The Pre-Columbian Pavilion as Postmodern Design
James N. Carder
"I thought they might enjoy the Byzantine."
As early as 1964, several art critics, as well as Philip Johnson himself, used the term “postmodern” to describe the design of the Dumbarton Oaks Pre-Columbian pavilion. “Pre-Columbian Art in a Post-Modern Museum,” Architectural Forum 120 (March, 1964): 106–11; “Philip Johnson and Post-Modern Architecture,”Architectural Review 136 (July, 1964): 4; Andrea O. Dean, “Mr. Johnson's Hidden Jewel of a Museum: It Was Called Postmodern in the Middle 60s and Has Been Neglected Ever Since,” Architecture: The AIA Journal 69, no. 5 (May 1980): 52–57. In 2001, Johnson said: “Before Dumbarton Oaks, I had been a pure Miesian. That building may have been my first postmodern building!” Johnson, “Foreword,” 20. The term “postmodern” had been coined for architectural design that displayed a concerted reaction against the modernism of the International Style and continued the trends of the New Formalism that first came about in the late 1950s. New Formalist architecture had championed the “classical” design vocabulary of symmetry, balance, clarity, and the repetition of certain classical forms—columns, arches, and domes, as well as an attention to detail and carefully worked out proportions. Postmodernism championed this return to historical referencing in architecture and frequently employed quotations of historical sources in a witty, sometimes even humorous manner. By reinterpreting familiar historical styles, Postmodern architects also sought ways to mine them for their expressive and symbolic potential. Moreover, Postmodern architects frequently were concerned with the settings and surroundings of their buildings and the need to harmonize architecture with its environment. Philip Johnson’s Pre-Columbian pavilion at Dumbarton Oaks fits precisely this postmodernist definition.
Johnson has stated that the design of the Pre-Columbian pavilion was influenced by the sixteenth-century Sultan Süleyman Madrasa (school) (1550–1557) in Istanbul (figs. 36 and 37) across the street from the Suleymaniye Mosque, There are actually five madrasas in the Suleymaniye Mosque complex in Istanbul, four of which are “double madrasas” with domes: the Evvel (first) and the Sani (second) Madrasas, to the southwest, and the Salis (third) and the Rabi (fourth) Madrasas, to the northeast. a complex that had been designed by the Turkish architect Mimar Sinan (1490–1588). Johnson had visited Istanbul as early as 1939. Franz Schulze, Philip Johnson: Life and Work (New York, 1994), 278. This Islamic domical architecture, itself dependent on the tradition of Roman and Byzantine vaulted buildings, apparently interested Johnson not only for the repeated domical vault form, About this influence, Johnson has said: “Islamic architects would never have designed a nine-dome composition. There was really no attempt on my part to study Islamic architecture. I just remembered that the feeling of repeated domes was a rather delicious way to organize space in a module. It’s modular.” Lewis and O’Connor, Philip Johnson, 59. but also for its ultimate historical association:
In fact, the idea of clustered domes came straight from Istanbul. And, of course, it didn't hurt that Dumbarton Oaks is a Byzantine institute. It’s hard to remember how the layering came about, but of course, I had both in mind, the main collection and the Pre-Columbian collection—although I knew it was for South American objects. But I thought they might enjoy the Byzantine. Lewis and O’Connor, Philip Johnson, 54-55. Johnson claimed that “Mrs. Bliss was just delighted that I thought of that [the Byzantine/Islamic historical reference].” Ibid., 58.
He also noted that neither the Pre-Columbian pavilion nor the Sinan madrasa were meant to be appreciated from the outside or cause someone “to say, ‘Oh, there’s an interesting building.’ But the minute you take an inside view you see what the idea is. It’s a purely ‘inside’ building. It has no facades at all. The madrasa had none; all you could see were the floating domes over the walls. I would have had to use a wall if it hadn’t been in the woods.” Ibid., 56.
Not explained by either Johnson’s statement or by the Sinan model is Johnson’s use of very large columns that both support the domed vaulting as well as prominently define eight intervals in each of the cylindrical galleries (fig. 38). The size and opacity of these columns, made especially noticeable by their conjunction with the transparent glass walls, and their overall density in the plan is somewhat reminiscent of the columns of the so-called hypostyle halls of ancient Egyptian temple architecture (fig. 39). Johnson had visited Egypt in 1928 and he would return in 1966. Schulze,Philip Johnson, 41 and 300. In 1965, Johnson would reprise these large cylindrical columns on the exterior of his (and Richard Foster’s) Kline Science Center at Yale University, New Haven, Connecticut. In both buildings, the column’s size is strictly a matter of design and not of structural engineering. At Dumbarton Oaks, the Illinois marble-sheathed columns have at their hollow centers thin steel-and-concrete lally columns to actually carry the roof loads. Sherman, “Uncommon Ground,” 39. Johnson therefore used the size and visual weight of the “marble” columns to emphasize both physical mass and the repetition of form in a manner reminiscent of ancient structures. Thus, despite the transparency of the intervening transparent glass, one psychologically understands the building’s physicality and thereby its architectural coherency. The cylindrical columns, without tapered sides, bases, or capitals, mimic the cylindrical gallery spaces and perhaps, again psychologically, imply a rotational pattern for moving through the eight galleries (figs. 40 and 41). Indeed, Johnson said the columns were intended “to roll you into the next pavilion.” “Pre-Columbian Art in a Post-Modern Museum,” 110.
Johnson’s creation of a logical pathway for visitor progression through the eight galleries of the Pre-Columbian pavilion was itself a postmodern design element that was based on historical programmatic architecture. In the Dumbarton Oaks galleries, however, Johnson made the pathway both interesting and beautiful: interesting because of the exhibits that would arrest progression, if only momentarily, and beautiful because of the bronze- and marble-framed views of the gardens. Each of the eight “way stations” along the pathway offered the visitor both enclosure—one shaped by columns and dome—and openness, due to the glass walls and clear sight-lines to the next gallery. The axial location of the central fountain, viewable from all galleries, served as the centrifugal stasis point of the rotation of the building’s spaces and, due to its inaccessibility, kept the visitor on the circulatory path. In an interview, Johnson remarked on this: “[The building’s] very processional. At least you’re not going to have to double back. No dead ends.” Lewis and O’Connor, Philip Johnson, 58. In 1965, Johnson would expound further on this idea in his article, “Whence and Whither: The Processional Element in Architecture,” Philip Johnson, “Whence and Whither: The Processional Element in Architecture,” Perspecta, The Yale Architectural Journal 9-10 (1965): 167–72. where he would state: “Architecture is surely not the design of space, certainly not the massing or organizing of volumes. These are ancillary to the main point which is the organization of procession. Architecture exists only in time.”