How We Research: Inside Art Revisionist’s Methodology

Transparency in methodology is not a luxury in scholarship — it is a requirement. If our conclusions cannot be independently verified, they are opinions, not findings. At Art Revisionist, we believe that showing our work is as important as the work itself.

This article describes the research methodology we apply across our investigations, from European bronze authentication to African art provenance. The specific techniques vary by project, but the underlying principles remain constant: primary sources first, transparency always, and conclusions that follow evidence rather than assumptions.

Phase One: Archival Research

Every investigation begins with documents. We access civil archives (birth, marriage, and death records), commercial registries, notarial records, customs documents, and institutional archives across multiple countries. For our Valsuani research, this meant working in Italian municipal archives, French business registries, and Swiss banking records — each providing pieces of a puzzle that no single archive could complete.

We cross-reference every significant claim against at least two independent primary sources. The standard in the published literature — accepting a claim because it appeared in a previous publication — is explicitly not our standard. As demonstrated in our investigation into who actually founded the Valsuani foundry, even universally accepted attributions can be wrong.

Phase Two: Material Analysis

Archival research tells us what people wrote and said. Material analysis tells us what the objects themselves reveal. We employ non-destructive analytical techniques including X-ray fluorescence for alloy composition, high-resolution photography and 3D scanning for surface analysis, and ultraviolet and infrared imaging for detecting repairs and alterations.

For bronze authentication, we work within the framework described in our technical analysis of the Valsuani casting method, comparing suspect works against authenticated reference examples using both traditional connoisseurship and computational analysis.

Phase Three: Contextual Research

Objects exist in cultural contexts, and authentication requires understanding those contexts. For European bronzes, this means understanding the business practices of foundries, the relationships between sculptors and founders, and the evolution of the art market over time.

For African art, contextual research is even more critical. Our work on Senufo art requires deep engagement with Senufo cultural systems, including the spiritual practices, social structures, and artistic conventions that shaped the objects we study. This is not supplementary information — it is essential to understanding what the objects are and what they mean.

Phase Four: Computational Analysis

Our AI research program applies machine learning and data analysis to art historical questions. Computer vision models trained on authenticated examples can identify stylistic patterns, detect anomalies, and flag potential issues that warrant closer investigation. Natural language processing helps us search and analyze archival documents at scales impossible for human researchers alone.

We are careful to position these tools correctly: they are aids to scholarship, not substitutes for it. An AI model can identify a statistical anomaly in a foundry stamp, but a human expert must interpret what that anomaly means in context.

Phase Five: Peer Review and Publication

We submit our findings to scrutiny before publishing them. Internal review by our research team is followed by consultation with external specialists — curators, conservators, art historians, and, where relevant, representatives of source communities. This process is time-consuming but essential: conclusions that cannot survive challenge are conclusions that should not be published.

Our research services for private clients follow the same methodological standards as our published scholarship. Whether we are investigating a single bronze for a collector or mapping the provenance of an institutional collection, the process is identical: archival first, material second, contextual third, computational fourth, peer-reviewed fifth.

What We Don’t Do

Methodology is also defined by its limits. We do not offer authentication opinions based solely on photographs — material analysis requires physical access. We do not confirm attributions that our evidence does not support, regardless of commercial pressure. And we do not present speculative conclusions as certain ones.

Our commitment, as stated on our about page, is to evidence-based art history. This means accepting that some questions cannot yet be answered, that some attributions must remain provisional, and that honest uncertainty is more valuable than false confidence.