Anthology of Computers and the Humanities · Volume 3

Framing the Canon: A Computational Study of Canonicity in Danish Golden Age Paintings (1750-1870)

Louise Brix Pilegaard Hansen1 , Rie Schmidt Eriksen1 , Pascale Feldkamp1 , Alie Lassche1 , Kristoffer Nielbo1 , Katrine Baunvig2 and Yuri Bizzoni1

  • 1 Center for Humanities Computing, Aarhus University, Aarhus, Denmark
  • 2 Center for Grundtvig Studies, Aarhus University, Aarhus, Denmark
Download PDF Download Citation

Permanent Link: https://doi.org/10.63744/KTLpQIY247dD

Published: 21 November 2025

Keywords: art historical canon, computational art analysis, image embeddings, canonicity

Abstract

This paper investigates the mechanisms of canon formation in Danish Golden Age paintings (c. 1750-1870). We aim to assess whether we can link canonical status to quantifiable intrinsic aesthetic traits or whether the canon is the result of extrinsic processes. Drawing on recent studies of literary canonicity and earlier work in art history, we extend the methodological scope of canon studies by analyzing 1,656 paintings from the Danish National Gallery using image embeddings. In both synchronic and diachronic clustering experiments, we find no distinct separation between canonical and non-canonical works, nor any indication that canonical paintings prefigured artistic innovation. However, above-chance performance from a binary classification task of canonical vs non-canonical artworks suggests subtle intrinsic differences between the two groups.1

  1. See github repository, https://github.com/centre-for-humanities-computing/canon-paintings-smk, for data and code.↩︎