Anda di halaman 1dari 24

N E W C O N C E P T S

I N C O M P L E X I T Y T H E O R Y
a r i s i n g f r o m s t u d i e s i n t h e
f i e l d o f a r c h i t e c t u r e
a n o v e r v i e w o f t h e f o u r b o o k s o f
t h e n a t u r e o f o r d e r
w i t h e m p h a s i s o n t h e s c i e n t i f i c p r o b l e m s
w h i c h a r e r a i s e d
c h r i s t o p h e r a l e x a n d e r **
m a y 2 0 0 3
** For thosewho knowChristopher Alexander primari ly asan architect, it may perhapsbeuseful to drawattention to thefact
that hiseducation started in physics, chemistry, and mathematics, and that hespent aconsiderablepart of hislifeasaworking
scientist. Seeendnote.
In writing this short overview for a scientic audience, it was very helpful to read preliminary comments
made by Brian Goodwin, Ian Stewart, and Philip Ball who had just read selected pages from proof copies of
Book 1. Their comments contained ideas and reactions that other scienticreaders might share when rst ex-
amining the four books of The Nat ur e of Or der . They were kind enough to draw attention, especially,
tocertain diculties a scienticreader might have, in consideringtheproblemsintroduced, or in makingthem
useful toelds such asbiology, ecology, physics, mathematics, or computer science, and extending tomany mat-
ters currently covered by complexity theory. I have written this paper tomake the connection tovarious scien-
ticelds more clear, and toencourage comment and debate by working scientists.
PRE A M B L E
The four books of The Nature Of Order were tecture today, with Grigor Mendels garden of
sweet peas in .c. Sweet peas, then, were not written, originally, in order to lay a scientic
foundation for theeld of architecture. I n writ- part of science- merely apart of lifepotentially
containing questions, originally unassuming in ingthem, over thecourseof thelast t went yseven
years, I found myself forced to confront unex- their content. Yet they implicitly contained
questions and focused our awareness on new pectedlydeepproblems, touchingnot onlyarchi-
tecture, but other scientic elds as well. Some questions which becameyears later what
we now know as genetics. That is the role that of these questions go so deep that they raise
questions rarely, if ever, faced in the scientic architecture, with its peculiar problems and
challenges, might play for sciencetoday. communit y.
I therefore found myself trying to give an- Thesituation iscomplicated by thefact that
architecture itself (the eld where I have most swerstothesequestions; startingwith answersat
least adequatefor theeld of architecture. I was claimtoexpertise) hasbeen in anatrociousmud-
dle, intellectually. This muddle had to be never writing directly from the point of view of
physics, or mathematics, or cosmology, or biol- cleaned up. And that was my main task during
thelast thirt yyearsasascientist; andasabui lder ogy, or ecology or cognitivetheory. Yet all these
eldsarelikely, in oneway or another, to betou- of bui ldings and communities. (The huge
diculties in architecture were reected in the ched by someof thendingsI havemade.
We thus have a situation, perhaps new, ugliness and soul-destroying chaos of the cities
and environments we were bui lding during the where architecture, generally, in the past very
much therecipient of received wisdom from the acth century - and in themi xed feelingsof dis-
maycaused bythesedevelopmentsat onetimeor natural sciences, is now generating new mate-
rial, and newideasof itsown , which havedirect another in nearly every thinking person, in-
deedI would guessin avery largefraction bearingon thesolution of problems nowclassed
as "complexit y theory," and doing so in ways of all peopleon Earth).
Trying to come to grips with these di- which, though obviously helpful to anyonecon-
cerned with bui lding, havenot arisen beforein culties, required construction of new concepts,
able to cope with the massive and complex na- themother eldsof scienceitself. Tounderstand
exactly what I mean, we might compare archi- tureof thediculties, and ableto focusaratio-
a
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
nal searchlight on questions which were, it congurations(i.e. theprocessesbywhich bui ld-
ings are conceived and made). So, whether I seemed, largely beyond the reach of methods
previously invented in other sciences. These wanted to or not, I had to deal with thesedi-
cult matters, because they lieat the very root of dicultiesarose, in part, I gradually discovered,
fromwidespread but wrong-headed assumptions architecture, and cannot be avoided: even
though the scientic world view and establish- about the very nature of architectureand, in
considerable part, too, from the dry positivist ment had previously not encountered them.
But thisiswherethingsget turned on their view too t ypical of technical scientic thinking
in the most recent era. But they also required headwhereit isarchitecturethat informssci-
encerather than viceversa. Architectureplacesa newwaysof thinkingabout issueswhich hadnot
received much attention in thenatural sciences, newkind of searchlight on certain newscientic
areas of thought fundamental to the study of simply because there was no need for them in
such eldsaschemistry or biology. complexstructuresandthusbecomesrelevant
to alargeclassof problemsrecently beginningto Facing problemsof architecturefrankly, re-
quired conceptual breakthroughs in several ar- gain attention in thescientic communit y itself.
We therefore have the almost unprecedented eas, becauseonecould not honestly confront the
problemsof design, without facing fundamental case of architecture raising scientic concepts,
questions, and answers, that bear on matters of questions of human feeling, spirit, beaut y, and
aboveall t woareasof content: thenatureof con- hard sciencebut which have not, previously,
been entertained. gurations themselves, and the genesis of new
B A C K G RO U N D O N A RC H I T E C T U RE
What aretheessential problemsof architecture . Thereistheissueof ecological and sustainable
and biological connection to theland. that require a new focus, as it might be under-
stood by any scientist whoapplied himself tothe 6. Thereisthevital issueof social agreement re-
gardingdecision makingin regardsto acomplex questionsof architecture.
.. Thereareissuesof value, that cannot besepa- system: this arises naturally when hundreds of
people need to make decisions together - often rated from the main task of serving functional
needs. Thus, aestheticsdismissed as subjec- thecasein thehuman environment.
,. Thereistheissueof emergingbeaut yof shape, tivein much contemporary scienceliesat the
coreof architecture. asthegoal and outcomeof all processes.
a. There is the issue of context a bui lding
grows out of, and must complement, the place Considered carefully from ascientic view-
point, theseissueslead to certain questions, and whereit appears. Thusthereisaconcept of heal-
ing (or making whole) and bui lding into a to certain conclusions.
Architecturepresentsanew kind of insight context.
. Thereistheissueof design and creation - pro- into complexit y because it is one of the human
endeavors where we most explicitly deal with cessescapableof generating unit y.
. Thereistheissueof human feeling: since, of complexit yand havetocreateit not t ypical in
physics or biology, at least not yet. Creation of course, no bui lding can be considered if it does
not connect, somehow, to human feeling as an soft ware in computer science, is another such
arena; and organization theory isanother. I n or- objectivematter.

N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
der to succeed in this very dicult task, which adaptation; [i.e. they havenot had to put theory
into practice, beyond observing, what natureit- poses challenges quite unlike those raised in
physicsor biology, I haveencountered questions, self practicesfor us].
Architecture, because it is so ordinary, and given solutionsto problems, which I believe
are not only useful in architecture (where they aectsbi llionsof people, and coversahugevol-
ume of physical stu, is, from a study point of aredemonstrably so); but that in part thesesolu-
tions and new concepts are almost certainly view, and fromatheoretical point of view, oneof
the rst cases we have encountered collectively, transferabletohelpsolveproblemsin physics, bi-
ology, and perhaps other elds. Although com- asacivi lization, whereit really matterswhether
you do thingsright or not [with theemphasison puter science and organization theory are the
elds where this appreciation of complexit y has doi.e. thisisabout practicemattering].
And it is here, for the very same reason, rst made itself felt, biology cannot be far be-
hind: and even questionsin physics, though ap- that new theory is forced into existence. The
insights into complexit y raised in these four parently moresimple, wi ll (I rmly believe) ulti-
mately turn out to depend on thesamekindsof booksarerelated, without doubt, to theinsights
that haveoccurred in thelast decadesof biology, issuesof complexit y.
Let megivesomenumbers. Wemay under- meteorology, etc. But they aredierent in kind.
I n those elds the scientists are passive as to standbyusingtheconcept of mistakes. A t ypical
house contains about accc man hours of labor. the issue of creation. I n architecture, we are
the active proponents. We have more at stake. Studies sugegst that in these accc man hours
(including both design time and construction I f we are wrong, we create a mess. And the
insights we have gained, so far, though vaguely time) thereisapotential for some keydecisions
of adaptation, per hour. This means, if handled related to the insights gained in physics, chaos
theory, and biology, areunique, morepowerful, wrongly, there is opportunit y for as many as -
.c,ccc possible mistakes in the housedeci- morepractical and if I may say so, far deeper
in content than the insights gained in the sion points where an error can be made. Or, on
the side side, -.c,ccc cases where, if handled passive sciences.
That is why wemust start paying attention well, thehousecan havebeautiful and perfect t
among its parts, and to its environment, and to to architecture, as a major source of insight in
the eld of complexit y. The creation of ne- itsusers needs.
Of course an embryo contains a far larger tuned, well-adapted complexit yas encoun-
tered for example in architecturemust now potential for mistakes: a
50
or .c
15
athousand
tri llion possible mistakes. But so far in the his- takeshapeasamajor topicof theoretical science.
Our abi lit y, or fai lure, to master this science, is tory of science, people have not actually con-
fronted thenecessit y of generating mistake-free crucial to our survival.

N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
S O M E E M E R G I N G S C I E N T I F I C
C O N C E P T S B E A R I N G O N C O M P L E X I T Y
W H I C H C O M E F R O M S T U D I E S I N
T H E F I E L D O F A R C H I T E C T U R E
1 / W H O L E N E SS A N D V A L U E A S A N E C E SSA RY PA RT
O F A N Y C O M PL E X SY ST E M
What can be a measure or criterion of success good . Nor, frankly, can an ordinary activit y like
gardening. Genetics is plainly in a situation to- for acomplex system?I f aself-respecting scien-
tist wasto tackletheproblemof giving structure day, where problems of value are beginning to
surface. And soft ware design, has run into the to the world, in the largeand that is essen-
tiallytheproblemof architecturethenregard- very same problem, and soft ware engineers and
computer scientists have begun to realize that a lessof what shibbolethsmay say, theremust bea
shared criterion of success. I f science, as pres- senseof value, if objectiveand careful, isalmost
theonly thing that can get themout of thepres- ently conceived does not have one that is useful
for architecture, then regardless, we must, of ent mess.
Thevery rst thing any scientist would do, course, nd one. And for it tobeshared, weneed
to nd onewhich isessentially universal, yet ca- if trying to make a sensible theory of architec-
ture, would be to recognize that there must be, pable of being shared by people of dierent
faiths, cultures, and opinions. at thebottomof it, ashared notion of qualit y, of
what weare, collectively, aiming for. The positivistic, value-free idea of art,
which camefromscience, andthedesirethat sci- I f everyone is trying to do something
dierent in a town or communit y, dierent in encehad to createavalue-freescience, pervaded
most acth century thought, and nally infected kind, not dierent in detai l of execution, then of
coursetherewi ll bechaos; just, indeed, what we architectureitself - oneof thesi lliest intellectual
transfusions of all time - since of course archi- haveexperienced in modern urbanism.
Yet for thelast hundredyearsor so, therehas tecture - by it very nature -- cannot manage
without a common sense shared criterion of been ataboo in thescientic communit y which
virtually forbids a scientist (when talking as a good qualit y.
I ndeed, aswemay seeupon reection, ava- scientist) from talking about value or qualit y as
though they really exist. I nstead, it has been an riet y of other scientic elds would also benet
if valuewereunderstood tobeanecessarypart of article of faith that good science comes only
when we make abstract machine-like pictures thestudy of complex systems. Biological systems
cannot be viewed as value free, ecological sys- that do not let our feelings or judgments of
goodnessget in theway. temscannot, landmanagement and erosion con-
trol, and hydrological management in thelarge, This has been a useful article of faith, and
has served science well for four hundred years. alsocannot managewithout acriterionof what is

N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
But it cannot serveuswell now. Why?Because, this way, and go on to say there is no such real
thing as qualit y in architecture. This attitude although it is a feature of non-complex systems
that they can be studied without focusing on destroys truth to such an extent that it cannot
makeasuccessful environment. value, it isalso afeatureof complex systemsthat
they can not bestudied successfully in thisway. So, continuing to repeat themantrathat we
scientistsshould not messwith value, isbecom- Although science (and acth architecture,
too) managed to get through the acth century, ingshort sighted andsi lly. Thefact that in archi-
tecture - if we keep our common senses - we by refusing to come to grips with this problem,
in fact in thelong run wecannot get on without MUST deal with questions of value, does not
mean that sciencecannot benet from architec- solving it. I t MUST besolved in someform. I n
the spirit of science weshall not expect to solve ture. On thecontrary, it means, rather, that the
existence of such questions in architecture - if it all at once. But wemust makean eort, make
atentativestab at it .. and try it out, and then see sensible answers are given to them - is likely to
beasourceof inspiration and encouragement to how we are doing, and improve what we have,
unti l weget something workable. other sciences which are suering from the
sameproblem. Of course architecture in the acth century
also contributed to the taboo on talking about But of course, acknowledging that it would
bedesirableto haveashared criterion of valueis qualit y or goodnessasthough it really exists. I n-
fected by positivism, by postmodernism, and by only the rst step. That, by itself, does not get
you to an operational process for establishing deconstructivism, and exhi larated by a phony
pluralism of anything goes, architects tried to shared value, or to asharable, operational proce-
durefor evaluatingapart of theenvironment, in get by, by saying everyone should do their
thing. . . each person isentitled to hisview, her away that can get shared results. I t seemsto me
that wewould do best if weagreeto keep thisan opinion, and so on. All true enough as com-
mentson thefreedom of human beings. But not open question, andnot closeo thepossibi lit yof
it being solvable, merely according to scientic away to doarchitecturesuccessfully. You cannot
throw the baby out with the bathwater in quite taboos.
2 / A N I N T U I T I V E M O D E L O F W H O L E N E SS
A S A RE C U RSI V E ST RU C T U RE
I f we ask ourselves what kind of criterion of of agood system would bethat it helpsboth the
systems around it and those which it contains. valuewemight beableto rely on, and especially
what kind of criterion we might wish to rely on And thegoodnessand helpingtowardsgoodness
is, in our ideal complex system, also reciprocal. asastandard for thegoodnessof acomplex sys-
tem, it would be rather reasonable to say some- That is, our good system, wi ll turn out to benot
only helping other systemsto becomegood, but thing along thefollowing lines:
I n a good system, we would expect to nd also, inturn, helpedbythegoodnessof thelarger
systems around it and by the goodness of the the following conditions: Any identiable sub-
systems, wewould hope, would bewell that is smaller oneswhich it contains.
I sthisaplatitude, perhapstoo vagueto be to say, in good condition. And we would hope
that thelarger world outsidethecomplex system taken seriously, or worse, tautologous?Not at all.
Asweknowfromrecursivefunction theory, sur- is also in good order, and well. Thus, the mark
6
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
prisinglysimpleideas, whenappliedrecursivelyat about something, not merely a collection of
empt y words. avariet y of nested levels, can haveprofound and
eective consequences - and, often, surprising I ndeed, if the world were marked by sys-
tems, large and small, of which this criterion ones. Both I an Stewart and Brian Goodwin ac-
knowledgetheimportanceof recursiveideas, and (that each system helps the other systems, in
concreteand discernibleways) could besaid, the say that in their view, too, recursivenessof quali-
tiesislikelytobeafeatureof all livingstructure. world would obviously be a much better place.
Water, foodproduction, vegetation, social condi- So, although it appearsto becircular to use
goodness as a concept within the denition of tions, fami lies, education, roads, parks, the
rooms in a house even, the very windows too, goodness itself, this apparent circularit y is only
apparent, not real, and a recursive structure of would all bebetter. Thiscriterion isadeep one,
and it behooves us to nd a precise and reliable thiskind, if followed through, can haveremark-
able and deep results. way of ascertaining what it means (in precise
terms), andof applyingthecriteriontoreal cases, And, as a word of caution to the reader: I s
this denition trivial, when applied in practice? so that we can judge their successes and
deciencies. I ndeed it isnot. Wehaveonly to imaginearow
of houses, in which every househelpsthestreet; Theonlyproblemisthat wedonot yet havea
powerful mathematical representation powerful and in which every garden helpsevery house, to
seethat eventhissimpledescriptionalreadytakes enough to achieve this, just as we do not yet
have a satisfying mathematics of embryonic us far beyond present day architecture. Clearly
contemporary housing estates or tracts do not growth. I n The Nature Of Order I have taken
rst stepsvery tentative onestowards just achieve this ideal, even according to the most
intuitive judgments. So this seemingly obscure exactly such arepresentation, incompletethough
it may sti ll be. yet actually concrete statement is very much
3 / A M A T H E M A T I C A L M O D E L O F W H O L E N E SS
I D E N T I F Y I N G W H O L E N E SS A S A W E L L - D E F I N E D
RE C U RSI V E ST RU C T U RE O F A N E W T Y PE
There isa relatively long-standing tradition of and Ko

hler s discussions of gestalt phenomena


in gure recognition and cognition. The term talkingabout wholenessof spatial congurations
and situationsin theworld. hasalso, in recent years, been used in avariet y of
religious and therapeutic contexts. However I t has been recognized informally, some-
timesmorestrongly, that wholenessisthekey to those, almost always well intentioned, have
rarely, if ever, been clear. many naturally occurring events, phenomena,
and aspects of system behavior . For example, Asaresult of experimentsI conductedat the
Center for Cognitive Studies at Harvard in the Bohrs insistence that the key to quantum me-
chanicsliesin thedependenceof themovement early .6cs, I becameconvinced that wholeness,
"the wholeness we see," is a real, well-dened of electronson theconguration and behavior of
thewhole: Bohmsdiscussion of thewholenessof structure, not merely a cognitive impression.
That thething werecognizeasthe"gestalt of a a quantum experiment as the origin of the be-
haviorsof electronsin theeld; Goldsteinsdis- gure, the pattern of ows in a hydrodynamic
eld, the "something" about an individual hu- cussion of the human organism; Wertheimers
,
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
man face which seems like that persons whole- wholeness, then, of aparticular conguration, is
an ordering on the dierent overlapping and ness, and which we recognize instantly, is - in
each case -- a describable mathematical nested wholes and systems, according to their
degreeof coherence- in short therelativecoher- structure.
However, therewasno then-existing math- enceof theentiresystem of setsand subsetsin a
part of space. I becamesure, slowly, that thissys- ematical structure I knew of, which was able to
capture this "something" or which could em- tem of sets with relatively dierent levels of co-
herence, was the clue to the kind of structure body it.
After several yearsof thought, astohowthis which would capture"the" wholeness.
The wholeness is that global structure structure might be represented, I came to the
conclusion that thecrucial issuelay in thenested which pays attention to, and captures, the rela-
tivestrength of dierent partsof thesystem, pay- systemof wholesthat cover thespace. Byawhole
I mean any relatively coherent spatial set, with ing attention both to theway they arenested in
one another, and how the pattern of strength the understanding that dierent wholes may
haverelativelydierent degreeof coherence. The varieswith thenesting.
4 / O B JE C T I V E M E A SU RE S O F C O H E RE N C E I N C O M PL E X
SY ST E M S, A N D T H E U N A V O I D A B L E RE L A T I O N SH I P
B E T W E E N ST RU C T U RE , F A C T , A N D B E A U T Y
Thereisasensein which Phi lip Ball and I dier in their behavior. But there is a very thin line -
in fact, I would arguethereisno substantial line profoundly. The substrate of his view of sci-
enceat least as it comes across in the inter- at all -- bet ween theissuesof relativecoherence
of subsystems in a physical-mechanical system, viewseems to be that science is about facts,
and therefore it cannot be concerned with aes- and themorecomplex distinctionsof coherence
in an aesthetic entit y - thephrasing of apieceof thetics, because aesthetics is inherently con-
cerned with matters of subjective human judg- music for example.
We routinely study relative coherence in ment, except insofar asaestheticsisconsidered a
matter of cognition. Possibly I an Stewart, too, crystals and economic systems. For example we
can analyze the cleavage planes by seeing that shares some such viewat least some of his
commentson Jenckssuggest that may beso. My some portions of the crystal are relatively more
coherent than others, and that fractureswi ll oc- view is that aesthetics is a mode of perceiving
deep structure, a mode no less profound than cur bet ween the more coherent parts. We can
analyzethesubsystemsof aneconomicsystemby other simpler forms of scientic observation
and experimentation. studiying inputs and outputs and decomposing
thematri x. Such thingscan beanalyzed by ava- How true is it really that aesthetics is non-
factual?And, I would ask, especially, how true riet y of mathematical techniques all depending
on numerical analysisof relativedegreesof con- can thisbe, asour scienticeortsmoveintothe
newterritory of highly complex structures? nection within, and bet ween thesubsystems.
The relative coherence of more complex Consider the relatively simple question of
coherence. Within complex structures the rela- entitiestherelativebeaut y of onecolumn in a
bui lding, versusanother, uglier column issus- tive coherence of dierent parts, dierent sys-
tems, isparamount, and playsaparamount role ceptibleto preciseobservation, and can bemade

N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
apart of sciencebynewkindsof experiment, us- model for thisidea. But, albeit preliminary, it is
in anycaseamathematical model of anewt ype- ing the human observer as a measuring instru-
ment. I f we can construct these experiments in and one suggested by architecture.
Wholeness itself, for example, cannot be such away that weget agreement among dier-
ent observers, and thusobtain hardnosed obser- discussed without makingevaluativestatements.
So where wi ll science be, if it cannot eectively vationsof relativecoherencein thesemorecom-
plex cases, in what way isit helpful to call these discusswholeness. Wolfgang Ko

hler recognized
this problem about sevent y years ago . . . but judgmentssubjective?
I believe it isretrogressive, and wi ll merely hardly anyone reads Ko

hler any more.


Scientists speak constantly as if there is close the door on st udy of more complex
phenomena, to statethat weshould ignoresuch somekind of great dividebet ween fact and aes-
thetics -- the one the province of science ; the observations as necessari ly subjective. Rather it
seems to me that they must be st udied, if we other the province of subjectivit y and art. Yet
thewholepurposeof myfour books, istodemon- are to understand the newly complex systems
we aspi re to deal with withi n a.st-cent ury strate that we cannot have an adequate world
view without a single view of science that em- science. At the very least we should leave the
possibi l it y open. braces both what we now think of as fact, to-
gether with what we regard as aesthetic facts I ndeed, as I have suggested in the books,
wholeness itself is a cousin-like structure to to- and observations.
Whereafter all, did theideacomefromthat pology - akin to structures in topology where
we have a system of nested overlapping sets, aesthetic judgments are subjective?The ancient
Greeksdid not think of them assubjective. Nor some "open" and others "closed." I n the case of
topology, there is a t wo-valued measure for the did the Romans. Nor did the ancient Chinese.
Nor did the great artists of I slam. I ndeed the dierent sets, . or c, open or closed. I n the
denition of wholeness I have oered, we have idea that aesthetic judgment is subjective is a
relatively recent arrival on the scene of human systems of nested, overlapping sets which can
take an innite set of coherence-values from c thought, and one which was recently fueled by
the positivist and mechanistic way of thinking to . along the continuum of least to maximal
coherence. Thewholeness, so dened, describes lessthan .cc yearsago - which scientiststhem-
selves are now rejecting. a vast fami ly of structures that comes from
dierentiated relative coherence, and shows (at There is no need for such arbitrary pro-
nouncements. I ndeed, such pronouncements least aimstoshow) howthat structureof relative
coherencethen creates(by recursion) thecoher- wi ll ki ll genuine scientic investigation in ad-
vanced complexit y theory, not help it. ence of the larger structure.
Obviously, thisstructurecannot bestudied- After all what is science?I t is the study of
what reallyhappens, howtheworld works. Done or even thought - without introducing the idea
of coherenceasan objectiveconcept. I t isin this in such a way that agreement can be forged by
clear thought, andbyempirical procedures. That sense - possibly disturbing to scientists unused
to the idea of recursion - that the conceptual is the picture I have provided. There is clear
thought about structure; and there is empirical framework is recursive.
Further, the issue is greatly complicated by basisand procedurespecied, which allow peo-
ple to form agreed on shared observations, and the fact that the relative coherence of one set,
dependson arecursion of valueswhich aregiven thereby to reachat least tentatively and
roughly at rst shared understanding, and re- to its subsets. I am only too aware that we do
not yet have a nicely worked out mathematical liable results.

N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
5 / F I F T E E N G E O M E T RI C PRO PE RT I E S A S N E C E SSA RY A N D
I N E V I T A B L E G E O M E T RI C F E A T U RE S O F RE A L I T Y
I N A N Y C O M PL E X SY ST E M
The possibi lit y, which is set out in The Nature (the transformations described in Book a), we
gain insight into the dynamical emergence of of Order, that wholeness is bui lt, essentially,
from fteen features of space, comes very close new structure and new behavior.
Let me give an example. Boundaries, and indeed, to Brian Goodwins science of quali-
ties. These fifteen features are described at especially thick boundaries with substance, can
play a role in helping the goodness of a center, length in chapters and 6 of Book ., where
they are described as they arise in artifacts, or in strengthening a center. This happens be-
cause, if t wo systemsareinteracting, thebound- and as they arise in natural systems. Goodwin
has made a compelling argument that qualita- ary condition is often turbulent or a source of
possible confusion. When the boundary zone tive features are observable, and objective in
the sense that they are apprehended by many itself hasdimension, it can then takeon an "in-
bet ween" structure, whichmitigatesor smoothes observers. He implies, but does not exactly say,
that these features - macro-features of systems out thepotential interacting processesin thein-
ner and outer zones. Fami liar examples are to which are not necessari ly to be described by
numerical parameters - do control vital aspects be seen in the very thick boundary around a
living cell (which contains so much vital func- of behavior, interaction, and dynamics. Thus
they are not only important because they are tionalit y), in the edge ecology bet ween a forest
and a lake, or in the corona of the sun which there, but also becausethey often play acontrol-
ling or decisive role in the behavior of the mitigates the interactions of the suns interior
and theprocessestaking placefurther out in the systems where they occur.
This corresponds closely to my own view. near vacuum beyond.
Theboundary playsahugerolein theeect I n the description of functional behavior given
in Books ., a and , again and again, it is and behavior of any system made of other sys-
tems, since the system wi ll literally be riddled the fteen properties which play a decisive role
in the way things work. This is, I believe, with such boundary layersand boundary zones.
Although onecannot say that every center must because the . properties describe the way that
centers are made more alive. Any interaction, haveaboundary of thiskind, it iscertainly one
of the ways in which a living center gets its in which one coherence interacts with another,
wi ll often circle around the way that centers stabi lit y and strength, and capacit y to interact
with other systems. in mutually interacting systems support one
another, or modify oneanother - often to create Not surprisingly, then, a transformation
which gives a given entit y such a boundary anewwholeness. That thisshould happen when
the properties come into play, is only natural, zonenot a very dicult kind of transforma-
tion to induce mechanically as part of any since it is these properties which cause the
functional behavior of the aggregate. There is, developmental processis likely to create a
niche for desirable eects. The transformation in my view, a link bet ween the larger, more
qualitative aspects of systems, and their func- which preserves and enhances structure, by
introducing boundaries, is likely to bring with tional behavior.
Thus, in the properties described in Book it a variet y of positive eects. Thus evolution,
ontogeny, planning, bui lding, and design, are ., and in thedynamic aspect of theseproperties
.c
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
number of thesetransformations, and that there all likely to benet (at the very least in a
heuristic or probabi listic fashion) from such is a calculus of these fteen transformations
as the driving force of all emergence, must of transformations.
The idea that there can only be a limited course be a matter of enormous interest.
6 / A M E E T I N G PO I N T B E T W E E N
C O G N I T I O N A N D O B JE C T I V E RE A L I T Y ?
Ball especially, and to someextent Stewart too, tion useful, mainlybecausewhen wethink of the
resultsthisway, it isthen quitecertain that they come again and again to the notion that what I
havedescribed isreally all about cognition; that all make sense within a fami liar mechanistic
modeof thinking. I t isafai lsafeway of looking is, about the structure which appears in our
minds - not the structure which appears in the at thetheory, becauseit isunassai lable, veriable,
and poses no deep and unpleasantly disturbing world. Assuch it mayhavesomethingtodowith
cognitive theory, but sheds little light on the problemsof ontology.
But that doesnot mean that it istrue, or that "hard" sciences as a commentary on how the
world ismade. it is the most interesting or deepest way to un-
derstand the scientic meaning of the facts I This is a very deep issue, and in some re-
spects it is the central kernel of my claim that have presented. Obviously, if the facts are facts
about theuniverse, theywi ll indeedalsoshowup The Nature of Order is about science and about
thenatureof theuniverse, not merely about hu- in cognition, and the cognitive interpretation
wi ll hold up. I t is, therefore, an entirely safe man cognition or psychology.
Let us begin with the idea that it is in any interpretation.
And of course, onecould also haveatheory case indeed also about cognition. Here Ball and
Stewart would agreewith me, I think. Theidea of architecturewhich iscognitivein origin, and
based on cognition for its foundations. This of wholenessasarecursivestructuremadeof lo-
cally occurring centers, that centersaremadeof would, however, be rather narrow -- even arbi-
trary. After all, why should we pick a cognitive other centers, and theideathat thefteen prop-
erties are the main glue that makes sense have theory of architecture? Why not an anthropo-
logical theory, or an ecological theory, and soon. coherence. . . theseareall legitimateconceptsfor
cognitivetheory. So, too, istheconcept that the More important, thetheory sheds practical
light on issues which have no connection with morecoherent athing is, cognitively, themoreit
wi ll be seen as a picture of the self, or of the cognition. For example, structural design is
made easier and better, when viewed from the soul - as a subjective experience of the knower.
And indeed much of this material had its point of view of this theory. The ow of forces
in acomplex system of structural members, can beginningsin work I undertook, fort y yearsago,
in theCenter for CognitiveStudiesat Harvard, hardly bedismissed asacognitiveproblem. The
forceshavereal behavior and real existence, out- where I was then working experimentally on
problems of cognition under Jerry Bruner and sideof ourselves. I f thistheory of wholenessand
unfoldingleadstogood results, and enablesusto GeorgeMi ller, with Bi ll Huggins, HarrisSavin,
Susan Carey, and others. nd structureswhich elegantly and cheaply re-
solvetheforces, wehavecrossed over into ques- I myself, when I am in my most sober and
pessimistic mode nd the cognitive interpreta- tionsof physical realit y. Yet, it is just so. Accu-
..
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
tive landscape, is also helped by concepts from mulated evidence from my laboratory shows, in
caseafter case, that it isso. this theory. Once again, this subject, in recent
yearsan object of considerablestudy in ecology, Simi larly, problemsof trac ow aremade
moresolvable, from within thisperspective. Al- cannot beconsidered acognitiveproblem. I t isa
problem about thecomplex living system which though tracflowis, remotely, acognitiveissue,
this is once again stretching the point. Here occurs in a hi lly terrain on the earths surface.
Yet, again, there is convincing evidence to sug- again wend thetheory giving ususeful, some-
timespenetratinginsightsintorealisticproblems gest that theconceptsof wholeness, centers, f-
teen properties, and structure-preserving trans- of design in a physical eld that is largely inde-
pendent of human cognition. formations, sheduseful light on waystoorganize
water and riparian areasin aterrain. The ow of water in an ecologically sensi
7 / A N E W , E X PE RI M E N T A L W A Y O F D E T E RM I N I N G D E G RE E
O F C O H E RE N C E , D E G RE E O F L I F E , A N D RE L A T I V E V A L U E
I n The Nature Of Order, an entirely new empiri- tureof Order, andespeciallyinBooks. and, there
arepowerful reasonsfor thinking that thevalue cal procedure, very dierent from traditional
formsof experiment - hasbeen proposed. I t has which inheres in wholeness reects on physical
realit y. I t is not like the kind of trivial social threecharacteristics:
(.) Theprocedureasksaperson to evaluate, agreement weget whenahundredpeoplesay, Yes
weall loveBigMacsbest, somethingwemight experimentally, through subjectiveself examina-
tion, the degree to which a certain system, or looselycall merelyintersubjectiveagreement. I t is
a dierent kind of agreement, which reects on thing, or event, or act enhances the observers
own wholeness. real physical systems, and is more akin to the
agreement several dierent cancer specialists (a) I t turnsout that peopleareableto carry
out thisprocess. might sharewhen they say that acertain persons
haggardfeaturessuggest thepresenceof anunde- () I t turnsout that thereisavery consider-
abledegreeof agreement in their ndings. tectedtumor. Thisisnot at all liketheagreement
shared by the Big Mac enthusiasts. I t is a judg- I t appears then, that after centuries, there
may exist a reliable and profound empirical ment, not an opinion, and is a judgment about
realit ywhichcanbetiedtothepresenceof den- method for reachingshared judgmentsabout the
degreeof valueinherent inacomplexsystem. For ableunderlying structures. Just so with thecases
I describein The Nature of Order. reasonsthat arediscussed extensively in The Na-
.a
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
8 / T H E SC I E N C E O F C O M PL E X I T Y M U ST M A K E RO O M F O R
SU B JE C T I V I T Y , N O T I N T H E SE N SE O F I D I O SY N C RA C Y
O F JU D G M E N T , B U T A S A C O N N E C T I O N
T O T H E H U M A N B E I N G.
Phi lip Ball says my denition of life as given what I have to say, fundamental to the idea of
wholenessassomethingnot merely present in an in The Nature of Order is utterly subjective.
What doeshemean by this? objectivematerial system, but also present in the
judgment, feeling, and experience of the ob- To untangle this statement, one must dis-
tinguish sharply, bet weent womeaningsof sub- server. I n short, cognitive/subjective experience
isarmed by objectiverealit y. jective, t wo quite dierent ways in which the
word isused. I n accusing me, if that istheright word, of
subjectivit y, Ball implies that some bad science (.) We can call a judgment subjective, and
mean that it isidiosyncratic: that is, it isaprod- has crept into Book . (subjectivit y of t ype .);
when in fact it isonlyin thesecondsensethat my uct of onepersonsmind or ideas, and not part of
sharedcanon, or capableof beingpart of ashared commentsaresubjective, but not in therst.
Possibly one of the most important notions canon. That of course, isavalid criticismof any-
thing purporting to be scientic, since the es- in avalid theory of architecture, isthat thejudg-
ments of fact, about qualit y, reside in reachable senceof scienceistheachievement of judgments
that can beshared, and established according to feelingsin anyhumanobserver. I ndeed, theneu-
tral observationsweneed, in order to reach ade- well-dened experiment.
(a) We can also call a statement subjective, quate discussion and comprehension of whole-
ness, areobservationsof at ypewhich can only be if it engages, or includes, thepersonal subjectiv-
it y of theobserver, theI -nessor consciousnessor obtained when we agree to use the observers feeling
of his or her own wholeness, as a measuring instru- feeling of an observer. This is fairly common-
placein science. I t occursfor example, in Chom- ment. Yet, subjective as it sounds to our mecha-
nist ears, this is nonetheless objective. I t opens skys famous opening of structural linguistics,
when he used his own perceptions of what is thedoor to anewstandard of observation, and a
new methodology of measurement. I n architec- grammatical, knowing that others would make
roughly the same judgments he made: and that ture, anyway, where my observations have been
most careful and extdned over several decades, I structures perceived therefore had objective
standing, even though subjectivein theway they cansaypositivelythat validandprofoundresults,
and ndings, cannot be reached without meet- wereexperienced.
Now, it is certainly true that The Nature of ing thiscondition.
I strongly suspect thesamewi ll turn out to Order islled with examplesof thissecond kind,
sinceunion of system behavior with thesubjec- betruein theother scientic disciplinesdealing
with complexit y. tiveexperienceof theobserver isfundamental to
9 / L O C A L SY M M E T RI E S A N D SU B - SY M M E T RI E S
in a coherent structure we are likely to see a sub-symmetries. Thus, complex systems wi ll be
marked by apreponderanceof local symmetries, well-developed system of local symmetries and
.
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
usuallyappearingin aframework of larger asym- Armed with the results of these experi-
ments, I set out to nd a common factor which metries. I have always been interested in I an
Stewart s discussions of symmetry, symmetries, explained therank orderingof coherenceamong
thedierent congurations. I t took t wo yearsto and symmetry breaking, and have myself spent
quitesometimemakingcalculationsabout sym- discover it. Finally, it turned out that when you
count thetotal number of sub-symmetriesin the metries, and trying to nd out how they appear
in complex systems, and how they inuencethe pattern (not theoverall symmetry, but theset of
all local symmetriesin connected sub-regionsof structure of complex systems. Moving on to
mathematics, I wi ll now give an example from the conguration), the most coherent patterns
are those that have the largest number of local symmetry, which shows something of the kind
of power, in very exact mathematical terms, sub-symmetries within them. This does not
means they are globally symmetrical congura- which the wholeness structure has potentially
within itsscope. tions. I t isatotally dierent kind of qualit y.
Now, what is interesting about this qualit y, A number of years ago, I made a series of
carefully controlled experimentsto study simple isthat it isplainlyadeepstructural featureof the
congurations - not something about the way congurationsof black andwhitesquares, andto
obtain estimates of their relative simplicit y and they areseen, but something about theway they
are. coherence. To do this I used experiments de-
signed tomeasureeaseof perception, easeof giv- Further, thepresenceor absenceof thiskind
of coherence is strongly correlated with appear- ing a name, speed of recognition, abi lit y to re-
member, and so on - a variet y of cognitive anceof structurein nature, in bui ldings, in crys-
tals, in fluid ow, in plant colonies and so on. measures, each susceptibletopreciseexperiment.
Several ndings: And, indeed, it is not far from that observation
to theobservation that transforming astructure First, the strong correlation bet ween all
these measures, although they are cognitively to increaseitsdensit y of local symmetries, isone
example of the kinds of structure-preserving quitedierent in character and process.
Second, very strong overall correlation transformations I have described in Book a.
There is no way this can be dismissed as among subjects: meaning that what dierent
people see as simple or coherent is measurable, cognitive.
I t is mathematics: and it is mathematics of and does not vary enormously from person to
person. real physical structures that unfold in three di-
mensional space. But it departsin an interesting Third, that dierencesof perception disap-
pear altogether when we induce people to see way from present conceptions of mathematics,
and onceagain providesinsight intothekindsof congurations in their wholeness. Experiments
show that such a holistic mode of perception is developmentswhich may beexpected, when one
starts working with the model of wholes and achievable, natural, and that onceit isattained it
isstableand reliable. wholenessthat I havedescribed.
10 / D E E P A D A PT A T I O N A S A C E N T RA L C O N C E PT
I N C O M PL E X SY ST E M T H E O RY A N D I N A RC H I T E C T U RE
I dene deep adaptation as the t ype of spatial elements and systems, and which ultimately
causestheharmoniousappearanceand geomet- adaptation which occurs bet ween neighboring
.
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
cally look like in a landscape? What physical rical cohesion wend in all living matter. Deep
adaptation istheprocesswhereby thelandscape, structuredoesit have, t ypically, when it hasoc-
curred in a system?Well of course, that is what or asystem, or aplant, or atown, proceedsby a
seriesof spatiallyorganizedadaptationsinwhich my eort to describeliving structurein Books.
and a and of The Nature of Order isall about. I each part is gradually tted to the parts near it:
and is simultaneously tted by the whole, to its havetried to focuson thephysical character of a
highly adapted or co-adapted system. But thisis position and performance in the whole. This
concept, greatly needing elaboration, is possibly a rst attempt, hardly paralleled at all, by con-
temporary writing in physics, or biology, or the most fruitful point of contact bet ween the
theory of complex systems, and the problem of ecology.
Asaresult, not only isour understandingof architecture(it isthesubject of anewbook, now
in preparation). I nterestingly, neither biology, adaptation limited: wearenai

ve, almost likein-


fants, when it comes to inventing an adaptive nor ecology, nor architecture, nor cit y planning,
so far haveaprofound or i lluminating model of process which creates suitably complex, beauti-
ful, and sophisticated well-adapted structure in this kind of adaptation: mutual adaptation
among thepartswithin asystem. almost any real-world system: among others,
highlyadaptedstructuresinafarmerseld, or in Adaptation , asageneral idea, isavital con-
cept, for example, in John Hollands writing on atown, or in astreet, or in aroom.
I n WolframsA New Kind of Science, for ex- complex adaptive systems. But sophisticated as
Hollandswork is, theadaptation hedescribesis ample, fascinating as it is, and ostensibly about
complexsystemtheory, thereare.acc pagesdis- nearly always described as the process by which
systems of numerical parameters are brought cussing the richness of step-by-step recursive
systemsof rules. Yet thereishardlyaword (actu- within certain numerical ranges. Complexadap-
tation is then described as adaptation for many ally, there is not one single word, I believe), on
the question of how such rule systems, for all variables, at once, often interacting. But littleof
thiskind of thinking hasyet allowed usto form their richness, might beaimed at theproduction
of good structure. agood mental pictureof what an adapted system
really is, structurally, when it occurs, nor how we How can weeven say that wehaveatheory
of complex systems, when wehaveso littleto say might pictureit in detai l for ourselves.
What does adaptation among parts t ypi about themost crucial point of all?
11 / T H E A B SO L U T E N E C E SSI T Y F O R SU C C E SSF U L
A D A PT A T I O N T O B E A C H I E V E D B Y G E N E RA T I V E M E A N S
Phi lip Ball remarked in his discussion, that he The idea that complex structures can only
bemadesuccessfully by generativetechniquesis believes I may be right that the processes of ar-
chitecture (and construction) would need to be obviousin biology, but not yet obviousin archi-
tecture. Nor isit obviousin organization theory, dramatically changed, in order to help create a
living world. or in computer science, hardly even in ecology
whereit hasperhapsmadesomeheadway. Yet in I found this comment reassuring when I
read it, sinceit seemsto meto i lluminateoneof all these cases generative methods must in the
long run beapplied if weareto succeed in creat- the deepest points of contact bet ween architec-
tureand science. ing living structureon thesurfaceof theEarth.
.
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
ories, ideas how to do it, oating about in the I had an extraordinary discussion recently
with a consultant in management theory, who most advanced circles, I asked. No, no onere-
ally triesto do anything likethat, hesaid. I t is was interviewing me. I discussed with him the
idea that adaptational complexit y - hence the just atheoretical idea.
So to himit wasobvious, necessary, true. . . richnessand depth of structureneededin acom-
plex organization - can only beachieved by gen- yet for all that, it hasnot yet been placed on the
agendaof practical action in thebusinessworld, erativemeans. Wewerediscussingthecaseof ar-
chitecture, where it is also something almost a place where innovation is usually rapid and
inventive. hidden from view, and has been replaced with
thesi llyandimpossibleideathat gooddesign(on Wolframhasdoneagreat serviceby placing
attention on the impact of generative methods, the drawing board) can make up for step-by-
step adaptation. and on the extraordinary richness of generative
schemes, and generated structures. I have made My interviewer wasenthusiastic. Wespoke
about human organizations, and I asked if it was simi lar inroadsin adierent sphere. I n Book a, I
havedescribedanewclassof generativesequences commonly understood that acomplex organiza-
tion would only becreated step by step, that isto for architecture: and have argued, I think truly,
that living structure cannot be attained in any say, generated; and asked how thiswasworking
in contemporary American corporations. Oh, sphere, without such generativesequences.
The dierence bet ween my generative weonlytalk about it, hesaid. Even though it is
obvious, almost no one actually does it, or tries schemes and Wolframs is that mine are uni-
formlybasedononetarget: thetarget of generat- togeneratealivingstructureof ahuman organi-
zation by thesemeans. ing living structure. They are not morally or
ethically neutral. I tried to push it. Aretherenot at least the
12 / T H E E F F E C T O F ST RU C T U RE - PRE SE RV I N G
T RA N SF O RM A T I O N S O N T H E W O RL D A N D T H E I R RO L E I N
T H E U N F O L D I N G O F W H O L E N E SS
I n my view, possibly the most signicant single I n any case, complexit y theory itself cer-
tainly has the knowledge and vision about the scientic ideain The Nature Of Order isthecon-
cept, rst presented in Book a, of a structure- importance of dynamic approaches to adapta-
tion. Genetic algorithms, annealing algorithms, preserving transformation. This concept arises,
naturally, from the concept of wholeness. Once and theentiretheory of computationally derived
dynamic structure, all attest to it. Yet, where in wehaveaconcept of wholenesswhichisnot vague
mumbojumbo, but acoherent andin-part mathe- complexit y theory, is there a straightforward,
common sense exposition of the general princi- matically denable structure for any given con-
guration, wearethenabletoask, of anychange, ples underlying the successful adaptation of a
complex structure, in real time, asareal practi- or modicationof thisstructure, or for anyevolu-
tion of that wholeness, whether thenew whole- cal matter?
Possibly the most important lesson of the ness emerges and continues naturally from the
previousstateof thestructure, or if it isin some discussions in The Nature Of Order, lies in the
way that the concept of a structure-preserving senseaviolation of itspreviousstructure.
.6
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
transformation or wholeness-preserving trans- impossible, of course, even to ask how hard
it is to nd the good congurations. formation fundamental to the proper design
or planning or construction of any bui lding However, in The Nature Of Order I have
established, I believe, that the goodness of an which haslifemayturnout tobeafoundation
stone, in the end, of the whole science of com- environment isamatter of fact (you would need
to read the whole of Nature of Order to under- plexit y theory.
I believe it is clear to us, intuitively, that stand why), not of subjectiveaestheticjudgment.
This has therefore allowed me, perhaps for the it is very hard to reach a well-adapted state
of any system we are trying to meddle in, or rst time, to ask concrete questions about the
t ype of processof design, or of construction, bui ld. I t is certainly clear as a topic in theoreti-
cal biology where Stuart Kaufmann, for exam- or of planning, or of step-by-step urban re-
newal andtoask what kindof processesmight ple, has tried to show how this happens in a
t ypical biological system. To my knowledge, enableusto get ahigher rateof successin reach-
ing good structures in our surroundings. the dicult y of nding good congurations in
a landscape, or in a street, or in a bui lding Here I have had some considerable suc-
cessand, as in other cases mentioned in this equally dicult problemshave not yet been
widely acknowledged by architects. This is paper, it seemsto methat thescientic commu-
nit y might learn a great deal about complexit y, mainly because in architecture the goodness
of dierent congurations has not yet been by focusing on the character and technique of
this success. accepted as a matter of fact. That makes it
13 / T H E H U G E N E SS O F C O N F I G U RA T I O N SPA C E A N D T H E
W A Y T H E T RA JE C T O RY O F A C O M PL E X SY ST E M
C A N RE A C H A D A PT A T I O N
I t isverydicult tondor design, or plan, com- I n Book , I havegiven anumerical estimate
of the relative number of good (well-adapted) plex structures of the t ype of complexit y t ypi-
cally encountered in bui ldings, neighborhoods, congurationscompared with thenumber of all
possible congurations. I n one calculation I gardens, even rooms. There are many possible
congurations. Only a few of them work well, reached theconclusion that theratio of success-
ful, well adapted, congurations to all possible andonlyveryfewof themshowthesubtleco-ad-
aptation amongthepartswhich createstruehar- congurationsisastaggering. in .c
12,000
. Thisis
so sparse (remember, there are only .c
44
mole- mony, or truly good functional behavior.
Kaufmann hasspoken eloquently about the culesin theocean, and only .c
80
particlesin the
known universe) that onecan hardly imagine tnesslandscapeand theproblem of nding the
good solutions. This problem exists because the how a system ever nds these few isolated
congurations. good solutionsareso tiny, likespecksof dust in
the vastness of a conguration space. The rela- Of course, the way it works i n nat ure is
not by search. The system does not wander tively rare living structures, viewed as points in
conguration space, are so small and so far about i n conguration space, looki ng for these
ti ny and very rarely occurri ng congurations. apart, that thechanceof ndingthemby search,
by design, isalmost vanishingly small. I t isto all I nstead, it just goes there. The grai n ripens.
The corn forms. The stal ks are threshed. The intentsimpossible.
.,
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
our is mi l led. The st ubble is ploughed back tions, when calculated properly, and when car-
ried out in adisciplined manner, havethepower i nto the soi l. One thi ng fol lows another, and
i n a particular way which leads from one good to reach the very rare and hard-to-nd good
solutions by stepwise transformations, where a conguration to another, and i n such a way
that any nat ural process gradual ly leads to- search procedureor adesign procedurejust sim-
ply cannot reach it in aniteamount of time. I n wards and homes i n on the good, wel l-
adapted congurations. other words, design cannot succeed in producing
optimal or sub-optimal solutions. Systems of Kaufmann hasbegun an eort to makethis
process precise by means of autocatalytic sets, well-oriented transformations performed over
time, can work, do work, and aretheonly tools and by showing the kind of path such a system
takes, and that it leads, morelikely than not, to for creating deeply adapted living structure in
towns and bui ldings. the well-adapted congurations. Of course his
work is incomplete, and heleaves uswondering What this means, in practice, is that as a
structureevolves, weguideitsevolution by par- just how, exactly does this work.
Myownview, basedonthirt yyearsof trying ticular sequences of structure preserving trans-
formations: these are the transformations de- to solve this problem with bui ldings, is that the
technique which must be used, is a new tech- ned by thefteen properties. I n caseafter case,
I have shown that eective adaptation occurs niquethat focusesonemergenceviawell-dened
structure-preservingtransformations. I t takesus when it isguided by carefully chosen sequences
of these fteen transformations, applied one to the sweet spots in conguration space by a
series of transformations based on the fteen after another, to the product of the previous
transformations. properties identied earlier. These transforma-
14 / M O RE O N A D A PT A T I O N A L SU C C E SS A S A SPE C I A L K I N D
O F T R A JE C T O R Y T H R O U G H C O N F I G U R A T I O N SPA C E
Therehasbeen mention of Stu Kaufmann, and What is common to these cases, is the ex-
traordinary numerical problem which every hisinnovativeideas. I t may help put thepresent
discussion in context, if I make a few remarks adaptivesystem faces.
I n The Nature of Order (Book appendi x) I about thenatureof hisachievement, asI viewit,
and how it pertains to the scientic problems havemadeacrudeestimatecomparingthenum-
ber of possiblecongurationsin agiven bui lding raised in The Nature of Order.
I n all complex systems, the key question is: design problem, with thenumber of thosepossi-
blesolutionsthat arelikelytobewell adapted How doesthe complex system receiveits order?
I t occurs in the growth of an organism, and its hencetohavelivingstructure. Theratio of these
t wo numbersistruly astonishing. trajectory asit becomest for itsenvironment; it
arises in the breaking of a wave to produce the I n my estimate there are, in all, .c
2,000,000,000
possible congurations; and of these there are beautiful congurationscaptured by Hokusai; it
arises in the evolution of organisms, and in the approximately .c
1,998,000,000
good congurations.
The absolute number of congurations both in attempt to nd agenomewhich iswell adapted;
it occurs in architecture, and the attempt to the good pi le and in the all pi le, are im-
menseimmense beyond imagining. There is make a bui lding, or a place in a town become
well-adapted. therefore no shortage of good solutions to any
.
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
givenproblem. But it istheratioof thet wonum- these.c
40
. Let usimaginenow, that westand on
thesurfaceof that onemolecule, and onceagain, bers which staggers the imagination. The ratio
bet ween the t wo numbers is, in rough terms, weimaginethat onemoleculelikeearthsocean,
with .c
40
of its own miniature particles, and about .c
12,000
. Further, although there are huge
numbers of possibly good congurations, these again wehaveto nd oneparticular tiny particle
among the .c
40
, which is a good one. Now we good onesaresparsely scattered throughout con-
guration space, they are certainly not nicely havesearched for somethingwith ararit y of . in
.c
80
. Let us now jump down again, and again grouped in any one part of conguration space.
What this meansisthat theproblem of nding treat thissecond earth asthereal earth, and once
again, now nd our way to a particular particle therelativelygoodcongurationsis, in principle,
aproblemof staggeringdicult y. I t isnot merely which hasararit y of . in .c
120
.
Let us now continue this procedure again likending aneedlein ahaystack. I t isnot even
likending asingleparticle, among all thepar- and again and again. I n order to nd a particle
with rarit y of . in .c
12,000
, we have to perform ticlesin theknown universe; that would merely
beaproblemof nding oneparticleamong .c
80
. thisextraordinary jump into asyet smaller uni-
verses, nolessthan threehundredtimes: wehave This problem is inexpressibly large by compari-
son. The compactness of the written arithmetic to perform this jump into a domain as large as
the molecules in the earths ocean, no less than expression .c
12,000
belies the true immensit y of
theactual number. three hundred times, one after the other. Only
then do weget near to our objective. Thistask issohugeastobealmost unimag-
inable. But wemayimagineit likethis. Consider And remember, thissearch for theneedlein
thisgargantuan haystack isnot an extraordinary the number of water molecules in the earths
oceans: about .c
40
. Suppose then, that we initi- task. Thisis, arithmetically, what happensin all
adaptation. I t is a process which happens every ated aprocesswhich allowed usto nd oneof a
mi llion specially marked molecules, among timethat successful adaptation takesplace.
15 / W H O L E N E SS- PRE SE RV I N G T RA N SF O RM A T I O N S A RE T H E
PRI M A RY W A Y S T H E T RA JE C T O RY O F A C O M PL E X SY ST E M
I S A B L E T O RE A C H SU C C E SSF U L A D A PT A T I O N.
How can a complex system nd its way to the uration space! What this means, in common
sense language, is the following. There are far good congurations? I n a theoretical sense, we
may say that thesystem walksthrough congu- too many possiblecongurationsfor agiven de-
sign problem. We cannot hope to nd good, or ration space, taking this turn and that, and al-
waysarriving at awell-adapted conguration. well adapted designs, merely by looking for
them. I nstead, wemust haveprocesseswhich The huge question, of course, is How this
walk iscontrolled: what aretherulesof thewalk, when applied to a given starting point for a de-
sign problem, or for a planning processwi ll that make it lead to good adaptation?Although
afew, very preliminary answershavebeen given takeusto good answers.
Nature has a waybui lt into the majorit y to this question, no good ones have yet been
given. Thisis, perhaps, THE scientic question of systems, of nding its way to a well-adapted
statefor any given complex systemat least for of our present era.
I n particular, in architecture, it is essential most cases. We do not have a way. For bui ld-
ingsand indeed for any complex system that we nd practical waysof traversing cong-
.
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
thisisthemost fundamental practical issueof all. conguration space- sincewecan nowadaysdo
virtually anything, and natural processesarenot And for all the complex systems on our planet,
that isthemost vast, andmost signicant problem steering us towards regionsof the conguration
space. Thefact that we, ascreatorsof bui ldings, of our era.
I n general wemay characterizethistask, as have to nd the tiny, nearly invisible needles
in the vast conguration haystack, makes our a task of walking through conguration space,
unti l we reach good results. The assumption practical task sti ll harder.
I have therefore spent much of the last is that there are (indeed, there must be) some
kinds of paths through conguration space t went y years, trying to nd out what practical
methods there are, for helping us traverse con- which can get asystemtothegood places. When
conguration space is smooth like a softly hi lly guration space, and for nding genuinely pro-
found and well-adapted bui ldings. I t turns out terrain, one can get to the peaks, generally, by
walking uphi ll unti l you get to the top: if I get that thefteen propertiesassociated with whole-
ness, described at length in Book ., provide a to thetop of onehi ll and it isnot high enough,
I walk to another hi ll, and go uphi ll further. substantial part of theanswer. I n Book a, I have
shown how living structureariseswhen reached These hi ll-climbing procedures only work on
smooth hi lly terrain, with not too many hi lls. by aseries of movements in conguration space
which are"structure-preserving" paths. Thisin- Kaufmann argued correctly, that real congura-
tion spaceisnot nicely behaved likethis, and so volves the use of the fteen properties as trans-
formations, not merely as geometric properties. he attempted to give an answer as to how an
adaptive system does get to the innitesimally When wehavearandom conguration, and are
trying to improve it step by step, we are most rare points that represent good adaptation.
Stu Kaufmann has made simi lar calcula- likelytoreach zonesof livingstructure(thegood
spots in conguration space), as we apply these tions, andpublishedafewyearsago, inhismonu-
mental work on theoretical biology where he transformations successively. The repeated use
of thesetransformations-- intensifying centers, showed that in principle, anyway, certain kinds
of movementsin thetnesslandscape(ashecalls emphasizing alternating repetition, increasing
densit y of local symmetries, and so on -- have conguration space), which may indeed home
in on "good" areas - and has oered a number immediate benecial eects. These are real ex-
planations, which have practical eects in real of in-principle explanations suggesting the
emergenceof living systemsisprobable, not im- practical bui ldings. And what it amounts to, in
informal language, is that the transformations probable, because of constrained movements in
congurationspacewhichleadautonomouslyto- represent a coded and precise way that aesthet-
ics- theimpulsetowardsbeaut y- playsadecisive wards well-adapted systems.
I applaud his work, and have been inspired role in the co-adaptation of complex systems.
I f I am right, the consequence of my argu- byit. However, that said, onemust alsoacknowl-
edgethat Stu hasnot yet given specic explana- mentsgofurther, andwouldseemtosuggest that
adaptation - the successful movement around tions, which describe, through such and such
detai led mechanisms, how any specic features conguration space - cannot succeed unless it
usesthistechnique. I ndeed, I believethestruc- that appear in living organisms actually arise.
Hisexplanation isat averyhigh level of general- ture-preserving transformations are likely to
have real practical eect on our understanding it y, and though convincing, leavesthehard work
of guring out how it really works, to createthe of evolution and ontogeny. I n these elds, too,
I believe it wi ll turn out that these concepts geometrical congurationsweobserve, toothers.
This is of particular concern to me. I n the are indispensable, and give answers to presently
open questions. world of bui lding onereally doesfacetheentire
ac
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
by step, in an incremental manner. I t might be said that thesefteen transfor-
mations coupled with the ideaof structure-pre- This might be powerful enough, in princi-
ple, to help us act as nature doesto create servingtransformations, doinpart nallyaccom-
plish what StuKaufmann hasnot yet attempted. adaptation and beaut y in complex systems, not
only eciencyand to restoreto thelandscape Theygiveusreal (andpracticallyworkable) indi-
cationsof how to reach thetiny good zonesin of Earth what wehavebeen busy, for t wo centu-
ries, unintentionally taking away. conguration space, by traversingthespace, step
A SU M M A RY :
T H E RO L E O F B E A U T Y I N T H E SC I E N C E O F C O M PL E X I T Y
All in all, to wrap up, this might be said: The by theoperation of thesamefteen transforma-
tions, which generatesthebeaut y of theower. I beaut y of naturally occurringpatternsand forms
hasrarelybeen discussedbyscientistsasapracti- believethat it isthesamefteen transformations
which mitigate and channel the crumbling and cal matter, as something needing to be ex-
plained, and aspart of scienceitself. Yet thef- heavingandbendingof thegeologicstratawhich
generated thebeaut y of theHimalaya; and these teen transformations, if indeed they provide a
primary thrust in theengineof evolution, and in fteen transformations, too, which mitigate the
action and swirling of thevorticeson Jupiter, or themany enginesof pattern formation, giveusa
way of understanding how beaut y - aesthetics - the rippled piebald congurations we call a
mackerel sky. plays a concrete role, not an incidental role, in
theformation of theuniverse. I knowthismust seemafantastic claim, es-
peciallysincewehavelearned somuch in thelast I believe the fteen transformations I have
discovered wi ll turn out to be naturally oc- t wocenturies, byinvokingpuremechanism, un-
guided and un-channeled. But we should re- curring, and necessari ly occurringin all complex
systems. Thelawsleadingtotheir existence, wi ll member that our current claims for the success
of contemporary methods, are indeed only turnout, I think, tobeinevitableor necessaryre-
sults of the unfolding of wholeness, under the claims, not yet proven tobesucient. I ndeed, in
thewritingsof each of thethreescientistsinter- right conditions. And I believe, too, that our
acth-century notion that mechanical eects, viewed -- Brian Goodwin, Phi lip Ball, and I an
Stewart -- thereisfrom timeto timevery frank without the guiding inuence of these fteen
transformations, can create the beautiful struc- acknowledgement of certain subtle, unsolved
problems, usually residing in the more holistic tures we encounter in the universe, is simply
wrong. I n other words, it is the action of wave aspects of the emergence of certain truly com-
plexwholes. All thesesubtleproblemshavetodo motion, mitigated by the fteen transforma-
tions, that creates the beaut y of the breaking with certain phenomena for which one cannot
quitegivebelievableoperational rulestoexplain, wave; it istheoperationof natural selection, mit-
igated by theaction of thesefteen transforma- or predict their occurrence. Nor can one t ypi-
cally create the truly beautiful new congura- tions, which generates discernible and coherent
formsin theplay of geneticsand evolution; I be- tions - except when we know already what they
are- in which casewecan of coursealwaysgive lieve it is the operation and unfolding of the
most ordinary ower or stem of grass, mitigated after-the-fact explanations of how they got
a.
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
there. Asan architect, I amparticularly awareof ural selection, and is the crucial missing part of
current explanations: a vital component in the thisproblem, since, by trade, I amalwaystrying
to get to new, beautiful congurations, which gamut of selective pressures. We need more
frankly to acknowledge such a possibi lit y, and have not been seen before. And I have learned
howto do it successfully. in my view scientists who aspire to realistic ex-
planations, like Dawkins, should stop dueling I am thereforeparticularly interested in the
fteen transformations (which I have described with creationists (which is far too easy), and
instead try to focuson thisgeometrical problem asthe"glues" of wholeness), asthemost power-
ful heuristicsin conguration spacethat I know at its root (which is much harder). I believe the
fteen transformationsI havedescribed go some of, because it turns out that these transforma-
tions do have the power to help reach new, and distance to laying a path toward the solution of
these diculties. truly beautiful congurations, and I believethey
doalsohavepredictiveforcein helpingtounder- Most scientists, and most lay people, share
intuitions (not always acknowledged) which stand how naturally occurring complex adaptive
systems nd their way to truly beautiful new ascribe something great to the action of the
universe. Roughly expressed, these intuitions congurations.
Why do the creationists keep on making rest on intuitive assessments that some deeper
coherent, and morewhole-oriented transforma- their fussabout evolution?I donot think it isonly
because of religion, but rather because some of tions, coupled with the action of the ordinary
mechanisms we understand, and strengthening them are aware that this problem of emergent
beaut y is not really solved. Why does Dawkins and reinforcing the wholeness which exists,
can give birth to new and beautiful congura- engage in such intense hand-to-hand combat
withthecreationists- somethingonewouldthink tions from the wholeness which exists. What
the arguments in The Nature Of Order attempt, hardlyworth theink?I sit not becauseof hisown
fai lure to acknowledge, more frankly, that the isto maketheseintuitionsprecise, and suscepti-
ble to experiment. larger question of emergenceof new, and beauti-
ful congurationsin evolution isnot yet solved - And let meunderlinethepoint: Thefteen
transformations dened in The Nature Of Order at least not inthesensethat computer simulations,
usingthealgorithmsof selection ascurrentlyun- arerooted precisely in thequalitieswhich make
thingsbeautiful. That ishowI found them. And derstood, could yet arriveat truly beautiful new
congurationsandthusdemonstratethetruth of that is why they work.
I s it not preferable (and more likely) that theideasof evolution aswecurrently understand
them? Approximations to beautiful congura- some relatively straightforward process of the
kind I have described, rooted in beaut y - yet tionscan besimulated, yes- just asin thecaseof
snowcrystals. Thereal thing - just asin thecase denable and in principle mathematizable, as
these transformations are -- is acting to help of snowcrystals- not quiteyet.
The successful evolution of new biological produceglobal, whole-orientedstructures, rather
than ascribing their appearance to creation, or forms is, in my view, undoubtedly modied by
transformations able to move toward structures to luck, or to blind chance, or to the action of
purely local, over-simplied, equation-driven that areinherently- that istosay, geometrically-
coherent. I believethisprocessaccompaniesnat- processes?
aa
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
C O N C L U SI O N
Therichnessof theseconcepts, their associated I n such a world, scientists would do better,
theprofound questionsof health, wholeness, na- questions, and theinherent interest of someten-
tative answers to these questions, do indicate a ture, ecology, and human joy, would bepart of a
single world view, in which it would be recog- new source of stimulation for the classical hard
sciences - all of which, for the moment, ema- nized aspart of science- scientia- that isto say,
knowledge - and in which scientists and artists natesfrom architecture.
I wonder if it may not beworth making the together, speaking a common language, would
take part in this joy, to the benet of all following comment, on themental world which
scientists inhabit, and on a possible way of ex- humankind.
But, of course, this could not be the dry- tending that world, to thebenet of usall.
I said at theoutset of thispaper, that scien- stick view of mechanizable questions such as
trac ow, or strength of materials. They would tists only rarely make things, and even the talk
about complexit ycurrent in thelast t wodecades, have to embrace the real questions, the hardest
questions, of the relationship bet ween human choosesverylimitedformsof complexit y, looked
at from very limited pointsof view. joy and health, and thegeometrical organization
of the planet, as a source of life, at every scale. Yet therich sourceof scienticconceptsand
explanation I have sketched, suggests that it is That would, indeed, changesciencefor ever.
Of coursetherearesomewhowouldsaythat thereal adaptational complexit y of theeveryday
world around us, which is potentially a rich work of thiskind iswholly inappropriatefor sci-
ence, and that they prefer a vision of science source of science: and also a profoundly tting
arenafor scientic eort. which ismore modest, small in scale, and deals
only with potentially and immediately answer- I f the house, the garden, the street, cities,
landscapes, worksof art, wereto becomenormal ablequestions. Assomeonewho wasmyself also
nurtured in that English empiricist tradition, I objects of our interest, and that the creation of
such things, instead of being split o asart or havemuch aection for that view.
But in any case, even if one takes Phi lip planning were to be given the deep aection,
passion which it deservesif, in short, theaims Balls sensible, modest and empiricist view, it
cannot be denied, I think, that the questions of science would move from analysis and hy-
pothesismaking, toalarger view, in which mak- raised in this paper, are important scientic
questionsby any standard at all, that themeans ing were also to be includedwould we not
then have a more beautiful science, one which of solving these questions exist, are new but
workable, and that, to makeprogressin existing really deals with the world, one which not only
helps us understand, but which also goes to a and nowcomingeldsof scienticenquirythese
questionsmust beanswered. Even thetentative, deeper level, and begins to encompass the wis-
domof theartist, and beginstotakeitsresponsi- and necessari ly partial answers to these ques-
tionswhich I havegiven, doopennewdoors, and bi lit y in healing the world which uninteionally
it hasso far created, and which it has, sadly, and mark new paths of enquiry and oer empirical
solutionsto problemsin thenatural sciences. unintentionally, so far helped to destroy.
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y
a
not e
Professor Alexanderseducation startedin thesciences. Hewasawardedthetopopen scholarshiptoTrinit yCollege, Cambridge
in ., in chemistry and physics, and went on to read mathematics at Cambridge. He took his doctorate in architecture at
Harvard (the rst Ph.D. in architecture ever awarded at Harvard), and was elected fellow at Harvard Universit y in .6..
During the same period he worked at MI T in transportation theory and in computer science, and worked at Harvard in
cognition and cognitive studies of wholeness and value. He became Professor of Architecture at Berkeley in .6, taught
there continuously for years, and is now Professor Emeritus at the Universit y of California. He is widely recognized as
the father of the pattern language movement in computer science. He was elected Fellow of the American Academy of Arts
and Sciences in .6 for his contributions to architecture.
a
N E W C O N C E P T S I N C O M P L E X I T Y T H E O R Y

Anda mungkin juga menyukai