As noted above by R.E. Somol, Eisenman was among a number of young architects who were particularly interested in the diagram (as opposed to the drawing), and its burgeoning importance for architectural theory. Eisenman, probably above all others, has continued this shift in focus, and has developed its theoretical implications for a number of years. He writes:
Generically, a diagram is a graphic shorthand. Though it is an ideogram, it is not necessarily an abstraction. It is a reresentation of something in that it is not the thing itself. In this sense, it cannot help but be embodied. It can never be free of value or meaning, even when it attempts to express relationships of formatin and their processes. At the same time, a diagram is neither a structure nor an abstraction of structure. While it explains relationships in an architectural object, it is not isomorphic with it (Diagram Diaries 27).
Here, I think, we can begin to see shades of a somewhat more sophisticated version of Eisenman's earlier (mis)readings of Derrida. The diagram is caught up in the endless reference to an architectural object (even if that 'object' is not actualized). Thus, the 'essence' of the diagram (if it can properly be said to have an essence) is always slightly out of step, out of joint, out of synch. It isn't simply a drawing or a "picture", nor is it simply a plan or a schematic. The diagram is no longer anti-structural or anti-functional for Eisenman, but between the structural and the non-structural. Eisenman writes that "unlike traditional forms of representation, the diagram as a generator is a mediation between a palpable object, a real building, and what can be called architecture's interiority" (Diagram Diaries 27). That is, the diagram spans the gap between the object (building) and that which is proper to architecture itself (i.e. the design process, the theoretical aspect). Architecture's "interiority" (it's history, theory, practice, its proper "being" above all else) is, according to Eisenman, always in a relation with the diagram. The diagram, which moves back and forth between architecture's "exteriority" (its products: buildings) and its interiorty, is what enables interiority as such.
I'll try and put it in plainer language, as I find Eisenman's terminology somewhat dense and confusing. The diagram is both a representation of an actual object (a building) and the 'essence' of architecture (though the word essence might be a little misleading here, its probbly helpful). So, the diagram occupies an odd position: what defines it is something that cannot be defined in itself. What is 'proper' to the diagram is this sort of oscillation between the interior and the exterior, the object and architecture. What Eisenman is proposing is that the diagram's position in fact determines what architecture, as architecture, is. Because architecture is not simply theoretical but bases its 'identity' on its objectification in buildings, it requires a go-between or a mediator (the diagram). Eisenman believs that this go-between limits and constitutes the possibilities of architecture. The limits of the diagram, then, are also at the same time the limits of architecture.
Let's draw a diagram then:
**************diagram***
**D*********************
***I****Exterior********
**A***d*****************
***G**i*****************
R a D
di A g RAM I
M r A
******a**************G***
******m***************R**
********Interior*****A***
**********************M**
***diagram***************
You get the picture.
So, yeah, the deal here is that the diagram, because of its curious ontological position (one that shifts constantly) isn't simply a product of architecture but, in fact, conditions architecture's limits. And Eisenman uses a number of historical examples to prove this. These are a little too specialized for someone with little real architectural knowledge to say too much about, so I'll just quote a chunk:
The diagram is not only an explanation, as something that comes after, but it also acts as an intermediary in the process of generation of real space and time. As a generator there is not necessarily a one-to-one correspondence between the diagram and the resultant form. There are many instances, for example in Le Corbusier's Modulor where the diagram is invisible in the building, yet it reappears as a repetitive element that occurs at many different scales, repeated in little segments of houses to large segments of urban plans, yet it is rarely an explicit form (Diagram Diaries 28).
And, regarding Andrea Palladio:
Wittkower's nine-square diagrams of Palladio's projects are diagrams in that they help to explain Palladio's work, but they do not show how Palladio worked. Palladio and Serlio had geometric schema in mind, sometimes explicit and sometimes implicit, which they drew into their projects. The notations of dimensions on the Palladian plans do not correspond to the actual project but to the diagram that is never drawn. A diagram implicit in the work is often never made explicit (Diagram Diaries 27).
So, though I'm unfamiliar with the intimate or diagrammatic details of either Le Corbusier's work or Palladio's, I think I can extract something of theoretical value here. What Eisenman is trying to get at is that the diagram, itself in constant flux, is also in a fluctuating relation with architecture. It isn't that the diagram has one sort of mediatory role, but that this role can be expanded or take on multiple characteristics. In Le Corbusier's work, for instance, the diagram plays the role of a kind of segment or building block: a repeated form that is not a totalizing determinant on the building as a whole, but which determines, in part, certain aspects which in turn relate to the whole. In Palladio's work, the diagram (which might never be drawn) remains implicit: a geometric scheme which may or may not help determine the dimensions and shape of a building. The geometry never precisely refers to the building, but refers back to itself, to a scheme, a diagram, a plan. The diagram, again, moves between building and architect: it avoids totality and only determines "in part". Diagrams jammed together do not play the role of calligram.
The diagram is embodied, but never fully.
This begs the question: what role (or what roles) the diagram plays in Eisenmen's own work. Well, unsurprisingly, the diagram is of great importance for his work: he describes the diagram as existing in a relation of anteriority to the interiority/exteriority of architecture and the architectural object. How's that for jargon? He writes:
The diagrams of the Fin d'Ou T Hous worked on many different levels. They evolved from the relationship of the diagram to the house, that is, the way the diagrams are marked in House IV. Essentially the idea was to produce a set of diagrams so that any attempt to trace its transformations from some origin was problematized. The diagrams indicated a possibility to be read and traced, but as in a bad mystery novel all of the clues, when traced to their origin, proved to be false. Instead, these clues led only to another system of tracking that would take one in another direction. This was similar to the process of decomposition that had begun with House X (Diagram Diaries 93).
Here the diagram is used as a method of destabilization and "unfounding." The diagrams for Eisenman's Fin d'Ou T Hous are used in such a way that each diagram (and there are a number of them) gives the impression of a clue or a trace: it refers back to some previous diagram, some previous 'origin'. But, when attempting to discover this origin, one only manages to happen upon yet another diagram, yet another trace. The process is similar when we encounter the House itself, or the model of the house: its 'features' seem to refer back to a plan, or a blueprint: but there are only endless diagrams. Each diagram enmeshed in a variety of processes: nesting, scaling, artificial excavation, etc. So instead of providing the project with a metaphorical foundation (i.e. a rigid plan) Eisenman's clue-like diagrams question the very notion of such a foundation. Even more literally: the Fin d'Ou T Hous itself questions the foundation: it is 'based' in/on an excavated space: its foundation is more absence. So, Eisenman's diagrammatic approach seems to undermine both the interior (theory) and the exterior (building).
Eisenman wants to go further than this destabilization.
While the foundations of architecture and the building are questioned (and, of course, the diagram itself), Eisenman moves on to question the very notion of architecture and the architect. Along with Bernard Tschumi's work (specifically Parc de la Villette, Paris) Eisenman hopes to undermine the traditional architectural hierarchy "ARCHITECT/building". By using processes like computerized iteration, Eisenman uses points, lines and planes (endlessly and often randomly reproduced) in order to achieve a final product that is not 'designed' in any traditional sense. Thus, the building is not simply the result of an architects planning: through the diagrammatic process, the architect's position becomes unstable. This process is much more clearly described by eversion in an excellent writeup entitled Superimposition: a pharmakon for architects. Though said writeup deals more directly with Bernard Tschumi, Eisenman often adopts similar terminology and processes. eversion writes:
...the act of superimposition represents Derrida's in-between or denial of hierarchy. One of the most important shifts in ideology to be recognised here is the way that the position of the architect himself is also questioned. The role of the creator is an implicit part of the systemic hierarchies that Tschumi is trying to disrupt. By devaluing his own position within the system Tschumi makes the traditional understanding of his presence tenuous.
By understanding how Derrida's undecideability has led us to the question of an architects loss of self we can begin to make the first connection between this early example of deconstruction (arguably one of the few to have a clear understanding of it's philosophical background rather than just becoming a style or -ism) and contemporary hypersurface theorists such as Greg Lynn. (from: Superimposition: a pharmakon for architects
Unlike Tschumi's, Eisenman's conception of deconstruction is somewhat muddled. But, in the main, I think we can extract some salient theoretical implications from it. In the wake of Tschumi and Eisenman (among others) we can't simply take the architect's position for granted. We also have available to us new conceptual tools that allow us to get beyond architectures old dogmatisms: tools like iteration, superimposition, superposition, diagram-based architecture, etc.
In Diagram Diaries Eisenman begins to move away from the "purely" deconstructive architecutre of his earlier works towards a new "post-deconstructive" architecture. I am uncertain of his success, but this new position relies mainly on the notion of a superposition (in contrast to superimposition). This new technique no longer relies on the forces of collage and the elimination of 'design' / the subject from the architectural process. Eisenman writes:
The distinction between Deleuze's idea of superimposition and my use of the term superposition is critical in this context. Superimposition refers to a vertical layering differentiating between ground and figure. Superposition refers to a coextensive, horizontal layering where there is no stable ground or origin, where ground and figure fluctuate between one another (Diagram Diaries 29-30).
I think one could make the argument that this new move is simply an updated version of superimposition. While superimposition, in Eisenman's early works, attempted to get beyond the "ground", it failed because it was always a detailed reaction to that ground: it took the ground for granted even when it went beyond it. In his later work, Eisenman attempts to recover this "error" by actually working over the ground itself (literally). The site is diagrammed and "superposed" over and within the other diagrams, which are themselves superposed. So, Eisenman's work is no longer simply a collage dealing with predetermined structural elements that are iterated out of one's control. Instead, even those given elements are called into question: collage (superimposition) is replaced by something even more unstable, something which works over and questions the framing of the collage as much as the collage process itself. I think this move reflects a more sophisticated, more detailed, reading of deconstructive thinking (i.e. Derrida) than his earlier work and "saves" him somewhat from the gaping maws of mis-understanding.