Architectural Association School of Architecture

4^th Year Histories & Theories Submission

ANTI- [ANTI -] RATIONALISTM

Martin Self

Rationalism in Contemporary Practices

This essay starts from Patrick Schumacher’s critique of UN Studio in AA Files 38, “Rational in Retrospect: Reflections on the Logic of Rationality in Recent Design”[1]. His reconstruction of the Arnhem Central master plan [Fig. 1] design process through a rationalist lens, serves as a good template on to which further analysis on to rationalism applied in architecture can be established. This essay seeks to compare and extend this informed critique to UN studio’s design process for Arnhem into current architectural practices and design technologies. Past Fordist and Posthumanist models of rationalism in architecture are compared to diagrammatic-based practices and into the subject of contemporary computation in architecture and how it contributes to further advancement towards a theory for non-linear rationalism that Schumacher indicates.

[Fig. 1] UN Studio Arnhem Central

Schumacher’s considers “A rational decision, in order to be described as rational, needs to be reconstructed as a section from successive, finite sets of alternatives. The minimum requirement for any act to qualify as a decision would thus be that it preceded by evaluation against at least one specifiable alternative.”[2] From this base definition, six models for decision-making are formulated: hierarchical (discrete chain and self-sufficient decision order, with no loops or iterations), ranked (presupposing a finite and stable list of ranking objectives), comprehensive (eat each level, the space of options is finite and comprehensive), decidable (at each level solutions are measurable and relative to performance), coherent (primary decision cannot be sacrifice to another reason at a later stage without subverting the whole process), decomposable (parallel objectives addresses in parallel trees, independent solutions and decisions.)[3]

By defining these schemas, Schumacher lays the points from which UN Studio scheme deviates from complying with these definitions. Specific models, such as the decomposable one, have been present in modern design and engineering practices. Fordism based on the segregation and specialization of task and one were each part is reducible to a series of decisions and calculable parts coming together in the moment of assembly. A model where the individuation of parts and mass serialization production leads to architectural projects avoids rational models that would include choice and performance (comprehensive and decidable), but become a model independent to context. This approach is post-humanist, expressed by Walter Rathenau: “mechanical production has elevated itself to an aim in itself.”[4] Ludwig Hilberseimer [Fig. 2] worked on projects were new urban organizations encouraged endless repetition and modularity, “For capitalism itself is but an obscured form of reason, and its productive attributes – rationalization, Fordism, Taylorim, planification – along with its corresponding massified subjectivity constitute what we might call Hilberseimer’s “concrete utopia”.[5] Hilberseimer used repetition as tools to communicate this era where the architecture of the metropolis revolved around the cell, the single room. Sachlichkeit ideas sought to rationalize between part and whole, molecular and molar, between the elementary cell and the totality of the physical.[6] His projects impose models that followed the cellular regime above all and interactions from within were reducible to the overall ideology of new citizenship and inhabitation vision for the time.

[Fig. 2] Ludwig Hilberseiimer, Hochhausstadt project

“The chaos of the city of today can be opposed only by attempts at theoretical systemization, having the purpose of enucleating form actual situations – in a totally abstract way – the fundamental principles of urban planning, thereby arrcing at the formulation of general forms that then permit the solution of determined concrete problems. Only the abstraction of the specific case enables us, in fact, to demonstrate how the disparate elements that make up a large city can be placed, in an orderly way, in relationship with the whole.”[7]

The response from the Sachlichkeit movement derives from Fordist influence at the time and maintains the same segregation between object and whole, between cell and building that is identifiable in the decomposable model. This interpretation got represented by the segregation of tasks between design process and manufacturing. In contemporary economies, however, one considers this model to favour interdisciplinary practices. The shift of Fordism to Postfordism leads to this possibility [Fig. 3]:

[Fig. 3] Ford’s assembly line, CAD-CAM systems, BIM modelling

“New computer-based production technologies developed the ability to offer product diversity (small runs) without the enormous relative cost of handicraft production that had previously limited deviations from the mass-product to the realm of luxury. This is the crucial material factor in the whole process: the micro-electronic revolution offering a productivity leap in the production of the desired economies of scope (rather than scale).”[8]

Advancement in CAD-CAM construction has bridged that decomposable model into one where whilst the independent of thought could be maintained, it s now encouraged a multidisciplinary singular approach. In terms of rational models, the decomposable model (parts within their own historical Fordist manufacturing sequence) have points of intervention during design and manufacturing, decidable rational models, which are included in order to benefit the overall outcome (Coherent Model).

The utopian models Modernism bequeathed to contemporary understanding of a design process derives not a revolutionary preposition that affected directly the built environment as response the radical socio-political changes sprouting from the 1920s. Considering the historical precedence on which modernist rationalist architecture developed and until the instance in which UN Studio reclaims a revision to the term, architecture underwent a series of shifts in attitude to the way design process is undertaken and the intent and inference of the designer in the outcome. Contrasting from the totalitarian utopia of the modernist and strive to equate architecture to veritable science, Van Berkel and Bos situates the project as a response to a site condition and the flows and nodes present in the interchange. Diagrams are used as means of responding to site dynamics. The discourse of the diagram is a direct response to modernism failure which tossed architectural thinking into an identity crisis, where experimentation lead to new models of appropriating the design process. At this stage, architecture incorporates the diagram as tool in re-examining its urban context as means of responding to it and designing with it. R.E Somol in his essay “Dummy Text or the Diagrammatic Basis of Contemporary Architecture” traces the post-modernist use of the diagram sprouting from the time of critique of the modernist decline and through ways of rearticulating past conventions into new dynamics.

Stan Allen defines diagrams as “a diagram is therefore not a thing in itself but a description of potential relationships among elements, not only an abstract model on the way things behave in the world but a map of possible worlds.[9]” This definition overlaps with Schumachers point on rational decision-making tree diagrams that ultimately lead to options during the design process unforeseen at the time but that by working with a series of constraint the event crystallises during the process. The diagram is a tool in this respects for the fostering of this discoveries by operating with relational logics and transformations that when applied systematically produce new unforeseen outcomes.

[Fig. 4] Eisenman diagrams for Church for the Year 2000

In the Church for the Year 2000 [Fig. 4] project two parallel premises govern the design: “the first I the relationship between proximity and distance implicit in the concept of pilgrimage and in the idea of communication media, the second is the new relationship between man, god and nature. In fact, t is a natural form that is chosen to symbolize this condition, both of proximity and distance: the condition of liquid crystals, their suspension between the static crystal and the liquid state.”[10]

The systemisation of a diagrammatic technique to represent a dynamic that would embody the metaphorical intention that the building is to reproduce will be equivalent in terms of rigorous in identifying traceable forces on the site the UN Studio’s Arnhem network diagrams and movement studies. The church project evokes a more religious character to the efficiency driven infrastructural project from UN Studio. This however, by no means means that Eisenman does not treat the building throughout the same rational transformational procedures the grid deformations and superimpositions of planes lead throughout the design process:

“Fundamentally a disciplinary device in that it situates itself on and undoes specific institutional and discursive oppositions (and that it provides a projective discipline for new work); that it suggests an alternative mode of repetition(one which deviates from the work of the modernist avant-gardes and envisions repetition as the production of difference rather than identity) and that it is a performative rather than a representational device (i.e., it is a too for the virtual rather than the real).[11]”

Diagrammatic processes allow for rational decision making in the sense that “according to formal decision-analysis, any motivation or justification would have to take the form of a criterion-based selection from a field of possible configurations.”[12] Advancements in computational modelling allows for increased capacity in testing variations and potentials of performance (comprehensive, decidable and coherent models). Parametric and computational simulation [Fig. 5] and areas where a feedback loop of a multiplicity of procedures are interlinked and encouraged to relate, are the current tendencies from which a pure rationalist attitude to design could be extracted from, but is in a state of higher complexity. The development of tools that translate variation into sets of designs is also a fulfilled criterion to an underlying rational standing to design. “The minimum requirement for any acts to qualify as a decision would thus be that it preceded by evaluation against at leas one specifiable alternative.”[13] The catalogue of variations and parametric relationships allow for such evaluations.

[Fig. 5] Marco Vanucci, SmartGeometry 08, AKT parametric modelling unit, p.ART

The catalogue informed designed and parametric iterations are at times superseded by design choices, In the case of Arnhem, the development of the network diagrams into a single folded surface that negotiates the hierarchy of circulation and ramp system is linked, by Van Berkel and Bos themselves, as sprouting from the initial movement studies and a Klein bottle [Fig. 6]. Schumacher here comments: “Without rules there is no determination (the amorphous shape of the scheme – obviously not a preconceived figure – inclines us to give crecedence to the suggestion that this figure emerged form movement studies.”[14]

Schumacher criticises this move as one that would in retrospect be a moment where post-rationalisation of a design choice ended up being coherent to the overall scheme, however, it is also the case that inclusion of further parameters such as folding and continuation of the surface allowed for a higher degrees of complexity to become available However way it is interpreted, during the design process “leaps and loops” are taken that push forward the design and its is an easy task to visualize them all as non-linear spot rationalization of the process when the project goes into publication, however, it is only when such turns are accredited to the rigorous of the whole process as assessment where such turns are entirety consistent with rational models of a combined nature: coherent with decisive, etc.

[Fig. 6] UN Studio Klein bottle and folding of layered floors for Arnhem

It depends, but what is certain is that, according to formal decision-analysis, any motivation or justification would have to take the form of a criterion-based selection from a field of possible configurations. Yet such rationality is open to subversion when as yet unthought-of possibilities emerge, for example, when van Berkel and Bois speak of creating a hybrid of a centralized system in an all-linear system.

[Fig. 7] Ron Arad and Buro Happold, Mediacite Liege, Belgium

Parametric modelling is calibrated by the inclusion of further parameters that arise from previous iterations that lead for such new inclusions. This is consistent with the model in as much it further deforms it. It is the capacity of the computer model to reassess an old model with new parameters maintaining coherency of the design process and able to streamline it throughout the design to manufacturing stage. Contemporary techniques of manufacturing and computer modelling are close to the point where architects, engineers and manufacturers potentially operate with the same model throughout the building process. Van Berkel and Bos’ leap from network to surface is now plausible in terms of rationalization of a same computer model in the sense that transformational logics that are unrelated (network to surface) can be assessed and quantified as rigorously.

Current building case studies, such as the Buro Happold / Ron Arad project in Liege [Fig. 7] is an example where Arad’s design is streamlined with engineers at Buro Happold and manufacturers and testing in London for the reduction of the strips that constitute the roof for the Mediacite got tested for structural efficiency and for alteration in fabrication logistics and material quantities. This process for design links the linearity and decomposable segregation of the disciplines involved and in streamlining the process allow for a higher degree of complex decision to be taken as it is feasible as the modifications are updated throughout the whole engineering and fabrication line.

This new relationship between architects and engineers moves a long leap from the Fordist model of early modernism. This allows seeing the whole design process as streamlined rational steps that are at any given moment modifiable and would update iterations down the line automatically and feasibly. Parametric modelling and digital fabrication allows for a true coherent model rational approach to the design model that includes leaps and loops that are consistent and reviewable within the entirety of the model.

At a level that does not endorse a particular ideological allegiance, testing and simulation are also tools that systematise, and rationalises design intent (decidable model). The use of Ecotect, ANSYS, finite element analysis, [Fig. 8] and other such tools, tests the performance of design choices (derived from a catalogue of variations, for example) as means of selecting the most suited design offspring. This approach, already, takes produce from a non-linear process, into a systemic rationalist one, where a series of steps are introduced to validity of a certain process, that in turn, points out which steps from the previous process lead to unsuitable design options and successful ones. A finite set of results feed into the rational model that in turn triggers new decision-making, leading to further bifurcation ns of the increasingly complex tree-diagram.

[Fig. 8] Ecotect Radiation Analysis, ANSYS CFD testing

The history of architectural production over the last forty years can broadly be characterised as the desire to establish an architecture at once “autonomous” and “heterogeneous” in contrast to the “anonymous” and “homogenous” building associated with the iterate rhetoric and post war experience of the modern movement.[15]

[Fig. 9] Generative Components symbol diagram showing parametric relationships

Homogenous fields and parametric model endorse a new idea for spatial production which could be taken as purely derived from a rationalised model of parameters, variations, testing of components, that when organized in an already parametric system gives spatial formations that are rational in the structure of the component relationships and are not credited by the inclusion of adding further transformational processes..

Applications to Schumacher’s category of a rationalist model in the design process of contemporary computational tools can be seen in parametric modelling. Parametric software’s such as Grasshopper and Generative Components [Fig. 9] use diagrams and symbols that help visualize the path of parametric relationships established. This path also builds up on hierarchies of relationship at multiple levels, where one element when modified changes all other parts lower in the hierarchical tree, but, when elements that are controlled in

For Schumacher, a pure rational model that acknowledges the high adaptability of changing between rationalist type models from the other (from hierarchical to comprehensible, etc, etc,) and feeds into the new discoveries taken from the sidetracking the design process sometimes leads to. Establishing a feedback loop from the original design process with the new design discoveries and rationalising them through the first rationalist approach is contended to be a possibility, however, Schumacher would argue that it has become, by that stage, a much more dynamic relationship.

A complex model of such dynamic relationship with an embedded theory of relationships is cybernetics. Architects tapping into this discourse therefore already seek the establishment of loops within their design goals. Emergence taps also into design process from within a system that relies on behaviours and relationships. These follow a series of rules that when iterated new models arise. These models resemble the design process in the fact that unknown variables appear that would then be reconsidered, though they stemmed from scripted finite steps. The capacity of computational power permits these new experiments in rational logics to drive design potentials and the sheer array of options is already a latent “world of worlds” just as Post-modernist diagrammatic practices were initiating in exploring through grid-based transformational de-deformations . Schumacher comments on this: “All we can do is to experiment with a variety of languages and hope that viable repertoires will crystallize through competition in practice,”[16]

The use of rational models independently through the design stages that are called upon the leaps and loops stages of the design processes follow a great rational intent to the overall project. It is in moment of design discoveries where the rational goal is revised, and potentially altered altogether, however, when a feedback loop is established and the rational models that were taken from the leaps and loops are re examined, this new cross examination of the design path becomes a new rational model in the sense of Deciding model of an overall coherent system that got revised precisely by a series of deciding model moments throughout the process. Schumacher points out:

“A previously abundance branch might afford help. A new tree might cut the whole path into doubt but by now it is too risky to jump and too late to return to the roots. At first we seek synergies but eventually we have to cut as much as is necessary to keep the process manageable.”[17]

The coming of age of computation techniques applied in architecture facilitates the possibility to explore the myriad of branches available during the design process Schumacher points out. The capacity of processing design iterations is economised by the reality of construction, and as Schumacher points out, that it’s in research where the explorations of a potential of nonp0liearn relational scope to architecture can be undertaken. Linking this to current economic context, companies have included parametric and computational research as means to bridge the gap between design, consultancy, and fabrication as means to revert costing and streamlining the building process. It is among the midst of an environment of collaboration where non-linear rationalism can be explored beyond the environment of academia. It is my personal opinion that, having advanced from an age where utopian schemes were heroically projected, diagrammatic processes attested vehemently by iterations and moments of intervention from the designer, and now computational informed systems for field-based designs, that rationalism can be called again as s systematic way for design in as much the capacity for establishing relationships along the design path and feedback loops is all too plausible and encouraged in times that call for extremes new models for inhabitation, whether information-based or environmental- (green) based. At the current moment in time, the benefits harvested from the computational experiments have come to define how can environmental disasters be mitigated through design and economical crises be circumscribed by streamlining the building process, bridging costs and maximising efficiency with an increase in limited resources and increased environmental regulations. Schumacher considers that we live in a time of bounded rationality, due to the nature of economisation of the design process, especially with tools that promise infinite iterative capacity. The possibility of getting lost in the wrong tree branch is curbed by the trimming of the tree from becoming a rhizome.

Ultimately, the role of the architect and the design process within the understating of how it can be described design in rational terms has come to terms with the underlying premise of architectural design beyond single authorship and heroism, iconicity and style. Solomon points out:

The role of the architect in this model is dissipated, as he or she becomes an organized and channeler of information, since rather than being limited to the decidedly vertical – the control and resistance of gravity, a calculation of statics and load – “forces” emerge as horizontal and non-specific (economic, political, cultural, local and global). And it is by means of the diagram that these new matters and activities – along with their diverse ecologies and multiplicities – can be made visible and related.”[18]

The diagram is extendable to the rational models that Schumacher describe and that are traceable to various moments of the design process from the Posthumanist to parametric modelling, but, the interrelation of the inputs and outputs and the architects capacity to intervene throughout the process with equal control of the growing rhizome, “with branches sprouting above and below”[19] leads us to consider that long before BIM modelling and generative algorithms the new potential for non-linear rational models was there. At the moment of the realization that the architect was channeler and not a utopian hero, non-linear rationalism becomes a possibility. It is in our current situation, however, where the tools that have been developed and the architects capacity to harness them and the willingness of collaboration with multiplicity of disciplines that truly leads to an undertaking of non-linearity in design which is not accidental or anecdotal, but entirely controllable, calculable and encouraged.

BIBLIOGRAPHY:

SCHUMACHER, Patrick “Rational in Retrospect : Reflections on the Logic of Rationality in Recent Design”, p. 32 – 36, in AA Files 38, Architectural Association, 1999, England.

SCHUMACHER, Patrick, “From Fordism to Postfordism” in “Productive Patterns”, http://www.patrikschumacher.com/Texts/Productive202.htm

GALOFARO, Luca, “Digital Eisenman : An Office of the Electronic Era”, Birkhauser, Basel, 1999

EISENMAN, Peter, “Diagram Diaries”, Thames and Hudson, London, 1999

HAYS, K. Michael, “Modernism and the Posthumanist Subject : The Architecture of Hannes Meyet and Ludwig Hilberseimer”, MIT Press, Cambridge – Massachussets, 1992

ALLEN, Stan, “Diagram Matters”, in ANY 23: Diagram Work: Data Mechanics for a Topological Age, 1998,

SAKAMOTO, Tomoko, FERRE, Albert, “VERB Monograph: From Control to Design: Parametric/Algorithmic Architecture”, Actar, New York, 2007.

---

[1] SCHUMACHER, Patrick “Rational in Retrospect : Reflections on the Logic of Rationality in Recent Design”, p. 32 – 36, in AA Files 38, Architectural Association, 1999, England.

[2] Ibid., p. 33.

[3] Loc. Cit.

[4] HAYS, K. Michael, “Modernism and the Posthumanist Subject : The Architecture of Hannes Meyet and Ludwig Hilberseimer”, MIT Press, Cambridge – Massachussets, 1992, p. 249.

[5] Ibid., p. 251.

[6] Ibid., p. 173.

[7] Ibid., p. 252.

[8] SCHUMACHER, Patrick, “From Fordism to Postfordism” in “Productive Patterns”, http://www.patrikschumacher.com/Texts/Productive202.htm

[9] ALLEN, Stan, “Diagram Matters”, in ANY 23: Diagram Work: Data Mechanics for a Topological Age, 1998, p. 23-16.

[10] GALOFARO, Luca, “Digital Eisenman : An Office of the Electronic Era”, Birkhauser, Basel, 1999, p.31

[11] EISENMAN, Peter, “Diagram Diaries”, Thames and Hudson, London, 1999, p. 8

[12] SCHUMACHER, Patrick, Op. Cit. p., 34.

[13] SCHUMACHER, Patrick, Op. Cit. p.34.

[14] Loc. Cit.

[15] EISENMAN, Peter, Op. Cit., p.9.

[16] SCHUMACHER, Patrick Op. Cit. p.35

[17] Ibid., p. 36.

[18] EISENMAN, Peter, Op. Cit. p. 24.

[19] SCHUMACHER, Patrick, Op. Cit. p.36.