Reconstructing Historical Events with CIDOC CRM

Historical sources frequently contain uncertain, approximate, and ambiguous temporal information, posing a significant challenge for their representation in computational systems. This paper addresses this challenge by demonstrating how semantic web technologies and an event-centric ontology can systematically represent the pervasive temporal uncertainty inherent in primary sources. Using the CIDOC CRM ontology, this study outlines a structural approach to transforming linear historical narratives into explicit, machine-interpretable graphs. Drawing on the professional and commercial networks of the Georgian and Victorian British music trade as a case study, the paper identifies five distinct forms of temporal uncertainty: circa dates, bounded dates, time-spans, relative dates, and inherently fuzzy historical eras. Through the strategic interplay of CIDOC CRM’s four temporal boundary properties (P82a, P81a, P81b, and P82b), the paper demonstrates how varying levels of temporal precision and modeling-imposed constraints can coexist within a single interval-based representation. Ultimately, this research argues that formalizing temporal uncertainty into explicit semantic statements does not flatten humanistic nuance. Instead, it provides a structural framework that makes complex historical data actionable for advanced workflows, such as automated reasoning and large-scale pattern analysis.