Disaggregating the aggregate : the question of measures in quantitative art history

1 The discipline of art history has been slow to recognize the potential value of quantitative approaches to the study of art. Moreover, the habit of systematic generalization has become increasingly rare among art historians. Even when they discuss economic behaviors, which for obvious reasons have been central to existing quantitative approaches to art, many art historians tend to treat the object or individual under study for its or his/or her unique characteristics rather than to look for shared characteristics and/ or behaviors.

2The relevance of quantitative research to art history has also been handicapped by the tendency for this research to treat art and artists in the aggregate, as economists treat commodities. Two basic elements of art history get lost in such an approach. The first is the historical dimension. Art acquires its value over time. There are real and vital differences between the value attached by a culture to artworks at the time of their making and the value such objects have today. Second, prevailing economic approaches to art and art markets tend to account in only a limited fashion for the variety of artistic behavior. The problem for quantitative art history as I see it is where do we find measures that are more consistent, more relevant, to the discipline as it is currently organized.

3Both qualitative and quantitative art scholars fail to take into account the strongly hierarchical nature of the discipline, especially as it pertains to Western art since the 15^th century. At least since Giorgio Vasari published his biographies of great Italian artists in the middle of the 16^th century, art history and the market for important art have been structured around a small succession of great masters, surrounded in their respective periods and locales by a great many more minor masters. In this discursive pyramid (fig. 1) Western art history has been divided between a few important artists and artworks at the top, commanding the most scholarly attention, and a mass of art objects and art journeymen below, with a considerable distance separating the two strata.

Fig. 1

4Qualitative art historians (and the general public) tend to regard masterworks and master artists as unique. Quantitative art historians have worked primarily with archival materials largely devoted to economic transactions, which yield their best results about the mass of art works and artists at the bottom of the pyramid. One would say about the art at the top of the pyramid that one could not put a price on it. At the bottom of the pyramid, however, the situation is reversed and art is treated as a commodity like any other.

5Art historians have failed to appreciate the size of the gulf that separates important art and artists from the rest. Let me demonstrate this point by looking at the canon of 17^th-century Dutch art. To create table 1, I counted all the illustrations devoted to leading Dutch artists reproduced in 18 European and American textbooks that treat 17^th-century Dutch painting2. Three artists, Rembrandt, Vermeer, and Hals received considerably more illustrations than all the rest.

Table 1. Dutch artists ranked according to number of illustrations in 18 European and American textbooks (minimum of two illustrations).

6Using this table as my guide I then consulted the Getty Research Institute’s library holdings of monographs on all the artists in table 1. I chose the Getty library for its inclusivity, owing to its perhaps unrivalled acquisition budget for monographs and exhibition catalogues published throughout the Western world.

7I counted all the monographs listed under the Library of Congress subject heading for the respective artists in the Getty collection. While the order at the bottom end of Dutch artists is rearranged somewhat from table 1 (which is to say that one artist is not significantly more important than the others at the base of the pyramid), table 2 is still dominated by the same three names. Rembrandt, who had three times as many illustrations as any other Dutch artist in table 1, appears in table 2 with more than twice as many monographs than all the other Dutch artists combined. Hals and Vermeer, though far below Rembrandt in the number of citations, remain substantially higher than other Dutch painters. The gap between the top and the bottom in table 2 is even larger than in table 1.

Table 2. Dutch artists ranked according to number of Library of Congress subject citations listed in the Getty Research Institute Library collection (October, 2008).

8What the discursive pyramid (fig. 2) represents in fact is the very close correlation between the discipline’s appreciation of the particular importance of individual artists and works of art and their market evaluation. We see these differences in valuation expressed in the quantity of literature devoted to an artist. Scholarly attention corresponds to an artist’s long-term reputation, reputations often acquired over several or more centuries. Artists’ reputations are reinforced by the world’s museums, which privilege some art and artists over others, especially the great museums, which possess holdings far larger than their available exhibition space. Not surprisingly, since the discipline of art history arose out of the art market through connoisseurship – putting names to art – market values closely follow scholarly and museological appraisals.

Fig. 2

9What are these valuations based on ? I would argue that the pyramidal structure of Western art exclusively reflects the Western cultural tradition of valuing innovation as the soledeterminant of artistic importance. What I am saying is that insofar as post-Medieval Western art history constitutes a set of interlocking diachronic narratives, the single element that privileges a work of art or an artist within these narratives has been the perception of innovation. At the bottom are the journeymen who work within and maintain relatively stable artistic traditions. At the top are the masters who are perceived to have made important changes to traditions and practices so that they redefined the behavior of subsequent generations of artists.

10In pre-Renaissance Europe, innovations were most often collective and anonymous. Such innovations were likely to represent the accumulation and refinement of practical knowledge, comparable to the unknown medieval inventors of the windmill, or what in English we would call « know how ». In the 15^th century innovations began to acquire authorship, as artists and art theorists struggled to raise the status of artists above the level of mere craftsmen. The rise of the artist as a creative individual is reflected in the corresponding rise of copyrights, patents, and the development of what we would call today « intellectual property rights3 ».

11It is important not to confuse artistic innovation with more widespread notions of artistic originality and artistic progress. Innovations are never original ; they are always creative recombinations of existing ideas. While innovations may progress, innovations (and innovators) may also revert to older stylistic and behavioral practices (which is why artistic innovations are fundamentally different from innovations in science, where scientific theories, if sufficiently proved to be consistent with nature, cannot be abandoned unless more powerfully explanations are discovered). In art, innovation represents change within a specific disciplinary practice, which is picked up and modified by subsequent innovators. Innovations can be technological, altering the means by which works of art are created. But, following the rise and dissemination of Renaissance humanism, innovation primarily came to be perceived as an intellectual attribute of art, or « intellectual property », rather than simply the display of craft skills. The importance of any innovation is not intrinsic to the invention of the artist, but arises from how much, how long, how widely that invention is valued by a larger community of practitioners (fig. 3). Innovations change the conditions under which and through which later innovators work. Sometimes the importance of an invention or inventor is immediately recognized. However, because an innovation’s significance is only realized through use, it can sometimes take a very long time before other important innovators take up an invention. Finally, it is important to note that present innovations are always refashioning our perception of past innovations.

Fig. 3

12Art history’s failure to recognize how the perception of innovation has structured the post-Medieval Western conception of artistic importance may be attributed, I believe, to the fact that art history has largely been read as a history of style. Because of the discipline’s fixation on style, art historians, as well as the general public, often make the mistake of thinking important art is the same as the most beautiful works of art or equate important art with that which most powerfully expresses a culture. In practice, while beautiful or powerfully expressed ideas often do constitute an artwork’s innovation, they are not necessary requirements for important art. Marcel Duchamp’s readymades forced the realization that artistic innovation need not have an aesthetic dimension at all. This is also an example of how more modern art redefines our conception of older art. And it is often the case that very beautiful works of art are not very innovative and therefore are not regarded by the discipline as a whole as being very important. Similarly, minor cultural products have often been shown to better reflect a culture than its highly exceptional masterworks.

13Given my claims about the discursive pyramid in art history and about how the perception of innovation is what establishes an artist or an artwork’s importance, I want now to extend my arguments in reference to qualitative and quantitative arguments that are to be discovered in recent literature on 17^th-century Netherlandish art. I begin with a qualitative art historian, one of my former teachers, Svetlana Alpers. In her fine book on Rembrandt, Rembrandt’s Enterprise, Alpersproposed to explain much of the artist’s behavior and the qualitative features of his art in relationship to his market4. But because Alpers hoped to explain what was unique about Rembrandt’s art, that is to say, how and why Rembrandt was exceptional among 17^th-century Netherlandish artists, she repeatedly represented Rembrandt’s market behavior as if it too was unique. So, for example, she argued that the youthful collaboration/competition between Rembrandt and Jan Lievens, with whom he may have once shared a studio, not only was unique to 17^th century Dutch art, it was so remarkable for the quality of the art produced and for the ambitions expressed by both artists that Alpers could think only to compare their relationship to that of Picasso and Braque’s during their discovery of Cubist practices5. To make this claim Alpers chose to ignore all the known instances where important Dutch artists worked in close proximity to other important Dutch artists, either as teachers or colleagues or both. To put it even more strongly, Alpers appears not to have realized that it is always or almost always the case that important, innovative artists work in close proximity to other important artists.

14Communities of important artists create strong ties that provide mutual support, establish research agendas, and provide the competitive push to innovate. Any reader of Vasari can hardly fail to observe the repeated close relationships between important artists described in his narrative. Yet perhaps we should not fault Alpers, because even quantitative researchers have failed so far to measure this phenomenon, or even, to ask the question, why do great innovators always grow out of environments of important collaborators ? Imagine the general resistance among humanists to the idea of quantifying art, if even a great scholar like Alpers can be blind to simple patterns of artistic behavior.

15For the other side of the equation I take the example of one of the most influential quantitative scholars of 17^th-century Dutch art, John Michael Montias6. In Montias’ work, and subsequent scholarship he inspired, both the strengths and the limitations of current quantitative practices are exposed. As an economist turned art historian, Montias made several important contributions to our knowledge of the period. He identified Vermeer’s patron and established quite precisely the environment in which Vermeer worked. More generally speaking, it is possible that it was largely because of Montias that the richest quantitative studies in Western art history relate to the 17^th century or earlier. Certainly scholarship on early modern European art and artists since Montias has been much less prejudiced in regard to the economic lives of artists than, say, scholarship on Western, post-17^th-century art.

16Montias’ work is impressive for its archival depth and breadth. But Montias’ contributions have also been limited by his data. At best, he was able to account for short-term reputations and value. In what I take to be a typical essay by Montias, he wrote about the role of art dealers in the 17^th-century Dutch market, in which he fixed to some degree the kinds of works dealers sold and the variety of dealer relationships with artists7. More importantly, we learn from Montias how, in wealthy Holland, possessing multiple art centers, art dealers primarily served their local market. This tendency became even more localized as the century progressed. Over the course of the century art dealers responded to increased demand for works by named artists. But even late in the century, dealers continued to sell in large quantities copies, forgeries, and works by effectively anonymous artists. Subsequent scholarship on the Dutch art market, based on Montias’ findings, has suggested the presence of a two-tiered market system, divided between works by named artists and those belonging to low-level artistic production8. These two markets appear to have possessed distinct price scales, modes of promotion, and audiences. Current scholarship also strongly suggests that the large majority of named artists in 17^th-century Netherlands strove, with recognition, to find one or two patrons to underwrite their production, so that the artist might achieve a measure of financial security. While a Netherlandish autonomous market might offer a forum for experimentation and innovation, as some scholars have suggested, it created greater risks for artists than the guild system it was gradually replacing.

17If we combine all this evidence, what it suggests is that dealer influence over artists’ careers remained small during the 17^th century. So while Montias’ work on dealers is interesting in itself, it does not go very far to explain the behaviors of important Dutch artists, the named artists within the discursive pyramid of canonical art.

18What are the potential alternatives to both the Alpers and the Montias’ models ? I would argue that even more important than the collection and quantifying of data is the prior task of asking the kinds of questions that might yield new and interesting results. To ask such questions one must acquire the habit of systematic generalization. And one must recognize both the historical market conditions of artistic behavior and our current understanding of the innovative art or artwork’s achievement, as established through use over time.

19Working backward from what we prize today runs strongly against the view that a work of art’s meaning can only be understood within its original temporal and spatial context. Yet the very decisions on what art we choose to study are implicitly, at the very least, based on modern judgments, not those of the period under study. For example, in 1721 Arnold Houbraken published his The Great Theatre of Dutch Painters upon which much of our information about Rembrandt, among others, has been based. His history of Dutch art, containing over 600 biographies, however, mentions neither Vermeer nor Hobbema9. And he proclaimed Adriaen van der Werff « the greatest of all Dutch painters10 ». I could not find a monograph on this artist, and the American digital archive, ArtStor, possessed only one color photograph of the artist’s work, compared to the thousands of color digital images they have for Rembrandt.

20If we attend to what we value today, we can open up avenues for new research that can then be put back into historical context. Let me offer a model for this kind of work. Several years ago, at the Metropolitan Museum in New York I was surprised to discover a peculiarity about the two pictures by Nicolaes Maes the Met chose to hang of the seven paintings overall they own by the artist. One picture was dated to about 1655 and the other to about 1653. I noted that Maes’ birth date was 1634, which meant that the artist painted these works sometime between the age of 19 and 21. Consulting the Metropolitan catalogue of their entire collection, I discovered that four of their seven pictures were painted between 1653 and 1657. The other three Maes paintings the Metropolitan owns are portraits, one perhaps from as late as 1670 and two from 1676. If we expect artists to improve with age, which is a common assumption, why did the Met prefer to show such early works ? After all, how many, if any, other artists hanging on the walls of the Metropolitan Museum are represented by works completed at such an early age ? I doubt there are any.

21Was the hanging of these very early Maes’ paintings a curatorial aberration ? To check, I did a rough survey of Maes’ paintings belonging to numerous major European and American museums. When I calculated the median year of all these pictures : it turned out to be 1656, when Maes was 22. The artist painted the Met’s pictures at an even earlier age, but the subjects and style were consistent with the important paintings belonging to the other museums I examined.

22I have not yet surveyed all the major 17^th-century Dutch artists to determine whether anyone matured as early as Maes, but my impression is that no other Dutch painter produced his most highly regarded work at such an early age. Many are represented by works made considerably later in the artists’ careers. Consider Jan Steen, born probably in 1626, about eight years before Maes. There are many more Steen paintings in European and American museums than Maes’, which might indicate a higher regard for the artist’s work over the centuries, or reflect the artist’s higher rate of productivity (although Maes, I understand, was a very productive artist). In any case, the median year of all Steen’s pictures that I surveyed was 1665, when the artist was 39.

23How do we explain such radical age differences in the perception of an artist’s peak productivity ? Of course, many art historians might say that specific, individual circumstances make such comparisons impossible. Training, family life, personal temperament, job opportunities and other factors might contribute to different maturation rates. This is all well and good, but this kind of thinking also prevents scholars from seeing interesting questions. What qualities as an artist did Maes possess that led him to produce such significant work at so early an age ? Maes, if you will, was a prodigy almost rivaling Picasso, yet no one seems to have noticed.

24There is not an extensive literature on Maes. Nonetheless I found in a recent book by Martha Hollander concerning the representation of space in Dutch genre painting an entire chapter devoted to Maes’ genre pictures, that is to say, to the artist’s early paintings11. Hollander never mentions Maes’ age. She notes that Maes made these pictures having recently left Rembrandt’s studio, but she did not then think to compare Maes’ genre pictures to Rembrandt’s, even in regard to her specific subject, the representation of space. Nor does she explicitly compare Maes’ production during these years to other, recent Rembrandt-trained artists.

25If we look at Rembrandt’s genre paintings from around the early 1650s, when Maes worked in the older artist’s studio, we discover that Rembrandt’s method of handling space and compositions was radically different than many of Maes’ early, post-Rembrandt genre pictures. By the early 1650s Rembrandt had already begun to favor compositions of single or several figures, isolated against a dark background and occupying a large share of the painting’s surface. Maes’ genre pictures are often smaller than Rembrandt’s figure compositions, more spatially complex, and depict small domestic virtues or moralizing scenes that never interested Rembrandt. Maes did learn from Rembrandt the habit of using strong contrasts of light and dark, evident in almost all his genre pictures, which sets his work apart from contemporary genre painters like Jan Steen. And yet when Maes worked in Rembrandt’s studio he produced paintings that, like many other Rembrandt pupils, have been permanently confused with the master’s, so that positive attribution continues to elude scholars. This work may or may not have been painted by Maes, or by Rembrandt, but Maes was the youngest of all Rembrandt’s famous pupils and thus best positioned among Rembrandt’s more talented pupils to imitate the large-scale format of Rembrandt’s post-1650s painting. Moreover, there are paintings by Maes that closely resemble in format and style of painting his master’s work.

26Upon leaving Rembrandt’s studio, Maes returned to his native city of Dordrecht, where he seems to have self-consciously positioned himself by offering a kind of painting then unique to Dordrecht, but popular in other Dutch cities. Throughout the 1650s he primarily painted genre pictures. And he displayed a surprising capacity to work in multiple styles at what was virtually the same time (another kind of behavior we first associate with Picasso). Finally, around 1670 Maes gave up genre painting as well as the Rembrandt manner in favor of portrait painting – more in keeping with international, especially French-influenced tastes, in portraiture. Maes devoted the rest of his career to portrait painting, at which he is said to have made a lot of money. Maes’ portraits are clearly considered less important today than his genre pictures, but at the time portraits were a more reliable and profitable way to earn a living.

27Given these facts Maes may exhibit a form of market behavior unnoticed by Hollander. Although the Netherlands is a small country and the distance between its cities are not very large, as we have seen from Montias’ work on Dutch dealers, the art markets in Holland were primarily local and became more so as the century progressed. Hollander in her book illustrates 16 significant Dutch painters of domestic interiors. Of these 16, only one, Emanuel de Witte, spent his entire career in Amsterdam, and he specialized in church interiors and rarely painted domestic spaces, so he is not actually very representative of domestic genre painting. Gabriel Metsu spent the last years of his life in Amsterdam, but established his career and his characteristic style while working in Leiden. The remaining 14 artists all worked in four other Dutch cities : Delft, Dordrecht, Leiden, and Rotterdam. Amsterdam was considerably larger than the others in the middle of the 17^th century, with over a quarter million inhabitants. By comparison, Leiden had a population of around50 000 at mid-century, Dordrecht about 24 000, Delft 23 000, and Rotterdam 20 000 inhabitants12. At mid-century, Amsterdam was Europe’s most important shipping center, it possessed the first stock exchange, and thus was the world’s leading financial center. Amsterdam’s concentration of wealth and population rendered cities like Leiden and Delft provincial, and this carries over into their respective art worlds. It is possible then to cast Maes’ decision to ignore significant aspects of his training under Rembrandt and to adopt genre painting on his return to Dordrecht as an immediate response to provincial taste. My recommendation to Hollander then would be that we need to know more about why the painting of domestic interiors appealed more to provincial tastes than to those Amsterdam patrons of the arts.

28Maes’ career trajectory, important early work, interesting but negligible late work is also diametrically opposed to that of his teacher, Rembrandt. In creative behavior, Maes exhibited strikingly different manners of working, responding to different markets. In Rembrandt’s corpus there are no such radical breaks in style or accommodations to different markets. Rembrandt’s art evolved slowly and the transitions over the decades are subtle and difficult to described. Rembrandt scholars have barely remarked on such important modifications to Rembrandt’s art, such as changes in size, format, and even supports (Rembrandt abandoned painting on wood in mid-career), because they arrived so gradually. And even the best Rembrandt literature has been ill equipped to describe the differences in behavior between someone like Rembrandt and someone like Maes, because these behaviors are always treated as if they were unique. What, we need still to ask about, enabled Maes’ facility to move from style to style as his market required ?

29In advancing this model, I have not resolved why, for example, we value Maes’ early work more than his later production, only that we do, nor why the artist displayed a chameleon-like ability to adjust to different markets at such an early age, only that he did, nor why the smaller Dutch cities might have preferred domestic genre pictures more than collectors living in Amsterdam, only that this seems to have been the case. What I have been able to do is to ask questions about Maes’ and Rembrandt’s behavior vis-à-vis their markets that the best scholars working in the period have overlooked. By attending to both local data and the subsequent perception of artistic innovation, we can formulate new ways to measure what we value in the Western tradition and why we value it. At its best, quantification shows us what is notapparent in the particular and refashions the questions upon which specialized research can then be focused.