Data: Image & Object

Le Journal pour rire, 25 June, 1853, cover

Le Journal pour rire, 25 June, 1853, cover

By Julia Langbein

On June 25, 1853 the Parisian Journal pour Rire published this colourful cover: each little square caricatures a specific painting concurrently on view at the Salon, the annual or biennial exhibition of fine arts in Paris.  This is one of my “images”– it is “evidence” for arguments about the relation between, say, press technology and painting in the nineteenth century, about the role of the comic in cultivating a critical beholder, and about the lore of a laughing public so essential to avant-gardism.

At some point in the recent past I started referring to images like this as my “data.” Why?  This kind of image was produced in an expanding popular press, and I have a thousand like it, in digital form, in my own database.  Art historians from across the field will sympathize: Research possibilities have been cracked wide open by the digitization of print material and the disciplinary embrace of “visual culture,” but hard drives are groaning.  We have practical problems of processing and storage, but there are deeper problems of method.  How do people trained in the close visual analysis of single images make arguments with massive amounts of stuff? Is “data” a thing, or a way of thinking, and if it’s the latter, do I need (yet more) training? What kind (and will it take another ten years)? Am I still doing art history?

These were precisely the kinds of problems addressed by participants from a range of disciplines at this year’s Image and Object workshop on the theme of “Data.”  Remind me not to complain about data storage around Felix Hofmann, from Engineering Science, whose elegantly graphic X-ray diffraction image of the structure of a crystal can be made by the terabyte daily.  Classical Archaeologist Alexandra Sofroniew reminded us that though “data” often implies a particularly modern kind of information–digital bits and bytes–archaeology’s data problem is as earthbound as it gets.  A superabundance of potsherds raises ethical questions about what to keep, and methodological issues when sheer quantity is mistaken for comprehensiveness (the problem of too many potsherds doesn’t solve the problem of gaps in the knowledge they produce.)

This tension between “data” as abundant and as significant has something to do with the question that this conference raised most clearly for me (and for many others, I gleaned over coffee.) “What is the difference between something like ‘evidence’ or ‘archive’ and ‘data’?”  Perhaps the key image of the conference, a touchstone for many presenters, was the one shown by convenor Mirjam Brusius of the archaeologists in the colonnades of the Alte Nationalgalerie in Berlin picking over the bricks hauled back from Babylon in 1927/28, trying to piece together what would become the famous Ishtar Gate of the Pergamon museum. What is so compelling about this image is the contrast between the heaps of jagged, fragmented earth, and the promise– or fantasy– of its ultimate architectural coherence.  This idea of the completed wall touches on the shift in expectations that occurs when one starts calling one’s material “data.”  Evidence is targeted and selective, archives are scattered messes, but “data” casts a ghostly shape of future or virtual or unseen coherence, even if one doesn’t have all the pieces.  Even the collective singular noun hints that these fragments, digital or dug up, are part of a conceptual wall; the “data set” is an archipelago nation.  Although as Brusius reminded us, while the Ishtar Gate may have come together as an impressive architectural whole, it differs significantly from an original we’ll never know completely– it remains a fantasy of coherence.

It’s counterintuitive to think of data as a subtle smuggler of fantasy– don’t we often use the term when we want to sound empiricist?  It strikes me that my adoption of the term “data” for my archive began when I started talking to people outside my own field (e.g. economists, librarians) about how to build a searchable database of this comic press material.    “Data” signals a desire for convertibility outside a given discipline– it converts my colourful drawing into the stuff of economics– and it frees up my images to be “raw” data for someone else’s argument.   “Evidence” locks the image into my own interpretation, but creating searchable “data” seems like a gift to the field.

Ethics came up explicitly only briefly in Brusius’ opening remarks, but ethical ambiguity suffused the conference.  Which potsherds should be kept, which discarded? Benjamin Henning argued that his cartographic visualization of layers of complex data can penetrate the text-weary eyes of bureaucratic “decision-makers,” implying that they’ll be pushed in more ethical directions.  But in what direction does data’s promise of significance, utility, and quantifiability push art historians?

Jacques-Louis David, The 'Envois de Rome'

Jacques-Louis David, The ‘Envois de Rome’ of 1808, 1808, pencil on paper, Ecole Nationale Superieure des Beaux-Arts

Camille Matthieu (Art History) presented an 1808 drawing by Jacques-Louis David in which he copies figure studies sent back by his students from Rome to Paris.  A “data-driven analysis” of these studies, one that quantified gender, would plot the shift from male to female figure studies. And with this data to ground us, we could appreciate better the “revolutionary” nature of the irruption of a single female figure in the top right corner of this particular drawing so early in the century.  This is David’s record of a single female submission by Ingres, who originated this academic turn toward the female nude.  In the end, Matthieu teased out the drawing’s strangeness and significance–the way the female nude was pure contour as opposed to the shaded men, her quarantine in the upper right corner, her muteness and near-androgeny– without much recourse to data.  Data-driven analysis, while tempting, is hypothetical in this case anyway, in that the archives of the École Nationale Superieur des Beaux-Arts are hopelessly patchy.  There’s no question art historians have much to gain from learning how to manage, shape, and show large quantities of (often digital) information; but we should also be aware of wanting to invoke “data” because of pressure to promise the coherence and significance of an Ishtar Wall (a trophy for the Pergamon), when there is so much to get from a brick.

Julia Langbein is a Junior Research Fellow in Art History at Trinity College, Oxford