What's Happening?
Texas State University (TXST) is leading a research initiative, "Sondondo Valley. Population and Territory in Rural Andean History," that employs artificial intelligence (AI) to analyze digitized Peruvian parish records from the 18th to early 20th centuries.
This project, funded by a $23,400 Research Enhancement Program grant from TXST's Division of Research, aims to overcome the historical challenge faced by researchers in connecting scattered handwritten records of baptisms, marriages, and burials. Historians previously found it nearly impossible to trace the lives of individuals and families over time due to the sheer volume and disorganization of these documents. The TXST team, including graduate students Vamsi Krishna Samavedam and Sneha Tulshiram Nitnaware, is developing computational programs to interpret these records. These programs are designed to identify when multiple entries, such as a baptismal record, a marriage record, and a burial register, belong to the same individual, even with inconsistencies in spelling and handwriting. The project began in 2019 with Peruvian researchers digitizing the parish record books, and in 2023, a team processed some of these records into a computer database.
Why It's Important?
This project holds significant importance for historical research and the broader application of AI in humanities. By enabling the linkage of individual records across generations, it will allow historians to ask new questions about migration patterns, family relationships, disease prevalence, demographic changes, and community formation in rural Indigenous Andean communities. The ability to trace an individual's complete history from baptism to burial will provide unprecedented insights into the social and cultural dynamics of the region during the Spanish colonial and modern national periods. Furthermore, the methodology developed by TXST is designed to be replicable and adaptable, potentially benefiting other researchers studying similar historical records globally. This initiative showcases the growing impact of digital humanities and the transformative potential of AI in making vast, previously inaccessible historical data interpretable, thereby enriching our understanding of human history and societal evolution.
What's Next?
The TXST team plans to continue refining its AI model to improve accuracy in identifying individuals across various records, accounting for factors like temporal feasibility and genealogical constraints. The long-term vision for the project includes creating an open-access website. This platform will allow researchers, educators, and the public to search for individuals, places, surnames, and historical events within the digitized records. The website is also expected to feature interactive maps to visualize migration patterns across the Andes and network diagrams to illustrate family relationships spanning multiple generations. This will not only make the research findings widely accessible but also empower other scholars to utilize the data and methodology for their own studies. The project's lead, De la Puente, has already explored applying this model to records of Native Americans who lived in the San Antonio Missions, indicating a potential expansion of its application to other historical datasets.
Beyond the Headlines
The TXST project delves into the ethical and methodological considerations of using AI for historical research, particularly when dealing with sensitive and personal data from past populations. The challenge of accurately linking records despite variations in handwriting and spelling highlights the need for sophisticated AI algorithms that can infer connections probabilistically rather than relying on rigid pattern recognition. This approach acknowledges the inherent imperfections in historical data and seeks to build models that learn from these imperfections. The project also underscores the collaborative nature of modern historical research, combining the expertise of historians with data scientists and machine learning specialists. The creation of an open-access platform for this data raises questions about data stewardship, accessibility, and the potential for new interpretations and narratives to emerge from a broader engagement with these historical records. It represents a significant step towards democratizing access to historical information and fostering interdisciplinary research.
