Parliament's Debates about Infrastructure: An Exercise in Using Dynamic Topic Models to Synthesize Historical Change.

Area of Science:

• Digital Humanities
• History of Technology
• Computational Social Science

Background:

• Traditional historical research relies on secondary sources and manual analysis.
• The potential of "distant reading" and digital tools for uncovering hidden patterns in historical data remains underexplored.
• Nineteenth-century British parliamentary debates offer a rich dataset for studying imperial infrastructure development.

Purpose of the Study:

• To investigate whether distant reading and digital tools can enhance the history of technology.
• To compare the efficacy of digital tools against traditional secondary sources in historical analysis.
• To identify hitherto undetected patterns in the history of British imperial infrastructure.

Main Methods:

• Application of topic modeling and dynamic topic modeling to synthesize large historical datasets.
• Analysis of digitized British parliamentary debates from the nineteenth century.
• Comparative analysis of results generated by digital tools versus established historical accounts.

Main Results:

• Digital analysis largely corroborated known turning points, key players, and technologies in British infrastructure history.
• Digital investigation revealed novel insights in specific areas: River Shannon improvements, political party alignment with technologies, and the chronology of public space construction in imperial capitals.
• Topic models were validated as a reliable source for periodizing technological history over time.

Conclusions:

• Distant reading and digital tools offer valuable complementary methods for historical research, particularly in the history of technology.
• Computational methods can uncover new historical narratives and refine existing understandings of technological development and imperial history.
• The study validates the use of topic models for periodizing technological change and identifying nuanced historical patterns.