Answers are grounded in this paper and cite their sources.
Historical archaeology lies at the intersection of the written word, the spoken word, and material things, yet each dataset documents the past unevenly. To accommodate the inherent biases in incomplete historical and material data, we incorporate machine learning algorithms to create more dynamic assessments of places documented on historical maps. By creating a modern dataset based on a triangulation of hybridized data sources, we propose methods to digitally fill in gaps within traditional data sources to identify probable locations of past human activity. Embedding the in-between spaces in our model of historical coastal plantation spaces allows us to more fully engage with interpretations of liminal, disenfranchized spaces, notably the rival geographies embedded within antebellum places. Ultimately, we argue that machine learning is most useful not as a tool to discover archaeological sites, but rather as a tool to identify and highlight subtle spatial relationships, especially at sites where overt cultural practices were surveyed and controlled, such as antebellum plantations.
No takes yet. Share an insight, caveat, or question.
Cochran et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: