Are We All Archivists Now? Provenance Provocations for Information Visualization

Authors

Anamaria Crisan (University of Waterloo)

Keywords

Position Paper, Provenance Capture, Graphical Histories, Human-AI Teams, Information Visualization

Abstract

Provenance is commonly understood as a record of the processes by which an analysis was produced, and provenance research has accordingly focused on faithfully capturing interactions, workflows, and computational steps to support reproducibility and reuse. We argue that this framing overlooks a more fundamental question: what evidence of an analysis should be preserved in the first place? Drawing on archival science, history, and human sensemaking, we reframe provenance capture as an epistemic and archival design problem rather than a purely technical one. We examine it through three questions, concerning the sources of provenance, the selection of observable traces, and the timing of capture, and offer a set of provocations that expose assumptions embedded in current provenance systems, with implications for both human- and AI-assisted analysis. Provenance systems, we argue, do not simply record analyses; they shape the evidence from which analyses are later interpreted, reproduced, and understood.