Italian Trulli

Unthinking Rubens and Rembrandt: Counterfactual Analysis and Digital Art History

Harm Nijboer (, Huygens ING, Netherlands, The and Judith Brouwer (, Huygens ING, Netherlands, The and Marten Jan Bok (, University of Amsterdam

Netherlandish art from the seventeenth century is closely associated with a few famous painters from that era: Rubens, Rembrandt, Van Dyck, Vermeer and a few others. This is not only true for the general public, but to a large degree for the field of art history as well. A great many books and articles have been written on Rembrandt and Rubens alone. Worldcat, for instance, lists 20 984 publications on Rembrandt and 11 589 publications on Rubens. The enormity of these figures becomes even more striking when we compare them to those of lesser known seventeenth century Netherlandish painters like Hercules Seghers (303 publications), Cornelis de Vos (62 publications) and Daniel Vertangen (1 publication). And when something is written about these lesser known painters it is often in relation or in contrast to Rembrandt or Rubens. To illustrate this: for his recent book (2015) on all history painters from mid seventeenth century Amsterdam, Eric Jan Sluijter could think of no better title than Rembrandt's Rivals.

The omnipresence of Rembrandt and Rubens in our thinking about seventeenth century Netherlandish art has without doubt biased art historical documentation on this period. Data on these two famous painters are readily available, even in digital formats. In an effort to counter this bias towards the famous the authors of this paper have collected and structured biographical data on all (known) painters from seventeenth century Antwerp and Amsterdam in the ECARTICO prosopographical database. If available, we have also collected biographical data on their parents, spouses and often other direct relatives. Person in the data set have been extensively linked to external resources like VIAF, Getty ULAN, Wikidata and more. As a result ECARTICO provides a (partial) virtual model of the social networks that constituted the Antwerp and Amsterdam art worlds in the seventeenth century.

In this paper we present the outcomes of a network analysis on the data derived from ECARTICO. The analysis shows that Rubens and Rembrandt – maybe as one might expect – were indeed key actors in the social networks of their artistic communities. We could be satisfied by this outcome because it once and for all shows that Rembrandt was not a “lone genius.” But we also have to raise the question whether this outcome is not affected by the fact that Rembrandt and Rubens are somewhat over-documented. Furthermore, we should ask the question to what extend the central positions of Rubens and Rembrandt shaped the structure of their artistic communities as a whole.

To address the latter and related questions we have to deal with an important feature of (virtual) models that are not well understood by many historians, namely that models allow for experimentation. Network analysis packages allow users to delete nodes in the network and to rerun the analysis. For this paper we have done such an experiment and deleted Rubens and Rembrandt from their respective social networks. This experiment shows that despite the centrality of Rubens and Rembrandt in the art worlds of respectively Antwerp and Amsterdam, the social structure of these milieus was not entirely centered around them. Furthermore it will highlight the independent centrality of some painters that operated outside the sphere of influence of Rubens and Rembrandt.

We are aware that such an experiment will meet skepticism by many historians (art historians included) or – even worse – will be qualified as “iffy history.” However, we argue that such a counterfactual analysis is not intended to get an impression of a world without Rembrandt and Rubens. On the contrary, the whole experiment is intended to get a better understanding of the impact of Rubens and Rembrandt on their artistic communities. Furthermore, it might provide us some assistance in assessing potential bias in the underlying data. Last but not least there is an epistomological argument. Sunstein (2016) recently argued that all historical explanation implicitly involves counterfactual reasoning. Due to the ongoing digitization of sources and data, (art) historians will increasingly have to deal with models of the past. It is a great opportunity to make our counterfactual reasoning explicit.

Appendix A

  1. Sluijter, E.J. (2015). Rembrandt's Rivals: History Painting in Amsterdam (1630-1650). Amsterdam/Philadelphia: John Benjamins.
  2. Sunstein, C.R. (2016). Historical Explanations Always Involve Counterfactual History. Journal of the Philosophy of History. 10(3): 433-440.