Styles Revisited: From Iconology to Digital Image Studies


What is style? Even though the concept of style is ubiquitous in art history, there seems to be no single definition on which art historians can agree, and despite many attempts at clarifying its uses and connotations, its meaning remains elusive. As James Elkins concluded in his article for the Grove Dictionary of Art: “The further the concept of style is investigated, the more it appears as an inherently partly incoherent concept, opaque to analysis.[1]”

The historiography of the term, at least in the West, is well-known albeit complex. Following Johann Joachim Winckelmann’s use of style to discuss ancient Greek art,[2] the history of art became to a certain extent a story of successive rises and falls of the classical ideal (from Ancient to Medieval art, from Renaissance art to Baroque, etc.), and the organic vision of early, middle, and late style came to be applied not only to period styles but also to artists’ individual styles. Discussing and arguing about styles, and disagreeing about them, became a habitus of the discipline, so much so that it is still difficult today to say whether one is making a de re or de dicto statement while referring to style, be it of a period or individual.

The way in which historians have studied styles has also changed their understanding of the term. In the second half of the 19th century, photography transformed the investigation of styles by enabling the careful examinations of visual markers present in a work of art.[3] The examples of Alois Riegls’s Stilfragen: Grundlegungen zu einer Geschichte der Ornamentik (Problems of Style: Foundations for a History of Ornament, 1893) and Heinrich Wölfflin’s Kunstgeschichtliche Grundbegriffe (Fundamental Principles of Art History, 1899)[4] steered several generations of scholars towards the study of stylistic diffusions in strictly visual terms, from artefact to artefact, from image to image, and even from form to form. This approach led some art historians to think about form as an autonomous quality, detached from historical and social processes. Aby Warburg’s iconology is certainly the most interesting outcome of such thinking.[5] Warburg advanced the hypothesis that Pathosformeln circulate in images and styles, and that they conceal information on collective phobias, desires, and expectations.[6] His iconological studies resulted in a theory of symbolic forms that was further developed as a philosophy of symbolic meaning in Ernst Cassirer’s Philosophie der symbolischen Formen.[7]

Compared to earlier investigations, Warburg’s iconology ultimately had a limited impact on art history, and since his time the field has yet to see a new general theory of style, or of the circulation of style. The reticence of Warburg's successors in developing a science of styles was not unreasonable. The problem at the core of a general iconology was its usage of and dependency upon extensive chronologies, which were so consequential that they made it humanly impossible to deal with sources other than images. As a result, not enough attention could be paid to how the development of styles was shaped by historical, social, economic, and political factors —factors that can be found in sources that are not strictly visual.

This is probably the reason why art historians tend to approach the question of style and form through the singular object (or person). In an essay on style published in 1961, Meyer Schapiro recognized the gap between the study of form through the singular object on the one hand, and the approach through "broad similarities" spread over space and time on the other.[8] Though the study of form through singular objects is possible for individual scholars, the study of the circulation of forms on large scales represents a titanic undertaking. Beyond the reach of a discipline that chose qualitative monography over quantitative studies, the question of style has long been abandoned to customs and common sense.

Despite this, the dialogue around style has continued to emerge across many aspects of art history even when it is not figured as the main object of study: here we need only think of the relationship between pictorial style and history “meant to show how the style of picture is the proper material of social history”, as Michael Baxandall put it;[9] we could also look to the frequent adaptation in the discipline of methods based on renaissance-centric stylistic analysis.[10] Each of these new investigations has reshaped the meaning of style, without tackling the problem of what composes a stylistic unit as such.

Considering the ambiguous character of the concept of style, is style best left alone? However tempting it might be to eliminate such an ambiguous notion from our vocabulary, the imperative of constructing a more global history of art compels us to return to styles, their circulations, and their transformations. In recent years, many art historians have attempted to study styles (in particular modern ones) on a global scale.  A host of projects and exhibitions on international cubism, abstraction, or pop have looked to shed light on the contributions of lesser known artists and regions to these well-established styles. While those projects are incredibly valuable, one cannot help wondering if, through their very reliance on style and stylistic diffusion, they do not ultimately reinforce the center-peripheries model and the pervasive hierarchies they are trying to undo. When we use period styles as a form of classification, is there not a risk of neglecting important microvariations which comprise a corpus of images? With this in mind, rather than leaving style alone, perhaps we ought to revisit it in a more critical and rigorous fashion.

Style, and specifically style classification, has also been the subject of many recent undertakings in the domain of computer vision and machine learning. Recent developments have made it possible to classify images with respect to their style, clustering representations based on a set of recognized dimensions. This computational synthesis escapes the logocentrism of previous approaches, instead concentrating on the recognition of salient elements repeated across a corpus of images. Such an approach obviously opens a new array of challenges and questions that need to be met and answered. What is style for an algorithm? Which algorithms are currently available to classify images according to styles and what are the relevant dimensions they record? How is the saliency of these dimensions encoded and used? How dynamic is a style, and how to record it? How reproducible are the results obtained by algorithms? What are the corpora on which these algorithms have been trained? Who describes these corpora, and how can we deal with differences in quality between them? What are the biases of these corpora and how do they affect the final results?  Are aesthetic qualities the only components of a style, or we should consider a larger spectrum of possibilities (e.g. representational tropes, technical methods)?

Despite (or because of) all these blind spots and limitations, these recent investigations open up many new opportunities, and not least the possibility of attempting to construct the general theory of styles Warburg called for in his work. It is in this context that Artl@s/Visual Contagions intends to devote its 2021-2022 research seminar to the ubiquitous, elusive, but thought-provoking notion of “style.”

The seminar will take place online every other Mondays at 2pm (GMT+1). The dates a priori considered are: September 20, October 18, November 15, December 13, 2021 and January 24, February 7, March 7, April 11, May 16, June 13, 2022.

[1] James Elkins, “Style (2003),” Grove Art Online

[2] J. J. Winckelmann, Gedanken über die Nachahmung der griechischen Werke in der Malerey und Bildhauerkunst, Dresden, 1755, and Geschichte der Kunst des Alterthum, Dresden, 1764.

[3] On photography and art history, see Wolfgang M. Freitag, “Early Uses of Photography In the History Of Art,” in: Art Journal 39 (1979-1980), 20-119; and for a synthesis: Ingeborg Reichle, "Photographie und Lichtbild: Die 'unsichtbaren' Bildmedien der Kunstgeschichte", in: Anja Zimmermann (ed.), Sichtbarkeit und Medium. Austausch, Verknüpfung und Differenz naturwissenschaftlicher und ästhetischer Bildstrategien (Hambug: Hamburg University Press, 2005), 169-181.

[4] Alois Riegl, Stilfragen: Grundlegungen zu einer Geschichte der Ornamentik (Berlin: George Siemens, 1893); Heinrich Wölfflin, Kunstgeschichtliche Grundbegriffe: Das Problem der Stilentwicklung in der neueren Kunst (Munich: F. Bruckmann A.-G., 1899). Alois Riegel, Historische Grammatik der bildenden Künste, posthume edition by K. M. Swoboda and O. Pächt, 1966 (Sesto San Giovanni: Mimesis Verlag, 2017, introduction Andrea Pinotti).

[5] See (consulted 19 December 2019).

[6] Georges Didi-Huberman, L'image survivante : histoire de l'art et temps des fantômes selon Aby Warburg (Paris:  Les Éditions de Minuit, 2002).

[7] Ernst Cassirer, Philosophie der symbolischen Formen, 3 vols. (1st ed. Berlin: Bruno Cassirer, 1923–1929. Last ed. by Claus Rosenkranz: Ernst Cassirer / Gesammelte Werke, vol. 11–13 (Hamburg: Meiner, 2001-2003).

[8] According to Donald Preziosi, 142. Quoting Meyer Schapiro, ‘style’, in Morris Philipson, Aesthetics Today (Cleveland and New York, 1961), 97.

[9] Michael Baxandall, Painting and Experience in Fifteenth Century Italy: A Primer in the Social History of Pictorial Style (Oxford: Oxford University Press, 1988).

[10] Svletana Alpers, The Art of Describing: Dutch art in the Seventeenth Century (Chicago: University of Chicago Press, 1983).