Suddenly a shift occurred, a shift in the path and the Diamond Projects appeared. (Hejduk 1985: 34)
In November 1967, architect and educator John Hejduk (1929–2000) exhibited a series of drawings and models in New York at The Architectural League that explored the architectural implications of the forty-five degree rotation of bounding elements relative to an orthogonal system. The work was the result of a six-year investigation into the problem.
Diamond House A, Diamond House B, and Diamond Museum C are the three developed projects in Hejduk’s Diamond series (1962–1967). The drawings and models in the exhibition also included paintings on the same theme by Robert Slutzky. The Diamond Projects have been partially published and are regularly referred to in writings by and on Hejduk, and they have also been the object of retrospective considerations by him, but they have received only limited consideration in secondary writing.1
The Diamond Projects occupy a unique place in Hejduk’s work. In the quote that opens this paper, Hejduk reflects on the sudden drift from the Nine-Square (Texas) series that happened around 1962, identifying a shift, if not a full break, that these projects mark in his interests and style. The Diamonds follow his near decade-long work on the Nine-Square or Texas Houses and are themselves subsequently followed by a number of smaller studies (the 1/4, 1/2, 3/4 Series among others) and, in time, the Wall Houses.2
To take Hejduk at his word, at the time, the three projects arrived out of a desire to explore the implications of ‘the diamond canvases of Mondrian for architects of today’3 (Hejduk 1985b: 48). Following Hejduk, therefore, their role is primarily formal, the projects exploring formal-spatial possibilities in diamond configurations by means of columns (House A), planes (House B), and then biomorphic shapes (Museum C), to make only the most reductive of interpretations. These three conditions — columns, planes, biomorphic shapes — are the most basic form questions addressed, as Hejduk’s synoptic table diagramming his work between 1955 and 1972 suggests.4
I would argue that the impact and importance of the Diamond Projects, however, is more than a search for new form relationships or a transmutation of architectural space concepts. In addition to these, the Diamond Projects can be read as an index of two other considerations. Firstly, they can be seen as an effort to overcome a number of architectural biases Hejduk traces to the Le Corbusier of Villa Garches and the Visual Arts Center at Harvard University.5 In particular, Hejduk tries to escape from the biases of frontality and the horizontal stratifications of space idealized in the Domino diagram and traced by Hejduk to the Villa Garches.
Secondly, I believe the Diamond Projects are also manifestations of a specific temporality. And while I will examine the first consideration, it is this latter aspect in particular — that of a uniquely architectural temporality — that I will focus on.
The temporality rendered in the Diamonds, I suggest, is time that is always already compressed, different from that which is composed by a line of images or vignettes and thus is reliant neither on futures nor pasts. Importantly, in terms of it’s potential for architecture, the temporality revealed or explored in these projects is independent of movement.
Different from Hejduk’s emphasis on the formal and spatial importance of the projects, I argue that a close examination of the drawings and period writing reveals a specific notion of time, one whose potential for practice and theory has yet to be examined. This immanent temporality may be a general characteristic of modernist architecture, and thus its interrogation may contribute to our understanding of parallel and different manifestations, whether architectural or from other domains.6
To begin to frame the concept of a temporality specific to certain works of modernist architecture, I turn to the notion of direct time as theorised by Gilles Deleuze (1925–1995) and deployed most fully in Cinema 2: The Time Image. In Cinema 2, Deleuze proposes that in some mid-twentieth-century films a new relationship of movement and time is made visible. Time is no longer subordinate to movement, he writes, and a reversal occurs such that ‘time ceases to be the measurement of normal movement, it increasingly appears for itself’ (Deleuze 1989: xi).
A phenomenon similar to that which Deleuze discerns in the realm of philosophy and cinema and characterized as pure time, one independent of motion’s measure, can be claimed for architecture. Certain kinds of temporal relationships are uncovered in buildings that release or make concrete a Deleuzian direct time, one not bound to a vision in motion, or to a promenade architecturale (Deleuze 1989: xii). An examination of such projects from the point of view of time will reveal a range of formal moves and composition devices at work along with their resultant formal-spatial effects. A preliminary analysis of the Diamond Projects will be used to explore these propositions.
In order to frame the larger question to which the present study contributes, one might ask: how are buildings and projects composed such that time is confronted? Is there a specifically modernist concept of time, a modernist mode for the creation and expression of time? Just as there are different space concepts, are there different architectural concepts of time, and if so, how do they work and what are the distinguishing characteristics of each?
In this paper I work from drawings held in the John Hejduk Archive at the Canadian Centre for Architecture in Montréal, published interviews and writings of Hejduk, and secondary writings on Hejduk. The drawings form the primary analytic material. These are supplemented by key documents, including the Three Projects (1969) ‘Diamond Thesis’, published by Hejduk to accompany the 1967 exhibition; his 1965 review of Le Corbusier’s just opened Visual Arts Center and published as an homage following Le Corbusier’s death in August of the same year; and interviews with Don Wall published in Mask of Medusa. Secondary writings from the period and more recent scholarship on Hejduk’s work provide historical and critical context.
The first thing that is striking when looking at drawings of the three projects is that the rotation of the bounding field relative to a horizontal grid produces much more than just a ‘hypotenuse moment’, as Hejduk characterises the densest effects produced in the Diamond Projects. The new relationships of form have at least two major consequences: peripheric tensions of the edge and field extensions beyond the building volume that render an expanding space (Hejduk 1985c: 90). Rotation creates conditions which challenge potential bay readings and disrupt frontal/oblique factors, as everything in one sense is or has become diagonally engaged in a perimeter condition. And unlike six of the seven preceding Nine-Square (Texas) houses, there are no site plans and thus the ground as figure and force factor is not available. In other words, there is no outside, no house-site relationship.7
Diamond House A (Fig. 1) is a thirteen-column, ten-grid-line, square-bay plan. Unlike the ‘Texas’ house investigations, bay counts is not a useful way to characterize the series, the number of regular four-sided bays being highly limited. In House A, for example, there are four full bays, eight ‘three-fifths’ bays ringing the perimeter, and eight minor ‘triangular’ bays at the points. Column or grid lines are counted both vertically and horizontally. The columns are round, introducing a ‘centrifugal force and multi-directional whirl’ into the plan (Hejduk 1985a: 73).
Grid lines are inscribed in the floor plan as narrow bands, and there is a straight run stair in the second bay from left. Elements (fireplaces, furniture, partitions) generally abutt the column/beam bands, though not consistently. Certain partitions on the second and third floors are located well off the grid. On the second and third levels, full floor-height brise soleil bars provide a continuous agitation of the light, to use Hejduk’s characterization of the effect.8 Their rhythm varies from floor to floor and there is no clear method for placement of the brise soleil, with one exception. The bars on the grid-line extension to the perimeter are consistently doubled, and the glass is pushed out in these instances, including at the top and bottom points, to the outside of the slab. This is most easily seen in model photographs and introduces a direction in the plans to the extent that left-right grids are not treated in the same manner either in the mid-bay positions or at their points.9
In Diamond House B (Fig. 2), another four-level project, parallel rows of walls replace columns, generally running vertically according to the plans as published. Unlike House A, the stairs have shifted: there is a switch-back stair at the bottom and two spiral stairs framing the spine. Right-angled relations inside the diamond plan are constant, until the final floor is reached and curved elements appear. Counting from point to point, this is a twelve-grid line, square-bay plan. Floor scoring continues as in House A: wall-width scores in the regular bay pattern. If House A was an exploration of the formal implications of right-angled conditions within a diamond configuration via columns or walls in shifting right-angled or oblique relationships, then House B tests the potential using walls alone.10
As in House A, revealing a lingering tendency that might be described as cubistic, a direction as well as an eccentricity is introduced by placement and delineation of walls of different dimensions. Thicker walls are located on top-to-bottom grid lines three and five (counting left to right), establishing a dominant direction in organizing spatial flow and a shifted center onto grid line four. The perimeter agitation of light in House A that Hejduk found so troubling twenty years later is not present in House B.11 A constant floor slab overhang continues, interrupted in this case only by glass blades extended along eight of the grid line extensions, all continuing the top-bottom alignment of the major internal walls and further reinforcing the direction.
Diamond Museum C (Fig. 3) begins again with a square bay, now in a larger eighteen-grid line, forty-one round-column configuration. Like Diamond Houses A and B, a greater concentration in the vertical (top-bottom) alignment is palpable in the plans as published. All three vertical circulation elements — the switchback ramp and two switchback stairs — as well as the cluster of biomorphic-shaped walls are all aligned in a top-bottom arrangement. As in House A and House B, the top-bottom bias is reinforced by perimeter brise soleil blade walls. The brise soleil blades are on a two-thirds grid-line rhythm that continues without interruption around the full diamond perimeter, aligning with every other grid line. The rows of display cases in Diamond Museum C are all on a horizontal left to right alignment. In this, they could be interpreted as explicitly following the general disposition of Mondrian’s unfinished Victory Boogie-Woogie. In its final state, Victory has eight horizontal rows of small color swatch tapes, and the display cases in Museum C roughly align with these. Other shared features include the openness of all four points and the use of relatively large areas of neutral color (open field) on the top/bottom spine. Major walls are more or less on a pinwheel configuration. Scoring of the plan on the column-grid with variously spaced double lines follows a left–right pattern reinforcing this directional bias. A second level of single line floor scoring at right angles to this divides each bay into three vertical rectangles. This could be interpreted as introducing a Cubistic gesture (directional) in a Neo-Plasticist (non-direction, square grid) realm. Plan tensions are introduced between a bias to the vertical against uniform extensive or explosive forces, not at the points but out to the edges of the diamond.
Having now an overview of key drawings for each of the three developed projects, it is appropriate to look at the larger context. To understand the impact or potential of the Diamonds, it is helpful to review Hejduk’s period writing and examine what was happening around him. In the air and occupying Hejduk is the experience of Le Corbusier’s Visual Arts Center, the specific suggestions and sensibilities of that ‘last realist’ — the Mondrian of Victory Boogie-Woogie (Hejduk 1985c: 128) — and Rowe and Slutzky’s explication of Villa Garches as an exemplar of flat space in their ‘Transparency: Literal and Phenomenal’ paper (1963).
Certainly this cluster of events can be conjectured as a primary cause of that shift in his thinking and practice announced in the opening citation: Hejduk’s experience of the Visual Arts Center in 1965 and Le Corbusier’s death in August of the same year; the desire by Hejduk to get out from under or go beyond the influences of Mondrian and the Le Corbusier of Garches and in particular the limits of frontality; the Domino spatial idea; a chronological notion of time.
Perhaps these last traits are what in the end Hejduk endeavours to get beyond, to work out architecturally. The Diamond Houses are privileged in the interviews with Wall as the site of working out certain Corbusian devices, but a close review of subsequent project drawings — 1/4 Series, 1/2 Series, Extension House — reveal lingering traces. As Hejduk himself states in an interview with Wall, the Diamond isometrics ‘reminded me of Le Corbusier. So I had to get rid of that, by working it out, by exorcising the images. Corbusier, and then Mondrian in a way’ (1985c: 36).
To further unpack the formal moves and motivations, consider Frampton’s description of Hejduk during the Diamonds phase as trapped, unable to adopt the ‘multi-directional spatiality of the De Stijl or Suprematist movements’, ‘unable to abandon [… an] emphasis on frontality’ (Frampton 1975: 142). There is evidence of these two conditions in the Diamonds’ general reliance on the free plan, column-slab or wall-slab model, and all three projects are horizontally layered within a regular volume and thus are indebted to or rely on the space concept of Domino. Hejduk was never concerned, however, with multi-directional spatiality, convinced as he was at the time that the already dense two-dimensionality of architectural space was the only one he should pursue. This belief is in evidence when he writes at the time about architectural space’s actual ‘two-dimensionality’ (Hejduk, 1985b: p.49).12
Form relations, concepts of space, and perhaps, as I am proposing, notions of time are all being worked on here. At first blush, and as suggested above, two compositional devices are constantly explored in the Diamonds. The first, an expansion ‘or exfoliation’ toward the periphery and beyond.13 A second composition strategy is that of a condensation of space and mass towards an always voided center. Hejduk, in an interview with Wall, states, ‘All my houses have voided centers… Maybe my contribution to architecture is the voided center’ (1985c: 131).
This idea of the voided center confirms a move in Hejduk’s attention away from a trabeated logic of post and beam frames, of singular columnar surfaces, and of articulated roofs (clearly and methodically treated in the Nine-Square (Texas) series) toward a ground-less architecture of round columns, freestanding partitions, and flat-slab floors and ceilings. In other words, a critical work on Le Corbusier’s Domino world of column-slab construction, would fully engage with the implications and effects of two kinds of architectural freedoms, those of ‘liberated space [and] liberated structure’, as Hejduk writes in a manuscript sheet from the period.14 It will take a few more projects, a few more steps, to get to what he claims was that ‘real break’ of the Wall House (1985c: 36), but certain problems explored in the Diamond configurations — voided center, frontality, perimeter warps — are evident. And there is perhaps an underlying dimension related to time. To track these conditions, I will start with the voided center.
In House A, the center is occupied by a round column, in House B the center is held by a shallow room bound by walls in shear, and in Museum C it’s another round column, here marking a spring-point of pin-wheeling elements. There is no tension; rather there is a state of ambiguous equilibrium, all the architectural energy having been moved to the perimeter. This is consistent with the form studies underway, by which Hejduk tries to establish an architectural condition that made manifest a phenomenon of ‘all-over kinetic equilibrium’ in the same pendulum arc as he tried to realize the condition for a neutral container, the two together mimicking what he described as a Michelangelo effect.15
In terms of frontality, here is Hejduk drawing to get over or to abandon the curious limits he sees in the Renaissance vision still evident in Le Corbusier’s Garches (Hejduk 1985b: 48). In the Diamonds, he works through that reliance on frontality by at least two moves already touched on above, that of voiding the center and charging the perimeter.
Linked to all these, and even if a first move, the forty-five degree rotation of the bounding envelope allows him to move beyond the Le Corbusier of the Visual Arts Center and of Villa Garches. For when confronted with the receding or encompassing walls — depending on the observer’s position, as was made clear in diagrams 8 and 9 of Hejduk’s ‘Diamond Thesis’ — the projects are always ready to place the observer in an oblique state, even without moving (Hejduk 1969).
In the rotation appears a curious, ambiguous notion of time, which Hejduk characterized as a ‘moment of the present’. There is nothing fantastic or imaginary about it, and in this Hejduk, like Mondrian, is a pure realist. This perhaps explains his unease at re-approaching decades later any of the Diamond Projects, whether A, B, or C. For the state they capture is the ‘flattest… quickest… fastest… most extended… most heightened’.16 It is not just an entry or threshold condition of walls, but an entire project working on, and intensely occupying, the present.
Perhaps he really did work it out, or was close to working Le Corbusier out of his system, by 1967. As evidence, Hejduk appears to have overcome frontality in favor of the always oblique, and moved beyond or away from a horizontal layering of regular volumes — the Domino model — toward a flat-space world of constantly vibrating and animated planes.
This latter state is evident in the three Diamond Projects. House A places L-shaped and single plane walls on the perimeter in a manner that torques the regular horizontal space volume. House B introduces double height volumes that cascade up the building and thus also disrupt a single horizontal space idea. Museum C achieves this by the intensity of plan figures such that the possible experience is as a section idea that goes beyond a simple horizontal layering.
What came next were projects that continued the form investigations of the Diamond Projects.17 Subsequent projects took on similar or related ideas and themes, replacing frontality, for example, with shear (3/4 Series, Extension House), the pinwheel (such as 1/4 House C, 1/4 House D), and the echelon (1/4 House B). Compressive space is taken even further, whether on the vertical (1/4 House A, Wall House 1), or the horizontal plane (Grandfather Wall House).18 Evidence of the explorations made possible by the efforts of the Diamond Projects can be seen in these and other developed projects. If Hejduk had not yet fully overcome Le Corbusier’s influence, he was on his way. Consider as evidence two drawings situated somewhere in that whirl of 1960s work on Nine-Square (Texas) House 5 (1958–60), the Diamond series (1962–67), the 1/4 Series (1967), and the ‘Out of Time’ (1965) and Three Projects (1967) texts (see Figs. 4 and 5).
On a sheet containing several plans, all in diamond configurations, the most developed disposes four round columns, thirty-six feet on center, in a single square bay set inside the mid-points of bounding walls, with all other elements — partitions, fireplaces, and furniture — in right-angled relations (Fig. 4).19 Internal partitions are arranged like a pinwheel, with the four arms at approximately similar 9, 1, 3, and 5 o’clock positions. Some of the effects of spatial warp, the elements and their relationships needed to get there, are being explored.
To take another example of what was made possible, consider what appears to be an early study for 1/4 House C. Having been set aside during the investigations into diamond configurations, the ground has suddenly returned, here in this study, as a diamond-shaped site20 (see Figs. 5, 6). Each quadrant bounded on one side by pin-wheeling walls is occupied by a volume (rectangle, diamond, quarter circle, quarter square) and wall positions and fireplace elements are tested. In the developed 1/4 House C, the site remains, gripped by those four walls whose direction — are they moving into or out of that impossible center, impossible because so compressed — is one of the many ambiguities at work. And thus other questions come into view, to be examined in the opening created by the Diamond Projects, and leading to the 1/4, 1/2, 3/4 Series, and the real break of the Wall Houses, as discussed above.
To conclude, and as an opening for further research, I will return to the initial reference to Deleuze’s concept of a pure time in Cinema 2. As Deleuze succinctly characterizes it, there are certain conditions rendered in works of cinema — and by extension, I would claim, in certain readings of the Diamond Projects — such that one is plunged directly into an experience of time irrespective of a reliance on movement. As Deleuze stated, certain kinds of aesthetic works are conceived in such a way that ‘we are plunged into time rather than crossing space’ (Deleuze 1989: xii).
According to Deleuze, a reversal in the relation of movement and time can be discerned in certain works. For philosophy, the reversal occurs slowly and only in some thinkers over hundreds of years. In cinema, again according to Deleuze, the reversal of the movement–time relation, affording time the capacity to manifest itself independent of motion, occurred after 1945 and thus at a much more accelerated pace. Deleuze’s concept of direct time is one independent of movement and he finds examples of this condition in films directed by Renoir, Fellini, and Welles, among others (Deleuze 1989: xii).
Is there similar evidence in Hejduk’s Diamond Projects? Could a similar reversal be said to have occurred more generally in the realm of architecture and if so, how would we recognize it? What kinds of devices would be at work to give rise to a pure time, one different from a past-present-future time, that purely empirical succession of things that, for example, the promenade architectural manifested in Le Corbusier’s Visual Arts Center gives expression to?
Though tentative, and calling for further development, I believe there is evidence of a concept of direct time at work in the three Diamond Projects, briefly considered here and whose interpretation I have sketched out above. Returning to the opening questions, it can be claimed that two theses have been accompanying the Diamond Projects of Hejduk, a thesis of simultaneity and a diagonal or diamond thesis.21 The first, that of simultaneity, announces a multi-planar, explosive world replacing a world of frontality and contained horizontal stratifications, the world, that is, of the Villa Garches and the Visual Arts Center. This thesis assumes there is an oblique, always lateral spatial order endeavouring to supplant or at least get away from Renaissance frontality. The second thesis concerns the concept of a temporality characterised by a concept of direct time.
These two theses work in parallel on the problems of the observer in motion, and thus of time (the thesis of simultaneity) and of space (that collapsed, ‘actual’ two-dimensional space of the diamond). They are given further expression and exploration in the projects that follow the Diamond series and in Hejduk’s writing. The thesis of simultaneity and of time independent of an observer in movement, for example, is explicit in the diagrams that Hejduk used to illustrate the history of space in architecture and give expression to the ideas in the ‘Diamond Thesis’ (Hejduk 1985b: 49, diagram 7 in particular).
Providing a kind of open unity to the form relations and spatial orders at issue — voided center, peripheral tension, exploded field, volume transmuted into plane and thus creating the conditions for flat space to appear — Hejduk’s specific notion of time revealed in the Diamonds is that dimension which ensures a single whole is never completely given. To go further, it might be that in these projects there is evidence of a pure or direct time in the sense framed by Deleuze. And in this instance, direct time would be that function or operation that holds it all together. As produced by the Diamond Projects, diagonal time has the strange power to affirm parts that do not make a whole in space, nor form a succession in time. Time is exactly the diagonal of all possible spaces made possible as a result of those two freedoms, as noted above, that Hejduk found in Le Corbusier — liberated space and liberated structure. The Diamond Projects therefore can be read as giving expression to certain characteristics of a direct or pure time and in this they constitute an appropriate interpretive category alongside the formal-spatial one emphasised by Hejduk.