other in Travelogues 1500-1875 - A Computerized Analysis

Travelogues are a highly important source for historical research. They provide information about regions, cultures and religions but also enable insights into the authors cultural self-conceptions. However, the backgrounds of travelers, the aim of their journeys as well as the intended audiences vary considerably. This makes this source rather difficult to engage with. As a consequence scientific interest tends to concentrate on single texts. Austrian and German researchers from History, Computer Science and Book History cooperate to analyze a digitized collection of several thousand travelogues, covering the period from 1500 until 1875. In close collaboration, the research team will develop novel digital techniques. These include the automatic detection of travelogues within the complete historical holdings of the Austrian National Library (ca. 600,000 volumes, which are currently digitized in the Austrian Books Online Project). Furthermore, the automatic identification of passages of text that relate to otherness and the creation of a database containing the identified travelogues and associated data, such as the name, date and place of birth of the author. The developed tools will be designed and made available to allow researchers from other fields to analyze similar questions based on other documents. Therefore, this project will lay the groundwork for further research and provide the base for follow-up projects. The developed digital methods provide the tools to seek for answers to the following sets of guiding questions: What are the characteristics of travelogues? What is specific about them? How can they be automatically detected? How did authors express otherness in travelogues? Did it change over time, e.g. by switching from fear to mockery, or from antipathy to fascination? How was the Orient depicted in travelogues about this region? How do these relate to the general discussion of the Orient and its inhabitants? Was there any change over time? We specifically focus on how otherness was expressed in travelogues about the Orient and how it changed over time. This is not a purely historical issue. Every one of us is constantly confronted with otherness due to processes such as global mass tourism, the internationalization of consumer culture, transnational migration and globalization. By critically engaging with the historic dimension of otherness within travelogues, it is possible to raise more awareness and it will help to generate historic contextual knowledge and orientation for the present and future.

This interdisciplinary project focused on German language travelogues in the collections of the Austrian National Library (1500 to 1876). To analyze perceptions of "the other" and "the Orient" (Ottoman Empire, Persia) in a large-scale text corpus, algorithms for the semi-automatized search for, and evaluation of texts were being created. Scientists from the Institute for Habsburg and Balkan Studies (IHB) at the Austrian Academy of Sciences, the Austrian Institute of Technology (AIT), the Department of History of the University of Salzburg, the Austrian National Library (ÖNB), and the research center L3S at the University of Hannover took part. Results: 1. The project team identified almost 5,000 travelogues in the digital and analogous holdings of the ÖNB. Based on title and keyword searches, the team semi-automatically created sets of likely travelogues that were sorted out manually. This data served as "ground truth" for algorithms for the automated recognition of travelogues, which worked out successfully. By applying these algorithms, the team identified further travelogues, many of which had previously been unknown to the research community. The method is absolutely new and proved the usability of mashine learning in the identification of specific types of texts, even if the conversion of the text into machine-encoded text (OCR) is of poor quality. For this, the project was rewarded with the Lee Dirks Award for Best Full Paper. In the meantime, other projects already tried and applied this method. 2. As another result the project secured a large amount of European cultural heritage since it created the largest corpus of (German language) travelogues ever. The largest part of the sources is now digitally accessible via a search marker in the online library catalog of the ÖNB. 3. The metadata, that have been checked intellectually and have been created, improves the visibility, and explorational potentials regarding the text corpus for the humanities and the computer sciences. 5. We successfully identified identical text passages ("text reuse") of the travelogues on the "Orient" by applying BLAST (Basic Local Alignment Search Tool), an algorithm originally established in the field of bioinformatics for fast DNA sequence alignment. The results are promising, but show, however, that the OCR quality is decisive and that ways to balance out different spellings, synonyms and verbal variations of the same and similar content need to be tackled in future research on the topic. 6. The semi-automated identification of perceptions of "otherness" proved to be currently not as feasible as hoped (reasons: heterogeneity of the corpus; great need for "ground truth" for machine learning; insufficient OCR quality). However, related questions have been addressed using classical humanities methods as part of a collected volume, that will be published in 2022.