On January 18-19, 2024, the workshop on Digital Humanities and the Study of Mediterranean Mobilities was held for the first time. This event took place at the University of California, Berkeley and allowed attendees to join via Zoom. The workshop was held as part of the IstanΠόλις Collaborative, a collaborative public and digital humanities project aimed at the reconstruction and analysis of Greek Orthodox communities in the broader context of late Ottoman Istanbul/Constantinople (1821-1923). Linked researchers and those collaborating with the project presented their research progress so far.
As a member of Storymodelers Lab and the Web Science and Digital Libraries (WS-DL) research group, I had the opportunity to participate and present my analysis of historical data from the Ottoman Empire on a dataset provided by Berkeley historians. This was a presentation that took place as part of the collaboration between the Storymodelers lab, led by Dr. Erika Frydenlund and Dr. Jose Padilla, and the IstanΠόλις Collaborative project, led by Dr. Christine Philliou.
 |
| Campus University of California, Berkeley |
During the two-day workshop, there was a varied agenda of presentations ranging from history experts to modelers and data scientists. This diversity of disciplines and approaches enriched the experience of all who attended. The main topics and discussions of the workshop were the need to advance the use of artificial intelligence and data science to increase the efficiency of the transliteration and transcription process and to improve the processing, analysis, archiving, and mapping of historical data.
The workshop began with a welcome and opening remarks by Dr. Christine Philliou and Dr. Dzovinar Derderian, who explained the dynamics of the workshop and detailed that the presentations to be given were research in progress, so there were no publications yet and collaboration and exchange of ideas among participants was welcome. The main objective was to enrich existing knowledge, encourage the application of new methodological approaches, discuss current challenges, explore innovative solutions, and establish networks among researchers with similar interests to complement the work of historians.
 |
| Welcome and Opening Remarks by Dr. Christine Philliou |
Panel 1. Research Projects on Mobilities
The format of Panel 1 was developed to include a total of five presentations on research advances in mobility in the Mediterranean. Each presentation allowed attendees to ask questions and express their ideas on the topics presented.
1.1. On Vegetable Oil, Digital Ottoman Studies, and Mapping Migrant Smuggling Networks in Late-Ottoman Anatolia [Dr. David Gutman, Manhattanville College]
The first presentation was led by Dr. David Gutman of Manhattanville College, who shared his progress in mapping smuggling networks in late Ottoman Anatolia. Following and analyzing manuscripts and documentation from the period, Dr. Gutman has been able to uncover nicknames of individuals linked to criminal networks, as well as their movements between cities. His goal is to capture these movements on a map; however, there is the limitation that current APIs are not designed to represent cities that existed many centuries ago. Faced with this barrier, Dr. Gutman created a map by hand, taking advantage of the skills he acquired from watching his father, who is a cartographer. Dr. Gutman highlighted the need to advance the digital representation of historic cities because, for him, people do not always need to find a narrative in a book but instead find a visualization through maps. This research breakthrough highlights the need for historians to collaborate with other disciplines, particularly with experts in technology, geolocation, and data visualization.
 |
| Dr. David Gutman shows his handmade map of the late Ottoman Empire. |
1.2. Armeno-Turkish Translation Mobilities - Canon Christoph von Schimid’s Influence on 19th-Century Turkish Christian Juvenile Literature [Dr. Hülya Çelik, Ruhr University Bochum and Dr. Ani Sargsyan, University of Hamburg]
Dr. Hülya Çelik from the Ruhr University Bochum and Dr. Ani Sargsyan from the University of Hamburg gave the second presentation. These researchers have advanced in analyzing and monitoring published and translated documents in multiple languages. Specifically, they have focused on documents of Schmid and devotional literature. In their presentation, the researchers mentioned that the study of translation trajectories provides a form of mobility. This means that translations into multiple languages show a flow of texts but also how these texts are received, reinterpreted, and adapted by various cultures over time.
This advance in research shows a need for interdisciplinary collaboration. Integrating knowledge from computer science could enhance the work of historians by increasing the efficiency of the reading and analysis process. Specifically, incorporating artificial intelligence tools in analyzing historical texts would allow for more accurate and fast identification of linguistic patterns and differences.
 |
| Presentation by Dr. Hülya Çelik and Dr. Ani Sargsyan |
1.3. The Emergence of Music Recording Industry and Armenian Sonic Connections in the Eastern Mediterranean (1900-1914) [Dr. Yektan Turkyilmaz, Central European University, Vienna]
The research progress by Dr. Yektan Turkyilmaz is about the emergence of the music recording industry and Armenian sound connections in the Eastern Mediterranean. During his presentation, Dr. Turkyilmaz illustrated how the production of musical recordings was shaped by cultural diversity and how these recordings can reveal underlying musical networks and cultures in the Mediterranean region. He pointed out that in the Mediterranean, the production of Armenian music was small compared to the large population of the Ottoman Empire. However, the visibility in the musical context of the region was remarkable.
Despite Dr. Turkyilmaz's advances in obtaining late musical records, the challenge of recording quality remains, as evidenced by playing back a song recorded in 1901 that sounded unclear. This aspect leads us to consider the potential for artificial intelligence to improve the sound quality of such historical recordings. This possibility could offer new perspectives for this historian's work. Also, one aspect that I found interesting was his initiative (although he does not know how to do it) to create an archive, considering the amount of musical information he has. The above opens the door to those interested in collaborating with this researcher. Creating a technological platform that allows interactive access to this type of historical recording with high sound quality would be something innovative.
 |
| Dr. Turkyilmaz reproducing an Armenian song from 1901 |
1.4. Marriage Mobilities in post-Ottoman Rhodes, 1912-1943 [ Andreas Guidi, INALCO Paris, Centre de recherche Europes-Eurasie]
Dr. Andreas Guidi presented his project focused on marriage dynamics in post-Ottoman Rhodes. He aims to analyze manuscripts and wedding certificates from that period to find marital connections between individuals. He is currently in the transcription phase and has developed an Excel database with detailed information on 74 records out of 750. Dr. Guidi emphasized in his presentation that, despite basing his analysis on authentic historical documents, there is a particular bias about the reliability of the connections identified. This is due to the possibility that the marriage certificates could contain narratives or testimonies that are not wholly truthful.
Dr. Guidi invited interdisciplinary collaboration to take his research to the next level. Joining efforts between historians and data scientists in this area of study by improving the efficiency of transcription and analyzing the information with data science methodologies can broaden the understanding of social structures as well as marital characteristics and networks of the past.
 |
| Dr. Andreas Guidi shows an example of transcription of Wedding certificates |
1.5. Digital Humanities and the Study of Mediterranean Mobilities [Zeynep Ertugrul, EHESS – CETOBaC]
Dr. Zeynep Ertugrul's research progress focuses on the extraction and analysis of information contained in a collection of documents related to the expropriations that occurred in the 1930s within the late Ottoman Empire. In addition to this collection of documents, Dr. Ertugrul has a large amount of ethnographic and demographic material at her disposal. Her main objective is to delimit the boundaries of towns and villages within the Ottoman Empire, carrying out a comparative analysis with present-day demographic extensions and identifying specific plots within historical settlements. Although the use of maps is an integral part of her methodology, Dr. Ertugrul highlighted a significant limitation in visualizing the boundaries of historical settlements because the available map APIs are not designed for this purpose. This forces her to have to perform visualizations by hand. It is clear to note that this work requires collaboration with other disciplines, especially with experts in digital mapping and geospatial analysis.
AI, VIRTUAL REALITY, MODELING METHODS
During the second day of the workshop, I had the privilege of opening the session. This was my first public research presentation since arriving in the United States. I shared my analysis on a historical dataset provided by one of the historians at the workshop. My presentation integrated exploratory data analysis and AI-generated images of the physical characteristics, professions, and mobilities of Greeks living in the Ottoman Empire. As part of the confidentiality of historians' information, I have not placed access to my slides in this report; however, I can say that as a Ph.D. student, participating in this event represented a unique opportunity that broadened my academic horizons in two main aspects. First, it allowed me to expand my professional network by interacting with renowned researchers in several disciplines. This opened the door to possible future collaborations and knowledge sharing. Second, the experience enriched my perspective and understanding of the importance of interdisciplinary work, especially between historians and computer scientists.
 |
| Jhon G. Botello shows AI-generated images of people's physical characteristics in the late Ottoman Empire. |
 |
| Presentation by Dr. Jose Padilla |
Finally, Dr. Adam Anderson from UC Berkeley closed the series of presentations on day two. His presentation was about the project he leads called "FactGrid Cuneiform." Cuneiform is one of the oldest known writing systems used in late Mesopotamia. FactGrid Cuneiform aims to develop an ontology that provides linguistic support for all languages written in cuneiform while creating interactive web applications, 3D virtual game environments, and archiving of historical information to make this type of data more accessible and attractive. The project hopes to feed and integrate four key databases with cuneiform sources: FactGrid, Wikidata, Zenodo, and GitHub.
After the presentations from day two, a discussion section was opened to debate the topics addressed in the workshop and encourage ideas that would motivate interdisciplinary collaboration to broaden the research horizons of the participating historians. The discussions focused on three aspects that were common to all the presentations and are therefore highlighted as crucial and a priority: 1) challenges and efficiency in transcriptions and transliterations, 2) data processing, analysis, and storage, and 3) overlapping and mapping of information.
The application and advancement in technologies such as optical character recognition (OCR) and natural language processing (NLP) tools, including the use of large linguistic models (LLMs), are avenues for addressing transcription and transliteration challenges in historical studies. These technologies not only have the potential to speed up this process, but also offer the ability to collaboratively correct and refine interpretations, so historians can focus directly on model-enhancing interpretation and analysis rather than getting lost in mere transcription. Although there are studies in this field (like the research of Dr. Adam Anderson), other historical application contexts still need to be explored and improved. Adapting models to particular challenges, such as those related to linguistic diversity and unique historical details, is certainly a promising field for future research and development, highlighting the need for collaboration between different disciplines.
Moreover, in the scenario where historical information is rapidly accumulating as historians work intensively to extract meaningful data in the form of text, audio, or maps, there is a need for advanced tools and software to process, analyze, and store large sets of historical information. This calls for computer scientists, computational social scientists, and data scientists to integrate methodologies and develop tools to analyze patterns of mobility, discover physical features that describe past societies and networks of interactions between people, increase the efficiency of data mining and analysis, and store data that will endure over time.
The workshop highlighted the relevance of FactGrid for archiving historical information. This collaborative online platform is designed to collect, manage, and visualize information from the past, focusing on historical events, people, and places. It uses the same architecture as the Wikidata tool, a collaborative database of interconnected data that powers the free encyclopedia Wikipedia. I talk in a bit more detail about this in my recent blog post for the Storymodelers lab.
Although FactGrid uses an intuitive user interface that allows users to explore, add, and edit data, and considering that Adam Anderson is advancing research with his FactGrid Cuneiform project, there are still challenges. The quality of the deposited information and the technological difficulty it represents for historians becomes a way for other disciplines, such as computer scientists, to find opportunities to contribute.
On the other hand, considering advances in historical data archiving through interdisciplinary collaboration, we can look to the future and two crucial questions emerge that I highlight in my blog for the Storymodelers lab: How can we analyze this wealth of historical data? And how can we transcend language barriers? Historians may face the difficulty of not having adequate methods to extract the information contained in their data, while data scientists may face the challenge of not understanding the language, running the risk of omitting valuable information or misinterpreting it during a translation process. This again underscores the importance of unifying efforts to maximize the potential for extracting information from past data.
Finally, the overlapping and mapping of historical information such as cities and edges of villages that once existed in the past were the workshop's critical and central discussion points. All presentations highlighted this topic as a primary need for historians. These professionals are deeply interested in visualizing and understanding the migrations and connections of the past and recognize the importance of locating historical events and societies geographically. In particular, the available information shows a prominent demand for mapping places and constructing geographic representations of historic towns, neighborhoods, or cities. This approach transcends the traditional narratives to which historians are accustomed and establishes itself as a path toward joint interdisciplinary collaboration. It is a fascinating challenge that represents an opportunity, especially for researchers interested in providing innovative technological solutions that bring the past to life and allow for a deeper exploration of social and geographical dynamics.
Conclusion
In conclusion, my attendance and participation in the Digital Humanities Workshop and the Mediterranean Mobilities 2024 Study gave me a new perspective. It took me out of my comfort zone and allowed me to appreciate how, as computer science researchers, we have a vast field to explore and contribute to historical research. Applying our tools and methodologies can significantly complement the work of all historians, not just those present at the workshop. From my experience, collaborating with historians not only motivates us to look for innovative ways to apply our technical knowledge so that we can communicate our findings to them but also provides us with new ways of application that we might have previously thought were non-existent.
I hope this blog will motivate readers to cooperate with historians and other disciplines outside the comfort zone. It is about more than just creating tools and improving models that contribute to the promising future of artificial intelligence but also about discovering historical events in little or never-explored contexts.
Written By Jhon G. Botello