Recent research from Valentine Figuroa at the Massachusetts Institute of Technology (MIT) highlights how advancements in machine learning can enhance the analysis of large-scale visual data, particularly in the field of historical political economy. The study emphasizes that paintings housed in museums and private collections represent a significant yet underutilized resource for understanding political narratives and cultural transformations.
To effectively leverage this visual data, Figuroa proposes a comprehensive framework that establishes criteria for assessing the information encoded in paintings. This framework is grounded in the traditional humanities, addressing critical questions about the assumptions made during interpretation. The research applies this methodology to a database comprising 25,000 European paintings spanning from 1000 CE to the First World War.
Three Applications of the Framework
Figuroa’s study presents three distinct applications that illustrate how paintings can convey various types of information. Each application focuses on a different aspect of cultural and political change during the early-modern period.
The first application revisits the concept of a European “civilizing process.” This term refers to the internalization of stricter behavioral norms that coincided with the expansion of state authority. By analyzing paintings depicting meals, the research investigates whether there is an observable increase in the complexity of etiquette over time. This exploration could provide valuable insights into how social customs evolved alongside political structures.
The second application centers on portraits, examining how political elites crafted their public personas. The analysis reveals a long-term transition from chivalric representations to more rational-bureaucratic depictions of men. This shift reflects broader societal changes and the ways in which power dynamics influenced artistic expression.
The third application focuses on a significant trend of secularization, as evidenced by the declining proportion of religious paintings. This trend began prior to the Reformation and accelerated in its aftermath. By documenting this process, the research sheds light on changing cultural values and the diminishing role of religion in public life.
Figuroa’s innovative approach not only redefines how visual data can be interpreted but also opens new avenues for understanding historical political contexts. As machine learning continues to evolve, the potential for further discoveries in political history through art remains vast. The findings underscore the importance of integrating traditional humanities perspectives with modern analytical techniques.
Through this research, Figuroa contributes to a growing body of work that seeks to bridge the gap between visual culture and political history, offering fresh insights into the complexities of human behavior and societal evolution.
