You are on page 1of 17

Landscapes of Change: Boccioni's "Stati d'animo" as a General Theory of Models Author(s): Sanford Kwinter and Umberto Boccioni Source:

Assemblage, No. 19 (Dec., 1992), pp. 50-65 Published by: The MIT Press Stable URL: Accessed: 30/01/2010 16:01
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Assemblage.






t' v











Landscapes of Change:


of Models





11r r1. ^^ _ _~si

Umberto Boccioni, study after che partono, 1912 ~~~~Quelli

assemblage 19

Sanford Kwinter hastaughttheory and the Unidesignat Harvard University, of Illinois, andSCIArc.He is versity of twoforthcoming (1992),andauthor bookson the roleof timeandthe life in relation sciences to modern architecture.
coeditor of ZONE 1-2, The Contemporary City (1986), and ZONE 6, Incorporations

A versionof this essaywas previously publishedin England.WhileAssemblage will not normally reprintmaterial,the editorsfeel thegeneralprojectestablished hereis important for the journal's emerging program.

Matter,accordingto Henri Bergson,is made up of "modifications, perturbations, changesof tension or of energyand nothing else."'The formsof life differfrom this only in their and their capacityto overgreatercomplexityof organization come torpor,2 for both are immersedwithin the same universal streamof durationand constitute not differententities, but ratherdifferentmodalities,of a singleelan vital. Yet even as Bergsonwrote,life was no longerso surely,nor by so great a magnitude,the most complex nor the most autonomous entity in the universe.For duringthe same years,the mathematicianHenriPoincarewas discovering,to his own horror, that the mechanicsof just three moving bodies bound by a - and interactingin a single relation- gravity single isolated system producedbehaviorso complex that no differential equation,neitherknownnor possible,could everdescribe it.3Poincare's discoveryshowed that evolvingsystemswith even veryfew parameters may quicklybe deprivedof their deterministicveneersand begin to behave in a seemingly independent (random)fashion.What this meant was that it was no longerpossibleto show that one state of naturefollowed anotherby necessityratherthan by utter caprice.Time, in other words,reappeared in the worldas somethingreal, as a destabilizingbut creativemilieu;it was seen to suffuse everything,to beareach thing along, generatingit and degeneratingit in the process.Soon there was no escapingthe fact that transformation and noveltywere the irreducible qualities that any theoryof formwould need to confront.4 It was no wonderthat futurism- the socialmovement most - was obdeeply sensitizedto cataclysmicperturbations sessed with complexes: delirious,infernal,and promiscuous. For the veryethics and physicsof the futuristprogram, conceived as an open, far-from-equilibrium system, responsiveto and willingto amplifyeverydestabilizingfluctuationin the environment,necessitatedits multiple impregnationboth in and by the social,material,and affectivesystemsthat surroundedit. The futuristuniverse- the firstaesthetic system to breakalmost entirelywith the classicalone - could properlybe understoodonly in the languageof waves,fields, and fronts.The type of movements it was obsessedby were those that carvedshapesin time not space;it studied the stabilities achievedthroughhomeostaticknots of force in perpetual strife,it embracedthe beautyand evanescenceof becoming.

Assemblage19 1992 by the Massachusetts Institute of Technology



Yet futurism'sprofoundestgift to our centurywas its seeminglyhubristicattempt to link the biosphereand the mechanospherewithin a single dynamicalsystem. Umberto Boccioni'sthree-paintingseriesStati d'animobelongs to this projectand as such comprisesthe firstpurely modalpaintingsin the historyof art since the late medieval The spatiotemporal locus of the trainstation scene period.5 is here splinteredand kaleidoscopedinto so much elementary like matter,but only the better to be redeployedintensively, sounds in a musical continuum or topologicalflows on a twodimensionalplane- scattered,accelerated,accreted,collided into three entirelydistinct surfaces,or developmental fields. One scene, but three modalitiesof inhabitingmatter. As primeexemplarsof modal complexity,it was naturalthat railroad stations should playa privilegedrole in futuristpractice; they were the firstliteral,complex systemsof material flows manifestedat a phenomenalscale whose associated formscould be apprehendedas such, understoodand actively engaged.The dynamicaland morphologicalphenomena associatedwith this type of multiple convergence of flows have alreadybeen developed in relationto this work.6 But the middle panel in the Stati d'animoseries,Quelli che partono, seems to belong to an opposite but relatedproblem,and one that deservesseriousattention. Quelli che partonono longer describesa convergenceof flowsbut ratherthe event of their breakingup, or bifurcation. What does it mean, then, when something stable and continuous ceases to be so?What does it mean when the unfolding of a dynamicalprocesssuddenlyshifts into a new mode, when an ensemble of units and forcesbreaksup to form two or more independent,more highly organizedsystems?The paintingQuelli che partonowedges its own diagonalcascades and chevronformsbetween its two neighborpanels:on one side, the undulating,orbicular, systolic-diastolic processesof and embrace in Gli on the addii, and, organicism depicted verticalstriationsof other, the inertial,gravity-subjugated, Quelli che restano.The fullnessand roundnessof the first workis not simplyone field of shapesamong three, but rather the veryplenitude from which the other two arederived. Between the firstpanel and the other two, there has taken place a catastrophe.

But beforewe can understandwhat this means it will be necesand saryto understandpreciselywhat a form is, how it arrives, why the "formproblem"has been so difficult to handle. Most classicaltheoriesof form arelimited by a majorshortcoming: they areunable to account for the emergence,or genesis, of formswithout recourseto metaphysicalmodels. One of these classicaltheories- perhapsthe paradigmatic one - is the so-calledhylomorphicmodel. Accordingto this model an independentlyconstituted and fixed form is understoodto be combined or impressedwith a certainquantityof hyle,or matter, itself conceived as a fundamentallyinert,homogeneous substance.Once broughttogether,these two abstract elements are said to form a thing. Yet, as we will see, a form can no more be fixed and given in advance (in what space would this workof formingbe done?) than can "matter" seriMuch ouslybe consideredto be either static or homogeneous.7 of this perennialmisunderstanding found itself recapitulated throughoutour modern scientifictraditionbecause it lent itself well to reductionismand controlledquantitativemodeling. Reductionismis the method by which one reducescomplex phenomena to simplerisolatedsystemsthat can be fully controlledand understood.Quantitativemethods, on the other hand, are relatedto reductionism,but they aremore fundamental,because they dictate how farreductionismmust go. Accordingto them, reductionismmust reducephenomena to the ideal scale at which no morequalitiesexistwithin a system,until what is left areonly quantities,or quantitative relations.This is, for example,the basis of the Cartesiangrid system that underliesmost modernmodels of form.8 The classicalgridsystemdoes not, strictlyspeaking, limit one to static models of form,but it does limit one to linearmodels of movementor change.A linearmodel is one in which the state of a systemat a given moment can be expressedin the very same terms (numberand relationof parameters) as anyof its earlieror laterstates.The differential calculusof Newton is preciselysuch a model describingflowson the plane (differential equationsaremechanismsthat generatesets of continuous numericalvaluesthat, when fed into Euclideanspace,appearas linearmovement). But if the standard calculuscan successfully model the evolutionof successivestatesof a system,it can do so only insofaras it plots the movementsof a bodywithin that that the system system,and neverthe changesor transformations


assemblage 19

2. Boccioni,Gliaddii (Farewells),1911

4. The cell, or blister,form of the sand domes that appear on a beach at low tide are the result of an "exfoliation" (the emission or release of a new surface or fold) triggered by a conflict of regimes (an encounter of forces whose sum will deform the system in a particular direction) in the neighborhood of a so-called butterflycatastrophe. The butterflyacts as an organizing center for a shock wave that "knits"the three evolving fronts into a pocket as it passes through them. The blister,or dome, is the morphogenetic "smoothing"of the introduced by the original ~~instabilities *0^~~~~ conflict.

in its young, 3. SpiralNebula NGC4530 still spheroid, ringlessstate



5. Christaller model showing symmetry breaking and the resultant complexity that arises in an initiallyhomogeneous (point) field through even the most rudimentaryfeedback mechanismsbetween the individualpoints. The diagram models economic activityas it distributesitself in a geographical space, carvingup the field almost randomlyinto centers, epicenters, and satellite regions. Thisis due to the proliferationof nonlinearitiesin the evolutionary mechanismand its extreme sensitivityto purelychance factors that are continuallyrecycledback into the system, magnifying their effects.

*1* 4'1

6. Coleoptera larvaeself-aggregating. A gradient field (a field of graduated differences registered by chemicalconcentrations or some other effector-substance) naturallyarises as the larvae begin to emit pheromones into the environment in direct proportionto their level of nourishment. The larvaethen begin to migrate toward regions of greater pheromone (and food) concentration, which, in turn, both increasesthe concentration and steepens the gradient until a definitive cluster is formed. Ifthe field is initially homogeneous but very dense, diffusion of informationwill be very rapidand will soon result in a single large cluster. If the field is initiallyhomogeneous but sparse, signals will be weak and not oriented, resulting in no definitive clustering.For values in between, any number of clusters may be sustained, though only if they are established at the outset. The arbitrarily largerthe initialsize, the greater the chance of a given clusterto persistover time.

7. Escherichia coli bacteria in petri dishes clustertogether into regular,radialpatterns of bunched cells to protect themselves from noxious chemicalsor stimuli (for example, antibiotics).The chemotactic signals that trigger the formation of the patterns are amino acids secreted by each individualbacterium.The clustering is induced by feedback mechanismsbetween the bacteriathemselves.


assemblage 19

9. Gleitbrettershearing in sandy slate. The faulting pattern is caused by the superimpositionof two simultaneous laminarcatastrophes,shearing and folding. 8. Boccioni,Quelliche partono (Those who leave), 1911 11. Development of a spiralaggregation wave in the dictyosteliumslime mold. A remarkable,complex series of events takes place that gives shape and organization to an initiallyhomogeneous field of individualamoebae. The first break in symmetry- a cell-free space in an even lawn - becomes the focus for a global spiralwave that first orients the cells, then gathers them into "streamers," drawing them toward a doughnut ring that surroundsthe initialcenter. Position in the field and responsesto chemotactic pulses causes the cells to differentiate functionallyfrom simple relay elements to complex, self-entrainingoscillators. Suddenly,organized and synchronized waves pass through the field, directing the amoebae to form a single semispherical mound (on the spot of the original void) and a single differentiated multicellularorganismwith a foot, capable of migratingsignificantdistances in search of new sources of food.

* .G




a giant, brilliant 10. Nebula NGC4594, but much older galaxy breaks into two distinctstellar populations- the old stars forming a sphericalhalo, the new ones collecting on the much less dense central disk.



12. Boccioni,Quelli che restano (Those who stay), 1911

14. Cavityproduced behind a sphere dropped in water. A reaction splash (above surface) illustratesThom's elliptical,or filament, catastrophe.

13. Cuspson a beach illustrategeneralized, periodic,cascadecatastrophein water and sand.


assemblage 19

Indeed,not only the systembut also the body itself undergoes. that moves throughit is condemnedto perpetualself-identity: for it, too, can changeonly in degree (quantity)and neverin kind (quality).9 Further,these typesof smooth continuous changesarenot true changesat all, at least not in the deep qualitativesense that we wouldneed to explainthe genesisor of a form. appearance Moderntopologicaltheory,largelyintroducedby Poincare, offereda decisivebreakthrough with respectto the limitations of these systems.On the one hand, it entailedthe revival of to one visumethods studydynamics,permitting geometrical whose complexitysurpassed the allyto model relationships it limits of algebraic on the other, permittedone to expression; not translational the study only changeswithin the systembut the qualitativetransformations that the systemitself undergoes. The classicalcalculusof Newton and Leibnizwas developed alongthe lines of a ballisticmodel, the plottingof and of realbodies againstan inert,featureless, trajectories decould be exhaustively immobilespacewhose coordinates scribedin purelynumericalterms (x,y). Topologyinstead events(deformations) that introdescribestransformational duce realdiscontinuitiesinto the evolutionof the system of a given itself. In topologicalmanifoldsthe characteristics substrate the are not determined by quantitative mapping by the specific"singularspace (the grid)below it, but rather ities"of the flow spaceof whichit itself is part.These singucriticalvaluesor qualitativefeaturesthat arise laritiesrepresent at differentpointswithin the systemdependingon what the systemis actuallydoing at a given moment or place.It is just this variability and contingencythat is of greatimportance. In a generalsense, What exactlyarethese singularities? in continuous process(if one designatepoints any singularities in the time is then the dictum that real, every point accepts universecan be saidto be continuallymappedonto itself) wherea merelyquantitativeor lineardevelopmentsuddenly of a "quality" resultsin the appearance (that is, a diffeomorfailsto map onto and a arises suddenly eventually point phism in a complexflow of materials is what itself).1A singularity makesa rainbow appearin a mist, magnetismarisein a slabof or convectioncurrentsemergein a iron,or eitherice crystals beardesignations pan of water.Some of these singularities at "zerodegreesCelsius,"forexample,denotes the singularity

whichwaterturnsto ice or ice backto water- yet most do not. Thus matteris not in any sense homogeneous,but contains an infinityof singularities that maybe understoodas that emergeundercertain,but veryspecific,condiproperties tions." What is crucialaboutall of this is the following: both "ice"and "water," as well as "magnetism" and "diffusion," areforms,and they areall bornat and owe theirexistenceto that is not Indeed,there is no formanywhere singularities. associatedwith at least one (thoughmost likelymore than one) singularity. In topologysingularities of flowson the planearemorelimited and specificbut can give riseto enormously complexand behavior. These have already been classifiedin variegated various and separatrices whose ways,most often as attractors varieties and combinations give riseto specificqualitiesand behaviors: sinks,sources(repellors), saddles,and limit cycles. Eachof these describes a particular wayof influencingthe movementof a point in a givenregionof the systemor space.'2 Now clearly, a planeis a verysimple,even rudimentary space. A flowin the planecan essentially be described by two paramor "freedom." Most systems eters,or two degreesof variability in the realworld,that is, most formsor morphogenetic fields, areclearlymorecomplexthan this. Yet it is enough to underto gain standhow formsemergeand evolvein simple"2-space" of how morecomplexformsevolvein more an appreciation theory complexspaces.What is centralhere is the dynamical of morphogenesis, whichcharacterizes all formas the irruption of a discontinuity, not on the systembut in it or of it. Fora formto emerge,the entirespace (system)must be transformedalongwith it. is called a This type of local but generalizedtransformation in which a sysA describes the way catastrophe. catastrophe tem - sometimes as a resultof even the most infinitesimal - will mutate or jump to an entirelydifferent perturbation Now it is a basic tenet of the level of activityor organization. laws of thermodynamics that in orderfor something to happen within a system,there must firstbe a generaldistribution of differences within that system. In dynamicsthese arecalled or gradientsand their essentialrole is to link the "potentials" points in a systemand drawflows from one place to another. A potentialis a simple concept:13 anythingsitting on one's



desk or bookshelfbearsa potential (to fall to the floor)within a system (vectorfield) determinedby gravity. The floor,on the other hand, is an attractorbecause it representsone of of the potential in the system.Any state of several"minima" the system at which things aremomentarilystable (bookon the shelf or on the floor) representsa form. States and forms, then, areexactlythe same thing. If the flow of the book on the shelf has been apparently arrested,it is because it has been capturedby a point attractorat one place in the system. The book cannot move until this attractorvanisheswith its basin and anotherappearsto absorbthe newly corresponding releasedflows.The destructionof the attractor(and the creationof a new one) is a catastrophe. Now before developingthis theoryfurtherit will be necessary to make a few observations.It appears,in a certainsense, that the concept of formhas been defined as a state of a system at a particular point in time. In fact, formsrepresent stablemoments nothing absolute,but ratherstructurally within a system'sevolution;yet their emergence (theirgenesis) derivesfrom the crossingof a qualitativethresholdthat a moment of structural is, paradoxically, instability.This is possiblebecause formsare not simplysystemsunderstoodin the classicalsense, but belong to a specialtype knownas "dissipative systems."A dissipativesystem or structureis an open, dynamicalsystem.By "open"one means that it is an evolvingsystem,like a pot of coffee or the local weather,that has energy (information)flowingout of it, and likelyinto it as well. From wheredoes this energycome and to where does it go? It comes from other systems,both those contiguous to it and those operatingwithin it or upon it: that is, at entirely differentscalesof action.We will see what this means in a moment. For now, one need only note that it is the continual feeding and siphoningof energyor informationto and from a system that keeps the system dynamic- simultaneouslyin continual transformation locallyand in dynamicequilibrium globally.The flow of energythrougha system ensuresthe following: 1. That informationfrom outside the system will pass to the inside. The effects of this simple operationare actuallyvery complex:the outside of the system becomes slightlydepleted in the processand transformed in its capacitiesand potential

its energies;the operationaffects the inside by perturbing flowsever so much awayfrom their equilibriaor attractors, the system for potentiallycreativedisturbances "priming" (morphogenesis).It also carriesenergyor informationfrom inside the system to outside, producingthese same effects now in reverse. 2. That informationfromcertainlevels in the system is transportedto other levels, with resultsthat may be verydramatic.14What one means by dramaticis simplythis: certain parts,or strata,of the system may alreadyhave absorbedas much energyas they can hold in theircurrent stabilizedconfiguration.Any change at all, no matterhow tiny, will precipitate a catastrophe(a morphogenesis),forcingthe system to find a new equilibriumin the newlyconfiguredfield. The effect of these liberatedand capturedflowson the neighboring systemscreatesan algebraicproblemtoo complex (because full of nonlinearities)to predict.Qualitativemodeling has a chance, however,because at the veryleast it offersanaboxes"of lytic precisionwherebefore there were only "black forces. mysterious,irreducible of everydissipativesystem perpetually It is the property to seek a rest state or equilibriumwhere it will remainuntil anotherthresholdin the system'sdynamicis crossed.Again, stabilizationgatheraroundsingularities figuresof structural that themselvesare defined dynamically, for these, too, can be maintainedonly at a certainenergycost. Everyrealsystem is made up of other systems,and they are all continuallyleaking informationto one anotherin such a way as to link them acrossa single "continuumof influence."'5 All the formsof the universeare producedas by-productsor maps of particular evolutionarysegments of one or anotherdynamicalsystem. Indeed, formsare not fixed things, but continuous metastable events. Catastrophetheoryis one method for describingthe evolution of formsin nature.It is essentiallya topologicaltheory that describesthe behaviorof forcesin space over time, but its techniqueshave been extended to many realworldphenomena, such as the formingof tools, the capsizingof ships, and psychology(anorexianervosa,fight-flight embryology, theory).This is possiblebecause the behaviorof realforcesin realspace (forcesappliedto a beam, weight poorlydistributed


assemblage 19

in a ship'shull) followsexactlythe same rulesas forcesmodeled in complex (topological,parameter, or "phase") space.16 that Catastrophetheoryrecognizes everyevent (or form) enfolds within it a multiplicityof forcesand is the resultof not one, but many differentcauses. Let us look at how this is done. Heraclitean"science" Catastrophetheoryis a fundamentally in that it recognizesthat all form is the resultof strifeand conflict. It showsthat the combinationof any two or more and disconconflictingforcesmay resultin entirelyirregular tinuous behaviorif allowedto interactdynamically. This means that if one plots these forceson a plane as intersecting at a point, each forcewill be affected unequallyas the point is moved in any direction.The effects of this initial difference producedin one of the forcesmay simplybe compensatedfor, or absorbedby, a proportionate gain in the opposite force;but it may also happen that a smalldropin the firstforcewill trigger a gain in a thirdforce that will diminishthe second force to an even greaterdegreethan the diminishmentundergoneby the firstforce.This will then set up a feedbackcycle between the firstand thirdforcesthat may in a shorttime overwhelm the second entirely.In this case, the second force could actuallybe said to have been fated for demolitionby its own initial strength.Had it been weakat the outset a completelydifferent scenariomay have ensued, one that might have allowedit to dominate in the end. The point here is that conditionson the dynamicalplane areveryerratic,and mere position means far less than the pathwayby whichone arrives there.17 Catastrophe in for these situations. It is intertheoryspecializes accounting ested in the effects of forcesappliedon a dynamicalsystem fromoutside, forcesthat it then becomes the taskof the system to neutralize,absorb,or resolve.As the resultantpoint begins to make its wayacrossthe plane (phase space), it will, accordingto the theory,encounter (nonlinear)regionswhere its behaviorgoes haywire,wheregradual,continuous inputs producesudden, discontinuousresults.Here the system flips - a catastrophe- and gives riseto a whole new state or form. It is the way in which catastrophetheoryresolvesor embraces conflict and differencethat constitutes its radicaloppositionto hylomorphictheory.For catastrophetheorygrantsa certain realityto all virtualforcesin a field, even those that have not been actualized,but remainenfolded until a singularity can drawthem out. A form arisesfrom somethingcalled a

15. Catastrophesurface showing control space, event space, fold, and its projectionas a cusp (the catastrophe set). The plane below representsa Cartesian parameterspace uninflected by any singularity.When a given trajectoryis projected onto the space above it, both continuous and discontinuousbehaviors become manifest. The fold representsan area of special interest and complexity because, for one thing, it is "bimodal," meaning that a single point in parameter space maps onto the fold twice, in two different modes (representedas upper and lower plateau).18


universel("universal deploiement unfolding"),a dynamical in which pathway everyvirtualityis activated,even though some chosen.19 Forms are alwaysnew and unpredictonly get able unfoldingsshapedby their adventuresin time.20 And, as we will see, only a fold offersthe properconditions to sustain anotherunfolding. The idea that everyobject in the worldcan be associatedwith one or anotherdynamicalsystem is not new; indeed, D'Arcy Yet a dyThompson had alreadyarguedthis back in 1917.21 namicalsystem is much more than a substratespace, it is in fact an "evental" complex. Now a catastrophe,as I have alreadysuggested,can occur only in the regionof a singularity. The regionson the plane (of parameterspace) that give rise to catastrophesusuallyoccupybut a small portionof the availablespace and they alwayshave a regular and beautiful form.This form is what is knownas the "catastrophe set" classified Rene (the sevenelementary Thom). catastrophes by This form- the cusp, or catastropheset - is a form indeed, yet it is of a slightlydifferentnaturethan the formsdiscussed till now. Though the cusp fullybelongs to the dynamical system, it is only a two-dimensionalprojectionof the higher dimensional"event-form" unfoldingas a catastropheon the event surfaceabove it. Here the catastropheis actuallya three-dimensional irruptionon a two-dimensionalsurface (note that the action of folding is alreadya passagetowarda higherdimension). What is interestingis that the catastroeven phe set alwayshas the same form (geometrically) though the catastropheevent-form (the specific unfolding) is unpredictableand open-ended.The catastropheset is, in fact, an example of a virtualform.22 Virtualformsare real"folds"(not symbolic,not ideal) in real n-dimensionalspace that can give rise to indeterminatemorphogenetic events in the n+ 1 space (the space one dimension higherup).23 A genuine freedomand indeterminacy 1 in the n+ event space (the catastrophesurface) reigns where formsare actualizedor unfolded, since the precise number,quality,and combinationof realforcesconverging on the fold is quasi-random and unknowablein advance. Indeed, it is more trulythe taskof historiansand theoreticians to reconstitutethese after the fact than for science to predictthem before they happen.24

Among the examplesthat Thom gives of geometricalentities that function like virtualor enfolded formsarehis concepts of or "geneticforms." "charts"

Predator 16. The capture morphology

17. One half of a predation loop

These figures,such as the capturemorphologyillustrated here, are said to exist virtuallysomewherein all biological beings, waitingto be unfolded in a varietyof situations. These are,however,not at all fixed engrams,"butare defined The dynamically, by a kind of never-endingembryology." chartsaretriggeredby so-calledperceptioncatastrophes the sudden appearance, for example,of an object of preyin the visualor olfactoryfield of the predator(note that this event is alreadythe projectionof a fold embedded in another, contiguous space) - that is, by the sudden eruptionof particulargeometricconfigurations in the outside worldthat and a virtual matrixwithin the animal. correspondto, trigger, But the (predator-prey) need not be conceivedas a loop instead, it can be seen as a correspondence phenomenon;25 chance encounterof two flowson the same fold that causes their mutual, spontaneousgeometricizationand common The capture unfoldinginto a single form:the "capture." chreod- the moving template throughwhich virtualforms are actualized- is once againthe n-l "space" that guides, but does not entirelydetermine,morphological events playthemselves out on another linked but ing closely higher dimensionalsurface.26


assemblage 19

19. Epigeneticlandscapeseen from below. The complex relief features of the epigenetic surface are themselves largelythe expressionof a prodigiously complex network of interactionsunderlying it. The guy-ropes are tethered not only to random points on the overhead surface, but to points on other guyropes as well, and to pegs in the lower surfacethat themselves represent only semistabilizedforms, thus multiplying exponentiallythe nonlinearitiesflowing through the system. Not to be diminished in importanceeither is the tension surface above as a distinctdomain contributing its own forces to the field. No change in any single parametercan fail to be relayedthroughout the system and to affect, in turn, conditions all acrossthe event surface.

18. Epigeneticlandscapeseen from above. The evolution of a given form represented by the ball'strajectoryfrom a higher to a lower point - will likely join one of four pathways corresponding to the successivedifferentiations of the rivuletson the valley floor. Yet the introductionof any exogenous forces at any time in the system'sevolution will perturbthe ball from its determined trajectoryand cause it to evolve a unique and original form. Thusthe epigenetic landscape is far from deterministic:on the contrary,it actuallyabsorbsand renderscreative all contingency.



Among the most powerfulgeometricalconcepts invented to depict the relationbetween phenomenalforms (phenotypes) and the morphogeneticfields in which they ariseis Conrad The Waddington'sconcept of the "epigeneticlandscape." is an surface undulatingtopographical epigenetic landscape in phase space (and thereforea descriptivemodel, not an device) whose multiplicityof valleyscorresponds explanatory to the possibletrajectories(shapes)of any body evolving on it. [figure18] (appearing) Assumingthat there exists at all levels of naturea principle to the path of most economic action or least corresponding resistance(whichis only a misguidedlynegativeexpressionof the deeper principlethat everyaction is nonetheless accompanied by its own sufficientconditions), the rivuletsand to builtmodulationsof the epigenetic landscapecorrespond in tendencies, or default scenarios,that would condition the evolution of formsin the hypotheticalabsence of supplementaryforcesacting overtime. But one should not be fooled into takingthe "form" of the epigenetic landscapeas itself
"essential," fixed, or predetermined. For it, too, is only a template, or virtual form, assembled in another dimension, as a multiplicity generated by an extremely complex field of forces. [figure 19] Once time is introduced into this system, a form can gradually unfold on this surface as a historically specific flow of matter that actualizes (resolves, incarnates) the forces converging on the plane. These are the phenomenal forms that we conventionally associate with our lived world. What we have generally failed to understand about them is that they exist, enfolded in a virtual space, but are actualized (unfolded) only in time as a suite of morphological events and differentiations ever-carving themselves into the epigenetic landscape. We would not be unjustified in saying, then, that in Boccioni's Stati d'animo series, what we find depicted are three evental complexes, or three morphogenetic fields, each arising within the same complex system of real matter and forces. Their startling morphological variety can be accounted for by the fact that each is triggered by a different singularity that, in turn, binds it to a specific attractor - farewells: turbulence, aggregation; parting: bifurcation, declension;

The inchoate qualitiesof the staying:inertia,laminarity. that traditionally we are conditionedto see form "fragments" here are, in fact, nothing else than the manifest workof time plyingthe folds of matterto releasethe virtualformswithin it. Each panel defines a unique field of unfolding,a section througha distinct epigeneticlandscapein which formsexist only in evolution or equilibrium,that is, as event-generated the multiple conflictualplayof forces diagrams, incarnating acrossall the dimensionsof space and their modalitiesof convergenceat a single specificinstant in time. Notes
I would like to express my gratitude to The Getty Center for the History of Art and the Humanities whose invitation to be a researchfellow in the fall of 1990 providedthe opportunity to undertakethe research and writingof this essay. 1. Henri Bergson,Matterand Memory,trans. Nancy Margaret Paul and W. Scott Palmer (New York:Zone Books, 1988), 201. 2. Henri Bergson,CreativeEvolution, trans.ArthurMitchell (Lanham,Md.: UniversityPressof America, 1983), chap. 2. 3. Henri Poincare,"Surle probleme des trois corps et les equations de la dynamique,"Acta Mathematica13 (1890): 1-270, and idem, Les Methodesnouvellesde la mecanique celeste, vols. 1-3 (Paris:GauthiersVillars,1892-99). On the homoclinic functions that Poincare discovered,compare IvarsEkelund, Le Calcul, l'imprevu(Paris:Le Seuil, 1984), and RalphAbrahamand Chris Shaw,Dynamics:The Geometryof Behavior,vol. 3, Global Behavior,and vol. 4, Bifurcation Behavior(Santa Cruz:Ariel Press, 1986-88). 4. The biological sciences were the first to turn their attention to a systematic description of morphogenesis: entelechy theory (Driesch), gradienttheory (Boveri,Child), embryonicfields (Gurwitsch),systemfield theory (Weiss), topological deformation analysis (D'Arcy Thompson), Umwelt theory (von Uexkiill), regulationtheory. Finally, in EnglandJoseph Needham, Conrad Waddington, and others created the Theoretical Biology Club to study developmental pathways in complex fields. 5. A combination of Christian neoPlatonism, eastern influences, and the eventual demise of Greco-Roman formalismsduring the Middle Ages freed pictorialrepresentation for centuries from its more static, analytic, spatializingtendencies. This partialemancipation from optical space introduced the modal (or musical) tendencies that we associate with the dynamic, surfacerather affirming,matter-organizing than space-projectingworksof Byzantine and Romanesque art right up to the time of Duccio. Futurism's returnto the modal, that is, to a space emphasizing both movement and "mosaic"qualities, does cubism one better by incorporatingreal, as opposed to derivedor abstract,time into the modern work. 6. See my "LaCitta Nuova: Modernity and Continuity,"in ZONE 1-2,


assemblage 19

The Contemporary City, ed. Michel Feher and SanfordKwinter(New York:Zone Books, 1986). 7. An important systematic critique of hylomorphismcan be found in Gilbert Simondon, L'lndividuet sa (Paris: genese physico-biologique PUF, 1964). See also JosephNeedham, "Matter,Form, Evolution and Us," This ChangingWorld6 (1944). 8. The grid system, just like the socalled Western metaphysicsupon which it is based, can be neither subvertednor deconstructed from within, despite the claims of practitioners who imagine they have introduced their own little cracksinto it. The only "weakforms"that we know of are those weak forms of thought that, unable to resingularize a homogenous, quantitative space, console themselves by attackingit resentfully.Far more interesting is the dishomogenizing act of mathematician Rene Thom, who sees in even the hyperrational, quasifeatureless Cartesian planethe appearance of irreduciblemorphological qualities:in the case at hand, a socalled capture chreod in a predation loop (a point being swallowedby the x and y axes as they scissoraround it). See Rene Thom, StructuralStability and Morphogenesis (Reading, Mass.:Addison-Wesley, 1975), 31415. 9. I am following Bergson'sdistinction between qualitative, intensive multiplicities and quantitative,extensive ones. The formerare defined as those that cannot be divided without changing in nature.This is because they possess creative inhomogeneities - their intensive nature follows from the fact that they are distributed in time (and thereforeunequally,cumulatively, dynamically)not space- and because they are virtualand pro-

ceed by continuous actualization/ differentiation,not repetition/ representation. 10. HasslerWhitney, "On Singularitiesof Mappingsof Euclidean Spaces:I. Mappingsof the Plane onto the Plane,"Annals of Mathematics62 (1955). Whitney is the inventor of singularitytheory and the true father of catastrophe theory. 11. The perplexingresults of recent cold fusion experiments suggest that one is on the verge of isolating a new singularityof water (this may not be cold fusion but it is almost certainlysomething else). How much of human historycould be written from the perspectiveof how cultures organizedthemselves aroundthe singularitiesof materials in general,and water in particular (the hydrauliccultures of the Middle Ages, the fundamentalrole of steam pressurein nineteenthcentury industrialism,etc.)? The question is always:Which singularitieshave crossedthe defining thresholdof human technics and which have not? 12. The terms systemand space had become indistinguishable,I would argue, alreadyin the workof Ludwig Boltzmann,who in the 1870s invented an entirely new type of space known as "phasespace."Phase space is a multidimensionalspace whose coordinatesno longer represent "places"but rather(possible) states of a system. The coordinates representindependent variables that characterizea real system at any given moment, such as its temperature,pressure,speed, direction, and all the possible combinations of these. As a given system evolves through time, it carvesout a precise figure in phase space, forming,as it were, its behaviorportrait.It is this dynamical,virtualform, unavailable

eitherto bruteperception or algebraic or classicalEuclidean expression, that is the proposof the present article. 13. Potential theory is in many ways the precursor of phase space and representsa crucialepisode in the dynamicizationof the classical coordinategrid.J.-L.Lagrange reformulateddynamics in the late eighteenth centurywhen he introduced the theory of the conservation of energy.Each point on the grid came to representa certain amount of availableenergy,or potential to change or move to another state. A few decades later William Rowan Hamilton added momentum coordinates(velocity times mass) to potential coordinates,alLagrange's lowing one to plot the system's total energyas a single quantity on the Cartesianplane. Seen in retrospect, it was but a short furtherstep to the nineteenth century'sprobability theory and Boltzmann'sstatistical mechanics. 14. Sometimes quantum, atomic, or molecularlevel phenomena manifest themselves globallyas in laser-phaselocking, Benardconvection patterns,or percolation transitions.At the biologicallevel dictyostelium amoebas, termite larvae,and neural networksare three verycommonly studied examples of spontaneouslyselforganizingsystems. 15. M. V. Berry,"Regular and IrregularMotion," in Topics in NonLinearDynamics,ed. Siebe Jorna (New York:AmericanInstitute of Physics, 1978), 16-120. 16. See StructuralStabilityand Morphogenesis, chaps. 9-11, for Rene Thom's exhiliratingbut exceedingly bizarredemonstrationsof the mathematics of organogenesis. The chapterthat develops the dynamical theory of genital "chreods"

carriesthe epigraph"Andthe wordwas made flesh" from the Gospels accordingto Saint John. Because there has been an almost systematic mistreatment of these and related ideas within architectural circles it is perhapsappropriate to registerthe following caveat:French catastrophetheory -that of Thom and certain of his Americanfollowers- is different, and to my mind far richer,than the apparentlymore accessible, and well-publicized,catastrophe theory of the English School that of ChristopherZeeman and others at the Universityof Warwickand elsewhere.The formeris primarily a philosophical, theoretical enterprise,indeed, an "artof models" applicableto natural phenomena in generalthrough an intuitive and descriptivegeometrical-mathematicalmethod. The latter, however,often called "applied"catastrophetheory, has recourse,all too often, to simplistic and insufficientlyconceptualized models; it is far more obsessed with predictiveaccuracy and quantitative verification,criteria that arguablyare partlyinappropriateand partlypremature though perhapsnecessaryfor the evangelism with which they disseminate the theory in popular and scientific circles. English catastrophe theory has, as a result, often overstatedits case and suffered many sterile ventures and witheringcriticisms,while the importance and suggestivenessof Thom's workcontinues to increase exponentially. 17. Trajectoriesin state or phase or evoluspace are developmental tionarypathwaysthat describethe transformativeaction of time on a system; they are no longer simply "formsin space"in the classical (reductionist,metaphysical)sense.



In this respect, the "history" of a system takes on an importance unprecedented in scientific method. 18. Clearlyreal objects are constituted in n-dimensional parameter space and inhabit folds so complex that their modality factor may be representedas n-1. For an explanation of this relation, see the ensuing discussion. 19. Evocationis another word often used here (as in biochemical evocators duringembryological development) to describe the simultaneous emergence at certain critical moments within a morphogenetic field of not one, but a variety of new regimes. See Thom, StructuralStability and Morphogenesis, 31-34, 95-96; idem, Modeles de la morphogenese mathematiques (Paris:Bourgois, 1981), chap. 3; and idem, Paraboleset catastrophes (Paris:Flammarion, 1983), 27-28. 20. This extremely interesting and complex aspect of dynamics cannot be dealt with here. I have developed some aspects of enfoldedness and virtualityin my "Drawingas Eros and Memory,"Steel Notes (National Galleryof Canada, 1988) and will treat the problem of "virtuality"in depth in a forthcoming study of formalism. 21. We owe to D'ArcyThompson the originalinsight that genetic programcan contribute nothing more than internal constraintson the evolution of a system or form, that much still depends on what accidental (exogenous) forces are deployed with it to actualize or unfold the individual. See D'Arcy Thompson, On Growthand Form (Cambridge:CambridgeUniversity Press, 1961). See also idem, "On the Shapes of Eggs and the Causes Which DetermineThem,"Nature78 In this same vein, Thom (1908): 111. himself has gone as faras to pro-

nounce his own workto be a type of "neo-Lamarckism." See Thom, StructuralStability and Morphogenesis, 205, 281, 293, and idem, Paraboleset catastrophes,44-45. 22. For Thom, however,the elementarycatastrophesare, on the contrary,pregivenPlatonic forms that somehow determine all morphogenesis in the universe, from the breakingof waves in the ocean (hyperbolicumbilic) to the formation of fingerson the hand (elliptical umbilic). But this formulation seems naive and incommensurate with Thom's other, generallymore nuanced positions. Preferablehere is Waddington's notionof "homeorhesis,"which describesthe by whichinexactmorphoprinciple genetic trajectories,or moving templates, guide the evolution of forms. Waddington's equations describingnonlinearityin embryonic development are among the clearest and most accessible to the lay reader.See Conrad H. Waddington, The Strategyof the Genes (New York:Macmillan, 1957), 16-22. 23. Clearly,the present study does not even attempt to exhaust the immense and peculiarmorphodynamic powerof the fold. It describes only its exfoliativeproperties, leaving to one side entirely its involutionary dynamics, such as those found in the infinite but bounded morphologyof the chaotic attractorand the stretch/fold dynamicsof mixing. 24. BetweenJamesClerkMaxwell and Poincare(and omitting for the moment Boltzmann'simportant contribution)the whole determinacyproblembecame one of the greatwatershedsof late-nineteenthcenturyepistemology.For Maxwell, the impossibilityof achievingabsolute exactitude in the descriptionof physicalsystemswould alwaysgive

instabilitiesthe upperhand and introduceunpredictability into the system. Indeed, Maxwellclaimed that no event ever occuredtwice. Poincareadded to this a mathematical proofdemonstratingthat even impossibleevents must recurinfinitely. Classicalquantitativemethods had simplyreacheda dead end in their attempt to providea rigorous descriptionof events unfoldingover time. These historicaldevelopments clearedthe way for the returnof geometry (qualitativemethods) as a mode of explanation.See "Does the of PhysicalScience Tend to Progress Give any Advantageto the Opinion of Necessity (or Determinism)over that of Events and the Freedomof the Will?"in L. Campbelland W. Garret,The Life of JamesClerkMaxwell (London:Macmillan,1882), and Poincare,Methodesnouvelles,vol. 1. 25. This does, however,seem to be the way Thom seeks to understand it. An alternate,and perhapsricher, generalmodel may be derived from the "enactionist"theories of U. Maturanaand F. Varela,Autopoiesis and Cognition:The Realizationof the Living (Dordrecht:Reidel, 1980). 26. The term chreodis Waddington's coinage, from the Greekkhre or "determined"and "necessary" hodos "route"or "pathway."

4, 13. W. H. Bascom,Wavesand Beaches (New York:Doubleday, 1974). 5, 6. I. Prigogineand I. Stengers,La NouvelleAlliance (Paris:Gallimard, 1979). Redrawnby Joseph MacDonald. 7. Dr. Elena Budrene,Harvard University. 9. E. Sherbon Hills, Elementsof StructuralGeology(New York: Methuen, 1963). 11. Alan Garfinkel,"The Slime Mold Dictyostelium as a Model of Self-Organizationin Social Systems," in Self-OrganizingSystems, ed. Eugene Yates (New York:Plenum Press, 1987). Redrawnby Joseph MacDonald. 14. G. Birkhoffand E. H. Zarantonello, lets, Wakesand Cavities (New York:Academic Press, 1957). 15. Drawnby JosephMacDonald. 16, 17. Ren6 Thom, StructuralStability and Morphogenesis (Reading, Mass.:Addison-Wesley, 1975). 18, 19. Conrad H. Waddington, Strategyof the Genes (New York: Macmillan, 1957).

1. The FuturistImagination:Word + Imagein Italian FuturistPainting, PoDrawing,Collage, and Free-Word etry,ed. Anne Coffin Hanson (New Haven:Yale UniversityArt Gallery, 1983). 2, 8, 12. Privatecollection, New YorkCity. 3, 10. S. Chandrasekhar, Principles of Stellar Dynamics (New York:Dover, 1960).