Toward Yona Friedman

a Scientific
Architecture
TOWARD A SCIENTIFIC ARCHITECTURE
al
TOWARD A SCIENTIFIC ARCHITECTURE

Yona Friedman

translated by Cynthia Lang

GEORGE £.MAYCOCK

The MIT Press

Cambridge, Massachusetts, and London, England
First MIT Press paperback edition, 1980
Chapters 1-7 were originally published in French under the title Vers une archi-
tecture scientifique by Editions Pierre Belfond, Paris.

English translation and revision copyright © 1975 by

The Massachusetts Institute of Technology

All rights reserved. No part of this book may be reproduced in any form or by
any means, electronic or mechanical, including photocopying, recording, or by
any information storage and retrieval system, without permission in writing
from the publisher.

This book was set in IBM Composer Univers

by Technical Composition,
and printed and bound in
the United States of America

Library of Congress Cataloging in Publication Data

Friedman, Yona, 1923-

Toward a scientific architecture.

Translation of Vers une architecture scientifique.

1. Architecture. |, Title.
NA2500.F7413 720.1 75-8769
ISBN 0-262-06058-2 (hard)
ISBN 0-262-56019-4 (paper)
CONTENTS
FOREWORD TO THE ENGLISH-LANGUAGE EDITION by
Nicholas P. Negroponte ix

PERSPECTIVE xi

1 DEMOCRATIZATION 1
1.1. The Many 1
1.2. Information Processing: The Information Circuit between User
and Planner 4
1.3. The Solution: Direct Feedback from the User 6
1.4 Architecture: A ‘‘Teachable’”’ Science 11
1.5 Architecture Can Be Taught at the Primary Level 13

2 ESSENTIALS OF THE METHOD 15

2.1 Objective and Intuitive Systems - 15
2.2 A Complete List of Solutions as an Acceptable Answer 18
2.3 Choosing an Appropriate Axiomatics 20
2.4 Axioms to Describe the Work of Architects and Planners 25
2.5 A Mapping for Architects and Planners 26
2.6 Constructing Combinatorial Lists 29

3 THE NEW TASK OF ARCHITECTS AND PLANNERS = 33

3.1. Like Eating Out 33
3.2. Building the Repertoire 35
3.3 The Path Matrix 40
3.4 The Supporting Network and Its Hardware Counterpart 43
3.5 A Technological Necessity: The Chance to Make
Modifications 47
3.6 Keeping a History of Modifications 49
3.7. An Application of the Repertoire: The Flatwriter 53
Conclusion 60

4 URBAN MECHANISMS 62
4.1 Complete Planar Graphs with n 2 7: Crossovers and Potential
Conflicts 62
4.2 Networks 64
vi Contents

4.3. Example of an Analysis Using Networks: Patterns of

Transportation 68
4.4 How Do You Observe a City? 72
4.5 The Urban Mechanism 74
4.6 Probability as a Frequency Concept or as a Gauge 76
4.7 The Effort Map: Its Characteristics and Use 78
4.8 A History Depicted by Urban Mechanisms 82
4.9 The Unpredictability of History 86
4.10 Applying the Method of Urban Mechanisms 87

5 GENERAL APPLICATIONS OF THE METHOD 93

5.1 Looking for Laws of Nature in Histories 93
5.2 Constructing the Field of Reference: Types of Human
Behavior Based on Abstract Categories 95
5.3 Constructing the Field of Distributions in Space
(Configurations) 98
5.4 The Reference Matrix: A Definitive Instrument for the City
Planning Laboratory 100
5.5 Possible Interrelations among Histories 104
5.6 Possible Categories of Observable Activities and Their
Interpretations 107

6 THECITY 110
6.1 The Private City 110
6.2 Problems Become Meaningful Only If We Have Adequate
Information: The Infrastructure as Medium 112
6.3. The Infrastructure Must Fulfill Certain Conditions 114
6.4 The Infrastructure Is Not the City 116
6.5 An Example of the Infrastructure as ‘‘Hardware”’: The
“Spatial City” 116
6.6 Modifying a Choice Carried Out within the Infrastructure 121
Appendix to Chapter 6: New Problems in the Industrialization of
Construction 124

7 SOME CONCLUSIONS ABOUT SOCIETY 128

7.1. The Noncompetitive Society 128
vii Contents

Ve2mmGrisisiel oz
Some Superficial Conclusions 135

APPENDIX: SOCIETY = ENVIRONMENT 142

A.1 Terminology 142
A.2 Critical Group 152
A.3 Critical Group and Environment 155
A.4 Realizable Utopias 158
A.5 Limited Effect of Technological Utopias 167
Conclusion 168
 Digitized by the Internet Archive
in 2022 with funding from
Kahle/Austin Foundation

https://archive.org/details/towardscientificOO0Ofrie
FOREWORD TO THE ENGLISH-LANGUAGE EDITION

In 1964, when | was a graduate student of architecture, | was sent to

the airport (because | spoke French) to pick up a visiting lecturer,
Yona Friedman. | was turned on by a soft-spoken but persuasive
argument,
for removing the architect as middleman between a user’s
needs and their resolution in the built environment. Friedman’s thesis
rested, in part, on the matter of who bore the risk in bad design. A
decade later, | find myself working with him (under National Science
Foundation support) and personally close to him and his ideologies.
For me the excitement in this volume results from its simplicity.
Yona Friedman has used a mathematical scaffolding to support philo-
sophical positions in a manner which affords the reader the oppor-
tunity to disagree with his utopian posture, but still benefit from his
techniques. | see this volume as an excellent example of scientific in-
quiry into design, so desperately lacking in our profession.
If you are a student of architecture you will find the paradoxical in-
tersection of two academic streams—participatory design and scientific
methods—too frequently held apart by the circumstance of our train-
ing. You will find that Friedman contradicts the scholastic insinuation
that, if you can understand graph theory, for example, you probably
lack the human compassion and breadth of thought necessary to appre-
ciate social goals. | believe that he offers the most genuine view of in-
dividualization and community development in architecture without
timidity and without political rhetoric.
If you are a practitioner of architecture you should expect to find
illustrations of the bitter conflict between good intentions and pater-
nalistic ideas of professional behavior. Nonetheless, Friedman provides
an early warning mechanism which is as applicable to the design of
airports and hospitals by professionals as the design of homes by
future users. His methods afford simple but essential predictions like:
This space must be small, this will have no light—without any commit-
ments to physical form.
If you are a researcher in architecture you can enjoy the exemplary
nature of this probe into the science of design. It epitomizes a style of
research that is based on personal experience and does not require the
paraphernalia of institutions and equipment.
x Foreword to the English-Language Edition

This book is one result of Friedman’s inquiry. For students, archi-

tects, and researchers, its benefits should derive even more from the
questions raised than from the questions answered.

Nicholas P. Negroponte
PERSPECTIVE

Since 1957, | have been working on a theory which would free the
client from the “‘patronage”’ of the architect, and at the same time, |
have been looking for a way to make the architect useful to the client.
This theory has introduced new terms (which | had to invent) such as
“mobile architecture” and ‘‘spatial planning,’’ and new techniques
which I’ve worked out, such as “spatial infrastructure,” a technique
that allows maximum freedom to the inhabitant. The thesis which
took shape from this theory was published in 1958 in mimeographed
form and deposited in the Bibliotheque Nationale in 1959.
The propositions contained in this publication (and in the following
ones) were poorly received by most professionals, but very quickly
imitated by the most far-seeing ones—that is, the very good architects
(young ones, as well as those who had already ‘‘arrived’’) such as
Schulze-Fielitz (1961), Kenzo Tange (1962), Kurokawa, the Archigram
group, Safdie, Bofill, MUhlestein and many others. | was personally
very proud and flattered, although, for the most part, their works were
too much inspired by the kind of visualizations (drawings or photo-
graphs of models) that have appeared in professional journals since
1959, journals which have published too many pictures without spend-
ing enough time on the reasoning which led me to these visualizations.
For the most part, these architects also did not understand that the
forms which | proposed were only the result of this very strict reason-
ing.
This book tries to provide some perspective by presenting nothing but
the reasoning, without wandering off into visualizations of the results
which this reasoning has produced.
xii Perspective

The inevitability of this strict reasoning has had one direct conse-
quence: many of my friends (and especially those | have named) have
had to come to the idea of “‘infrastructure,”” without completely under-
standing at what point it became inevitable . .. and it is, for it is the
only possible answer to the wish to emancipate the inhabitant, the only
possible conclusion from the idea that such an infrastructure should
not be too much a determinant of the chosen behavior of each inhab-
itant.
The fact that | can finally publish an architectural book that is noth-
ing but reasoning—without worrying about pictures—is a clear sign that
the spirit of a generation of architects has changed: now, having imi-
tated, it can learn why that imitation was commendable, being inevi-
table.

Yona Friedman
Paris, 1970
TOWARD A SCIENTIFIC ARCHITECTURE
 ee =
2172992 a GMO:
K
t

—— ;
— S

x
oe

‘ dl —
~

= ‘ é of 7
*
“ * as
‘ ; 14
a fs ‘

t
a
J a

- f) “re - a

oe
’

7 a
u
fom -
a

Bea)
= ns
oS a
CHAPTER 1 DEMOCRATIZATION

1.1 THE MANY

There are certain very important considerations of an epistemological
nature at the basis of any scientific method. They give an idea of the
subject it deals with and of the tools it employs and answer these ques-
tions: what can we know, and how, using these special tools, can we
know it? In other words, these characteristics define the method by
setting its boundaries.
It follows naturally from these statements that when the subject of
a scientific method undergoes quantitative changes (and thus implicitly
changes its boundaries) the method as a whole must be transformed.
This is what has happened in physics, mathematics, and in biology and
what is happening today in the behavioral sciences, including architec-
ture and urban planning.
Before, the classic progression in architecture consisted of a simple
chain of operations, which began with the future user of the architec-
tural product (we will call him “‘user”’ or ‘‘client’’). This client had
very specific needs and personally conveyed them to the architect.
The architect (or designer) drew up a plan trans/ating the specific
needs of the client into a piece of hardware” which was supposed to
fill those needs. The execution of the project was in the hands of a
skilled tradesman. When the piece of hardware was finished, everyone
was happy in the best of all possible worlds.
| have caricatured this process a little to underline the role of the
architect as translator of the specific needs of the client into a lan-

*| use the word hardware here to mean the finished product, or building.
2 Democratization

guage comprehensible to the skilled tradesman. The architect was an

indispensable middleman as long as the client and the builder had no
common idiom: there could be no doubt that at each stage of the pro-
cess all the decisions had been made exclusively by the client.
At first sight the situation today seems to be identical, except for one
factor, which is totally new: the number of clients has become enor-
mous. But alas, our initial considerations have shown that a single new
factor is more than enough to change the original situation completely,
and that is what has happened.
In our crude representation, the client spoke directly with the archi-
tect about his specific needs. As it often happened that these needs
were not very clear even in the client’s mind, the architect had to de-
vote a great deal of his time to acquiring sufficient information about
them. | could cite as an example the famous architect who declared
publicly: “’In general, it takes about six months to understand one
client’s way of life,”
The question | ask is: How much time would he need to understand
ten thousand clients?
Of course it is a purely rhetorical question. The time needed would
be longer than the written history of mankind (which is about five
thousand years). In spite of that, it is considered perfectly reasonable
for today’s architect or planner to work for ten thousand clients.
There are only two solutions to this problem:
1. Supply a large enough number of architects (or planners) so that
each of them can devote himself to a very few clients.
2. Reduce the period spent gathering information (between the client's
visit and the construction of the hardware.
The first solution would require unprecedented numbers of architects
(it would amount to the creation of about two million architects in the
next ten years, in the U.S.A. alone). Logically enough, the profession
has chosen the second solution.
That is should have followed this course was bad enough, but the
way in which it did so, which approached the absurd, was even worse.
We shall see why.
3 The Many

DECISION

USE

THE Racic PRocess INCLUDES

ONL) “THE USER AND THE PRODUCT
Figure 1

The architects and planners said to themselves: It is impossible for us

to find out out the specific needs of each user. Let’s try to find out the
average needs of future users, or, in other words, the specific needs of
the average future user. ’
| do not think that it is necessary to describe the results in detail. It
is enough to point to the massive discontent displayed by all users of
architectural hardware. The reason for this discontent is evident: the
average client (user) does not exist. Therefore, if all we do is satisfy
the needs of the average consumer then logically we are not fully satis-
fying the needs of any of the real consumers. So instead of satisfying
the real consumer, who does exist, we are satisfying one who doesn’t.
Obviously this argument is simplified to prove the point. But the real
situation is worse still: in my line of reasoning, | assumed that archi-
tects (or planners, or sociologists) were capable of objectively defining
average needs. Now actually it was not just the logic of the notion that
was fallacious; the very evaluation of the average user’s needs was inevi-
tably distorted by the preconceived ideas cf those making it.
This critical situation results from a fact that the profession is trying
to ignore, that is, that the number of clients has grown, which neces-
sarily implies that the whole process of which the architect is a part
must be changed.
The search for this new process is the object of this book.
4 Democratization

1.2 INFORMATION PROCESSING: THE INFORMATION CIRCUIT

BETWEEN USER AND PLANNER
Let us reconsider the situation. !n the traditional process, as we have
seen, the mechanism worked this way: the architect (or the planner)
and the builder were only the ‘‘channel”’ by which ‘‘information con-
tent,’”’ or the ‘‘message,’’ that is, the specific needs of the user, were
relayed to the finished building.
This process was a simple one, made up of a transmitting station (the
future user), a channe/ (the architect and the builder together), a rece/v-
ing station (the hardware, or finished building), and information return
or feedback (the usefulness of the product made available to the client).
This system allowed for no corrections, no adjustments in case the
feedback was unsatisfactory. Therefore, if the receiving station (the
finished building) had not received the message (the specific needs of
the client) from the transmitting station (the client), the responsibility
lay with the channel (the architect and the builder). Any adjustments
after the fact being impossible, the preliminary period, while the archi-
tect was getting to know the needs of his client, had to be very long
indeed, as we have seen. As a consequence few or no modifications
were necessary once the building was completed.
When this system was altered to serve an increased number of ‘‘future
users’ it became fundamentally different, as shown by the diagram in
Figure 3.
ARHITERT | oean SLATORS
BVILDERQ OF THE DECGSION

r€ os! on k Tq Mee
2o, A
Soregae

use PRODUCT

TRE TRANSLATOR ENTERS “THE PRocess

Figure2
 Information Circuit Between User and Planner

i ek

ARH ITECT
TRANSLATION

USE

od

JAMMED CIRCUIT
Figure 3

AST
peone
OT

.
4 .
ae

HoTWRE % mk TRANSLATION
USER.
(ARCH |TECT
a me Fictitious |
FYTURE vANGRAGE MAN
USER

BROKEN CRCUIT
Figure 4
6 Democratization

The diagram shows the ideal situation as the architect and planner
imagined it. Just for the sake of hypothesis, let’s admit that this could
work. The scheme has two bottlenecks: one at the architect's end and
one at the point of actual construction. The first bottleneck is so nar-
row that it restricts the arrival of information to the architect; the
second hinders feedback about the building once it is constructed. The
vulnerable points of the system are these: the architect’s handling of
information, and the finished building’s adjustment to uses as various
as his client’s.
The concept of the average man of course does not help the situation,
since it merely adds a new step (and therefore new possibilities for
error) without eliminating the bottlenecks. They stay the same, even
though their positions are changed.
| shall cali these two bottlenecks the architect’s and planner’s infor-
mational short circuits.

1.3 THE SOLUTION: DIRECT FEEDBACK FROM THE USER

Before continuing, | must try to define certain principles which | did
not mention in the first section, principles which underlie all my rea-
soning so far.
Every science, every discipline or system, is essentially based on infor-
mation. We can sum up the specific nature of a science or an art by de-
fining the following points:
1. how (or by whom) the significant message was sent
2. how it was transmitted
3. the message which reached the recipient.
Epistemology is essentially information theory in disguise, and the
decisive element in a discipline is the way it manipulates information
in a given context. Naturally, when the context changes, the way of
handling information must also change.
Now in the preceding section | stated that in the architectural process,
there was one part of information handling (the channel) which had
not changed as the context changed, so that the process had to readjust
itself at another stage, involving the modification of the hardware for
its uses. Unfortunately, this adjustment proved to bevirtually impos-
Z Direct Feedback from the User

sible. That is the cause of the present crisis in the planning disciplines.
Our goal, therefore, is to construct a new process, which will eliminate
information short circuits and therefore unreliability from the mes-
sage On arrival, in other words, ‘‘noise’’ (disturbance in an information-
al system). The process would look something like Figure 5.
Figures 5 and 6 build on the relationship between two loops; neither
one has a bottleneck like the ones in Figures 3 and 4.
The first loop represents the future user’s direct decision. This loop
includes only the user at one end, and, at the other end, a repertoire of
all physically possible combinations of hardware he could bring about.
The second loop includes only the user of the finished product.
In these two loops, the intermediary—the role played by the architect
or planner—has been eliminated. If it were necessary, we could put
him back in the second loop to interpret between the client and the
finished building (and thus act as artisan);
but never in the first loop.
We have eliminated the bottleneck which appeared in the diagram in
Figures 3 and 4 by dividing the original loop into two separate ones.
This operation allows us to line up, side by side, as many double loops
as are necessary for as many users, because each loop is distinct along

ENTAN On

OnE FUTURE ‘ ¢
USER SP WARN INS
OF THE
CON SE QUEN
CBS
FOR THE
INDIVR UAL,

4h
ARs
OTHER USERS FEEDBACK
Figure5
8 Democratization

its entire course.

| shall explain the significance of Figure 6 by using the example of an
imaginary but possible situation.
In our scheme, instead of an architect, the future user encounters a
repertoire of all the possible arrangements (solutions) that his way of
life may require. This repertoire, which is necessarily limited, must be
presented to him /n a form he can understand. Thus, for each item in
the repertoire there is a warning. It tells the future user—again, in terms
he will understand—the advantages and disadvantages, in terms of use,
of picking a particular item. (The warnings, we will see in Chapter 3,
are not based on any particular value system, but on the intrinsic pro-
perties and the logic of the projected solution; it may happen that
the same warning can represent an advantage to one user and an incon-
venience to another, since two clients may have entirely different ways
of life.)
Thus, the first loop is the user informing himself by means of the re-
pertoire. The way in which this repertoire may be put together is dis-

(Suse ve Ss Cy

KISSINDIVIDGAL my

TRE aes SYSTEM:

THE USER
ts soi ONY PERSON
IN THe GRIT
Figure 6
9 Direct Feedback from the User

cussed in Chapters 2 and 3. We will describe how it is linked with a

warning system in Chapters 3 and 4.
This first loop ends with the future user’s choice of one item from
the list in the repertoire.
The second loop resembles the classic process very closely; the future
user explains his choice directly to the artisan, who will produce the
building. To do this, the user indicates his choice by means of a num-
ber code, taken from the repertoire. The artisan (or the contractor)
erects the hardware chosen (or conceived) by the client, who then puts
the object of his choice to use. It is he who takes responsibility for the
entire process, since he has been forewarned of the implications of the
particular solution he has chosen.
In this remodeled version of the process, the two loops begin and end
with the user: the first with his decision,
the second with his communi-
cation of this decision and the results.
No special channel, no interpreter is necessary in these two loops. At
first glance, it seems that the architect or the planner, who has for cen-
turies kept his role as interpreter, has been eliminated from the process.
Actually it is a little different; in fact, the channel has not been elimi-
nated in the new process. The channel is the repertoire itself, or, more
precisely, the notation (mapping) used in the repertoire. This mapping
must be understood by any user, as well as by the artisan who con-
structs the building. Thus the remodeled process takes us back to the
scheme in Figure 1, which consists of three elements: the future user,
the channel (in Figure 2 the architect, in Figure 5 the repertoire) and
the finished product.
So it is really not the architect or planner who has been eliminated
from the process, but rather his old role. He has a place, a new role, in
the new system: he constructs the repertoire.
| am going to try to give an example of this process, an example taken
from a system that does not involve the environment directly, so sim-
ple that we are not even aware that a scheme exists. It is the telephone
system.
The repertoire: all the combinations of ten digits (in the United
States). Any ten-digit number is a potential telephone number.
 Democratization

THE 2 AXES CF COMPETENCE

THE
FUTURE USER
ADAPTS
THE MEGHAIE
TO HS PURPOSE.
te HAS ee
TO BE INTUITIVE.

t Genstaye
BSRShe REPERTARE,
AND on Saat
INPRACTRUCTUQE.
HE MOST RE ,OBJECTINE’,
Figure7

The warning to the individual: the telephone directory. It warns any

one who wants to make up a telephone number of the consequences of
his act: it tells him who will answer his call
The infrastructure: the technical network. This network, if it is con-
structured properly, can route a// calls, all conversations from any sta-
tion to any other. The telephone technicians have no obligation to tell
you to whom you should speak or what you should say. The jnfrastruc
ture, therefore, is impersonal.
Warning to the community: when you are on the line, your number
and the number of your caller will be busy, so that any third person
calling in will hear the busy signal.
Of course this example is very rudimentary; but it can point up how
a highly technical and therefore impersonal system like the telephone
can be used to express the most personal feelings of the people who
use it. This example also shows that the task of technicians is not to
11 A Teachable Science

deal with the emotional, intuitive aspect of use (in our example, the
job of the technicians has nothing to do with the content of the conver-
sations). Their job is to establish the repertoire, the instructions, and
the infrastructure.

1.4 ARCHITECTURE: A “TEACHABLE” SCIENCE

Let us look at architecture in this new light: architecture, the design of
complete repertoires that hold all the possible solutions to a problem,
thanks to a specific notation (mapping) and the elaboration of a meth-
od which will match up a warning with each item in the repertoire.
This brief statement completely overturns the old image of architec-
ture or planning, insofar as it still exists. To see what kinds of changes
this will bring about in the profession, we can start by examining what
kinds of equipment will be given to those-who practice it.
Today there are no strict rules, either in architecture or in city plan-
ning, which allow us to predict accurately the results of a particular de-
cision. Both professions make use of tricks of the trade. These tricks do
not necessarily work in every case, and it is often difficult to tell the
difference between a case where they will work and a case where they
will not.
This observation implies that tricks cannot be taught without getting
into awkward situations (because they are not generally applicable) and
there is no method of imposing objective control by reference to con-
text. A trick of the trade is basically a highly intuitive rule (see the
first section of Chapter 2).
How do schools function when they deal with highly intuitive sys-
tems? Learning is dispensed by ‘’masters,’’ who have their own special
tricks, which are usually impossible to communicate. Even though they
can use them, they can’t teach them. Masters are surrounded, there-
fore, by apprentices who do their best to imitate their master’s work-
ing style, hoping by one means or another to pick up his ‘‘touch.”
| call such disciplines the ‘‘prenticeable”’ disciplines. Here, the qual-
ifications of the student depend essentially on the qualifications or the
personality of the master.
12 Democratization

A TEACHABLE” DISCIPUNE
oe HAS STRICT RULES.
> ANY TEACHER CAN
cp TRANSMIT THEM -
wae - sete COXAMYE | ARITHMETIC)
‘————_

/\ KrowLlepe

TEACHERS,

x A , PRENTICEA BLE"
pie DISCIPLINE
eng) CaN BE TRANSMITTED
eNLyY BY IMITATION

AN AN EXAMPLE !
STUDENT CNeneC tune TODAY)
TEACHER
Figure 8

But there are other disciplines, for the most part called sciences, in
which strict rules have been established, at least for the most important
areas, rules which are valid in a well-defined set of instances. In these
disciplines the entire system of rules is as generalized as possible.
In these disciplines, schools operate according to an arsenal of rules,
which are available to the public (in books or films, for instance). Any-
body who reads and understands these rules can apply them himself,
without having to imitate the masters; and once he understands them,
he can communicate them to anyone else. They are stated in such a
way that in any instance it is clear whether they apply or not.
| call these the teachable disciplines. |n these areas of study, the per-
sonality of the teacher is of no importance.
And so, if the activities of architects and planners today derive from
learning by apprenticeship, their new work, as we have just seen, will
emerge by contrast as a science that can be taught.

1.5 ARCHITECTURE CAN BE TAUGHT AT THE PRIMARY LEVEL

The point of such a long introductory chapter is to insist on this fact:
13 Architecture Can Be Taught at the Primary Level

the power of choice rightfully belongs to the future user. That is why
this chapter is entitled ‘‘Democratization,”’ for the word democracy
indicates that everybody has his share in making the decisions (the elec-
tion of a plenipotentiary representative is no more than a very second-
ary, and a very distorted, form of democracy).
The act of deciding also implies that the one who makes the decisions
is the one who takes the risks. Any system that does not give the right
of choice to those who must bear the consequences of a bad choice is
an immoral system. But that is exactly the way that architects and
planners work. They make the decisions and the users take the risks.
To do away with this situation, | have looked for methods that will
guarantee to the future user the power of decision—after he thoroughly
understands the risks involved. \t is both immoral and dangerous to
leave choices to people who have not been properly informed about
the consequences of their decisions.
The problem, which was initially one of information, is still one of
information, but on a different level. The important question is: How
do we inform the future user? (Information /s politics, and the only
true form of politics.)
Let’s look again at the example in Figure 5, where the future user
had to make a choice from a repertoire in which he found all the neces-
sary instructions. To use this information (the repertoire and the in-
structions) he had to be able to read both of them, but not to set them
up himself.
The professional who sets up the repertoire and the instructions has to be
ascientist, practicing a teachable discipline. He will be the new version
of the architect or planner: the ‘‘redesigned”’ architect or planner.
The future user and the planner-scientist will need the same skills:
they must know how to read the repertoire, how to make choices from
it, and how to read the warnings. These are skills that should be taught
in primary school.
But the planner-scientist will have more to learn: ‘‘objective’’ rules and
their application to the construction of a repertoire and its warnings.
He will study these at the university level.
14 Democratization

There is nothing unusual about this situation: mathematicians and lay-

men use the same methods and the same rules, but on different levels.
Anyone can read a weather map, but only a scientist can draw one.
The key to the functioning of an environment structured and ruled
democratically rests in this twofold instruction.
We shall study both its aspects in the following chapters.
CHAPTER 2 ESSENTIALS OF THE METHOD

2.1. OBJECTIVE AND INTUITIVE SYSTEMS

{In order to understand how to build a repertoire we have to examine
the informational content of systems. [N.B. | use the term system,
here and later on, to mean a set of elements where all the elements are
linked (related) to the set by at least one link (relation) but where the
links do not necessarily have any regularity. Thus a structure, defined
as a set of relations arnong the subsets of a set where the elements are
put together according to particular rules, is implicitly a system, but a
system does not necessarily have a structure] .
Let’s suppose that | had to describe a system so that other people
could understand which one | was talking about. There are two ways
to do it: the first way would be to give a certain number of instruc-
tions that form a sequence, organized so that any person, whatever his
tastes, his culture, his race, his opinions, could carry out this sequence.
Once it had been carried out, the result would be the same, no matter
who the experimenter happened to be. | call this an objective, or scien-
tific, system. Naturally, the proper sequence of steps would have to be
laid out as precisely as possible.
For example:
First condition (implied): the experimenter is within the earth’s field
of gravity.
First step: he takes a glass in his hand.
Second step: he lets go of the glass a yard or more above the ground.
Third step: result: the glass falls.
Fourth step: secondary result: the glass could break (this possibility
16 Essentials of the Method

depends on conditions that are not defined here, such as the material
and shape of the glass, the kind of floor, and so forth).
This secondary result will depend on conditions not specified, and we
shall pass over it unless it is more precisely defined.
The diagram in Figure 9 represents the essentials of such a process,
which is the general scheme for all experiments of this sort, where the
personality of the experimenter has no effect on the result.
It is easy to see that, as long as the result of such an experiment is
known, we can always reconstruct an excellent approximation of the
prescribed sequence of steps; but there is no way of determining from
the final result who the experimenter was.
The opposite of this kind of description, which | shall call an objec-
tive description and which makes use of a series of steps, is another
kind, which | shall call ‘‘intuitive.’’
In the intuitive description of a system, there is no series of steps, but
rather a symbol, used as a code. The symbol may be all there is to con-
vey a message. It contains information that cannot be communicated
any other way (which is why a symbol is needed); and we can be almost
certain that each experimenter will understand the content of any one
symbol differently, depending on his taste, his culture, his race, and his
opinions.

AN OBJECTIVE SYSTEM
Seer nications
FOR THE
SEQUENCE oF STEPS

RECONSTRUCTION
OF THE PRESCRIBED
FRom THE RESULTS

3S RESoLT
OF THE EXPER NENT

sf EXPERIMENTER
Figure 9
{WZ Objective and Intuitive Systems

ST
AN INTUITIVE SYSTEM t A SPECTATOR

(WEAGE)
ND
Ge) : > nN 2° SPECTATOR

WORK
OF ART
A GREAT ASS BP SPECTATOR

Figure10

A good example of an intuitive system is art.

An artist creating a work of art considers that the message in his work
is the important element, not the object itself. (According to my defi-
nition, a work of art is an object that carries a message). But in reality
each spectator of an art object receives a different message; and if |
know how the message is interpreted by one spectator (which in this
example constitutes the result of the experiment) | will not be able to
reconstruct the original message, but it will certainly give me a good
idea of the personality of that particular spectator.
Clearly, this classification of ways of description has been simplified
a great deal. In reality, there are no pure cases that belong exclusively
to one of these methods. Any system can be described both ways, or,
more exactly, it should be described both ways at the same time. How-
ever, we can say that a system is described largely by the objective
method (and is thus an objective system) or largely by the intuitive
method (and is thus an intuitive system). In their classic form, archi-
tecture and planning are intuitive systems.
The theme of the first chapter was that the situation in architecture
and planning needs to change with respect to objectivity and intuition.
| divided the discipline into two distinct parts; the part that concerns
the planner has become an objective system, while the part that is in-
tuitive involves the user, as shown in Figure 11.
So, from now on | shall consider only the objective part of the dis-
18 Essentials of the Method

COMPARISON OF DISUPLINES

ORECTIVE
SYSTEMS

INTUITIVE
SYSTENS

Figure 11

cipline, not because it is more important, but because it is the part that
can be communicated, and the part that can be worked out. The reper-
toire that | am trying to establish must therefore be constructed as an
objective system.

2.2 A COMPLETE LIST OF SOLUTIONS AS AN ACCEPTABLE

ANSWER
We've seen that the specific criterion for defining an objective system
is the possibility of reconstructing its ‘‘history’’ (the sequence of steps
which led to a given and recognizable result). Therefore, we can extend
the definition of objective systems to systems which are characterized
by one or several series of steps, all of which can be reconstructed once
the final result is known. (All sciences work this way.)
Here | have to point out that for practical reasons the word “‘several,”’
which | just used, necessarily represents a finite number in my mind.
For if it also implied an infinite number, the criterion of ‘‘constructi-
19 A Complete List of Solutions

bility” would not be respected. Infinity is only potential; and nobody

has ever been able to construct it.
Now to clarify the expression ‘‘one or several series,’’ | am going to
try to give a simple example. Suppose | want to find out how a parti-
cular person arrived at the number 5 by adding two integers together.
It is unlikely that | could guess the way he actually did it, but | could
make a list of all the possible ways of doing this addition, and state that
whichever formula was actually used, it must be included in the follow-
ing list: 0+ 5, 1+4,2+ 3. These sums make up an exhaustive list.
If the question were: ‘‘How did someone get the number 5 by taking
the difference of 2 integers?’”’ the problem would be fundamentally dif-
ferent. | could start the list: 5 —0,6—1, 7 2, and so forth, but since
it is infinite this second list would never be exhaustive, and | could
never state for sure that the formula actually used by this person was
on the list. Therefore, according to the criteria of objective systems, this
second list does not constitute an acceptable answer.
To continue this premise with a more general statement: if there are
two finite sets, A and B, the union of these sets (A U B) satisfies the

OQ OF Go

SYSTEMS
ris USED
Finrreée
(ve
is useb
we
¢ a
NOT
ay
SYSTEMS
OBJECTIVE
REONECT
Ow IN
Zpemcrion
a
ge
UTHUS

Figure 12
20 Essentials of the Method

criteria of objective systems, but their intersection (A B) does not,

for A UB can be broken down into a finite list of all the possible com-
binations of A and 8, whereas A ™ B cannot (Figure 12).

2.3 CHOOSING AN APPROPRIATE AXIOMATICS

The preceding paragraphs were mostly concerned with the conditions
for the description of a system, whatever it may have been. Now we
come to the next problem: how do you describe the system?
Every description is made up of a number of statements. The number
can be any finite number (a description made with an infinite number
of statements is pointless, as we've already seen). If a system is made
up of relations (links) and elements to start with, the number of state-
ments that make up its description cannot be greater than the sum of
the number of elements and the number of links. We could simply say
that the most complete description of the system is the system itself.
Naturally the least complete description would be made by just one
statement about it; for example: system X exists.
A description of a system must have at least a certain number of state-
ments, the actual number being a positive integer somewhere between
1 and n, n standing for the number of elements pius the number of links.
If the system presents even the slightest regularity (if there is a struc-
ture, several similar elements or several similar connections between
elements) the most complete description of the system will have several
repeated statements. In order to simplify, we'll replace those repetitions
with a new statement that indicates that statement S is repeated /n times.
If these repetitions recur regularly, we can say that each time condition
C happens, statement S will be repeated.
All these statements, replacing the simple repetition of a series of iden-
tical statements, will allow us to reduce the number of statements we
need to describe the system. Moreover, the list of statements that make
up the description will be shorter. Statements displaying this regularity
will be called systems of abbreviation. (1 use this term and structure
interchangeably.)
It is obvious that the abridged description is more convenient to use
than the long one. It is also obvious that an abridged description, with-
21 Choosing an Appropriate Axiomatics

out losing its accuracy, does run the risk of being less complete than an
unabridged one.
As for the characteristics of the statements making up a description,
we could postulate the following: the statements must be clear to any-
one; they must meet the criteria | have set down for objective systems;
all their terms must be the ‘‘results’’ of a series of simple steps; and the
results must be the same for anyone, that is, for anyone relying on a
well-defined category of knowledge.
| shall call these statements axioms, and the terms of these statements
primitive terms. Consequently, a system can be described by means of
axioms. They will not be explained themselves (that is, described by
means of other ‘‘meta-axioms’’), but, given that they have been con-
structed as an objective system, they will be assumed to be universally
understood, accepted, and verified. =
These axioms must meet certain conditions:
1. An axiom cannot contradict the statement of another axiom used to
describe the same system. (The axioms must be consistent.)
2. An axiom cannot repeat the statement made by another one that
describes the same system (they must not be redundant).
3. Finally, all the axioms that are used to describe a system must be
sufficient, taken together, for an approximate reconstruction of the
entire system; that is, there must be no instance where an event (or
term) belonging to the system does not have a place, even if under a
collective noun, in the system of description (the axioms must be
complete).
It is very difficult to fulfill these conditions, especially if there are too
many axioms. Actually, these conditions convey the fact that the
axioms describing a system are all interlinked.
When | state the first axiom, its content does not matter, and | am
completely free to choose it arbitrarily. By contrast, | am no longer
free at all in choosing the second one. It has to state a new fact, not
covered by the first axiom, and it cannot contradict the first one. There-
fore, | must limit my choice for the second axiom by verifying that it
is both consistent and nonredundant with respect to the first. The pro-
cess of solution will be the same for the third axiom, but more compli-
22 Essentials of the Method

cated (because of the condition of nonredundancy and consistency

with respect to the first and second axioms), and so forth.
When | come to choose the last axiom, besides verifying that it is
nonredundant and consistent with respect to all the preceding axioms,
| have to verify whether an additional consideration is fulfilled and ask
myself: is the description of the system complete enough to satisfy its
purpose? For it is the purpose which determines the degree of simpli-
fication—or abbreviation—of the description.
To describe this process of verification, | should say that each axiom
must close back on the preceding ones. Once the axiomatics is com-
plete, | will no longer be able to tell which axiom was the first in the
series of operations (the order of construction of the axiomatics); the
axioms should close back on each other, whatever order they are listed
in.
An axiomatics (a system of axioms) is, therefore, a completely closed
system, in which all the links represent the same primitive generating
steps. The (entire) system therefore consists of the same primitive terms.
| now propose to use a device: let us represent an axiomatics graphi-
cally in the following way. A point will represent an axiom, a line join-
ing two points will be a link, and an arrow on the line will show which
axiom comes after which. We shall think of these diagrams (mappings)
as plane figures (no line can pass over another). The graphic one-to-one
representation (mapping) of a system containing 7 axioms will be
shown by a complete p/anar n graph.
Figure 13 shows that everything goes well in a system of axioms as
long as it includes no more than three axioms; all the conditions (con-
sistency, nonredundancy, and closure) can be verified. The trouble
starts with the fourth axiom. The last line, that of closure, has an arrow
whose direction cannot be determined without additional information
on the subject described by the system. So the axiomatics is not neces-
sarily complete. In other words, it is impossible to verify consistency,
nonredundancy, and closure for the lines of a system of four axioms
without introducing an additional statement.
There are even more difficulties in a system of five axioms. In such a
system, the problem is not the direction of the arrow (this occurs only
23 Choosing an Appropriate Axiomatics

A SET OF hh ELENEArTS
CAN BE DESCRIBED BY STATENENTS
ANYWHERE FRoM 4 TOW, IN NOM RER_

le A STATEMENT IS REPRESENTED By e
A RELATION BETWEEN 2 STATEMENTS BY
AND “THE ORDER IN WHICH “THE STATEMENTS FOULOW
EACH OTHER. By ——

THEN A SYSTEME OF Nn ELEMENTS

ON RE DESCRIBED BY STATEMENTS 4S FOLLOWS :
@) (b)
e e>e AM / ON a
4 2
stavenent STATEMENTS «STE.
(@) IN THE CASE OF 4 STATEMENTS
THE ORDER IN WHICH “THEY ParLow €Acn OTHER
CANNOT BE EASILY ESTABLISHED
(b) NORE “THAN 4 STATEMENTS RESULT IN A "CROSSOVER
THUS IMPLY AN UN@NUNCLATED
S ;
ADDITIONAL STATEMENT.
Figure 13

when nr is an even number). Here, there appears a crossover which is

not authorized by our mapping, for its literal significance would be that
of an unstated axiom; our system of five axioms would become an
incomplete system of six axioms, which we cannot allow becuase of its
character of incompleteness.
A complete 7” graph will always have crossovers when 7 is equal to or
greater than five. So a system of five or more axioms will always run
the risk of being incomplete.
This little exercise tells us (if the mapping was correct) that an axio-
matic description with more than three axioms would be vulnerable
(with respect to consistency, nonredundancy, and closure) in every
case where an experimental verification of the original system (which
was the point of the description) is unfeasible.
One could ask: is mapping a system of axioms by means of a complete
24 Essentials of the Method

planar graph really a proper one-to-one representation? For example,

could not the problem of the crossovers be eliminated by using a non-
planar graph for the mapping?
Mapping with planar graphs is, in fact, perfectly correct; for if we
were to allow nonplanar solutions, we would need an additional rule
which explained the significance of the additional dimension in setting
up the graph, a rule which in itself would have no representational
value.
Each element in the mapping (points, lines, labels) has a clearly de-
fined significance. But this rule of the additional dimension would have
none, so it would represent a foreign element which would correspond
to nothing in the system.
On the other hand, our mapping system does not exclude systems of
three or more dimensions; it only discourages them, because the ab-
sence of a statement explaining the significance of an additional dimen-
sion makes unworkable (incomplete) systems.
There follows this simple precept, on which | shall rely throughout
this work: the description of a system of axioms is most convenient
(easiest to verify with respect to consistency, nonredundancy, and
closure) when there are no more than three axioms.
These three axioms can be easily constructed as follows: the first iden-
tifies the system to be described by naming the operations at the origin
of the system; the second imposes a restriction (specification) on the
first, and the third restricts both the first and the second. For example,
we can take the step-by-step description of the mechanism of compre-
hension as a basic model for this kind of axiomatic system.
The first step in every process of human comprehension consists of
the identification of something. Let us think of this identification as
putting a mental label on that something so that it will be possible to
tell if it is or is not the thing we have labeled.
The second step consists of delimiting (of setting apart) the thing
labeled from all the other things. The labeled thing will be thought of
as unique.
The third step consists of making a comparison between the labeled
thing and all other things.
25 Axioms for Architects and Planners

| shall call these steps:

1. identification (of one thing among others)
2. establishing uniqueness (of the labeled thing)
3. establishing a relation (between the labeled thing and all other things).
Even if this system, and particularly this example, is a little off the
path | have decided to follow in order to discover a science of planning,
it is necessary to take the detour, because it will be useful later on.

2.4 AXIOMS TO DESCRIBE THE WORK OF ARCHITECTS AND

PLANNERS
In this essential section we can finally leave the realm of abstract rea-
soning and come back to questions of professional concern for archi-
tects and planners. | shall apply the preceding argument to this realm,
first of all in trying to find an axiomatic Gescription, as general as pos-
sible, for the work of these professions.
The general description includes three statements, in accordance with
a previously fixed condition. This previous condition will not, properly
speaking, be part of the axiomatics itself, but it will determine its range
of validity. (| have said that any operational result must be valid for
anyone, that is, any person belonging to a well-defined category).
The range of validity for our axiomatics will be any space where a
human being can physically survive (in practical terms, what we call
the biosphere).
The work of architects and planners deals with effecting separations
within the biosphere. In fact, if we examine any product made by an
architect or planner, we can state that it takes the form of an ‘‘envel-
ope,”’ or an “enclosure.” The interior of this envelope encloses a volume
or a surface separated from the exterior. The fundamental task of the
planner as well as of the architect is the activity of making this separa-
tion. The only difference is that the first works on two-dimensional en-
velopes like zones while the second builds three-dimensional envelopes.
But both activities produce enclosures.
So the first axiom will be: the work of architects and planners pro-
duces enclosures—separations in space.
But there are many kinds of enclosures that are not products of the
26 Essentials of the Method

architect’s or planner’s skill, because they are inaccessible. They are

things such as boxes, sardine tins, and the empty spaces inside objects.
So it will be practical to choose a second axiom that can state this
specification: a separation of space (an enclosure) cannot be the work
of architects and planners if it does not have at least one access. In
other words, any two spaces separated by an architect or a planner
must have at least one ‘‘path” linking them.
When architects or planners bring about separations in space, they
do it so that the new spaces will all be given different qualities. The
act of enclosing creates differentiation. For example, an enclosure
could have climatic or visual qualities that are very different from
those of the original exterior biosphere or those of another enclosure.
So | shall propose a third axiom: in a system of spatial separations
there must be at least one enclosure that differs from the others in
some respect, whether as a result of physical qualities or of others.
Thus the three axioms | propose are:
1. Architects and planners bring about separations in preexisting space.
2. Each of the separated spaces must have at least one access.
3. There must be at least two different kinds of separated space.
Without contradiction or redundancy, these three axioms describe
every step that architects and planners could think of. Therefore, this
axiomatics is complete.
Many people would be quick to point out that these axioms are trite.
But axioms have to be so by definition, for they must state a fact that
is acceptable to everyone, that is obvious, and so on. What is important
about the use of these axioms is that they allow for any solution imag-
inable proposed by any human being imaginable within the realm that
they define, except for those that are physically impossible.

2.5 A MAPPING FOR ARCHITECTS AND PLANNERS

By themselves, the axioms formulated in the preceding section are
not enough help: | could not actually construct a repertoire based on
them. To obtain the repertoire (the complete list of all the possible
solutions for one problem) we would have to set down all the solutions
in one way or another and then put the list in some kind of order. (The
D7, A Mapping for Architects and Planners

list has to be in order so that we can find any item on the list whenever
we need it. For example, we might use an alphabetical order, just as
the telephone directory does. In that case, the rule which determines
precedence, or which item comes after which, would be the rule of the
alphabet.) .
So the first problem we confront is which terms to include in the list,
without leaving any out and without repeating. To arrive at this list,
we shall apply the process of mapping, using a model which easily gives
us a general view, and where each element corresponds to an element
in the original system and only to that element (a one-to-one relation-
ship). This relationship must hold for all the links among the elements
in the two systems. Since the axioms cover both the elements of a sys-
tem and the relationships among the elements, there is the same one-
to-one relationship between the rules which determine the mapping and
the axioms of the system represented by the mapping. We shall first
try to determine the rules of the mapping. Once these rules are defined,
we shall construct a combinatorial list of all the images that the rules
of the mapping generate. And this list will correspond, element by ele-
ment, to the repertoire which we wish to establish.
The term mapping indicates that there is a code which enables us to
transcribe each step or element of the system represented (mapped),
into a step or element of the system which represents it (mapping), and
that in each of these two systems there is only one element which cor-
responds to any element in the other system. The mapping procedure
lets us execute any step at all on the map, and to transpose this step
into the real system. So there is a rea/ result which corresponds to the
results of each mapped step, and we can recognize the real result by
comparing the result obtained on the map.
Now | am going to introduce a mapping code, which will be a code of
connected and labeled planar graphs, and we shall transcribe the axioms
into the rules of the mapping.
First axiom: a separated space can be created inside a preexisting
space.
First rule: there is a point to represent the preexisting space, and each
enclosed space in the preexisting space is represented by another point.
28 Essentials of the Method

Therefore, any graph that represents an enclosed space has to contain

at least two points.
Second axiom: an enclosure has at least one possible access from
another enclosure, and there can be no enclosure that is inaccessible
from some other enclosure.
Second rule: a line represents an access. This line links the points re-
presenting the two enclosures joined by the access. In any figure, there
must be no single point without at least one line to the other points.
Third axiom: there are at least two different kinds of enclosures of
which one is the preexisting space.
Third rule: there is a label to represent each category of enclosure. In
each representative figure, there are at least two labels (one for the pre-
existing space and one for the enclosures).
These three rules constitute a complete mapping for the axiomatic
system which relates to the activities of architects and planners. They
allow us to construct a list of every possibility. The mapping is made
up of points and lines, arranged so that every point or group of points
is linked to all the other points. All these points carry labels, and the
labels are of at least two kinds (planar graphs, connected and labeled).

THE JOB oF ARCHTECKS

AND PLANNERS:
MAKING ENCLOSURES.

THE S AXi10OmMS
PERONS en NITIES
ARCHITECTS
OF ANB LANNERS
Re MAKE ENCLOSURES
nN bRe-“EXISTING SPACE
eSfog cmc ENCLOSURE
ERE iS AT LEAST ONE PATH
LEADING TO EVERY OTHER ENCLOSURE
@) THERE ARE AT LEAST 2 NE
FERENT HNDS
OF ENCLOSURES.

THE 3 RULES OF MAPPING

G) © REPRESENTSAN ENCLOSURE
Qo ee AN ACCESS
® REPRESENTA DIFFERENCE
Figure 14
29 Constructing Combinatorial Lists

Some people will inevitably object to this axiomatic description, along

with its corresponding mapping rules, as being a little simplistic . This
would clearly be wrong; in fact, the description of the axioms and the
rules of the mapping cover every imaginable physical possibility for
joining enclosures together, and, thanks to the use labels, the en-
closures or their connections can be as different or as alike in any re-
spect as we are capable of imagining them.
The system | am proposing here is richer (offers more combinations)
than anything architecture has produced to date, in all of human his-
tory. And if one thinks about the variety of repertoires that can be
built by means of this mapping, one is amazed at the results, seeing the
poverty of our imagination.

2.6 CONSTRUCTING COMBINATORIAL LISTS

Now we come to our goal: the setting up of repertoires.
First of all, in this section we shall look at the abstract model for con-
structing lists, and study the question of which rules must be respected.
All of Chapter 3 will deal with translating these abstractions into
reality.
The mapping which we have decided on uses planar graphs, connected
and labeled, each corresponding to a particular plan taken from the
list of possible plans.
Every list of graphs will be composed as a function of nN, 7 being the
number of points contained in the graph (that is, the number of en-
closures in the plan).
Actually, there is no algorithm for the construction of such a list.
Therefore we will proceed (manually or with the help of a computer)
with the induction of a new point into the list of known ones.*
Just as an example, we shall construct all the solutions (and therefore
all the graphs) for n = 2,n = 3,n = 4, We'll only use two labels in these
examples: the sign O, representing exterior preexisting space, and the
sign 1, representing the interior of a separated space. In this way there
will only be one sign Oon every graph, all the other labels being 1.

*| call ‘construction by induction” the construction of a complete list of (7 + 1)

graphs, obtained by adding a point outside the graph to each possible n graph, and
by linking this point in every possible way to the points of each n graph.
30 Essentials of the Method

THE QULES OF MAPPING PERMIT

CONSTRUCTION OF @MPLETE COMBINATORIAL USTS
OF ALL POSSIBLE SOLUTIONS

2 ENCLOSURES HAVE ONLY 1 POSSIBLE URKAGE

a HAVE S POSSIBLE UNWAGE SCHEMES

% ENCLOSURES

SO erES
A ENCLOSURES HAVE AA PoscIBLE LINKAGE SCHEMES

ALAEAA
AA aa
Ftc. @: intERion
©: 6XTERIOR
Figure 15

There are eleven figures in the complete mapping list for n = 4, these
figures being included in the list only when they are topologically dif-
ferent from the others (when a figure cannot be transformed into an-
other by the distortion of its links, without cutting them, as if, for ex-
ample, the graphs were made of wire, and could be bent in any direc-
tion without changing the pattern or linkage scheme according to which
they are joined). They could also be represented in a more convenient
visual arrangement, where the label O would be replaced by a white dot.
| could also establish this list by using, instead of graphs, their adja-
cency matrix. By this | mean a binary matrix which will show by the
symbol 1 if, between two points, noted one at the head of a column,
the other at the beginning of a row on the matrix, there is an imme-
diate link (adjacency) and will show by 0 if there is no immediate link.
The adjacency matrix is always symmetrical with respect to the prin-
cipal diagonal when the links do not have a directional arrow, that is
when each link permits movement in either direction. On the other
31 Constructing Combinatorial Lists

A LINKAGE SCHEME B &

ce) uA

BUS

CAN RE WRITTEN
AS A MATRIX A — AdNACENCY
(ADWACENCY MATRIX) O — NON-ADJACENCY

WHICH CAN BE WRITTEN AS WELLAS A

Sé QUENCE |
OA44 4040 - AAOO. Ct) 4000.

#* DENOTES A LABEL DIFFERENT FROK OTLERS

CEXTERION SPACE) Ss
Figure 16

hand, if a graph has directional arrows, the corresponding matrix is not

symmetrical.
The difference between different graphs rests only in the position of
the label O. Therefore we can put this sign in parentheses (O) before
the element of the sequence which represents the location of the point
marked O. In the preceding example, 0111 represents the point A, 1010
the point B, 1100 represents C, and 1000 represents D. Therefore for
the same graph we could write the position of the point marked.O
the point O in A: 0111.(0)1010.1100.1000 or 0111.(0)1010;1100.1000
the point O in B: 0111.1010.(0)1100.1.000 or 0111.1010.(0)1100.1000
the point O in C: 0111.1010.1100.(0)1000 or 0111.1010.1100.(0)
1000
In these notations, both the first and second expressions are exact,
for the points in them represent a complete row or column of the
matrix corresponding to the point in question.
32 Essentials of the Method

What is important in these matrix notations is that they constitute

a representation of the graph in computer language, and as we must
rely on a computer to set up all large lists, it is advantageous to adopt
a notation that does not have to be transcribed especially for it.
The combinatorial list of eleven figures (or matrices) in our example
contains all the possible solutions for linkages among any four elements,
as long as three of the elements have similar properties and there is no
more than one with different properties.
It is impossible to imagine any solution not included in this list, and
any other solution would be physically impossible. Therefore, this list
of eleven solutions to the problem is complete, or exhaustive.
It goes without saying that it is also a complete list of solutions for
establishing connections among four regions or four enclosures, where
one is different from the other three.
| am sorry to say that | have often shown so-called analyses of archi-
tectural or planning problems worked out by professionals, and that
in these analyses, which claim to compare possible solutions, certain
solutions were always purely and simply overlooked, even though some-
times the solutions that were overlooked were precisely the ones that
would have been most effective. Truly, sometimes our imagination is
limited!
CHAPTER 3 THE NEW TASK OF ARCHITECTS AND
PLANNERS

3.1 LIKE EATING OUT

The first two chapters described the problems of architecture and plan-
ing and proposed a solution based on the use of a repertoire. They also
showed how it would be possible to build a repertoire, at least in the
abstract. Now we’re going to see how a repertoire would actually work.
Let us take an everyday example. Restaurants the world over organ-
ize their service in terms of a repertoire. We call that repertoire the
menu.
The menu contains a list of all the dishes that the restaurant serves
(these are the preexisting elements). The menu lists each dish by its
name, along with a short description of how it is prepared and an indi-
cation of its price. We shall call the information accompanying each
item on the list the warning. The warning tells the client what the con-
sequences will be if he chooses a particular item on the list. In a res-
taurant, there will be economic consequences (involving the client’s
purse) or perhaps digestive consequences (involving his stomach). So if
he has taken the trouble to read the menu, the client is well informed.
Next, he will put together a combination of different dishes—in other
words, a meal. Probably no two clients will choose exactly the same
combination, even if they are having dinner together. A meal is a matter
of taste and personal choice, and there is no such thing as an indepen-
dent scale of values that says one gastronomical composition is better
than another one either in the absolute or on a scale of values.
The owner of the restaurant never interferes in the choice his client
makes, whatever it is, even if the choice disagrees with his own person-
al taste, or if he thinks the bill will be extravagant. His job is to serve
34 The New Task of Architects and Planners

the client, whatever he has ordered from the repertoire, and in accor-
dance with the information it contains. Actually, the restaurant owner's
creative work lies in setting up the menu.
This example demonstrates the method that | propose for setting up
a repertoire for the architect’s client to use. The repertoire will be
made up of a complete list of all the possible linkages and labeling (the
mapping of the problem itself). The client (earlier we called him the
future user) will actually have the freedom to choose any possible com-
bination instead of having to go along with the preferences of some
architect or other.
For each possible combination, the repertoire will have a correspond-
ing warning indicating its cost. The warning will tell the client the cost
expressed in effort; this is the intrinsic property of each combination,
that is, the advantages or inconveniences of this combination in terms
of the client’s particular mode of use.
There is a third problem to be solved, to which the architect or plan-
ner must find a solution: again as with the restaurant, he must provide
the support for all the possible combinations which a client is apt to
want. The architect guarantees the hardware, or the /nfrastructure, in
such a way that after the client has made a choice, he has the chance
to change or modify his choice (and consequently change the corres-
ponding hardware) when he thinks it desirable (see Sections 3.5, 3.6).
This example of the restaurant exactly describes the problems of
architects and planners.
The restaurant does not take into account one element, however, that
is very important for the architect: the fact that the warning given by
the architect or planner must be addressed not only to the particular
client who makes the choice, but also to the community (by means of
media, for example).
In effect, the community must be informed of the immediate conse-
quences for it of every individual choice. .
The following sections of the chapter will deal with the menu in archi-
tecture, and Chapters 4 and 5 will deal with a method for informing the
community about the consequences of individual choices.
35 Building the Repertoire

3.2 BUILDING THE REPERTOIRE

Now let us set up the menu: the list of choices from which anyone can
make his own environment. This is the repertoire we have talked about
so often in the preceding chapters.
The basis for the repertoire will be the complete list of diagrams of
possible combinations, which we spoke about in Chapter 2. These com-
binations will be described by planar n graphs, connected and labeled
by means of n — 1 (at most) labels of one kind, and one label (at least)
of another kind; they correspond to the possibilities of arranging
(n — 1) spaces which connect, directly or indirectly, with the exterior.
Elsewhere in connection with the rules of mapping | defined a label
as representing a specialization of space, a difference. \t goes without
saying that the label, as it marks a difference, can indicate completely
heterogeneous categories, depending on a given context (country, social
class, technical methods, language, and so on) in which the architect
supposedly works. Just to demonstrate, let us establish a code of
labeling valid in a context chosen at random. The first important

AN EXAMPLE CF NoTarric’)

CON FIGU RATION SHAPE POSITION OF THe EQU(PMENT| mui PMENT

273 4S. 6) 7
Cn oh | oa BOOGHMEoe| be!
ayVI V7 - Pl -

A LAREL PRECEDING A SECTION IN A SE

I ry (CORRESPONDING TO A LUNE IN THE ABYACE
Wc y
Lu 037 NATEUX
, THUS REPRESENTING 4 POINT OP THE
means SCHEME), SHows
OA HAP E
038 Pits MENT
Pes, TON OF THE EQUIPMENT
Figure 17
36 The New Task of Architects and Planners

decision will concern what hierarchy one wishes to establish among the
different categories. This hierarchy could be completely different for
each individual, and it is thus subjectively established. For example,
let us choose the labeling in Figure 17.
A point on the graph which carries the label 100 represents a square
room, not equipped in any particular way. A line joining two points
(which represent two rooms) and carrying the label 2 indicates that the
squares representing the volumes overlap by half a side.
Let us take one of the graphs in Figure 15, for instance, the last one.
We will give it an arbitrary label: Figure 18.
In a repertoire of solutions which deal with three rooms joined togeth-
er and having specialized functions in accordance with the labeling of
Figure 16, there are 11 possible basic combinations, (linkage schemes)
and, for each combination, 15 labels for the kind of their linkage to
each other, 18 for the shape of the rooms, 368 positions for 2 kinds of
equipment (kitchen and bathroom); and each apartment can have 64
different orientations. The ‘‘menu”’ that started with a single set of
labeling in fact contains 11 X 15 X 18 X 368 X 2 X 64 = 24, 330, 240
solutions in all.
Obviously that is an astronomical number (I hope that | have not
made a mistake in calculating it; | am not very good at arithmetic). But
this number does represent every possible solution for arranging these
three rooms when one can choose, for each room, from three different
shapes, two kinds of utilities, three positions, and four orientations. If
you wanted to write out this list, it would take two hundred and forty
volumes of a thousand pages, at a hundred lines a page (please see the
parenthetical note in the preceding sentence).
This figure shows, once again, what meager intuitive proposals archi-
tects actually make. These 24 million solutions are a// that are possible,
and all that could be desirable to a particular client. And no one could
decide ahead of time, without knowing this particular client, which
solution he would personally choose.
The difficulty with such a voluminous list lies not only in the fact that
the architect has practically no chance of guessing which solution a
client will choose; the other trouble is that the client would certainly
37 Building the Repertoire

AW. A 72)

4 6

200 °

THiS LABELED GRAPH

CORRESPONDS EXACTLY “TO THIS PLAN:

rr CduULD &LSO BE WRITTEN

IN THIS SEQUENCE '

(4 300 Olll A 112 lolo 6 72I loo © loool

Cc SEP M CSEPM €C SEP M EK M
C: ARRANGENENT
S + SHAPE
E : EQUIPMENT
P * POSITION OF E.
VF? MATRIX LINE
EX : EXTERIC?.
Figure 18

not know how to manage such an unwieldy catalogue in order to choose

his three rooms.
To make a choice possible, it would be useful to reduce the list, and
instead of using labels with five digits (representing the five categories)
to use an abridged list which would still be nonetheless complete.
We will make up a complete list where each label belonging to a point
on the graph would have just one digit. We could consider, just as an
example, the linkage of the rooms and their utilities, and ignore the
arrangement, the shape, the orientation, and so on. Then, the list of
labels, with each label corresponding to an appropriate set of equip-
ment, could be the following:
1: kitchen + void, 2: bathroom + void, 3: void + void, 4: bathroom
+ kitchen.
If in our example we allow for no apartments without utilities, and
38 The New Task of Architects and Planners

no apartments with two kitchens (and consequently admit that label

3 is the only one we would meet more than once in a given solution)
there are only six different ways to arrange each apartment: 123, 132,
231, 334, 343, 433. In certain graphs for the list of combinations, there
could be labels identical to each other because of the symmetry of the
graph (in these cases, the labels would be topological equivalents).
Indexes of the graphs in the list in Figure 15: 4 enclosures (3 rooms
plus exterior): 1234567891011
Number of possible combinations of labels: 665666666 6 5
The total of all the solutions (which covers, in less detailed form, the
24 million solutions which we mentioned before) will not exceed 64.
This list could be presented on a single page, and it constitutes a per-
fectly usable ‘‘menu.”
Instead of reducing the number of labels, another way to present such
long lists is to give only the labels, with the rules of composition to be
applied to them, leaving all the real work and the decisions to the client.
The methods are essentially identical. The only difference lies in how
the list as a whole is presented.
In Section 7 of this chapter, we shall see a working example of the
second method, thanks to the use of a very simple machine which
makes things easier for the client.
| should add here a remark to explain why | only use planar graphs
(I have already explained a similar choice with respect to the axioma-
tics). A nonplanar graph is a graph of five or more points—we shall
come across one later on—which, on paper, will show crossovers. But
these crossovers (crossing points) do not conform to our rules, and
they have no meaning by our conventions, according to which a point
represents a space or an enclosure. We might consider the crossovers
as representing spaces or enclosures not catalogued in the repertoire of
enclosures, such as hallways or stairs.
There is a general statement which indicates, for a graph, the greatest
number of real connections that it can contain (the saturation point of
the graph): L = 3(n — 2), where L represents the number of connec-
tions.
Every time the number of connections in the 7 graph is greater than L
39 Building the Repertoire

there are inevitably crossovers, that is, spaces not catalogued in the
repertoire of enclosures (in the set of enclosures mapped by the graph).
We find a typical example of a crossover, or an uncatalogued space,
in the solution of a small, insignificant problem like the following: con-
struct four rooms, each one of which is directly connected to each of
the others and all having an access directly on a garden (Figure 19).
However | try to plot this graph, there will inevitably be a crossover
(C), which could be interpreted as representing a stairway or a corridor.
Without the intervention of this sixth ‘‘uncatalogued’”’ room, the prob-
lem is insoluble.
Now let us recapitulate the steps we take to construct the complete
list of solutions for assembling n elements (the first part of our reper-
toire, without the instructions):
1. Construct the set of all the planar 7 graphs.
2. Choose positive categories in order of importance (according to per-
sonal and arbitrary criteria of appreciation) and rank each of the cate-
gories.
3. Assign a symbol as a label to each possible case in each category.

A FouR-Roon House
IN WHICH ALL THE ROOMS
eel ee IN TERCON NECTED
AN
EVERY ROOM HAC A DOOR TO THE GARDEN
\S IMPossiBLe
as CROSSOVER ,
A REPRESENTS A STAIRWAY
oR A coORRIpoR

Figure 19
40 The New Task of Architects and Planners

4. Set up all the possible labels according to the directions in steps 2

and 3.
5. Set up the complete labeling system for planar” graphs, using the
labels established in step 4.

3.3 THE PATH MATRIX

As | have described it, the repertoire contains more than just a com-
plete list of items. There is also a warning for each item catalogued in
the repertoire, which is made up of two elements: the first is simply
the purchase price (or the ‘‘purchase effort’’) for the combination in
question. The second element indicates the characteristics of the hard-
ware which corresponds to this number on the list—what the user can
expect from it (tentatively, we will call this element the use efficiency).
The path matrix (which will be important in indicating the use effi-
ciency of an item) shows the shortest distance from one point to an-
other, for each pair of points on the graph. Like the adjacency matrix,
it is symmetrical with regard to the principal diagonal for any graph
which does not carry a directional arrow.
The ia column shows the sum of all the distances from one point
given on then graph to all the other points on the same graph. SOdi
therefore, shows the average position of a given point (SSd/n — 1 will
be the term for the average distance from one point on the n graph to
all the other points). Thus ) can be used to measure the site advan-
tage which a given room has in comparison to the other rooms within
the same layout. In the example in Figure 20, | did not assume differ-
ent distances between any two adjacent points, and to make matters
simpler | proposed distances having the same length. In this example,
enclosure C has a “‘privileged’’ site compared to rooms A and B, which,
in turn, are privileged compared to room D.
But let us suppose that for his own reasons the future user does not
use these four basic rooms equally. Let us suppose, for example, that
in one day he were to go twice as often into room A as into rooms C
and D, and three times as often into room B as into rooms C and D.
We would call this frequency of movement to a particular room the
use weight of that room, calculated for a given time span.
41 The Path Matrix

A B
CORRESPONDS
To THIS PLAN
(FOR EXAMPLE)
Cc D

TwE PATH MATTIUX

Re TIVE ACCESS \QAUITY
OF GACH Coe

ve
EEQY DAY cere aes ps
: oa
opal
wi 2344.
TA BCD! Ldw
ERPOgT Loy
HE BXPENDS BVERY DAY seeder lee
IN USING HIS HOUSE . 2401 2
(w= paaquenc) A610 aH
Figure 20

We could write the use weight into the series representing a graph,
just as we did for the labels; for instance, we might write it in paren-
theses in front of the element in the sequence that represents a given
point. So if you indicated the use weight that way, the graph in Figure
20 would give the following sequence:
(2) 1100(3)=1105(1) 15111) 001
Now let us construct the path matrix in an improved form. Each term
of the new matrix will show the product dw of the distance by the
use weight (or how often the route is used) for any two points. This
product is obtained by multiplying the frequency of use (w) from the
point which is the terminus of the path.
To obtain the new matrix, we write in above the path matrix, in the
column which corresponds to each terminal point, the use weight of
that point, and we multiply each term in the column (representing the
distance) by this frequency (w). The matrix thus obtained will show
the real effort the future user will make during a given time span (in
this case, during one day) in using the set of enclosures in his particular
way. This new matrix will be called the deta//ed effort matrix for this
linkage of enclosures (apartment) or for the corresponding graph.
42 The New Task of Architects and Planners

After constructing the detailed effort matrix, we shall take the sum of
all the terms in each row of the matrix. The sum total of a given row
corresponding to a point of departure on the graph (and to a given en-
closure in the apartment), or the /oca/ effort va/ue, will measure the
site advantage of that enclosure within a given linkage scheme used in
a particular way. The formula for local effort value (E) for the n graph
will be:

y=n

E, =) Ay "Wy,
=1

This formula represents our warning system. |t makes it possible to

inform the future user what the consequences will be if he chooses a
particular linkage and uses it in his particular way.
To improve ‘‘readability,’’ whenever the graph representing the set of
enclosures has a large number of points, we shall indicate the local ef-
fort values E by writing them above each of the points that they belong
to. We shall call these lines the /ines of isoeffort, and the resulting map,
the map of isoefforts, or the effort diagram.
To give an ideal of how this diagram works, let us suppose that the
future user chooses the linkage of enclosures that he prefers. Once he
has made the preliminary choice, he assigns to each enclosure in the
set a use weight (w), reflecting his own way of using a particular room.
The diagram constructed from these data (the linkage scheme and the
“projected’’ use) will show the local effort value corresponding to each
element in the set, as a function of the client’s particular behavior.
So the effort diagram informs the future user about the consequences
of his choice (comfort, physical exertion, and so on) and lets him com-
pare these with the consequences of other preliminary choices he has
made, until he decides on one of the linkages, the one that seems likely
to give him the most satisfaction.
The effort diagram lets him experiment, in his mind, with several pos-
sible arrangements of space without having actually to build them all
(which would be impossible in any case). It lets him compare different
linkages from the point of view of a particular manner of use, whatever
43 The Network and Its Hardware Counterpart

it may be, without making the mistakes usually made in intuitive com-
parisons. (Anyone who has served on an architectural jury knows that
there is no common scale of values that would allow for a comparison
of two projects, which is not true if one uses ‘‘effort maps’’).
It is very easy to construct this matrix, even without the help of a
computer (as long as there are not a large number of points), and it pro-
vides useful information about the consequences of each choice, or of
a change in a client’s behavior.
In addition to local effort, later on we will also apply the concept of
global effort, under the formulae =u = aped, which represents
the sum of all the local efforts in a given linkage scheme for a given
distribution of use weights. This concept will be very useful to us when
the problem takes the form of a comparison, pure and simple, among
different distributions of use weights within the same linkage, or
among a large number of different linkages (Chapter 4).

3.4 THE SUPPORTING NETWORK AND ITS HARDWARE

COUNTERPART
Everything discussed so far concerns only the act of choice. Who must
make the choice, and why? What is a repertoire? How should a reper-
toire be built, and who should build it? What are the consequences of
a choice between the possible alternatives? We have dealt with such
questions as these.
Now, in this section and in Chapter 6, which covers the technological
considerations, we come to the following problem: what is the ‘‘hard-
ware” that makes it possible to construct any of the combinations in
the repertoire with the minimum of constraints?
Let us examine the characteristics which we must look for in this
undetermined hardware, which is necessitated by the fact that the ac-
tual construction, in the material sense of the word, cannot wait for
the final decisions of all the future users, and cannot be delayed by
long fruitless inquiries.
We have to find a kind of hardware that can be built before all the
choices are made, and which, once they are made, allows for any organ-
ization of rooms according to a particular choice, whatever it may be.
44 The New Task of Architects and Planners

We have already seen that individual choices were represented by

points, lines, and labels, and that we could obtain a complete list of all
possible choices by using these simple components and taking the fol-
lowing steps:
1. Take a set of N points
2. Make a combinatorial list of all the planar connections transforming
the 7 points into a unique and closed figure
3. Construct for each of these links the combinatorial list of all possible
m labels (m = the number of labels).
Actually, the first step presupposes as a previous condition that there
are at least 2 points with absolutely no links between them. Once laid
out, these” points can be used to build the representation of the entire
repertoire, simply by applying the rules of mapping which we have
already studied.
The simplest figure necessary to support the repertoire will therefore
be a completely nonconnected graph containing at least 7 points.
Naturally, we could trace any of the graphs in Figure 15 onto the com-

AN INFRASTRUCTURE
1S A PHYSICAL. SUPPORT
OW WHICH AU POSSIBLE PLANS CAN BE IMPLEMENTED
ae PARA Came FORM (MAPPING) OF THE [INFRASTRUCTUG
@ A SETOF @ A SET OF PoINTs
UNCONNECTED POINTS STRONGLY CONNECTED

Figure 21
45 The Network and Its Hardware Counterpart

pletely nonconnected graph, simply by adding the connections between

certain points (or the sign 1 in the adjacency matrix for certain points).
The completely nonconnected graph, therefore, represents an /nfra-
structure for the mapping of the complete list of possibilities (for ex-
ample, the 24 million possibilities in Figure 17). This ‘‘abstract’”’ infra-
structure is truly the most indeterminate one possible.
So the entire infrastructure is represented by this set of unconnected
points. We can decode this abstract image by inverting the rules of map-
ping for connected graphs (which were: ‘‘a point represents an enclos-
ure,”’ ‘a line is an access,”” and ‘‘no room can be inaccessible’’). De-
coded according to these rules, the set of unconnected points repre-
sents a set of inaccessible enclosures, as many as there are points.
When a future user makes a decision (a choice) the first question will
concern what connections to establish. They will be shown by a linkage
scheme for a certain number of points corresponding to the enclosures
he has decided to make accessible in a particular way. The linkage re-
quired by the client’s choice will of necessity be included in the reper-
toire, and this choice made from the repertoire can be carried out in
the actual construction (the set of enclosures with no access) by creat-
ing the corresponding openings in the previously continuous walls.
In a technical rather than industrial context this is like an artificial
sort of troglodyte habitat, such as certain cultures have come to pro-
duce, in which the ‘‘envelope”’ of the enclosure is built first and then
doors and windows are ‘‘punched out.”
Figures 22 and 23 show a model of this kind of infrastructure and its
corresponding mapping, along with several arbitrary examples of spe-
cific choices, represented in graphs and transposed onto an infrastruc-
ture.
This is only one example of an infrastructure effectively adapted to
a particular local industrial context. | have proposed this type of infra-
structure to solve construction problems for low-cost housing in an
African nation. This proposition made it possible to combine mass pro-
duction using local techniques with the very rich individual character-
istics of a sophisticated culture.
Another way of transposing the principle of the graph of unconnected
46 The New Task of Architects and Planners

THE (TROGLODYTE’ INFRASTRY CTURE

CORRESPONDS
TOA SET OF UNCONNECTED POINTS

(Ts PHYSICAL REPRESEWEO SENERAL 4QBITRAAILY CHOSEN PLangs

i A SET (CORRESPONDING TO THE THKK UNe
OF UNCONNETED ENCLOSURES GRAPHS) CARRIED OUT IN THE
INFRASTRU CTURE
BY OPENING PARTITIONS
Figure 22

wTROGLoDyTe”
INFRASTRUCTURE

FAG
PHY,
Nena
ARBITRARILY
PLANS
CARRIED ONT
CHaSeN

IN TRE INFRASTRUCTURE

Figure 23
47 The Chance to Make Modifications

points into hardware is to use the complementary figure: that of a

planar network having the maximum number of connections between
points, so that all enclosures mapped by this figure would be as inter-
connected as possible from the start (thus having maximum accessibi-
lity). It would follow that all the points in the map of the infrastructure
would be linked as often as possible in a planar graph or in a planar net-
work (| have already spoken of saturated graphs in Section 3.2 and |
shall deal with planar networks later on). A particular choice could be
schematized by blocking out one or more of the links from the maxi-
mum number, which would correspond to building a wall in order to
seal the access between certain enclosures, rather than punching out an
opening, as in the first example. The kind of hardware that permits the
use of this technique is called the “skeleton” type, shown in Figures
24 and 25.
The example in our illustration certainly displays a wealth of forms
(since a large number of labels was included). It is easy to obtain so
much variety within the infrastructure whether it is of the troglodyte
type or the skeletal type, but | should add that it is easier from a tech-
nological point of view with the skeletal type.

3.5 A TECHNOLOGICAL NECESSITY: THE CHANCE TO MAKE

MODIFICATIONS
Figure 6 in the first chapter showed how the future user made his de-
cision—a process which eliminated the communication bottlenecks. The
figure was made up of two loops: the first represented the act of
choosing, the second represented transposing the decision into hard-
ware. Now we are going to add a third loop—one for corrections—to
the diagram (Figure 26).
The significance of the diagram is this: the future user chooses an
initial arrangement of enclosures, which are built inside the infrastruc-
ture. After using it for a while, he discovers that he would like to
change certain aspects of the arrangement, for reasons of his own (per-
haps a change in his life style, a miscalculation, change in an exterior
factor). He will make a second choice (with or without the repertoire
this time) and will decide on a new arrangement for his enclosure.
48 The New Task of Architects and Planners

A SKELETON INFRASTRUCTURE
CORRESPONDS TOA SET
oF STRONGLY CONNECTED POINTS

‘al

ae

ITS PHYSICAL ASPRESEN ETO SEVERA ARBITRARILY CHOSEN

\S & SET OF ENCLOSVAES PLANS
STRONGLY CONNECTED CCORRESPONDINGTd THE FULL
UNE GRAPHS)
CARRIED OuT IN THE INFRA-
STR ctu
BY ABBING PARTITIONS
Figure 24

Naturally, his new choice will correspond to one of the possible solu-
tions; it will therefore be capable of realization within the infrastruc-
ture, and the opportunity of implementing it will be his due. (I shall
deal with the technological conditions for carrying this out in Chapter
6.)
Now, although the point of the infrastructure was a solution which
we could achieve by abstract reasoning, and although the reasoning led
us to define the infrastructure as a type of “‘hardware,”’ it is not the
same with modification, the necessary conditions for which cannot be
determined in this way.
Actually, the possibility of modification is not a problem of abstract
structure, like the repertoire and the infrastructure, but rather a simple
technological problem, at least in its principal aspects. However, we
can state in abstract form the conditions for making modifications by
saying: “Everything within the infrastructure (and therefore the whole
49 Keeping a History of Modifications

process of opening up or of shutting off rooms) must be reversible.”

The hardware must be changeable—only the infrastructure can be rigid
and fixed in place. If | wanted to make a diagram that conveyed this
statement visually, | would draw only the infrastructure in ink and use
pencil for the connections and labels. Modification would consist of
rubbing out some of the lines with an eraser.
| think that in thus separating the infrastructure, which corresponds
to a set of unconnected points, or to a saturated graph, from its ‘’filling-
in,’ which is the changeable hardware corresponding to lines and labels
that are added or removed, | have succeeded in articulating the funda-
mental conclusion of this chapter. | shall get to its technological aspect
in Chapter 6. Here we are interested onlyin the model.

3.6 KEEPING A HISTORY OF MODIFICATIONS

If it is possible to make modifications, then the forms of filling-in for
the infrastructure (for instance, opening or insertion of partitions) will
change with time.
All possible changes, along with all possible variations, will of course
remain in the complete list of possibilities (links and labels). It is inter-

SKHELETé:) INFRASTRUCTURE

EMPTY SEVERAL ARBITRARILY CHOSEN

INPRASTRUCTURE PLANS
CARRIED OUT INTHE (NFRASTRUCURE
Figure 25
50 The New Task of Architects and Planners

THE USER mobiFies “THE ARRANGEMENT CHOSEN AND

CARRIED OUT WITHIN “THE INFRASTRUCTURE, EVEN AFTER
\T WAS CARALED OUT. He DOES IT IN ACCORDANCE WITH
THE WAY HE PREFERS TD USE IT IN THE FOTURE.
THE SEQUENCE ARC |S THE HISTORY OF HIS MODE oF WE.
ABC Covub MEAN, FoR EXAMPLE, OW THE BASIS GE THIS LUST

fe tl eee)
A B c b € F
THAT “THE USER'S HiIstORy &

alee
A
GS)8 Se aieealS
Hig HisToDRY IS THUS GEPRESENTED FROM THE ComPLETE Ust OF
SOLUTIONS, WHICH SeRVES ALA REFERENCE UST.
Figure 26

esting, and important, to be able to follow the history of these changes,

to note the sequence of different modifications that users make over a
particular span of time, transforming one linkage represented by the
graph G* into another, represented by the graph GY, if planning has the
least hope of becoming a scientific (or simply a conscious) activity, it
will fulfill it only by recording these ‘‘histories’’ and by trying to detect
the regularities which govern them. It will not be an easy job. We shall
study the tools for such research in Chapters 4 and 5.
For now, we shall think about the problem in its simplest form.
Let us imagine that in a given context a series of modifications have
been made within a particular infrastructure. And to simplify the ex-
ample, let us suppose that only four elements (four kinds of filling-in)
were involved at each stage of the modifications. We shall use the com-
plete list in Figure 26 as a reference, and in the sarne figure we shall
describe, step by step, the sequence of different arrangements of link-
51 Keeping a History of Modifications

ages that a fictitious observer would have seen.

For the ‘‘configurational history” of a particular user’s four-room link-
age, we shall use a sequence likead bb d ed. We shall come back to
this idea later on.
There is another history that the observer might see: a change not in
the linkage, but in the use weights of the different enclosures. Figure 27
gives an imaginary example (taken, for purposes of demonstration, from
the diagram in Figure 26).
When the use weights change, the local efforts change with them (dw
being dependent on w), for each particular element of the linkage of
enclosures. r
We could describe this history in yet another way: first by drawing
up a complete list of all the possible distributions of use weight (all the
possible labels for a graph of w’s for a given graph), then by calculating
local efforts for each distribution. | shall demonstrate this procedure
using the same graph from Figure 27. Figure 27 shows a list of weight
distributions for this graph, where the weights allowed for are
We lion 2.
This table constitutes a complete and well-ordered list for all the dis-
tributions of w that are possible in a graph of four points of the type in
Figure 27, with densities of 1 or 2 (2 labels) and shows all the possible
local efforts on the same graph. Once we have this table, we can use it
to follow the history directly after observing the events (or use varia-
tions) in the corresponding linkage of enclosures for regular periods.
For example, we might get Table 1.

Table 1

step 1 2 3 4 5 Ge ee
distribution S 7 2 6 3 MOR 3
Dd Na 19 24 29 29 19 Oy ee

In order to detect an eventual underlying regularity, it is important that

the list of distributions be well ordered and drawn up according to a
certain rule (such as a combinatorial operation).
52 The New Task of Architects and Planners

A B =
THE QRH WS REPAGSENTS THE PLAN in
T

a,>d
Cc D
aie MR
K Xx
ave OFS To EACH ROOM ONCE OR TWICE
THE CO ere

ee APORTMENT
@ THE SURED BYor Cm
Molowma:
Use FOR.
MR OX wile
CTHE SIGN e@ For
Goes THERE Twiceon Hoda collet Puhalen wedi

Beestat
te ili
,2
\

t)
-~-

@
z5
2\

A CASE HISTORY FoR MA x CouLtd BE

THE SEQUENcR: J. IV, WLW. vu, vi.
HIS EFFORT HISTORY WOULD BE -
16 14, (4,0, 24,2825,
OTe

&TC
FoRT
EF
BERESR
Ke
Sst ThE

Figure 27
53 The Flatwriter

3.7 AN APPLICATION OF THE REPERTOIRE: THE FLATWRITER

Thanks to a machine which | call the Flatwriter, each future inhabitant
of a city can imprint his personal preferences with respect to his apart-
ment (flat) to be, using symbols which put in visual form the different
elements of his decision so that the builder as well as his neighbors can
understand what his choice is.
In other words, this machine contains a repertoire of several million
possible plans for apartments, knows how to work out instructions
about the characteristic consequences of the way each future inhabitant
would use an apartment, and finally, can determine whether or not the
site chosen by a future inhabitant will risk upsetting the other inhabi-
tants.
Along with the Flatwriter | foresee the actual building of an infra-
structure, of an empty, multileveled framework which contains a utili-
ties network: water, gas, electricity, and sanitation.
The “organigram”’ (working plan) of the Flatwriter is shown in Figure
28.
Let us explain this working plan in terms of a certain Mr. Smith’s
decisions, and see what answers and advice the machine would give.
Syste
Mr. Smith is presented with a keyboard. The keyboard contains
1. all of the connections and the configurations which three spaces
(just to keep the example simple) could have among them;
2. all the shapes which each space could have with regard to these con-
nections (and in a technological context defined at the start);
3. all the positions which the kitchen and bathroom equipment could
occupy in a given space;
4. all the orientations the apartment could have.
Mr. Smith types on the keyboard just as he would type on a regular
typewriter. He chooses a configuration, than a shape for each room in
his apartment; he installs a kitchen and a bathroom, and finally he de-
cides which direction his apartment will face. He has used eight or
nine keys, on a keyboard of fifty-three keys, to choose his apartment
out of the 24 million possibilities which the machine offers.
54 The New Task of Architects and Planners

Figure 28

SCHEMATIC PRESENTATION Of THE Flatweirer.

(PRESENT ROLE
OF THE ARCHITECT)

MR (PRESENT AcE
AUSER : OF THE
AND AN ; nf «=BuILDeR)
INDIVIDUAL, ; ons oF SA

Ma x's
COMMUNITY!
INDIVIDUALS WHO MEST
CORK IST 4ND/OR COOPERATE.

THE PLANN ’
THE ROILDE >:
BY BEGSIONS
MADE BY THE
PTRE INKHABITANT.
Figure 29
 ajalal
The Flatwriter

3 5 gl]a
(C)
FITTING & y 2 a
SHAPES

b
©:

A|«aa
Ss

Y
ZF
ay

w™~ z
4) : O

Figure 30

=O DoD Ae HAD
Figure 31
by]
56 The New Task of Architects and Planners

STEP 2
At this stage, Mr. Smith makes all the decisions and the machine simply
prints them out, without contradiction or criticism. If, for example,
Mr. Smith has touched the keys in Figure 31, the Flatwriter prints out
the plan in Figure 32.
STEP 3
The Flatwriter prints out the cost of the apartment he has picked, cal-
culating according to an appropriate classification for each item.
STEP 4
Mr. Smith moves on to another keyboard—the keyboard of weights.
What it does is show him how many times a day he is in the habit of
going into the first, second, and third rooms. One after the other, he
touches the keys which represent how often he would come and go in
the different rooms in the plan he has chosen.
Siiebe5
Now the Flatwriter’s first warning appears: it tells Mr. Smith the con-
sequences of his choice, based on the plan and his way of life: the
effort implied by his use of each room in his future home. If this
effort, calculated from facts he has supplied, seems to be a disadvan-
tage, Mr. Smith chooses another plan (perhaps by changing the distri-

IAN
Figure 32

EXEQUTION CATEGORY A4 AX
CONSTRUCTION PRKE ............ $
Figure 33
57 The Flatwriter

HOW MANY TIMeS A DAY

Bo You Gd TO EACH ROOM2

FUTURE

EEL
Room §ROON Rook
\ Ty TT)

Figure 34

bution of functions for each room) or he may even think about chang-
ing his own way of using the rooms (the number of comings and
goings).
STEP 6
Next, the Flatwriter reproduces on a video screen a picture of the
plan of an infrastructure, an empty framework (Figure 35). Each empty
space in the framework has a number. When these spaces are taken up
by an aprtment, their outlines show on the screen.
| should mention that the infrastructure has been planned so that:
a. any choice can be realized in the empty spaces (voids);
b. any choice initially realized in these spaces can be changed and cor-
rected, without entailing any change in the infrastructure itself. All the
changes within the infrastructure involved only the movable parts (par-
titions, ceilings, floors) that have been inserted into the framework.
Mr. Smith types out the numbers of the spaces (voids in the infra-
structure) where he wants to place the apartment he has chosen. He
choses the spaces in Figure 37.
58 The New Task of Architects and Planners

NO} UNDER CONSTRAINT

NOMBER : FREE
Figure 35

Rss Ss NeSty
HV
pet

BA Ne
Figure 36
eee
59 The Flatwriter

Figure 37

SER 7

The Flatwriter inserts the outline of the apartment into the infrastruc-
ture, checking to see that the site Mr. Smith has picked out does not
block the light, the air, or the access of one of the neighboring apart-
ments that is already entered. In the case that his apartment should
conflict with another one, the Flatwriter would answer:
Please Make Another Choice
STEPS
Once the control mechanism has accepted the apartment Mr. Smith has
chosen, on the site he would like, the Flatwriter recalculates the effort
diagram for the entire infrastructure (efforts corresponding to all the
apartments and public services already placed in the infrastructure).
This diagram, in color, is displayed on the screen, superimposed on the
infrastructure plan and the apartments already situated, or perhaps to
the side on another screen.
The effort diagram informs the entire population of the infrastructure
of changes implied by each new arrival for the utilization of the city or
neighborhood (the infrastructure). To give a rough interpretation, the
diagram shows whether the number of people passing by each dwelling
has increased or diminished in comparison to neighboring dwellings.
This information is very important: noise, quiet, commercial value,
accessibility, and so forth are all included in the parameter effort.

CONCLUSION
The Flatwriter allows each future inhabitant of a city
a. to choose the plan and the characteristics of his apartment (his own
individual environment), thus taking the place of the architect;
b. to choose the location of his environment in the city and receive an
immediate ‘‘construction permit,”’ thus taking the place of the planner
and the city council;
c. to be informed of the particular consequences, directly concerning
60 The New Task of Architects and Planners

him and his home, of each new choice or new decision being made
throughout the city.
The Flatwriter thus puts a new informational process between the
future user and the object he will use; it allows for almost limitless in-
dividual choice and an immediate opportunity to correct errors without
the intervention of professional intermediaries.

YY

\
Figure 38
/
AV. ‘a
CHAPTER 4 URBAN MECHANISMS

4.1 COMPLETE PLANAR GRAPHS WITH n 27: CROSSOVERS AND

POTENTIAL CONFLICTS
In all the preceding sections, we referred to linkage schemes as the base
for compiling a complete list of solutions to architects’ and planners’
problems. All other criteria were expressed by labeling (for example,
shape, fixtures, use weight, and so forth). The linkage schemes were
always presented in the form of convex planar graphs, and in all our
examples, the number 7 of points in the graph (the number of enclo-
sures) was very few.
We also talked about the impossibility of allowing crossovers in these
graphs, because according to our rules of mapping they had no signifi-
cance. Therefore, all the graphs which we studied were planar graphs.
| would like to come back to the idea of crossovers. Whatever system
| propose to represent by graphs, its elements will always be symbolized
by points, and the connections between the elements by lines. These
points and lines can be marked with labels, which indicate some differ-
ence between two elements or between two connections. But what does
it mean if crossovers appear within these graphs? They are points, but
they do not represent what | have just referred to as elements of the
system. They certainly do not represent connections. Consequently, we
must either invent a new mapping rule to specify the exact significance
of these points, or simply ignore them (which in itself would be a new
mapping rule), and indicate that at such a point, lines pass over one
another: in other words, that the graph is nonplanar. What, then, is the
state of affairs that this new rule describes?
| wanted to show through this reasoning that a crossover represents a
sort of uncertainty, as when a motorist comes to an intersection with
63 Complete Planar Graphs with n 27

no signs and does not know which road to take. A crossover implies
some idea of danger, for there could be a collision, or simply a turn in
the wrong direction, and there are no specific instructions that would
eliminate the danger. What | am describing is a point of general uncer-
tainty: where should he go? What may happen? And from what direc-
tion? | shall call the uncertainty represented by crossovers potential
conflict.
While there may not actually be a conflict, in our example of the
motorist, there is a good chance of one. However, the chances of con-
flict will be lessened if | receive adequate information about its proxim-
ity or if a regulatory device is installed at the crossover.
Earlier, this consideration led us to an interpretation of crossovers as
full-status points, but representing special enclosures, such as stairs, cor-
ridors, and so on, which we can think of as regulatory devices.
But these crossovers are like rabbits (so are the conflicts, for that
matter): they have a tendency to multiply at a terrifying rate. The num-
ber of crossovers in a complete graph (a graph where every point is
joined to every other point) is a function of n (which represents the
number of full-status points in the graph). The formula for finding the
number (a conjecture of Mr. Anthony Hill’s) has not yet been verified.
It is
1/64 (n—1)? (n—3)2 ifn is an uneven number
CKn = 2 er
1/64 n(n—2)* (n—4) ifn is an even number.

This means that the number of crossovers (and therefore of conflicts)

increases proportionally as the fourth power of the number of points
which are to be completely linked.
It also means that each time that there are more than seven points to
connect, there will be more crossovers than linked points. Therefore, if
we suppose that the crossovers represent stairs or corridors, a linkage of
seven main enclosures would necessitate nine hallways or stairways, if
every enclosure were to be connected directly to every other; and each
hallway and staircase would have four doors (see Figure 40).
If we want to construct the most complete possible 7 graph with no
crossovers, we will notice that instead of the LK =n/2(n—1) connec-
64 Urban Mechanisms

Ceapy OF MORE THAN 7 POINTS | IF COMPLETE,

wie eave. MORE CREKSOVERS THAN POINTS (SUMMITS),
UNKAGE oF HORE “THAN 7 ENCLOSURES
Glu NAVE MORE CORRIDORS (OR STAIRWAYS) THAN ENCLOSURES.

.
.

THE CompLeTE GeAPH Ky LAN) CORRESPONEIN

HAS 9 CROSSOVERS TO Ky %,
HAS 9°2*7 CORRIDORS Of”
STAIRWAYS
Figure 40

tions established in the complete graph, we can allow no more than

LS = 3(n—2). The number of connections missing with regard to the
complete graph will be: LM = 1/2 (n?—7n—12).
It is clear than when nr is a large number, LV becomes very large com-
pared to LS. In other words, if we are to assemble a great many points
(or enclosures) it is impossible to complete a perfectly connected organ-
ization.

4.2 NETWORKS
As | have already described, plotting a graph with a large number of
points soon implies a large number of crossovers, sources of potential
conflicts, and only some kind of regulatory system can reduce the
chances of conflict.
Since there are many crossovers, there must also be many regulatory
devices, which brings us to a most uneconomical situation. It would be
more economical, for instance, to install a few complex regulators in-
stead of many simple ones.
65 Networks

| have already proposed another way to make this system more eco-
nomical, and that is to use saturated graphs, which have LS connections,
instead of complete graphs with LK connections. This arrangement
allows us to have LM pairs of points indirectly linked. We could
present the same problem in still another way: as a set of 7 points.
The complete list of possible connections (direct or indirect) will
carry a number LK. If | plot all the connecting paths on a diagram of
LS connections, the LK connections will all still be possible (let us call
them itineraries instead of connections or paths). In this case, the
same number CK of crossovers remains, but the actual crossing points
will be regrouped on sites. The numberof regulatory devices could
then be reduced by a power of four.
When ann graph with LS connections supports a complex system of
itineraries, | will call this graph the /nfrastructure of the system of itin-
eraries (by analogy to the infrastructure with the maximum number of
connections, which | spoke about in Chapter 3). Figure 41 shows the
saturated 6-graph used as the infrastructure of the complete 6-graph.
When there are a large number of points to be joined, the infrastruc-
ture becomes what is known as a network. In my terminology, a net-
work is a graph which has a very large number of nodes. Here we will
use exclusively planar networks. Actually, we can always draw non-
planar networks on a planar network by resorting to the appropriate
system of labeling, which would show that there is a relation in space
between one link and another, at a point of the planar graph. We can
i!lustrate this process with the help of a photograph of a complex spatial
structure. If it is a clear photograph, it is possible, at each point of the
picture plane, to ‘‘read’”’ which line passes in front of which, or which
lines meet in a point. This photograph is an excellent example of a
labeled network, where the clarity of the image at each point on the
plane surface acts as a label.
The accuracy of this statement is easily demonstrated in Figure 40 by
replacing the labels of an arbitrarily constructed network with over-
passes and underpasses and /eve/ crossings. The labeled planar network
which results constitutes an exact mapping of a nonplanar network.
In fact, the labeled planar network cannot be a picture of a spatial net-
66 Urban Mechanisms

THE SATU Geary Cn)

HAS L,23tn-2 LINKS

TWE conpiere carn CK,,) THE SATURATED GRapH S,:

HAS bye
Be(n-t)
? UKs S, PUNCTIONS Ag AN
' ORE FOR Kk,

Figure 41

work unless it has at least two kinds of labels. If it has only one kind, it
can represent only a regular planar network or a set of lines bypassing
each other in space (Figure 43).
Obviously, we can easily replace a network, planar or nonplanar, by
the corresponding adjacency matrix.
Henceforth | shall use the term network to mean a graph with a large
number of points, where there are at least two different possible itiner-
aries having no common point between any two points. These networks
will serve as infrastructure for a large number of itineraries and will
assure that the system works even if one of the itineraries is interfered
with, for the others will make the necessary connections possible. These
networks will be described by a diagram or a matrix, like the planar
graphs, but if necessary it will be possible to label them so that they
could be interpreted as nonplanar. They may be regular, and in this
 Networks
67

A LABELED PLANAR Gey CNETWORK)

CORRESPONDS EXACTLY TOA GRAPH IN SPACE
LEGEND: © + LEVEL CROSSING
4 : glC ovearas
2s: a UNDERPASS

Figure 42

A PLANAR GRAPH
(S Not A SPATIAL NETWORK'S MAP
VNLESS IT HAS AT LEAST 2 KINDS
oF LABELS
Figure 43
68 Urban Mechanisms

case, a rule describing the construction of the network in one area will
be valid for the entire extent of the network.
The model for regular planar networks, which will be the focus of our
attention from now on (since all other types of networks are derived
from it) is the type called networks of homogeneous degree, that is,
networks where the number of lines meeting at any one point is always
the same. The degree of a point is the number of connections leading
to or from it; consequently, in an adjacency matrix the degree of a
point is given by the sum of all the values of its row.
There are four planar networks of homogeneous degree: those of
degree (3), (4), (5), and (6). (1 note the degree by putting the number
that indicates it in parentheses).
These networks will be the general models which we will use to ana-
lyze systems composed of a great number of points.

4.3 EXAMPLE OF AN ANALYSIS USING NETWORKS:

PATTERNS OF TRANSPORTATION
We have seen in Section 2.3 that before we analyze a system, we must
first describe it using axioms (three, if possible). Once the description
is complete, we assign a mapping rule to each axiom, and we compile a
combinatorial list conforming to these rules. To each term on the list
we attach a warning or a ranking on an arbitrary scale of values, or in
accordance with arbitrarily established criteria. It should always be
possible to express the corresponding warning numerically (for exam-
ple, by the cost or local effort value, as in the case of the Flatwriter).
Let us try to demonstrate this method by applying it to the problem
of transportation systems.
First, we look for three axioms, as in Section 2.3.
First axiom: a physical object moves on a line.
Second axiom: two physical objects can move along the same line only
if they are going in the same direction.
Third axiom: two physical objects can move on two different lines on
the condition that they do not arrive at the crossover of these lines at
the same time.
First mapping rule: a transport system can be represented by a system
69 Example of an Analysis

of lines.
Second mapping rule: the lines of this system must carry only one
arrow between two adjacent points.
Third mapping rule: all crossovers of the line system must be broken
down into subcrossings of the different possible itineraries.
These mapping rules allow us to represent any transportation system
by an appropriate network. These statements appear obvious (axioma-
tics always appear obvious) but they imply a very restrictive network
selection: although the first rule allows us to use any network, the
second requires that all the points in the network can be reached once
the lines have been marked by arrows. Finally, the third rule shows
each itinerary as an isolated component of the network (which leads us
to the concept of the complete graph).
Within the restrictions imposed by the axiomatic description of a
transportation system, certain requirements are not expressly stipulated,
for example:
a. it is not stipulated that a moving object will necessarily touch all the
points of the system in its path
b. it is not stipulated that a moving object must return to its point of
departure.
We can add restrictions a and 6 to our axiomatics if it becomes neces-
sary within a given technological context, but axioms are much more

THE 4 PLANAR NETWORKS Of HOMOGENEOUS DEGABE

(3) (4) (5) (6)

INDICATES THE ‘
NUMBER OF LUNES
oF A NODE
ened ANB ARWVING AT THE NODE.
oes
A NETWORK IS OF HOMOGENEOUS pecree
FALL nee NODES HAVE THe S4me DEGREE.
Figure 44
70 Urban Mechanisms

general (weaker) than contextual statements.

The mapping which conforms to these three rules will include planar
and nonplanar networks. Since the nonplanar networks are derived
from planar networks by means of labeling, we will examine only the
planar networks, reserving the designation p/anar networks carrying
more than one kind of label for the nonplanar ones.
Since networks of homogeneous degree are the most suitable type for
the study, we shall use them to map our axioms and then compare them
to each other with reference to the specific problems that each axiom
presents.
The first rule does not imply the selection of any one network over
another, for all the networks are composed of lines.
The second rule determines how we place arrows on a network. Here
we can add the “‘contextual’’ postulate (a) that reduces the number of
moving objects, but each point of the network must still be accessible
to at least one of the axes of movement.
The third rule concerns the nodes (the points) themselves, and exam-
ines the number of intersections clustered in one zone. All the itiner-
aries that pass through a given point depend on the lines and arrows
that meet at that point, and therefore on the degree of this point. So
the decisive factor for this third rule will be the degree (sum of the
values of one row in the adjacency matrix), broken down into groups
of arrival arrows (entrance) and departure arrows (exit).
For a given network, the analysis will work this way: First step: plot
arrows on the four homogeneous networks (of degrees 3, 4, 5, and 6)
and in such a way that certain itineraries can traverse the network from
one end to the other (axes) and that all the points can be reached by at
least one of these itineraries. We will be able to establish a comparison,
on the basis of the number of simple itineraries (or axes), for example,
and on the basis of repeated use of certain parts of trajectories.
Second step: plot all the paths that pass through a given point of the
network (we define as points of the network those of degree equal at
least to 3), and count all the subintersections among these paths in the
vicinity of the point.
Thus, we have compiled two complete lists (networks and paths), and
Wi, Example of an Analysis

ANALYSIS OF TRANSPO&Tt SYOTEMS

any FOR ANY TRANSPORTATION MaPOING RULES SYSTEM
TRANS PotrrATion FoR

A. AN oBIECT MONGS ALONG 9-4) budoR” UNES

JECS ON THE SAME =e ACH UNE HASA
2. a cau MOVE IN THE ee inate DIRECTIONAL
SAME DIRECTION cogiergg XZ ARROW
3 Two
iN THECWwects CAN E
SAME PLAC >
= a Leeere yee nated
AT THE SAME TIME. Al. CROES OVERS.

COMPARISON oF SOME NETWORKS

ae St Be
A. BWEARY NETWORK 18 A SYSTEM OF LUNES
2. THey DIFFER WITH REGARDTS DIRECTIONAL ARROWS:

3. THEY DIFFER WITH REGARD To SUBCRASSINGS:

A SUB CORE ing

| A SB AASSING 4 Ps
(3) 4 SvubcRess nes,
Figure 45

the two studies that we have made—of directional arrows and of sub-
intersections—have given us numerical results which we can use as the
basis for comparisons or as instructions. A network’s properties of
direction and of intersection do not represent values, advantages, or
inconveniences, but they are in a sense intrinsic properties of the net-
work. They will not be considered as advantages of inconveniences
until they are introduced into a specific context that spells out the in-
tensity, speed, dangers, and so on, of a particular transportation sys-
tem. This implies likewise that a transportation system is not necessar-
ily represented by the network itself (which has only the role of infra-
structure), but that the system depends on how the infrastructure is
used.
The result of this study of networks may lead us either to modify the
context of transportation or the network. Either course of action may
72 Urban Mechanisms

be applied by physical or nonphysical means (for example, by setting

up a traffic code).

4.4 HOW DO YOU OBSERVE A CITY?

Let us broach another subject.
| would like to examine the way we observe a larger complex, such as
a city, adhering to the postulates concerning objective descriptions which
| established in Chapter 2.
To begin with, | must put forth the improbable hypothesis (so often
unconsciously respected) that we are able to observe a city as if through
the eye of God. In other words, we assume, in this case, that with some
unknown instrument of observation we could see everything in this
city, everything that happens in the streets, in the houses, in the heads
of all the inhabitants, and of all the other living creatures, including
dogs and microorganisms. Naturally, such vision is impossible.
Instead, we could consider using a television camera that films the city,
night and day, from a helicopter. This camera would not see inside the
heads of people or microoganisms, but it would record all the move-
ments of all the people in the city. This record would show who went
where, and when. But even if we could devise it, this instrument would
be too costly to operate. In addition, we would need almost as many
people to interpret the record as there were people living in the city.
Therefore, instead of dealing with a city of 100,000 inhabitants p/us
30,000 image decoders (a sort of secret police) we would actually be
dealing with a city of 130,000 inhabitants of which 30,000 were police-
men. Therefore, the tool of observation which would assure us of re-
cognizing individual identities (and itineraries) would be inseparable
from the city under observation; consequently, this tool would implic-
itly falsify (transform) the thing observed.
Thus, bit by bit, we have had to leave complexity out of our informa-
tion because of practical problems: we abandoned all hope of record-
ing the individual activities which are the origin of specific movements,
and then had to forego knowing the identities of the people making
these movements.
We could doubtless observe individual itineraries (without knowing
us How Do You Observe a City?

specific identities), but even that would be very expensive and very
complicated, given the state of our technology. But the outward life of
a city is nothing other than the individual movements of its inhabitants,
precipitated by their motivations. All that is essential to the life of a
city is thus impossible for us to observe on the basis of our present
knowledge.
Then what is the best way to observe life in a city, using the tools we
have?
If we cannot observe people’s motivations, their identities, or even
their itineraries, we can at least count the number of people arriving at
a given destination, without worrying about knowing who they are,
their reasons for coming there, or the routes by which they came. This
is an observation—very limited in scope—which we can make, counting
mechanisms such as turnstiles and photoelectric cells having been in
use for some time.
Let us think of what constitutes ‘‘“movement” ina city. | can count all

How CAN WE OBSERVE A UTY 2

Aah MissY MRK MISS

et
ee
eee
coeace

A .60D's-E YES VIEW " A wKINGS- EYE View" SS-

AN . ACLOUNTANT
SHOWS SHOWS Kye view" SHOWS

4, THE IDENTITY THE IDENTITY |. THE NUMBER

meaye Useh.
oFOF a THE OF PEOPLE
arty city ARRIVING
2. HIS ITINERARY 2. HIS ITINERARY a Pie ean
3. WIS REASON ;
For. GoInc,

IMPOSSIBLE ! IMPOSSIBLE! POSSIBLE!

Figure 46
74 Urban Mechanisms

the people arriving at a given spot. But | shall never know why they are
coming (motivation) or from where (names, addresses, preferred itiner-
aries).
The only observations | can make about life in a city are thus
1. How many people arrive at a given spot and with what frequency
2. What constitutes the network for the transportation system that en-
ables them to reach a particular spot.
This brings us to a schema for the city which lends itself to the kind
of information we learned about in the preceding chapters: we will re-
present the city as a network (hence as a graph with a large number of
points). In every sector of this network, we will note the number of
people arriving during an arbitrary span of time.

4.5 THE URBAN MECHANISM

The picture in Figure 45 is a familiar one: it is a linkage scheme with
labeling that describes use weight (see Section 3.3). This use weight has
been calculated from the frequency of visits to an enclosure or to a
given complex. In this sketch which shows the city, it represents the
frequency of visits to an enclosure within a two-dimensional! complex.
We have used use weight to calculate local effort, represented in the
effort matrix by the sum of the values in each row. We can proceed in
the same way here.
The local effort diagram noted on this plan can be interpreted in a
very vivid way: The little person drawn on the diagram at point A lives
at A. All week long he must make trips into town for work, social life,
entertainment, shopping, and so forth. Even he does not know what
trips he will make this week, this month, in the coming year. A new
job, new friends, or the prospect of a bargain may take him to any part
of the city. Getting from his home to any point in the city will involve
a certain number of obstacles: distance, delays at intersections, detours
around one-way streets. The more numerous the obstacles, the more
effort he will have to expend to complete the trip.
Not knowing for sure where in the city Mr. A. has gone, we can work
only on the basis of probabilities. Most likely, Mr. A. has gone to a
place where many other prople have gone too. However, it is possible
743) The Urban Mechanism

FACTS WE CAN KNOW THESE Facts

Adour A City: ane SYSTEMAAT LED
THE }OCAL BPRAAT MATRIX
€ NETWORK OF PATHS ALORDING. TO THE PORMULA
Deineneecuerees E oe A W,
2. oe NUMBER oP PEOPLE A‘ 4x"
HO HAVE ARRIVED THIS ae Caw BE PACED
Sonic Acavew PER
heeob aresWH 1S Bashan FoR THE
THAN THEICHMATE)
BAS)
pap aes
AG) Bad 64) 0@)

iia ve
{pee
(fh) rerne“

X€2) : ARV
a NUMBER
AT POINTx

Figure 47

that Mr. A. did not go to this well-frequented spot. In that case, there
is a greater probability that Mr. B. went there. If Mr. A. and Mr. B.
have both gone elsewhere, then Mr. C.’s visit to that spot is even more
probable. And so forth. If there has been a crowd in one place, we are
sure to end up by finding someone who has been there!
This means that the number of arrivals (use weight) in one sector of
the network is proportional to the probability that any person chosen
at random in the city will in fact go to this sector. So we consider the
weight as proportional to this probability, and the distance (or time) as
proportional to the difficulty in getting to a given sector. Consequently,
we may say that the effort is proportional to the product of the diffi-
culty in getting to a given point, multiplied by the probability that
someone will actually go to that point.
Thus, BE =.) * Wy, assuming that there are only two points in the
entire town, a and b. (= is the local effort required for a person’s effec-
tive use of the city if he lives ina; dap represents the distance he must
travel and the difficulties he must surmount to get from a to b; wy is
76 Urban Mechanisms

the number of people arriving at b in one week.

N. B. Naturally the time span can be of any length, and | have chosen
a week only as an example. For practical purposes | could choose any-
thing from a second (almost continuous calculation) to a year. In any
case, if the calculated effort values are to be useful as a basis of compar-
ison, we must define precisely the time span on which the calculation
is based.
If our network includes points, the local effort for point a will be
given by the sum of all the effort necessary to get to any point in the
network:

Xen

Ea = aed mea
x=1

Thus the diagrams of effort calculated in Figure 45, which show the
relative ease with which a person living at a given point could adapt to
some unforeseeable change in the way the city is used, in the course of
his comings and goings. These figures have no absolute value; they are
only parameters which enable us to know exactly which of two homes
in the same city offers more advantages with regard to trips throughout
the city, no matter how the inhabitant’s situation may change.
To reintroduce an expression | have already used, the effort values
serve as a warning to the user of any set of enclosures that is very com-
plex and is used by many people.
In other words, this warning tells him about the mechanism of the
city. And the effort parameter is thus a characteristic of the urban
mechanism, which is composed, as are all mechanisms, of two elements:
the infrastructure (the network) and its mode of use (probable choice
of termini for itineraries expressed by the number of arrivals).

4.6 PROBABILITY AS A FREQUENCY CONCEPT OR AS A GAUGE

Again | must expand on a few thoughts of a theoretical nature relating
to the concept of probability as it is used in my model of an urban
mechanism.
Probability (which is not taken here in the sense of a mathematical
77 Probability as a Frequency Concept

entity) is based on the calculation of the number of positive results

(successes) obtained in the course of a stated number of attempts (tests).
This calculation indicates the frequency of a particular event among a
limited number of events.
But another interpretation comes closer to my way of thinking. Let
us imagine a sequence of events. Given complete information on the
order in which they happen, | can always predict which events will
follow which (if, for example, there is a cause-and-effect relationship
among them). The probability that my prediction is accurate will have
a value of 1: my information is therefore good. On the other hand, if
my information about the order of events is poor, my predictions will
be much less sure than they were in the first case. Thus | will make my
predictions for a fixed number of test cases, and | will attribute a fre-
quency probability to them (the number of correct predictions in rela-
tion to the total number of predictions made).
Now let us suppose that two people observe the same sequence of
events. One of them has good information and the other has poor in-
formation. The first person’s predictions will have nothing to do with
probability, because a probability of 1 is equivalent to certainty.
The second person’s predictions will be ‘‘reliable’’ to a degree ex-
pressed by the proportion of his correct predictions to the total num-
ber of predictions he makes. This is a frequency probability used as a
gauge of information.
But even if a third observer (1, for example) know the exact extent of
the second person’s information, it will only allow me to count ona
certain degree of precision or error in his predictions, a degree propor-
tional to the extent of his information. And even if the third observer
also knows the order of events with as much certainty as the first ob-
server, it will not help him guess whether the second observer's next
prediction will be accurate or not.
The quality of the second observer’s information will not depend ab-
solutely on the original series of events but uniquely and exclusively on
his degree of knowledge.
In this sense the frequency probability is not a property of the events
whose frequency is the point of the calculation, but rather a property
78 Urban Mechanisms

of the information the observer possesses concerning those events, for

it really represents the frequency of his accurate predictions. This
scheme shows how we can use probability as a gauge of information.
This probability will always be characteristic of the way a person has
obtained his information (and, implicitly, it will also be characteristic
of his means of observation).
In our urban mechanism model, the concept of probability will be
used in this second sense: use weight is proportional to the probability
that a person will go to a given spot (given the possibility—or the diffi-
culty—of observing where he actually goes). This concept of probabil-
ity allows us to gauge the quality of information furnished by the ‘‘tool”’
(counting) used to observe movements in the city.
Movements in a city are to some extent similar to the Brownian move-
ment of particles in a liquid. This similarity does not derive from an
analogy between the systems themselves, but from the way we obtain
information about the two kinds of movement: in the case of Brown-
ian movement, one can observe neither the ‘‘motivations”’ nor the “‘iden-
tities’’ of the molecules. The mathematical model that we use is like the
one that describes Brownian movement. We call it a mode/ of probabi!-
ity potential. Thus the effort map, which demonstrates the distribu-
tion of local effort values throughout the city, constitutes a map of use
potential, describing the situations created by people about whose
movements and behavior | can obtain only very poor information.
The mathematical model which describes the potential of any point
in the city is actually a sum of products, in essence representing the
results of operations of the ‘“‘union of sets’”’ type, the only kind of oper-
ation allowed in objective descriptions (see Section 2.2).

4.7 THE EFFORT MAP: ITS CHARACTERISTICS AND USE

The effort map is not merely a mathematical construction showing the
characteristics of information obtained by an observer of the city. Its ap-
plications are very practical, for it communicates to us, in very simplified
form, the facts necessary to make a decision, thus giving a// the users of the
city the chance to participate in this decision, in a manner consistent with
the process of democratization which was my original premise.
79 The Effort Map

Like all mathematical tools, this effort map ‘‘creates’’ information.

It yields more facts than it was originally necessary to put into it
to obtain this body of new information. This phenomenon is ex-
plained by the fact that data introduced into the map belong to
two distinct and virtually unrelated sets: the physical network on
the one hand and on the other our knowledge about behavioral
characteristics expressed in numbers of arrivals (or weights). The
map achieves a synthesis of these two sets by placing one in rela-
tion to the other, in the same way that the city itself constitutes a
“real’’ synthesis of the corresponding ‘‘real’’ sets that are its com-
ponents, its physical framework and the behavior of its inhabitants.
Naturally, the physical framework is the least important of the two
sets (the physical city, after all, is nothing but a ruin if it is not in-
habited), but interventions (transformations) are easy to bring about
in physical constructions and almost impossible in behavior be-
cause of the great inertia it displays.
The effort map derived from urban mechanisms represents a ‘‘mul-
tiple warning,’ based on the same principle as the local effort matrix,
that | have established for simple configurations such as a linkage of
spaces. The effort matrix has become a warning system that makes pos-
sible a simple numerical (therefore objective) comparison of the results
of different uses of a single configuration, or from the identical use of
different configurations. It plays the same role as the effort map, with
two fundamental differences:
1. It involves a complex system of configurations (a network)
2. The chart keeps a// the users of a complex (inhabitants) informed
simultaneously.
Let us proceed to an imaginary application of this method, step by
step:
The first question must deal with the meaning of the local effort
value; it represents the essential characteristic of a given site (address)
within the complex. Instead of describing a place of residence as ‘’cen-
trally located,” ‘“‘outlying,”” or ‘on a main thoroughfare” (concepts
which have no significance unless they derive from experience, and con-
sequently cannot be defined before the complex is actually built), we
80 Urban Mechanisms

will use the new method, which allows us to build and rebuild (modify)
the model instantaneously, by reference to the projected complex.
Thanks to the simple numerical value that expresses effort we can im-
mediately define the ‘effort situation’ of an address as many times as
necessary and constantly compare two addresses with respect to their
effort situations, a comparison which would be impossible to make if
geographic location were the only known fact.
The numerical data supplied by this model are more precise than any
expressed in words and allow us to translate effort values into simple
images that we can all understand. For example, if site A in a city has
an effort value associated with a particular behavior of the inhabitants
equal to 105, and if a nearby site B has an effort value of only 98, the
situation can be explained to a layman in the following manner: there
will be 105/98 times more people passing through B than through A.
That might be worthwhile for a shopkeeper to know. Or again, it will
be 98/105 times easier to go anywhere leaving from B than from A.
That might be interesting for a doctor to know. Obviously, if it is in
the shopkeeper’s or doctor’s interest to live in a place of small effort
value, the same does not hold true for a scholar, a musician, or an in-
valid who needs to live in seclusion. He might prefer the A address with
its higher effort value.
The choice between two possible sites with high or low effort value
is a private decision, different for each individual.
Consequently, the second important fact to note about the effort map
is that it establishes no preconceived scale of preferences. Every decision
concerning preferences remains exclusively in the hands of the indivi-
dual: the model does nothing but provide an objective basis for com-
parison, established from available information about the city or the
complex.
The third characteristic of the map is tied in with the idea of compar-
ison: if we follow the model, there can be no isolated act or interven-
tion in a complex. Every decision—such as construction of a building,
a change in business hours, the building of a new road—has an imme-
diate effect on every point in the city. Every intervention of this type,
therefore, can modify the effort situation of every inhabitant in the
81 The Effort Map

city. The map offers the opportunity to communicate these changes in

effort value brought about by each isolated intervention to a// the in-
habitants of the city, at the same time. We saw an example of this kind
of information feedback to everyone other than the client who was
making the choice—in the case of the Flatwriter.
The effects of every individual decision should be transmitted to all
the individuals making up a community—instantaneously, insofar as
possible.
| think of the effort map as a sort of meteorological map that shows
the fluctuations of the urban mechanism, and | think that it is as much
a maiter of public interest to everyone as is the weather report. All the
applications cited here are based on this analogy with the weather map,
from which the effort map differs only in that the city mechanism can
be influenced by the acts and choices of the inhabitants (for example,
by popular vote or by a change in individual or collective behavior),
while the weather, at least so far, is not subject to human will.
Before studying a real application, | must make an important observa-
tion about urban mechanisms: this model shows us a relationship be-
tween use effort (a parameter) on one hand, and, on the other, a set of
observable data: the physical layout of the city (the network), including
the number of arrivals at all points of the network (use weight). The
effort parameter is a sum of products, and consequently (as we have
seen in Section 2.2) there is always a finite list of ways to arrive at a
given sum and therefore at a given effort value. Thus any one situation
on the effort map may be obtained in many different ways, and ina
single network which corresponds to a single effort map, there can
therefore by a number of distributions of use weights, all different. We
will consider these distributions as equivalent, even though they are
visually different.
Another observation, just as important: ‘‘weight” represents the num-
ber of inhabitants who arrive at a point; the reason for these arrivals
does not matter. So there is no difference between a change in the num-
ber of arrivals brought about by some physical intervention (construc-
tion or demolition of a building) and a change due to nonphysical in-
tervention (some sort of attraction or restriction); either factor can
82 Urban Mechanisms

allow a larger or smaller number of people to get to a particular place.

For example, doubling the hours a store stays open doubles its use
weight as surely as if the owners doubled the size of the building (ob-
viously a highly simplified example).
Thus a change in use weight can be brought about by nonphysical as
well as physical intervention; and a city mechanism undergoes impor-
tant variations each year. A city like Paris, for instance, has an entirely
different mechanism during the tourist season than it has during the
rest of the year (the locations of termini in the network and the num-
ber of arrivals at each of these sites changes completely, even if there
is no physical intervention such as street repair).
Thus the model of the urban mechanism gives us information of vital
concern to the inhabitant; thanks to it, he can monitor the intentions
of the architect, the planner, or the politician.
Let us summarize the essential objectives of the effort map by stating
a few of its basic principles:
1. Establish a parameter which expresses the intrinsic features of a real
city (or of any way of using a network).
2. Find this parameter exclusively by means of objective observation
treated in the only way allowable for objective descriptions.
3. Use the parameter as a basis for comparison between two different
cities (or two organizations), thus making possible a conclusion as to
whether the systems are equivalent.
4. Give the parameter a form that is easy to interpret (a map).
5. It should be possible to obtain the same configuration on the effort
chart through a set of different interventions (physical constructions
or intangibles such as a political decision).
6. Plan the map so it can provide information about the ‘‘use situation”
of the city simultaneously to a// the inhabitants.

4.8 A HISTORY DEPICTED BY URBAN MECHANISMS

| have defined history (in Section 3.6) as a sequence of events in chron-
ological order. This definition does not imply that there must be regu-
larity (periodicity) in history, and the sequence of events may well
depend only on chance.
83 A History Depicted by Urban Mechanisms

To find out whether a history has some regularity (whether it involves

a structure) or whether it is completely subject to chance (in which case
it could be defined as a history with a very specific type of structure)
it is necessary to describe its chronological sequence by using the terms
of a unique and observable parameter. For example, we can describe a
city’s history in economic terms, or outline it according to an indivi-
dual’s description of his personal impressions. The first will be the his-
tory of the city’s economy; the second will be the history of the indivi-
dual. History thus interpreted is the history of a parameter, that is, the
sequence of variations of the parameter-in chronological order.
If we are looking for an underlying regularity, we will begin our re-
search by seeing whether or not the sequence of variations in chrono-
logical order has some sort of structure. We will call this structure,
which is governed by the rule of sequential composition, a /aw of na-
ture. \f we cannot discover a law of nature in the sequence itself, we
can always find a relationship between two histories that are distinct
but dependent on the same chronology (time scale). In that case, the
law of nature that we find does not refer only to the ‘‘elements’’ whose
histories we have used to discover a relationship; the relationship itself
creates a new parameter which may have its own history.
We cannot discover a “‘history’’ without making use of observation as
long as its underlying law of nature has not been established, nor can
we start with the hypothesis that such a law must exist. On the con-
trary, our initial hypothesis will be that every history may be the result
of chance. But once a law has been found by observing history, and if
we repeat this observation enough times, we no longer need to observe
histories of this kind, since we can reconstruct them all using the law of
nature that we have established. Everything that | have said about using
a model therefore means trying to establish the existence of some un-
derlying regularity by showing the relationship between the observed
history of a given city and a history which can be constructed, that of
the corresponding generalized model.
This is how | have interpreted urban mechanisms: the effort map is
the model, and the complete list of all the possible states of this model
can be constructed. The actual city constitutes a similar, though much
84 Urban Mechanisms

more complex, mechanism. Still, even admitting that complexity, as

soon as we have reduced the representation of the city to what we
know about it (Section 4.3) the actual city mechanism can be effec-
tively described using the effort map.
Once we have constructed the complete list of all the possible states
of the model, we must make the list a carefully ordered sequence by
defining the operation which transforms one state of the mechanism
into another state as the law determining this order (Figure 27 and
Table 1).
In connection with this well-ordered list of all possible states of the
model (which in itself constitutes a constructed history), we can note
the observed history of the mechanism represented by the model, and
start looking for regularities in this observed history by taking the con-
structed history as the fie/d of variants in the observed history. (In the
example in Figure 27 there is asymmetry in line E, for the five first
terms, and the differences) “E,, OE — 1are5,5, —5,5, and so
on, which shows a certain type of regularity. Of course, there can be
other types of regularity, and one would have to compare a great num-
ber of sequences to find them.
The principal steps that interest us are these:
1. To construct all possible states of the mechanism model.
2. To code, for the list, all these possible states.
3. To observe, over a particular span of time, different states of the
real mechanism (the city), presented in chronological order.
4. To record this history using the terms of the code established for the
list of possible states of this model.
5. To observe regularities in the ‘‘real’’ history as mapped on the ‘‘con-
structed”’ history (the list of all possible states of the model).
The “real’’ application of these operations can be depicted by an ex-
ample: let us suppose that the number of arrivals (weights) in an exist-
ing city has three different brackets (1, 2, 3). For purposes of simplifi-
cation, the city’s capacity would be measured according to these brack-
ets; each bracket would represent its number multiplied by ten thou-
sand. Thus, bracket 1 signifies ten thousand, bracket 2, twenty thou-
sand, and bracket 3, thirty thousand inhabitants arriving at a given
85 A History Depicted by Urban Mechanisms

point in the city during a certain time span. If the total of the brackets
is 36 (in that case the city has a population of 360,000), for twelve
principle arrival points (numbered one through twelve), the distribu-
tion of weight in the city forms a finite list whose terms are
D;: 1, 15 13 14 15 1g 17 1g 19 149 144 142 (10,000 arrivals at all points
in the city),
Dg: 2, 19 13 14 15 1g 17 1g 19 149 144 149 (10,000 arrivals at all points
from 2 to 12; 20,000 arrivals at point 1 only), and so on to

26244: 31 32 33 34 35 3g 37 3g 3g 349 314 312-

This list of all possible weight distributions for this case (the list of all
possible states of the urban mechanism, necessarily containing all the
changes that can be effected by all conceivable physical and nonphysi-
cal interventions) determines the ‘’field’’ of states of the mechanism.
Every conceivable event in its ‘“history”’ that has some repercussion for
the mechanism (an epidemic, demographic shifts, a boom in tourism,
economic depression, reduction in working hours, introduction of cul-
tural attractions) is under some code number D,. in the complete list.
Everything, absolutely everything in this city of 360,000 inhabitants
and 12 districts (except the death of a large part of the population,
which cannot be recorded since there is no bracket of zero value), is
indexed under some term in this list of 26,244 terms, each representing
a well-defined state of the mechanism. This is a list of very modest size.
for a computer.
Naturally, during any real time period, only a part of the complete list
of possible historical events will show up as real history. Since the
urban mechanism method uses only directly observable data (networks
and arrivals), the observer can note in chronological order all events he
has observed and then interpreted according to these data, and each of
the events will have its corresponding term in the list. This is how a
“real history’’ can be represented within the field of the model list.
History thus interpreted in terms of variations in local effort (dod *w)
or global effort (Dee or 3 dw) will express ‘‘tendencies’”’ such as
a tendency for effort to increase or decrease or to fluctuate according
to acertain rhythm, or to approach some constant value, and so on.
Whether or not a tendency is desirable is, once again, a matter to be
86 Urban Mechanisms

decided privately by each inhabitant (he may express his opinion by

popular vote, for example); moreover, he can delegate this power to
elected or appointed representatives (which is done today). As soon as
all the inhabitants accept a tendency as desirable (for example, value
24 in the table preceding Figure 27), it is easy to determine which states
of the mechanism (thus which distributions of weight) can produce a
tendency that approximates the desired value. There will almost always
be more than one distribution of weight which will give the same value,
and consequently more than one way of obtaining it (in Figure 27,
steps 5, 7, and 11 all lead to a global effort of 24). Generally one of
these ways is easier to implement than the others, and the political
decision always lies in choosing this way. The political decision which
settles the question can be made by vote, on the condition that all
those concerned are aware of all the possibilities and their foreseeable
consequences. This information about the results of planning policies
(warning) is made possible by the model | have described.

4.9 THE UNPREDICTABILITY OF HISTORY

In the preceding section, | introducted the hypothesis that the history
of an organization—the history of a city, for example—could be a se-
quence of chance events. Before arriving at this statement, we had dis-
covered that probability could be interpreted as a gauge of information;
thus absolute chance (the vaguest probability) is equivalent to the
lowest degree of information (absolute uncertainty).
As | have previously defined it, a law of nature implies the possibility
of reconstructing the next state of a mechanism from a given state of
this same mechanism, assuming sufficient reliability of observation.
Chance, on the other hand, implies that such an operation is impossible
and that a given state of the mechanism can be followed by any other
possible state.
But even this principle of uncertainty, or chance, is based on regularities
such as laws, structures an abreviation system, since it states that there isa
mechanism, that is, a system (of haphazardly assembled elements) capable
of passing through a finite number of possible states. {| have excluded in-
finity from our study as impossible on a practical level; see Section 2.2).
87 Applying the Urban Mechanisms Method

My whole method is based on the hypothesis that every organiza-

tion in some way constitutes a mechanism, and that we can discover
the method of observation that suits each mechanism, whatever it
may be.
The present impossibility of predicting how this type of mechanism
will function is the result of along history of treating these problems
in a nonscientific manner: it was “‘intuitive’’ observation (which claimed
to observe what is unobservable because uncommunicated and uncom-
municable, like personal motivations or individual values); it was mak-
ing the assumption that a mechanism could pass through an infinite
number of states (mistakenly ignoring the fact that it is impossible to
construct an infinite number); and ‘‘intuitive’’ parameters (‘‘good,”’
“bad’’)—all these, which, disguising the personal values of the observer,
got us to the state where prediction is impossible. An unforeseeable
event is unforeseeable only before it happens; once it has occurred, it
seems that it could always have been predicted. In fact a rough predic-
tion would be quite possible if only we limited ourselves to less ambi-
tious predictions (which would not attempt to give a detailed picture)
and more objective ones (made always in reference to a complete list
of possibilities). In this chapter my goal is to arrive at a better way to
handle knowledge and foresight about organizations, especially cities.
If unpredictability is in practice only the result of a nonscientific ap-
proach to knowledge, and if improving this approach can give us accu-
rate predictions, then once we have made these predictions, we must
apply them. Now, a prediction put to use becomes a means of in-
fluence; those who have the advantage of accurate predictions are
more powerful than those who do not. That is why, throughout the
history of mankind, we find the will to keep knowledge secret, for the
exclusive use of a ruling class (priesthood).
Today, we are rapidly breaking away from this “segregation” of infor-
mation; but different pressure groups still try to maintain information
scarcity in all the professional fields.

4.10 APPLYING THE METHOD OF URBAN MECHANISMS

The potential application of this method has two motives. The first is
88 Urban Mechanisms

the trend toward democratization, of which | have spoken at some

length, and the second is the attempt to arrive at a simple and strategic
way of predicting what will happen in a city.
Of the two, the most important use of the method lies in the struggle
for democratization, quite similar in application to the process | de-
scribed in Chapter 3 for the Flatwriter. It seeks a solution to the prob-
lem of how to arrive at a group decision, or at group control, when all
inhabitants belong to the group making the decision but act as indivi-
duals, both in creating situations and in making decisions about them.
This schema can be summarized in a few words: the individual behavior
of each inhabitant produces a new state in the mechanism; its charac-
teristics are fed back to him very rapidly (in real time), allowing the
inhabitant to modify his behavior if he is led to make a different de-
cision on the basis of this feedback.
Let us assume that an individual or a specific group proposes some
actual intervention in the city (for example, the physical construction
of a new point of attraction or a legislative change such as an extension
of working hours). The new intervention is easily expressed as a new
use weight at one or more spots in the city (the number of arrivals
caused by the new intervention). The algorithm used to calculate all
the new effort values (local or global) is expressed by sums of products
obtained by adding the products of the distance covered times the
weights for all points so that no local effort value is unchanged even if
only one weight has been modified.
Local effort values (parameters that change with variation in weights,
or in itineraries) have no absolute significance. Their role is to be useful
in comparing two states of a single mechanism or of two different
mechanisms. So the only decisive factors are the differences between
local efforts referring to a single point. For example, if before interven-
tion two blocks of houses, A and B, had a difference of effort values
E, — Eg = 105 — 100 = 5, this meant that block A had a situation that
was 5 percent better than that of block B, with respect to use efficiency
for a person living in either one of the two blocks. If, after intervention,
the difference between the two local effort values has gone to 3, then
this means that A’s condition has deteriorated or that B’s has improved
89 Applying the Urban Mechanisms Method

by a value of 2.
Now every proposed intervention can be evaluated in this way, for
each block of houses in the city, by reading this intervention as varia-
tions in local effort. Naturally, these variations will be different for
each block, for the effort map does not necessarily change into one
that is isomorphic to it. As we have seen in Section 4.6, these variations
can be interpreted in layman’s terms, such as the number of people
Passing by a given block of houses (sector of the network) within the
city, or the accessibility of this block. Effort values are proportional to
effective commercial values or to land values, or to nuisance factors
such as congestion and street noise. If inhabitants throughout the city
can learn, through variations in the parameter of local effort (by which
they will recognize fluctuations in commercial value and any foresee-
able disturbances), about the consequences of an intervention for the
sectors in which they live, and thus for their own blocks, then they can
take a position supporting or opposing the proposed intervention,
according to their personal interests.
Thus we might imagine the “‘intervention permit” (like a regulation
governing working hours or a construction permit) put to a popular
vote whenever the proposed intervention was thought to bring about,
or would bring about, significant changes (changes can be defined
numerically according to the number of people involved). This vote
could accept or reject the proposed intervention, or it could impose
penalties proportional to the damages and inconveniences incurred
by the community, as well as by individuals. (This is the function of
warning the community that | described in the case of the Flatwriter.)
The ballot would be different for each block of houses (since the
variation in the effort value would be different in each block, even if it
were brought about by a single intervention.) This ballot could contain
the list of all the alternative weights or situations the proposed inter-
vention would produce. And, for each alternative, it would note the
effort variation brought about in each block of houses. All the inhabi-
tants would have the right to vote for one or more solutions, whatever
they thought would give them the most favorable effort situations for
their particular cases. Thus they would act according to their personal
90 Urban Mechanisms

EXAMPLE oF A REFERENDUM
ABOUT THE CITY

AN |NTERYVENTION
with WEIGHT
W, MAY HAPPEN ON ALTERNATNE
SITES 4,2,3 AND A.

FoR EAGI Homme ADORESS A

THROURH P TRE pee
sHOws THE VEN
PEE G ep aes THE VALUES OF THE ALTERATIONS
OF GACH ALTERNaTIVE. (S SXAMPLE
wR Sex SRecceD AG ATTRACTION"
Ale E,- ©,
CININ@ THE NUNBE® OF PEOPLE
WHO pil RASS AODRESS A IF
OUT CN SITE 4.
k MR X, A SHOPKEEPER |VANTS CLIENTS-
HE WAS HIS SHoP IN A, AMD HE ULL
THERE FORE voTEe,iN CRDER oF pRE-
FERENCE, FoR ALTERNATIVE 4,
THEN ALTERNATIVE © on 3.
MRY, AN INVALID, VING IN. A, WANG
26S NCISE
HE wil yore Fou
ALTEQNATIVES 2 Of 4, BUT HE CAN
ACCEPT © QR Z AS WELU.-

Figure 48

preferences, but after having been properly informed of all the possible
consequences. Instead of being determined by a simple majority, an
evaluation of the results would choose the intervention that injured the
smallest number of people or that improved the situation of the largest
number of people.
This is one of the aspects of urban democratization.
Parallel to the actual vote, which involves calculating weights on the
basis of a rather long reference period (number of weekly, monthly,
or even annual arrivals) there is another, more up-to-date technique
that is based on continuous counting. Arrivals are counted around the
clock, and each arrival at each spot is immediately transmitted to a cen-
tral computer which just as quickly calculates a readjusted effort map
of the city. The effort values calculated can be transmitted instantan-
eously to any point in the city, for example by means of municipal
machinery which | will call c/ty barometers. At every street corner,
such a device would show the effort fluctuations at that spot at that
precise moment, something like the clock-thermometers that you see
91 Applying the Urban Mechanisms Method

on billboards. People would become as accustomed to checking for

effort fluctuations in their cities as they are to checking the weather.
Providing the user with instantaneous feedback on fluctuations in the
city has another consequence: if a user of the city can immediately
find out the probable result of something he plans to do (take a certain
road, stop at a certain place) then he can modify his initial intention if
he wants to, after he learns about current fluctuations (a traffic jam in
that direction, for example).
The implication of this is that the mechanism is self-adjusting and
may tend toward a stable state. But thatvis only a hypothesis. In any
case, once the information is sent back to the user, the mechanism will
be different from what it was initially. If we could confirm the hypo-
thesis that the mechanism tends toward a stable state (we would then
have a Markovian process) it would probably be the most important
contemporary discovery in the fields of architecture and planning, but
we cannot establish its validity without a real experiment. | am less in-
terested in defending this hypothesis than | am in suggesting the set of
tools we would need for the experiment.
Another application which promises to be of scientific interest (as
well as of practical value) is the use of signals for urban traffic. These
signals, which would show each person his chances of continuing on his
way by his chosen itinerary, would be installed far enough away from
detour points so that he would have sufficient time to make a decision.
A simple color code, for example, would show traffic jams on all seg-
ments of the road leaving from a specific detour point. Naturally, even
one car setting out on a segment would immediately change this rate,
and the next driver to come along would find a different distribution
of expectation rates on the signal. Having been warned, some drivers
would stick to their original routes, but others would change their
minds. The functioning of these signals would depend on the same
central effort map that we used in the preceding example, and each
driver’s decision would influence the central effort map by modifying
the traffic.
There is a temptation to see this fluctuating readjustment as a process
where game theory could easily be applied. Personally, | do not think
92 Urban Mechanisms

that game theory can be applied to either planning or automobile traf-

fic. For a game to be considered mathematical, there must be strictly
established rules. There is nothing like that in the process of using a
city, probably because we can obtain only very poor information a-
bout the individual, his motivations, and his behavior. Taking the pro-
cess as a whole, there are neither winners nor losers in the use of a city.
There is no game of competition, no exclusion of opponents from a
“territory,”’ and ‘‘individual strategies’’ do not affect the use of the
city but in general more limited personal objectives; and whatever the
objectives of a real, live user of the city, only he can know them in any
matter that concerns him.
Thus, the implementation of the urban mechanisms model provides
the user with immediate information about himself and other users.
He must make his own decisions, then and there, and not depend on
the advice of a planner. This is useful information, obviously, since it
helps us to understand what is really happening ina city.
CHAPTER 5 GENERAL APPLICATIONS OF THE
METHOD

5.1 LOOKING FOR LAWS OF NATURE IN HISTORIES

At the beginning of this book, we stated the principle that a science
should be compelled to use abbreviations that everyone can understand.
Obviously, these abbreviations can be used only in systems where cer-
tain repetitions are apparent (repetitions of elements, of operations,
and so on). These repetitions allow for the elaboration of rules: we
can call them rules (regularities) as well as ‘“‘laws of nature,”’ but they
are not necessarily the properties of the thing observed (of nature); in
fact, they are chiefly characteristic of the observer.
We also noted, in the preceding chapters, that the profession of archi-
tect or city planner, as presently practiced, is absolutely nonscientific;
it recognizes mere rules of thumb, instead of demanding rigorous reg-
ularities. We have seen that this is a grievous state of affairs, for the
new task that has fallen to this profession requires that it become
scientific.
The question that we shall ask ourselves in this chapter is: how may
potential regularities be discovered?
At present, other sciences, such as physics and biology, proceed by
experiment in order to discover regularities. Experimentation means
(as we saw in Chapter 2) the ‘‘repeatable”’ carrying out of a strictly
defined sequence of operations; anyone must be able to perform it,
and as often as desired; moreover, it must lead to the same result with
each repetition; if not, we must conclude that the operational sequence
is wrongly described, or else that it has been applied incorrectly.
In order to establish the ‘‘validity’’ of experiments in general, they
must be repeated a great number of times and, if possible, with varia-
94 General Applications

tions within the operational sequence. Therefore, it is difficult to apply

experimentation to disciplines where processes unfold slowly, at
great cost and risk, or even where the conditions prescribed by the
operational sequence cannot be created at will. Now the process of
constructing a building or a city is both long and costly, and therefore
the series of necessary experiments cannot be easily performed.
In similar cases, we have to use ‘‘models”’ or ‘‘imaginary’’ experiments.
The models are representations (mappings, with one-to-one relation-
ships)* of a reality, reproducing all the properties that have signifi-
cance for the observer. The model that we shall use to detect “laws of
nature” in architecture and city planning will be the one we used in
the preceding chapters (on models of urban mechanisms).
How would we conduct experiments with these models? We repeat
the operations upon them as often as we wish, each time introducing
one or several supplementary operations into the original sequence,
and observing the changes that this produces in the results. Each of
these experiments (carried out on models or even in some other way)
represents an element of a “‘history,’”’ and a series of consecutive ex-
periments represents the ‘’history”’ itself.
The exhaustive (complete) list of a// possible results (states of the
model) constitutes the ‘‘field’’ of a// the histories that may be pro-
duced for a given mechanism (a model or a real mechanism). The ‘’real
histories’’ observed can thus be represented by ‘’constructed histories”’
(the histories of the model), within the ‘‘field’’ of all possible histories.
If this field itself is constructed in a strictly ordered way, a history,
or set of histories, may present its own regularities (which can be math-
ematically formulated) within the field. (As we have seen, chance it-
self constitutes a regularity: it deals with an absence of relations among
any three arbitrarily chosen events A, B, and C; that is, the character-
istics which differentiate a pair formed by two of these events cannot
be found in the two other pairs that can be formed from the three ele-
ments. If, then, AB, BC and AC have no common characteristics, sets

*“One to one relationship’’ means that to each term of a set there corresponds
a well-defined term in the map of that set, and that to every element in the map
there also corresponds a well-defined element in the original set.
95 Types of Human Behavior

A, 8 and C constitute a fortuitous grouping (one due to chance.)

In order to detect regularities in this manner, | propose to follow the
following five stages:
1. Construction of the field of reference
2. Actual observation of a group of histories
3. ‘Mapping’ the observed histories within the field of reference
4. Establishing the common characteristics of the histories represented
5. “Decoding” the observed regularities, by transposing them from
represented histories into histories that cam be observed in actuality:
these regularities are the laws of nature that we are looking for.
| am going to follow this program in the following sections.

5.2 CONSTRUCTING THE FIELD OF REFERENCE: TYPES OF

HUMAN BEHAVIOR BASED ON ABSTRACT CATEGORIES
Let us imagine that | wish to describe human behavior with the help
of a complete list. To do this, | shall assume that certain categories of
activities exist, whatever their nature (at the present time, architects
use—with the cooperation of sociologists—categories of ‘‘work,”’ “‘lei-
sure,’ and “‘supply’’). The individual behavior of any particular per-
son can be described when we succeed in finding how often he per-
forms activities belonging to one or the other of these categories,
expressed, for example, by the fact of his ‘‘going to such a place”’ or
devoting ‘’so much time” to that activity; we thus establish the fre-
guency of an activity during a given time of reference. | could, for
example, describe my own type of behavior in the form of W:6, L:7,
S:2 and that of my wife in the form of W:5, L:7,S:1, and so on.
Naturally, we could construct a finite list, comprising all the types
of possible behavior, established according to these three categories
of arbitrarily chosen activities. The type of behavior that would begin
the list would take the form 0-0-0, describing the behavior.of a person
having absolutely no activity whatsoever (an invalid, for example),
and the last term of the list would be behavior 7-7-7 (that of a hyper-
active person), taking a week as the time of reference, and counting
activities by simply noting whether or not they have been pursued on
a given day, without mentioning the number of times they have oc-
96 General Applications

curred on that day.

This complete list of types of behavior, constructed from the hypoth-
eses mentioned (8 frequencies and 3 categories of activities) would
comprise 8° = 512 different terms, based on the model:

We Ss
B, Ome Onan c
Bo oeentep eae
B3 Oma)
B, oO 0

[EO a a
This combinatorial list contains any possible type of individual be-
havior, in the case of a description based on the categories ‘‘work,”’
“leisure,” and ‘‘supply,’’ and for a weekly frequency of activity.
Of course, these categories were chosen rather hastily, and | am using
them only because they are familiar. But one might very well choose
them on any distinctive criterion, as, for example, ‘‘variations in blood
pressure.’” One would then have category
A, activities having no in-
fluence on blood pressure; category &, activities lowering blood pres-
sure; and category C, activities increasing it. A list composed as a func-
tion of these categories A, 8, and C would be complete when it ap-
peared in the same exact form as the preceding one, and there would
be no activity pattern imaginable that could not be found somewhere
on the list, since all activity belongs to one of the categories, A, B, or
C, and can be carried out with different weekly frequencies, just as in
the preceding list. Thus these two lists equally well represent all pos-
sible types of behavior, provided that the number of frequencies and
categories is the same, although they are based on different categories
and both include 512 numerical expressions, which are exactly the
same in both lists, and noted in the same order.
We would be able to construct such a list on the basis of any other
imaginable categories, and naturally, the list would always be the
same. The difference between these lists would lie in the fact that the
97 Types of Human Behavior

LET A,® AND C BE CBSERVARLE CATECORIEg CF ACTIVITIES,

In WHICH
A can Be WORK" og SOMETHING THAT INCREASES Blood PRESGIOE
Soa » BUYING’, OR \ SOMETHING REBUCING Blood PREC RE"
Gar: «| LEISURE”, OR ~ SOMETHING HOLDING BLood RESURE STEADY”
WHATEVER MEANING IS AgsiGNen to CATECE
EVERY IMAGINABLE TYPE OF BEHAVIOR, WiLLR BE non i
THE COMBINATORIAL UST BELOW.
THIS UST SHOWS, FOR EACH BEHAVIOR, TrPE, How OFTEN
ALTIVITY A,B om c WS PERFORMED BURING 4 REFERENCE
PERLOBE.
ACT IVITIES

ae lokeye)
BEHAVIOR a Coe Sol
ee Gq OR C, : Ole

a 77
Ss
at Sta
HERE FREQUENCES OF
PRRFORMING AN ACTIVITY
RANGE RETWEEN © AND 7
Figure 49

actual observation and the ‘‘proportion of occurrence” of each of the

types of behavior would vary according to the categories chosen.
We shall call this list of 512 terms, each containing three symbols
chosen from among the figures 0 to 7, a complete list of types of be-
havior based on three abstract categories of activities: in other words,
such a list can very well constitute the ‘‘field’’ of any coherently ob-
served “‘history.”
If types of individual behavior expressed as a function of three ab-
stract categories, and observed according to their weekly frequency,
involve constructing a complete list of 512 terms, we must make a far
greater list when dealing with collective behavior. Let us consider three
abstract categories, 1,2, and 3, making up a similar list; if, within a
community, we decide to consider three fractions as significant, each
comprising a third of the total population of that community, we shall
have, as a complete list of the types of collective behavior, a list estab-
98 General Applications

lished according to the weekly frequency with which the abstract acti-
vities 1,2, and 3 are carried out:

Categories of Activity Fractions of the Population

1 2 $3 WS Woes
B, 0 0 0 B, 000 000 O00
Bo 0 6) 1 By 000 O00 O01
Be 0 1 0 Bz 000 O00 O10

5 RE ERE en ee eR: 8512 000 O00 777

8512 7 7 7 8513 000 001 O00
Types of individual BIPECU CREE: CoTIPO ULWCRC CMR OCMCR Cane ee. EPO

Types of collective behavior

in fact, these types of collective behavior would give us an enormous

list, but one which we could maintain within the limits of a computer's
capacity by a judicious choice of the number of categories of activities,
of maximal frequency in the time of reference, and of significant frac-
tions. Thus, we could construct the list of all possible collective activi-
ties, with reference to abstract categories of activities.
Aside from the fact that it is complete, this list is also well ordered,
since its rules of composition (combinatorial) permit the determina-
tion, for each term on the list, of the one to follow. The list is thus
well constructed according to the rules.

5.3 CONSTRUCTING THE FIELD OF DISTRIBUTIONS IN SPACE

(CONFIGURATIONS)
Next, for the experimentation that is the subject of this chapter, we
need a second list of a// the possible physical distributions of theaters
of activity, within a// the possible networks of 7 points, and therefore,
to use the terminology of the preceding chapter, the list of all possible
mechanisms.
How shall we construct this list? First of all, we know from the pre-
ceding chapters that any given network of 7 points can be represented
(mapped) by a binary sequence of (n — 1) signs (see Section 2.5).
99 Distributions in Space

Consequently, in constructing a combinatorial list of all sequences of

this type (with the exception of isomorphic sequences), we shall ob-
tain the combinatorial list of all networks of 7 points. This list will
have at most 27/2 (7-1) terms.
Moreover, we saw earlier (Section 3.3) that ‘‘use weights’’ could be
included in such sequences (representing a linkage scheme and there-
fore a network) in the form of “‘labels’’ placed, for example, above
each section representing a given point in the linkage. This means that
we can increase the list of binary combinations and insert into each
binary expression (and thus into each linkage scheme) all the ‘‘weight
distributions” (and therefore all the labelings) possible in a given net-
work. If, therefore, for the sake of convenience, we consider k ‘‘de-
grees of weight” (w,, having the vaiue 1,2, 3. ..k) to be significant,
then the possible weight distributions in the network of 7 points can-
not exceed the value: C = 2/2-1) « «”.

A UST OF AL 7¥PES of COWECTIVE BEHAVIOR B.,

BASED ON OBSERVARLE ACTIVITY PES A,B AND C, 4NB
ON FREQUENCIES with Witicd ey ARE BE cee ae
RANGING A THROUGH 7 (REFERENCE PERIOD: A WEE).
WE COLECTIVITY TO WHiey THE UST REFERS IS
CONSIDERED AS CONTAINING B BRACKETS , EACH
WAVING ITS CHARACTERIOTIC BEHAVIOR PATTERN.
BRACHETSS ARs SUPPCGED To HANE THE SAME SRE.
BaAckets

1 277\777) 777
THE ComPLETE UST CONTAINS ower ISd MILLION
POSSIBLE “TYPES OF COLLECTIVE REHAVION,
AND THERE IS No COLETTIVNE BEHayion IMAGINABLE
OA DBESCAIBARLE IN TERMS OF ORCERVARLE ACTIVITY
CATEGOMIES
THIc. List. A.B,
a) AND C THAT rT IS NOT INCLUDE=D IN

Figure 50
100 General Applications

EVEAY CONFIGURATION (DISTRIBUTION PATTERN OF

ARRIVAL NUMBERS — WEIGHTS— IN A NETWORK) D;,
CAN RE DESCRIRED BY THE ADJACENCY SEQUENG OF
THE NETWORK WITH EACH SECTOR LABELED WITh THE
WEIGHTY OF THE PoINT THE Secron REPRESENTS.
A) B@ ¢@) de)

THK CONFIGURATION
CAN BE DE&AaRIBAED
LG) BY THE Se€Quewce
(Fon EXAMPLE)
N@ oft) PG) a
@Molooloo.- (3) loloeloe.-.-© @) obobolo...0-(6) ete
FT SCQens ’
og BY
loo
(4) otto (3)om CBN (Zool... .- (6)
THUS IT \S POSSIBLE TH CONST@Ler —HeE COMPLETE UST
OF ALL POSSIBLE CONFIGURATIONS CONTAINING W Pants.
Figure 51

This list, like the preceding one (Section 5.2) can easily be construc-
ted by judiciously adjusting the values of n and k to the possibilities
of today’s computers.

5.4 THE REFERENCE MATRIX: A DEFINITIVE INSTRUMENT

FOR THE CITY PLANNING LABORATORY
Thus, in our preceding discussion, we have covered all considerations
for constructing two complete lists (the field of reference, as we called
it in Section 5.1):
a. The list of all possible types of behavior, based upon any types of
activities belonging to categories A, B, and C.
b. The list of all possible weight distributions in all possible networks
supported by 7 points.
It is clear that if we can form pairs that include one term from each
of these lists, such pairs will constitute the precise description of a
city, both as a mechanism and as a social entity.
101 The Reference Matrix

In order to form these pairs, we must add a restriction in advance: this

city has a definite population. The sum of the weights (number of
arrivals), which is necessarily a multiple of the total number of inhab-
itants and which depends on the number of units in the time of refer-
ence, and the sum of all the fractions of the population carrying ona
given activity (a sum which necessarily corresponds to the population)
must be coordinate.
If we call/ the sum of the frequencies of one type of collective be-
havior within the time span referred to; h, the number of categories of
activities; 6, the number of fractions; and p, the number of persons in-
cluded in a fraction, we shall be able to establish the following con-
straint:

sed fh)

iw? Hage =U apa

z=

This constraint assures us that the two descriptions of the same city
(description as a configuration and as collective behavior) are not con-
tradictory (that is, there is a corresponding pair that includes a term
from each of the two lists).
Thus, for each pair, there will be a specific value p, which can be
calculated. If we assign a fixed value to p, there will be terms in the
two lists which together cannot form a ‘‘type of city.’’ Therefore, we
shall necessarily select all the pairs which, satisfying this constraint, can
be formed using terms from the two lists.
Having thus constructed all the possible pairs of urban mechanisms
and corresponding types of behavior, we can obviously establish for
each mechanism the ‘‘global effort’’ related to it (Section 3.3):

Mill eon

swe Dien (1)

x=1 2z=1

Without exception, the following condition must be filled for all these
mechanisms:
102 General Applications

rota

Yo w,=piichsb (2)
z=1

Inevitably (the conditions imply it) every mechanism admitted is likely

to form pairs with a great number of types of behavior, and, inversely,
every type of behavior can be associated with a great number of mech-
anisms. We can represent all the mechanisms and all the types of be-
havior in the form of a matrix, putting, for example, all the types of
behavior in the principal row of this matrix, well-ordered, since they
result from combinatorial operations, and putting in the principle
column of the matrix (Section 5.2) all the ‘‘“mechanisms” (the distribu-
tions of weights in space) in the order stemming from the combinatorial
operations that produced them (Section 5.3).
Within the field of this matrix (the field of reference) we can note a
zero for any pair (mechanism and type of behavior) which does not
satisfy condition (2), and inscribe the appropriate vaiue of e (1) for the
pairs that fulfill it. Thus there will be, for each term representing a con-
figuration, a row containing values of e in positions (columns) corres-
ponding to the types of behavior accepted as pairing off with this parti-
cular configuration (mechanism) since they fulfill condition (2), and
zeroes in the positions belonging to the columns for types of behavior
not accepted as forming these pairs. Inversely, in each column of the
matrix corresponding to a given ‘‘type of behavior,” there will be an
appropriate value of e or a zero in the positions (rows) representing a
configuration (a mechanism) fulfilling or not fulfilling condition (2)
for this particular type of behavior.
This matrix is thus a simultaneous representation of all possible mech-
anisms and all the possible behaviors whether they exist or not, as well
as representations of all possible global efforts (e) which characterize
every urban organization resulting from a particular type of behavior
within a given structure,
Next, this matrix reconstructs in a practical manner the field of refer-
ence of all possible histories of all possible organizations for cities.
These histories may then be represented through
103 The Reference Matrix

THE TWO COMPLETE LISTS:

A) oF Au comectve BEMavion 8B;
2) OF ALL POSSIBLE CONFIGURATIONS OF W pants D;
MAILE IT POSSIBLE TO CONSTALCT A NaTRIX, WHICH,
= every ii 0; B; SATISFYING THe BQUATION
wy Fp). weHERe 2+ THE NUMBER
oF
ya J ) “i IN THE NEwoRE
W,* WEIGHT OF Point y
p = NUMBER
OF People
. IN BAAcKeT

"bs SE ALoon
= YM OF FREQUENCIES

b= THE
Ay psu laser or

CONTAINS A GL0BAL EFFORT VAL & ee THE FORM

ei % 2; é Any Wy
THS GLOBAL EFFORT MATRUX, Basen ON COMPLETE USTS OF
EON EAGURETIONS pr seed COUSECTIVE BEHANIOn -ryPES,
'S LABORATORY
TOOK
Figure 52

THIS TeoL wiLL Leck UKE THIS:

ON PLETE oF
CSCS pC HAVION Treet

D,
oe} 2,
fo] >
SEI 3
ai :
pg} :
Po}

aol:
5
Bae 4 he
S$ S|;
e

D..

Figure 53
104 General Applications

1. variations of weight distributions and configurations of infrastruc-

tures (mechanisms)
2. variations of types of behavior (distributions of activity of the total
population of the city)
3. variations in the global effort parameter, which is an intrinsic and
structural characteristic—however ili-defined—of a city.
The advantage of the crude values of the global effort parameter, in
relation to the more detailed ones of local effort, lies in the fact that
the number of possible cases, even choosing values as low as possible
for n and k, would vastly exceed the capacity of a computer, if it had
to store all the values of the local efforts in its memory. Therefore, it is
only by calculating the local effort values solely as a ‘‘foregettable’’ sub-
program, allowing for the calculation of global effort, that we can exe-
cute this program and remain within reasonable limits insofar as a com-
puter’s capacity is concerned.

5.5 POSSIBLE INTERRELATIONS AMONG HISTORIES

The practical importance of our matrix as a laboratory instrument
comes from the opportunity it offers for comparing histories, be they
abstract or observed.
Let us imagine, for example, a city organization corresponding to a
well-defined pair in the matrix, on a given date in the history of that
organization. At a later date, it will correspond to another pair in the
matrix, later still to another, and so on. By its variations, the ‘‘actual
history” will follow the plotting of three histories within the matrix
(three histories mathematically independent of each other). These three
histories will be the history of the configuration (vertical component),
the history of behavior (horizontal component), and the history of
global effort (the plotted curve itself). These three histories will all be
accurately represented /n the same system of reference.
We cannot guess in advance whether or not a regularity exists between
two or three of these histories that may constitute a law of nature; but
if one does exist, this matrix seems to be the best instrument for allow-
ing us to detect it.
105 Interrelations Among Histories

What facilitates the detection of a law, first of all, is the possibility of

plotting, within the matrix, the curve of a relatively long history for a
great number of existing city organizations simultaneously. Statistical
records allow us in effect to reconstitute the history of most cities over
a period of about fifty years, given that the numerical values that we
1
need for this research are very ‘‘global’’ (approximate). By representing
the tripartite history of all these city organizations, we shall obtain cer-
tain curves on the matrix, and the appearance of a regularity on one or
several of these curves can be a key to the discovery of a ‘‘law of nature.”
If, on the other hand, we did not find any law of this sort, that would
mean that city organizations (taking into account that our knowledge
of them is limited) are not subject to laws of the determinist type, but
that we must look for laws of chance. a
For better legibility we shall represent this matrix not in the form of
a simple numerical table, but, as we did in the case of effort maps,
graphically by means of a chart of ‘‘iso-global effort”’ lines which will
show the regions in which values for e belong to the same “‘bracket”’
(the same fraction).
The first regularity (or absence of regularity) will be distribution of
the lines of iso-global effort themselves. This regularity cannot be fore-
seen, Or even assumed, until the matrix is constructed in its entirety.
For if the iso-global effort lines in the local effort maps follow, in a pre-
dictable manner, the example of other models of the ‘‘distribution of
[i
potential’ type, by contrast the iso-global effort curves of our matrix
will not correspond to any known real model. We would therefore have
to look for some surprises.
{t may be that the iso-global effort lines will show the regularity or
the nonregularity of the e parameter. But we can also detect other regu-
larities within the two ‘‘complete lists’” themselves: first of all, equi-
valencies, that is, the same ‘‘configuration’”’ functioning in the same
manner (and thus having the same value for e) with very different types
of behavior, or even the same type of behavior functioning in the same
manner (the same value for e) within different physical configurations.
To carry this experimentation further, we can represent the history of
106 General Applications

THIS Too. CAN Show ‘THE HISTORY OF

SEVERAL CTIES §=CEXAMPLE)
PArgiggg Nuv.c. (A6q
2
° e

> :
° <
ry e
°
.

A WiSToRy OF
VOAIPLYA
WISTORY oF
WHAT Aas THE S PARIS
UNDERLYING RECULARITIES 2 C WISTORY om
New YorxX Ty
Figure 54

cities in the matrix; then, each time that the curve of a history shows
certain characteristic traits (peaks, inflections, and so on) seek out the
corresponding events in the actual history of the cities. We would thus
be able to discover whether or not certain characteristics of the history
of the mechanism correspond to certain types of events rather than to
others; however, the discovery of such relationships would be more a
surprise to me than an anticipated fact, for | am very suspicious of the
sort of intuitive explanations by which events are generally interpreted.
Therefore, we see that the construction of this matrix could be the
appropriate tool for a large number of experiments with models, and
that it would certainly lead us to a better understanding of the ‘‘natural
laws”’ of city planning, if such laws really exist.
Naturally, in addition to functioning as a tool for experimentation,
this matrix could also be used as a tool for planning. Let us suppose
that a city planner, working for a given society that has a definite type
of behavior, has to propose a certain number of planning alternatives.
107 Categories of Activities

The type of behavior of this society can be easily coded by an observer

who will note, in categories A, B, and C, the frequency with which cer-
tain activities, decided on in advance, are pursued. Once the type of
behavior has been defined, the city planner has only to look at the
matrix for all the configurations (mechanisms) having a value e of
global effort less than or equal to a threshold value arbitrarily chosen
and considered to be desirable. Thus the matrix—unlike ‘‘optimization”
models—provides an answer by selecting a set of equivalent solutions
(equivalent as to the global effort value that characterizes the mech-
anism), instead of a single and unique “‘best’’ solution.
This is how the matrix can be put to use by the “physical planner”’
(the urban planner, the builder). But there is asymmetrical path open
to the ‘‘social planner’: in his case, the physical configuration (the
mechanism) constitutes the observable datum (the network and the dis-
tribution of termini within). Once he has observed the existing config-
uration and situated it in the matrix, the social planner can search this
matrix for all types of behavior not carrying the sign O which, when
paired with the configuration in question, give a value e of global effort
less than or equal to a value arbitrarily determined in advance as a de-
sirable threshold.
Thus, this “‘tool’’ makes it possible to find, depending on the need at
hand, a set of solutions applicable to a given situation, with perfectly
predictable effects: either material configurations to be applied to a
particular social situation, or social improvements to be made in a given
physical context.

5.6 POSSIBLE CATEGORIES OF OBSERVABLE ACTIVITIES AND

THEIR INTERPRETATIONS
The matrix defined above could be an indispensable tool for another
kind of research: research on significant categories of activities.
As | showed earlier, the matrix constitutes a ““mapping field’”’ for pos-
sible relationships between the distributions of termini (mechanisms)
and types of behavior as defined by certain categories of activities.
The distributions of termini within a network are “hard” (concrete)
data, which anyone can observe and, as the terminus characteristics
108 General Applications

contained in these data account only for the situation of a terminus in

the network and its weight of attraction, these observations will be the
same whoever the observer happens to be.
Types of behavior are also hard data: the frequency of activity plotted
during a fixed time of reference can-be observed by anyone in the same
manner.
The only ‘‘soft’’ element (as in ‘‘software’’), and therefore the only
one that is arbitrarily chosen, is the set of ‘‘distinctive signs’’ by which
activities are taken as belonging to the same or to different categories.
Indeed, the matrix, which operates on abstract activities A, B, and C,
never has any direct connection with these signs.
Moreover, in our matrix, all the histories based on categories of activi-
ties A, B, and C, whatever the distinctive signs under which they are
observed, can be plotted simultaneously. Certain of them will show
regularities, others will not. For example, if we give the meanings of
“work,” “leisure,” and ‘‘supply’’ to categories A, B, and C, or if we
LT Hh
translate them by ‘activities raising blood pressure,”’ ‘‘not changing
blood pressure,” or ‘‘decreasing blood pressure,”’ the frequency distri-
butions, and consequently the histories obtained and their eventual
regularities, will doubtless be very different.
One of the problems of sociology lies precisely in this tacitly arbitrary
choice of categories of activity (work, leisure, supply, for example). |
do not see how a mere spectator could recognize, when an ‘‘observed”’
person pursues an activity, to what category this activity belongs, and
| do not even see that these categories are complementary (and conse-
quently that the list is complete). These categories are, therefore, a
completely arbitrary set, often the results of the semantic abuses of a
given civilization, and they are certainly not suited to projecting solu-
tions, or to the study of solutions in changing society. (Moreover, these
semantic preconceptions are one reason for extreme intolerance in every
society).
In any case, in the matrix under discussion, we can represent the
image of an observed history, with the aid of several constructed (map-
ped) histories, whose interpretation is based on different categories of
activities, defined by different distinctive signs. The ‘‘historical curves’’
109 Categories of Activities

which will result, even if they are different, will certainly have at least
one point in common: they will all refer to exactly the same sequence
of “‘actual’’ observed events, which they all represent, by means of
different modes of observation. One could certainly discover a regu-
larity among the different curves themselves, or between those curves
and the general distribution of values for e (global effort), or among any
sets of data in the matrix.
\f it were possible to detect certain of these relationships (regularities),
we would have a fundamental “‘law of nature’ for sociology, a law of
which we are yet unaware, and for which we may have great need.
Thus, for the sociologist (or for the specialist in group dynamics), the
matrix proposed here offers the hope of finally having an experimental
tool.
If, in the primitive state of affairs which presently obtains, the city
planner hopes for the aid of the sociologist in finding a way out of his
critical situation, in the future it may well be the city planner who
rescues the sociologist by providing him with both the tools and the
observations.
CHAPTER 6 THE CITY

6.1 THE PRIVATE CITY

| live in Paris, which is a very large city of eight million inhabitants. Of
course, to say that | live in Paris does not mean that | come into daily
or even yearly contact with all these people. In reality | encounter in
the city about two thousand persons in all. Among these, there are a-
bout twenty-five that | see every week, fifty that | see every two weeks,
and so on. The set of people whom | actually encounter, with varying
regularity, constitutes what | call my ‘private city.’’ So in effect | do
not live in Paris, but in my “‘private city’; and | would only be aware
of people leaving Paris if they belonged to this ‘‘private city.”
There are eight million of these private cities in Paris, since each in-
habitant has his own. Each “‘private city’’ represents a linkage scheme,
that is, the graph of the movements of an inhabitant making use of the
real infrastructure of the real city. This implies the existence of a great
number of crossovers, about 8,000,000%, or 4 X 1027. Since each cross-
over represents a ‘‘source of potential conflict,’’ such a number signifies
that | am living on a volcano!
The number of potential conflicts resulting from the peaceful coexis-
tence of a very large number of private cities is far too great for any
presently existing computer to manipulate. The problem of knowning
how a city functions is nowhere near solved, even in its simplest form,
and, except for the general application described in the previous chap-
ter, we must rule out the method of using a complete list, because of
the dimensions it may assume.
It is obvious that this complexity should prevent the observer (or the
city planner or the architect) from bringing to the problem a solution
 eo The Private City

THe REAL CITY 15 AN INFRASTRUCTURE

COUTAINING A PAlvaTe UTy FOR GACH INHABITANT

MR x’S$ PRIVATE ary

CONTAINS AL THE PEOPLE
MR xX vVisirs
DUAINS A PeRnion oF
REFERENCE
Figure 55 a

THe QeEAL CITY IS AN INEXTRICABLE

NETWORK OF ARIVATE CUTIES.
E oF THESE PRIVATE CIES ¢
i VERY IMPORTANT TO ITS wCREATORE

e A CRSSN SS. (IN A CIty

\ BE INTERPRETED
A SOVRCE OF POTENTEL
ConrucT
EN A Clty OF THE SIZE
& PARUS G MILLION) /NHAB
CREATES 4.407" conFucr
SOvUACES.
THAT IS A VERY Lance UOMBen ,
Figure 56
149) The City

based on a system of his own private values and affirming that such and
such a “‘private city’’ is preferable to all the others (the optimization
method). Now, this manner of judging the ‘‘value” of different private
cities (that is, different modes of living) is one of the gravest abuses,
even if unconscious, that a city planner commits, since it is a matter of
individual opinions, where no comparisons are possible.
However, if the city planner absolutely has not, and cannot have, any
moral right to make decisions concerning someone else’s private city,
he can still act as a ‘‘counsellor.’’ The model of the urban mechanism,
described in the preceding two chapters, in effect allows him to inform
each person of the consequences he may expect, both for himself and
for others, from living in such and such a “‘private city.’” The city plan-
ner himself cannot perform this task of warning each inhabitant indivi-
dually, but the inhabitant himself can be prepared, during his primary
schooling, to calculate these consequences on his own.
| should therefore propose teaching children, at the age of eight or ten,
to become accustomed to the “‘play”’ of a city: | mean “‘playing’’ and
not “‘playing a game,” for we have seen that game theory does not
apply to cities (there are no “‘rules of the game,” and certainly no final
outcome like the one that a ‘’game,”’ according to game theory, is
supposed to have). ‘‘Playing’’ here means simulating a system within
which the acts of others exert a definable influence on our own acts. |
am not a teacher, but | do not believe that it would be too difficult to
represent an “‘infrastructure,’’ for example, by a certain number of
chairs (let us say two chairs per child). and to let the children group
themselves on the chairs as they wish, in order to perform the activities
they like. Children also have their ‘private cities,’’ and one might give
them explanations that they could easily visualize: for example, of the
intrinsic characteristics of configurations that they have instinctively
chosen for themselves.

6.2 PROBLEMS BECOME MEANINGFUL ONLY IF WE HAVE

ADEQUATE INFORMATION: THE INFRASTRUCTURE AS
MEDIUM
The impossibility of finding solutions for all the conflicts brought
hs The Infrastructure as Medium

about by the existence of a large number of ‘‘private cities’’ results

from the extreme paucity of information obtainable about each ‘‘pri-
vate city.” Therefore, the first proposals mentioned in the preceding
section were (1) to find a simple way to code information concerning
private cities; (2) to teach the individual user to read this code.
Once the information has been made readable and has been read, it
allows a comparison among several possibilities. And the comparison
will give an individual user the means of profiting from his information
about his private city within a physical structure by finding the solu-
tion that he prefers, for he can find it only if he is familiar with all the
alternatives at his disposal, even in simplified form, that is, with a com-
plete list of possible solutions.
We cannot compare the solutions contained in this complete list if
they do not have a certain frame of reference in common; moreover,
frames of reference are more convenient when they can refer to a larger
set of possible solutions.
| have given the name “‘infrastructures”’ to these frames of reference
which can be used for all possible solutions. An infrastructure may thus
contain all the ‘‘private cities’’ (imaginable or existing) which | spoke
about earlier. In a preceding chapter we saw that the physical form of
an infrastructure is generally a network, for which rules of utilization
have been established (we saw several examples of these in Chapter 4).
An infrastructure is composed, therefore, of anetwork + rules. Con-
sequently, city planners as well as users are intensely concerned with
infrastructures and their rules of use. Once the common rules for using
a network are established, any user can have his private rules. We would
regard the best common rules are those which conflict the least with
private rules.
We can state this principle in another way, by saying that an infra-
structure must be as “‘uncommitted” (‘‘neutral’’) as possible, that is,
conceived with the individual user in mind, with the end of satisfying
his preferences, whatever they may be.
| stress this statement because it contradicts a highbrow superstition
which would have us believe that a city’s ‘‘personality’’ must lie in its
arrangement and form. It is, in fact, a grave mistake: indeed, the less
114 The City

THE INFRASTRUCTURE OF THE City

SS THE Mediunr
WHICH SUPPORTS INDIVIDUAL REHAVIOR

t kK
eee ik A e

5 r
< b
REHAViOR. ReHAVOA
OF MA x oF MAS Y
THE INFRASTRUCTURE
SyP oadlig ”
EHAVL
OF Batra ke y THE WFRACTRUCIURE OF AC
ANH “THAT oF A PART OF rey
CWEIGRROVAHOOD, BvILDING ETC)
WANE THE Sane MAppINd
THIS INPRASTRUCTUQE IW THE
Figure ABove
CA) BELong re A City,
TOA ROILBDING oR TOA plat
Figure 57

personality the “hardware” of a city has, the more possible it is for its
inhabitants to have their own. The city /s its inhabitants, and the phys-
ical city without inhabitants is nothing but a ruin.
Since the beginning of the book, | have insisted, first and foremost,
that information processes should be involved in city planning. After
all the theoretical:and general reflections which have been necessary,
we can now consider this fact as evident: the infrastructure is the prin-
cipal medium of information ofa city, in the sense that a medium con-
veys a message: the infrastructure is the medium and the usage schemes
are the messages. The usage schemes (private cities) superimposed on
this infrastructure are the real city.

6.3 THE INFRASTRUCTURE MUST FULFILL CERTAIN

CONDITIONS
There are two categories of ‘‘private cities’” superimposed on the infra-
115 Infrastructure Must Fulfill Certain Conditions

structure:
1. the ‘‘private’’ private city, used by an inhabitant most of the time:
his personal environment (flat, apartment, house, room).
2. the “‘public” private city, used less often by the inhabitant during a
given time of reference: his ‘‘errands network” (social life, public obli-
gations, trips for no particular purpose, and so on).
The difference between these two categories is solely quantitative: the
“private’’ private city is the one that we use every day, and the “‘public’”’
private city the one we use less frequently. Naturally, the ‘‘private pri-
vate city’’ of one person can be the “public private city’’ of another.
(For example, if | am invited to someone's home for a drink, my ‘‘pub-
lic’’ private city coincides with his ‘‘private’’ private city.)
We do not use the “‘private” private city of another person as if it
were our own; on the other hand, several ‘‘public’’ private cities can
have intersections or even be identical. A given place in the city can be-
long to several of these ‘‘public’”’ private cities; for example, in Paris,
the waiting room at the Gare Saint-Lazare, where | go to meet my girl-
friend, rather than to catch a train.
The physical infrastructure (hardware) of the city must consequently
fulfill the following conditions:
1. An inhabitant must be able to have, within the infrastructure, the
“orivate’’ city he prefers for himself, whatever its nature. (He must thus
be able to choose, according to his preferences, the cost of his home
and his situation within the infrastructure.)
2. He must be informed of the consequences of his choice as far as cost
and situation are concerned, and also of the consequences of other
people’s choices: therefore the infrastructure’s role as frame of refer-
ence (infrastructure = network + rules) must be clearly established.
3. An inhabitant must subsequently be able to modify any decision and
every implementation of his choice (cost and site of his home): the in-
frastructure must allow for ‘“mobility.’” (N. B. Since 1957, | have been
using the term ‘mobile architecture” for any solution allowing users to
make a direct decision and transform their environments directly as
soon as they decided to reconsider and modify their previous decisions.
The term, which seemed somewhat exaggerated when it was first used
116 The City

(and which | had chosen for want of a better one) today has become a
generally accepted expression in professional jargon, and even time-
honored to the point where is has transcended my personal jargon.)

6.4 THE INFRASTRUCTURE IS NOT THE CITY

In one of the preceding chapters, we saw that every 7 graph, including
the complete graph, could be plotted on the saturated graph (which
constitutes its support). If we consider the activities of the inhabitants
(which are symbolized by their ‘private cities’’) as a less connected graph
than the one representing the physical infrastructure of the city, which
(a?
supports the other graph, the ‘‘real’’ city will consist, for us, of ““hard-
ware” (physical infrastructure) and the modes of use within this infra-
structure.
This statement shows clearly why the expressions ‘‘urban design’’ and
“urban planning’ are so devoid of meaning: it shows us how utopian it
is to hope to find someone capable of projecting or planning modes of
use, or even specialized physical constructions intended for certain
modes of use, real but not detected or detectable. This same statement
leads us to the observation that the only logical act for an architect or
city planner must consist of designing the most ‘‘neutral’’ physical in-
frastructures possible in relation to modes of use.
| settled this problem in an earlier section of this book by proposing,
first of all, sets of points ‘‘with the minimum number of connections”
or ‘‘with the maximum number of connections” (Section 3.4); then,
any planar network of homogeneous degree (Section 4.2) as an appro-
priate mapping for infrastructures.

6.5 AN EXAMPLE OF THE INFRASTRUCTURE AS “HARDWARE”:

THE “SPATIAL CITY”
Just as an example, | am now going to try to describe to you an infra-
structure which | created in 1958 and used for different projects, most
of which have been widely published in professional journals.
This project was based on two sets of criteria, the first comprising
general criteria for infrastructures, the second made up of economic
considerations relating to the context.
AGW, The Infrastructure as Hardware

Let us draw up the list of these two groups of criteria:

The first set (1) concerns economy of use. The principle has been to
reduce use efforts, such as time lost or dissatisfaction due to the pre-
established and imposed modes of use, contrary to the personal habits
of the real user, and so on.
la. The infrastructure must be the most ‘‘neutral’’ possible network,
interfering as little as possible with arbitrarily chosen modes of use ap-
plied within it.
1b. The infrastructure must contain the fewest possible potential con-
flicts (crossovers).
1c. The infrastructure is the on/y fixed physical element of the city: all
the other ‘‘hardware”’ (elements inserted “into the infrastructure, floors,
ceilings, walls, partitions, fixtures, and so on) as well as nonphysical
elements (movements of people, climate, visual aspects, and others) are
interchangeable, transformable, movable: they are in effect ‘‘mobile”’
and can be adapted to any particular mode of use.
1d. The infrastructure requires a minimum of “‘safety’’ regulations, pro-
viding for the elimination of mutually disruptive or contradictory
modes of use.
The most important of these regulations concerns the relation between
height (and therefore minimum clearance for windows) and the ‘‘coef-
ficient of surface use’’ (the corresponding formula is given in Figure
58).
The second set of criteria (2) concerns economy of construction.
2a. Construction of the infrastructure must fulfill all the criteria of set
(1) concerning economy of use.
2b. Construction of the infrastructure must impose the fewest possible
elements of new or specialized construction, and must take advantage
of preexisting conditions.
2c. Construction must economize as much as possible on the costliest
elements in a given context: manpower on construction sites and trans-
port (raw material and manufacture being the least costly elements).
These two sets of criteria led me to the following solution:
1a. The infrastructure must be a planar network of homogeneous degree
or the “‘spatialization’”’ (through labeling) of a network of this type.
118 The City

PARAMETERS SF USE

Sa cot ee ae ee

So eG ee

GASSICAL COW CERT

oF mot Ua ene ela
NUMB OP Peo
R UNIT of POPACE

tptdtord tet d+ bid +p

THE NUMBER OF LEVELS
CescTiPucyTion or THE
SAME OCCUPANCY PATERN: GROVWD SLAPACE)
1§ UMITED By Vw
AN ADWCTTED MINIMUMOCCUPA
BSTAMGE RLIDENCE
BETWEEN OFPERERE NatObat Ae oF
GS THE PARAMETER MOS
THE CORFE ICIEtee SF GROUND
ALB MULTIPLICATION

TE AREA
Gabon svarace (Se
THe EQVAMON OF i CourFiiarr
(N “THE CASE 4 n LEVELS 1S

eeSs. oh <2
he
mem.
Figure 58
 119 The Infrastructure as Hardware

1b. This network must be, if possible, a planar homogeneous network

of fourth degree or the “‘spatialization’’ of this network.
1c. The physical infrastructure must be of the ‘‘skeleton’”’ type (repre-
sented by the graph with the maximum number of connections), not
of the ‘‘troglodyte”’ type.
1d. The number of levels of the infrastructure must equal at least 4 and
at most 8 if one bears in mind the minimum clearance for windows and
the building heights currently permitted in Europe or America.
2a. The infrastructure corresponding to the criteria of set (1) must be a
continuous multistory skeletal construction comprising 4 to 8 levels.
The beams and/or columns that constitute
the skeleton must contain
all the supply and disposal lines so that no space in the skeleton lacks a
connection or outlet for any kind of line (electricity, water, drainage).
2b. As in European or American locations (for which these projects
were designed), where spaces of essentially two sizes are used, the
spaces in our infrastructure belong to two categories: the small spaces,
~ characterized by a small use load (offices, apartments, hospitals,
schools), and the large spaces, with a large use load (public meeting
places, warehouses, stores, garages, industrial sites). We call the first
type “‘cellular’’ and the second type ‘‘communal.”’ It makes sense to
(22
locate the ‘‘communal” spaces; with their large volumes and their heavy
use loads, at the ground level of the infrastructure (the earth’s surface
being the least costly of the structures capable of supporting a heavy
load), and to assign ‘‘cellular use’’ to the small spaces within the skeletal
part of the infrastructure, above ground level.
2c. The two most expensive operations in construction are site work
and transport. So the infrastructure must be constructed of easily
transportable materials, and it should be assembled with the least
possible work at the site.
The infrastructure which | then proposed was conceived as a spatial
framework in which the total height of all the levels together was the
construction height (that is, height upon which the construction’s load-
bearing depends); the resistance of this increased construction height
made it possible to span the large spaces situated at ground level. The
120 The City

¢ ee
U

SUR a
SPANNING

\ POG x —
EDT as
e-— °°
oe
S 0 \ Aa e
wo TNO es SySMG OOS
EONS .
CROSSING #« COMImLeVsS BA
IDEAL CENTERS BY tae Suerte

MULTILEVEL
SUPPORTING PATRIA FLECORTING SPACE PRANE

cid CoMNEm «—«-TRANGULATION

Figure 59
121 Modifying a Choice Within the Infrastructure

HYPERBOLIC PAQ«BoLolbd
HYPE RBOLOID
OF RETRACTION
eei
Lp sr a
S @'
CON@ SHEET)

Figure 59, continued

structure included triangulated beams, whose constituent bars crossed

the structure from end to end and were attached by collar joints. The
spatial framework had been chosen so that the empty spaces within the
structure corresponded to the cellular spaces, and so that between two
cellular spaces there were the fewest possible diagonals in the separating
plane (such diagonals obstruct potential passages). Finally, there was to
be no diagonal, or any other beam, within the spaces themselves.
All these spatial structures can be developed according to a simple
archetype, which assembles hyperboloids of rotation of a single surface;
the beams then coincide with two families of straight lines plotted on
these surfaces (although the latter do not exist physically) using points
of a planar network of homogeneous degree.

6.6 MODIFYING A CHOICE CARRIED OUT WITHIN THE

INFRASTRUCTURE
The arrangement resulting from applying one or several choices and im-
plementing them within the infrastructure can become unsatisfactory
for one reason or another. In that case, it will be necessary for the indi-
vidual user (or group of users) who have made that choice to be able
to modify the arrangement. It will always be possible to carry out this
122 The City

modification if all the elements inserted in the infrastructure are mobile,

and if the infrastructure remains the only fixed physical component of
the city. As we have already seen, this material component must be de-
signed to permit any imaginable arrangement, that is to say, any labeled
linkage scheme of a collection of n distinct elements included in a com-
plete list (repertoire). Once the project has been set up in this manner,
there is no chance of obstacles to any imaginable modification except
perhaps for the conservatism of the users.
As the modifications can lead from any possible arrangement (labeled,
connected planar n graph) contained in the complete list to any other,
and as no correction can lead to an arrangement not contained in the
list (since only physically or geometrically impossible arrangements are
not recorded in it), the corrections can set off an endless cyclical pro-
cess: no one arrangement is ‘‘better’’ than the others, except in a given
context where great narrow-mindedness would prevai!. But | cannot
imagine that such a context, if it existed, would long obtain, and con-
sequently there is no final solution foreseeable for this cyclical chain
of corrections, and no arrangement that is absolutely ‘’best.’’
One implication of this conclusion is that ‘‘“game theory’’ cannot be
used to describe urban affairs, since by definition this theory must
operate toward a ‘‘final state’ (when one of the partners wins; or, in
the case of a tie, toward an indecisive final situation). An endless game
is meaningless, and game theory cannot be applied to unending pro-
cesses, since all its evaluations are based on the end of the game; for
example, how to reach the end the fastest, or how to terminate the
game under given conditions, whatever they may be. So | do not see
any meaning to the question: ‘‘How can you end an endless game?”’
The use of a city is not a game that follows game theory.
Let us return to the example of the user of the ‘‘Flatwriter’’ (see Sec-
tion 3.7). Let us imagine that one regulation of the infrastructure
obliges all users to leave an empty space for each space used within the
structure; in other words, that only 50 percent of the available places in
the infrastructure will be sold. |
From that premise, the first user to choose the site for his personal
environment in the infrastructure seems to have the most freedom of
123 Modifying a Choice Within the Infrastructure

choice, and the last user seems to have the least freedom of choice. In
reality, the situation is exactly reversed: the first user has chosen his
site with the greatest liberty as regards his geogrphic situation, but he
has the least influence on who his neighbors will be. Every person
making a choice after him can become his neighbor. On the other hand,
the last one to choose can become the neighbor of any group, according
to his preferences. From the viewpoint of his ‘private city,’’ he has the
greatest freedom of choice.
But, since the modification process is endless, there is no “‘last to
choose.”’ For example, the first user will be able to modify his choice
once the last one has made his, and he will be able to choose a location
where he will have friends that he prefers. After that another user can
make a similar modification, and this can continue endlessly. Thus no
one can predict what arrangement this chain of modification will tend
toward in the end.
Thus the infrastructure, appropriately chosen, makes possible the
birth of the ‘indeterminate city’ (which already exists today, but in a
latent form). The fact of giving it an infrastructure allows this dynamic
city to attain the characteristic rhythm of change appropriate to it.
Thus the city will become a sort of random configuration.
APPENDIX TO CHAPTER 6: NEW PROBLEMS IN THE
INDUSTRIALIZATION OF CONSTRUCTION

1. THE GENERAL CONTEXT

The world situation shows very different characteristics from that of a
decade ago: we are witnessing the passage of our world from a period
of local affluence to one of generalized scarcity.
Indeed, the known energy reserve is considered sufficient for at most
fifteen years, and the time span necessary for the development of new
sources of energy is estimated by the experts as at least twice as long.
This points to a period of energy scarcity of fifteen to twenty years.
World food reserves are insufficient to assure survival for more than
one year. If we happen to have two consecutive drought years in Asia
(which happens generally once every twenty-five years), food might be
scarce all over the world (famines are no longer local as they were in
the nineteenth century). We have no hope even of setting up some cen-
tral organization to coordinate measures to deal with these emerging
menaces: indeed, communication necessarily breaks down within organ-
izations larger than a certain ‘‘natural threshold.’”’ The disintegration of
large organizations (which are considered by Western-type civilizations
to be unique guarantees against poverty) accelerates the processes of
general impoverishment.
It is urgent that industry, agriculture, and organizations become
aware of this deterioration of the global situation, but obviously it is
easier to write about it than to do something. The tremendous inertia
of human habits, the interest of ruling groups in continuing the exploi-
tation of existing technological setups, and the difficulty of communi-
cation make it impossible for any warning to be efficient within short
range. We are thus sitting in a ‘‘spaceship’’ whose steering does not
125 Appendix

function any more, heading for a crash. The best we can do is to make
this crash softer.
To do so, we would need to find some new ways
1. of using less energy and fewer raw materials
2. of reducing food consumption to the biologicaliy necessary level in-
stead of wasting food as we do today
3. of adapting technology for relatively small groups, who would pro-
duce on a relatively small scale, instead of mass-production technology,
which involves mass transportation, loss of individuality, and so on.
We can presume that civilizations that do not adapt themselves to this
course might have a very reduced chance of survivial.

2. PARTICULAR CONTEXT OF BUILDING INDUSTRIALIZATION

| had to open with this review because the trend | have described
determines the future for the field of construction. This future seems
to be somewhat less dark than that of most industries, simply because
building industry is still not committed to any definite technology and
thus alternative ways are still open.
The problems, which will be the subject of this appendix, have to do
with these alternatives.
What features should the new building technology have?
1. Let us begin with the energy problem: building technology should
consume the least energy possible. A building should ‘‘cost’’ minimal
energy, both in assembly (in the factory and in situ) and in transport.
For example, one square yard of wood construction consumes less
energy than one square yard of steel structure, which in turn is ““energy-
cheaper” than a square yard of reinforced concrete building.
Let us presume, that in order to determine ‘‘energy costs,’’ we could
introduce a parameter indicating for any building technology the pro-
portion ‘applied energy’’/’‘lifetime of the construction.’’ Such a para-
meter could give useful hints as to which technologies are to be favored.
2. ‘Lifetime’ of a building means the time span in which it can be used
efficiently in any way which should be determined by its users. If users
cannot use a building any more in their own way, the building can be
considered as ‘‘dead” even if it is not a ruin or demolished. (Many
126 The City

buidlings today are ‘‘still-born.”’

A corollary to this statement would be that in order for a building to
“live,”’ its space allocation, the connection schemes between its spaces,
its organization, what we call generally its “‘plan,’’ must be determined
by its future users.
3. Another corollary to our statement about the “lifetime” of a build-
ing is that the user should be able to perform all repairs himself. \ndeed,
we see that a large number of buildings decay very fast, because the
technology used to construct them does not admit simple repairs, and
there is no regular servicing to perform repair jobs requiring skill.
4. The organization which would produce construction elements which
fulfill these conditions (low energy consumption in the production pro-
cess, adaptability to plans conceived by the user, ease of repair for the
unskilled handyman) must be necessarily a re/atively smal/ organization,
between industrial and artisanal size.
This desideratum foliows both from the ‘‘low-energy condition” (as
small production plants working for regional markets need less energy
’
and less transportation than large ones), and from the ‘‘low-production’
condition (small amounts of very varied stock elements permit a bigger
selection than very large amounts of uniform elements).
Beside these particular reasons, smaller industrial plants are called for
as well by the general context: the communication breakdown | men-
tioned before is essentially unfavorable to large organizations.
5. A last, but not the least important, desideratum for the new type of
industrialization process for buildings concerns the “agglomeration
characteristics’ implied by these technologies.
Let me explain this concept. If, for example, a technology does not
admit the construction of individual low-rise buildings, or that of collec-
tive high-rise buildings, such a technology produces a constraint for
collective use, urban design, and so on. Other consequences of con-
straints resulting from the implementation of a technology might con-
cern: distances between buildings, use of ground surface, use of over-
head space. Such constraints also influence traffic patterns and land use
characteristics. Appliances might have effects on utilization patterns
(for example, certain appliances involve community use, others favor
127 Appendix

individual use as it is in the case of heating appliances or washing de-

vices. These constraints are what | call the ‘‘agglomeration characteris-
tics’’ of technological objects. |
It is not enough to say that a technology such as that described in
Chapter 6 will make possible any plan and any design. It is necessary
to encourage the user about his own capacity both as designer and
repairman. It is necessary to give him the tools, analytical or material,
to act as such. It is necessary to make him understand that he himself
is the only ‘‘expert’’ who can solve his own problems, and that he
should trust himself more than he trusts other ‘‘experts.”’

3. CONCLUSION
| have sketched, by intention, a very radical image of new trends in
building industrialization. The trends | mentioned imply, even more
than a transformation of technology, a transformation of attitude.
There is a growing resistance among professionals (architects, engi-
neers, builders) against such a change of attitude. There is—as | tried to
point out—a growing pressure in favor of a new attitude among users,
a pressure resulting from emerging conditions in the world context.
| think that a new trend is necessary and feasible (feasible, because
building industralization is still not so developed that it will not
admit reorganization). But when | say ‘‘necessary,’’ | mean that the
alternative trend should completely supplant classical techniques and
attitudes. | think that both trends should coexist until a time when a
tradeoff equilibrating the two tendencies becomes possible. In order to
reach this situation, we have to work on the new trend, which has re-
search priority (the classical trend does not need it: it exists already).
It is the task of research to develop what does not exist but is imminent.
CHAPTER 7 SOME CONCLUSIONS ABOUT SOCIETY

7.1 THE NONCOMPETITIVE SOCIETY

There is certainly no lack of prophecies and projections today. And yet,
a new and absolutely unprecedented form of society is developing be-
fore our eyes, ignored by most of the prophecies.
| am going to try to describe it, not in the form of a prediction but as
a way of demonstrating how inadequate our imagination and intuition
are even when it comes to picturing so imminent a reality.
We all agree that survival demands a continual struggle for most (if not
all) living creatures. It is evident that this struggle takes the form of a
competitive game, with rules that vary according to the species, wherein
the loser ‘‘loses’’ the right to survive.
We also know that this struggle for survival affects only the areas
where a certain means of survival is not available to a species in suffi-
cient quantities. An arid region, for example, where all the animals
struggle for possesion of a water site, is different from a country where
water can be found practically everywhere. Another example: despite
the primordial importance of oxygen, no animal struggles to obtain it.
A first observation can be expressed thus: competition among indivi-
duals of a species is a function of the scarcity of the means of survival.
This first observation is valid for every animal species.
A second observation follows, which is valid only for man and certain
animals: the scarcity of the means of survival can be natural or artificial
(that is, a part of the species can induce an artificial scarcity with the
aim of imposing its will on the other part). All struggles, political or
military, among human beings, are characterized by the efforts of one
side to control one of the means of the other’s survival. This observation
129 Noncompetitive Society

implies the existence of another ingenious device: if it is impossible to

introduce artificial scarcity of the real means of survivial, there is always
the possibility of inventing fictitious means. These fictitious means of
survival are provided by religions, group loyalties, various legal concepts,
and so on. Their essential characteristic lies in setting up a rule of pre-
eminence.
Our epoch is approaching (in certain countries it has already reached)
sufficient supply of the absolutely necessary means of survivial. (Bread,
water, and so on no longer present problems, for example, in the
U.S. A.). These countries have therefore introduced ‘‘acute awareness
of preeminence” (status consciousness) into the scale of ‘‘public’”’
values and Americans today are struggling more for ‘‘status symbols”’
than for anything else. ,
Yet there is this interesting but no less obvious fact: the first signs of
a “noncompetitive” society appear precisely in these countries. | am
thinking in particular of all the tendencies called ‘‘hippie.’’ [t seems to
me that the outward signs of hippyism have only negligeable impor-
tance compared to the significance of the effort (perhaps unconscious)
to eliminate competition: refusal to work (to earn more money than
the necessary minimum), refusal to form competitive subgroups (non-
violence), sexual promiscuity (the sexual partner is one of the necessi-
ties of survival which, all through history, has been the most “‘artifi-
cially’’ scarce).
Despite the imminence of this development, neither | (an observer)
nor the hippies have the slightest idea of the form that a noncompeti-
tive society would assume.
By way of example, | am going to try to construct a representation
(of course many representations differing from my example are possi-
ble).
A noncompetitive society (according to this picture) cou/d be noth-
ing but a nonsociety. Beginning with the fact that the ties between indi-
viduals are automatically linked to competition, this new “society”
leads us to conclude that a noncompetitive society must take the form
of a group of individuals having no contact with each other.
To us today as members of a classical type of society, it seems that
130 Conclusions About Society

the members of such a mutually indifferent society wou/d be bored to

death.
But this representation of society can lead us much further. First | am
reminded of the strange metamorphosis of certain grasshoppers: for a
long time, entomologists thought that there were two species of these
grasshoppers (‘‘solitary’’ and ‘‘gregarious’’) until the day they found that
these two species were the same, and that the grasshopper mere/y
changed its physical aspect depending on whether it lived alone or in a
group. Could it be possible that man is a similar animal, that is, differ-
ent depending on whether he lives in society or in isolation? If so, that
would imply that the noncompetitive society might involve a biological
mutation of the human species.
A strange observation comes to mind in this connection. For a long
time, it has been known that longevity among the inhabitants of certain
countries is more common than elsewhere. These countries (Bulgaria,
Soviet Georgia, Iceland, and others) do not have many geographic
characteristics in common, but /n a// these countries men live in a very
isolated, solitary way or in very small groups. \s it, then, possible to say
that perhaps noncompetitiveness, the solitary life, and longevity are not
isolated facts? My brief reflections have led me far, much farther than
| wished. The only thing that | really wanted to demonstrate is that the
sociologist, the biologist, and the psychologist have neither valid expec-
tations as to the development of life styles, nor even any imagination
about the possibilities of development. All planning based on prognos-
tics thus errs not on account of too much imagination but on account
of too /ittle. Admitting the possibility of cumulative errors, | am sus-
picious of methods based on intuition. Any prediction that is not based
on an axiomatics must be highly erroneous. \n the preceding sentences
| gave an example of a line of reasoning intuitively constructed, whose
truth or falsity cannot be demonstrated or even evaluated. In any case,
| had fun doing it.
To be serious again, after this mere game of invention, | can admit
that | came in the end to construct an axiomatics that shows (following
the criteria specified earlier) that this noncompetitive type of society
is really possible. The subject of this axiomatics is the function of
131 Noncompetitive Society

“property.”’ From the axiomatics, | constructed the complete combina-

torial list that corresponds to it; then | interpreted each term in the list
as a possible “‘society,’’ and | added an example of a possible ‘‘society”’
from the different classes of possible ‘‘societies.’’
Axioms:
1. An individual can use an object
A1: exclusively for himself
B1: simultaneously with others
2. The attitude of the other individuals toward use of the object may
vary
A2: their consent and help are necessary
B2: their consent and help are not necessary
3. The possibility of transferring the object varies
A3: the object can be transferred to other individuals
B3: the object cannot be transferred to other individuals
The complete combinatorial list constructed with the aid of this axio-
matics takes the following form:
A1 A2 A3: property exclusive, consent necessary, transfer possible
(example: ‘‘classical’’ type of society)
A1 A2 B3: property exclusive, consent necessary, transfer impossible
(example: society based on prestige, a status society)
A1 B2 A3: property exclusive, consent unnecessary, transfer possible
(example: ‘‘primitive’’ type of society, based on violence)
B1 B2 B3: property exclusive, consent unnecessary, transfer impos-
sible (example: group of recluses)
B1 A2 A3: property commonly owned, consent necessary, transfer
possible (example: ‘‘collectivist’’ society)
B1 A2 B3: property commonly owned, consent necessary, transfer
impossible (example: noncompetitive society)
B1 B2 A3: property commonly owned, consent unnecessary, trans-
fer possible (example: a pack of animals)
B1 B2 B3: property simultaneously owned, consent unnecessary,
transfer impossible (example: religious community, ‘vegetative’ society)
| do not intend to pursue the entire analysis which can be developed
from this list (that is, to set up a schema of transpositions of one term
132 Conclusions About Society

on the list to another). | shall be satisfied with saying that the noncom-
petitive society is possible and observable and that | wait with curiosity
to see what form it will take.

7.2 CRISIS
We hear a lot of talk about pollution—air pollution, water pollution—
which has become a fashionable slogan. Politicians like to raise a smoke-
screen by using this slogan to conceal certain more “‘real’’ problems.
Personally | have the impression that one of the worst sorts of pollu-
tion threatening us is what | call ‘information pollution.” By that |
mean (and the concluding remarks of the preceding paragraph alluded
to it) that we are in the process of piling up an enormous quantity of
unrelated data, in a disorderly fashion, and in such abundance that it
is becoming almost impossible to unscramble it. Naturally this ‘‘infor-
mation heap” is both redundant and contradictory (most of it being
garbage) and because of the quantity alone it is impossible to get to its
usable part. Information pollution also means that the ‘waste products”’
of our present knowledge are polluting the usable part of that know-
ledge.
Let us consider an information system of n elements (transmitting
and receiving stations, messages, or anything else that can be meant by
“element’’). The highest degree of potential connection between the n
elements will be represented by the complete graph Kn, which assumes
that all the elements are directly linked to each other. Such a graph
runs the risk of being very complicated, but the itineraries linking one
element to any other will be, in this case, the simplest possible. The
only difficulty with this schema lies in the fact that the graph generates
CKhni= ken crossovers, which can all be considered sources of poten-
tial conflict (points at which errors originate or where a particular signal
is necessary to eliminate such errors). In an organization based on in-
formation disseminated through communications from the outside,
this implies that the field of possible errors is proportional as a fourth
power to the amount of useful information. Such a proportion repre-
sents totally inadequate efficiency.
The implications of this fact doubtless suggest why we are presently
133 Noncompetitive Society

witnessing a process of self-destruction in political entities of large

dimensions, with more than 200 million inhabitants, (such as the
U.S.A. and the U.S.S.R.). There seems to be a kind of natural
threshold, analogous to “critica! mass’ in nuclear physics (conflicts
become explosive beyond a certain determined number of elements
composing a mass).
Let us imagine an organization not based on information coming from
outside and then channeled, but where the necessary information is
contained in a “‘band’’—a memory set up in advance, a kind of a priori
knowledge. This organization would be able to function without pro-
ducing the great number of potential conflicts which we spoke about
earlier, but, on the other hand, it would become relatively rigid and
impossible to ‘‘readjust.”’
Let us now try to examine society considered as a group of individuals
possessing certain common characteristics. Either the characteristics
will be acknowledged as common by the members of that society—in
which case it will be a society of the “immediate cornmunication’”’ type
(for if these characteristics are lacking, the members of that society
will no longer recognize each other) or these common characteristics
will be observed by someone outside the society—and it will then be
considered a ‘‘memory” type of society (in which case, the outside
observer can very easily detect similarities—dictated by a memory of
the a priori type—which remain unknown to the members of the society
itself). | am using the terms “‘society of the immediate communication
type”’ and ‘‘society of the memory type” just as labels, whose meaning
will be indicated later.
Let us try to imagine, by way of example, a science fiction scenario
showing diverse organizations in the form of societies belonging to
worlds other than ours. We shall try to describe these societies by de-
ducing their characteristics from a certain number of supposed ‘‘funda-
mental traits.”
24 A TELEPRATAIC SOCIETY
Let us imagine a society where direct, immediate mental transmission
exists among the members. Thus all the members of that society could
pick up all the thoughts of the others, whether or not they wanted to.
134 Conclusions About Society

Obviously, such a society would not have a language, since any ex-
change of information in code would be superfluous: that would also
be true for art, science, and so on. Even acts provoking pleasure or dis-
pleasure could not be carried out by a limited number of members of
that society, without their being felt at the same time by all the other
members.
Another consequence: that society could have only a collective mem-
ory, with no individual memory or consciousness of self in any of its
members. No isolated event could be recorded so as to permit its later
identification by its effects since all the members of the society would
have recorded it at the same time, and therefore no system to detect
regularities could be constructed. Among other concepts, this society
would not have that of time, for, since its collective memory would be
incapable of distinguishing an ordered sequence of certain nonlocaliza-
ble events (for example, the case of a succession of events affecting
only one member of the society at a time, and affecting them one after
the other, could not be detected, for the collective memory would not
know if an event had affected the same member rn times, or 1 different
members a single time).
Such a society would thus be almost entirely cut off from our mode
of thought, and it could not understand any structure we use.
7.2.2 A “MASS-SENSITIVE” SOCIETY
This time, let us imagine the existence of a society of beings insensitive
to light waves, sound waves, or any other kind of waves, in other words,
to all ‘“‘“media’’ (of information) reflected by surfaces. However, these
beings would react to variations in mass, potential, and so on, and would
be capable of localizing them. Naturally the members of such a society
would never see a surface or a form (just as a scale is not sensitive to the
form, but to the mass of the objects placed on its platform) but they
would “‘see’’ the slightest variation in mass, or in the configuration of
potentials.
This society would also be ‘‘telepathic,’”’ in the sense that it would
register the totality or the largest part of the mental processes of all its
members, as well as all the physical processes, within a certain distance,
and simultaneously.
135 Some Superficial Conclusions

This society would thus have an organization analogous to that of the

telepathic society.
7.2.3 A SOCIETY OF LEMMINGS
To conclude, let us imagine beings who simply have no memory what-
soever. Naturally, these beings could not survive, since they could not
learn how to anticipate dangerous external influences, and would be
rapidly exterminated en masse by communal accidents such as falling
from a precipice one after the other.
For these beings, the sole chance of survival would lie in direct, imme-
diate communication among all of them: pain, acting upon one of them
at the moment of the accident, would thus stop all the others at that
same moment, and before the pain disappeared would prevent them
from repeating the act carried out by the.one already affected by the
pain. Although immediately forgotten after the accident, this telepathic
warning could nevertheless preserve the society from extinction, pro-
vided that the number of its members was sufficiently large so that
even an accident of long duration could not do away with these “‘lem-
mings’’ one by one.

SOME SUPERFICIAL CONCLUSIONS

Although they begin from different definitions, the three examples cited
are in fact one and the same case: all three describe an organization in
which the paired operations ‘‘communication/memory” function differ-
ently from the way they do in our society.
Now our society in its present form possesses a very well-developed
memory and a woefully weak communication system. It therefore
functions by means of “‘structures,’” or systems of abbreviation which
alleviate this deficiency somewhat.
In contrast, the imaginary societies that | have described have no
formal memory; but they compensate for this lack by a system of im-
mediate communication. Without such a system (as in the case of the
“lemmings’’) they could not survive.
There is not much exchange possible between societies based on pure
memory and societies based on pure communication: for the latter,
our “‘structures’’ have no meaning.
136 Conclusions About Society

In these terms, the noncompetitive society of which | spoke at the

beginning of this chapter would be a society based principally on mem-
ory, far more than our present society is.
| am now going to try to estimate the chances that such a society will
come to exist in the future.
On the practical level, until now we have considered organizations of
the ‘‘most connected”’ type (based on communication) and the ‘‘least
connected”’ type (based on memory). We discovered that the tendency
toward a ‘‘more connected” society, based on communication, led very
quickly and inevitably to information pollution, and that the tendency
toward a “‘less connected”’ society, based on memory, implied the ne-
cessity of a preestablished a priori memory content. Now, one could
not imagine that such a memory content could be established without
the existence of an earlier organization based on communication.
Such thoughts can have rather comic consequences: each of these or-
ganizations (societies), which are mutually complementary, constitutes
a deadly “virus” for the other. (\ call a ‘‘viral effect’ the destructive in-
fluence of an external structural rule which contradicts and breaks
down the existing structural rules in a system represented by a network,
in such a way that most of the operations that could be carried out on
the original network can be no longer once the “‘viral effect’’ has been
introduced.)
Thus the noncompetitive society, based on memory, which is in fact
a nonsociety, would necessarily be an agent of destruction in our pre-
sent society, which is based principally on communication and a prey
to information pollution. At any rate this society does in fact react
with instinctive hatred against ‘‘hippies,’’ who are the precursors of a
noncompetitive society.
(We can observe a simplified example of a case of viral effect, when a
traffic network planned for a specific type of organization proves un-
suitable for another type of organization.) These observations allow us
to give a new definition of infrastructure (defined in another chapter)
as the network that best resists viruses.

*The noncompetitive society is, in fact, a tendency to promote noncommunica-

tion (as | have said, it is a nonsociety).
137 Some Superficial Conclusions

In my predictions | dare to assume that the noncompetitive society

will destroy ours. It may be that by the time it becomes the predomi-
nant model, the remains of our society will in turn have the same viral
effect and produce the opposite tendency, and so on.
The question that presents itself is this: can this alternating develop-
ment be considered a Markovian process leading to a stable state?
To suggest a reply to this question, | must turn to explanations of a
general order: if | begin to construct (map) the representation of a set
of individuals connected in a certain way (that is, in a society, as we
have defined it), with the individuals represented by points and their
relationships by lines, | shall necessarily obtain graphs or networks.
Earlier | called the “‘private city’’ of an individual the figure constituted
by the point corresponding to that individual and by all the other points
connected to him by a single arc. The ‘“‘“mapping” of a society will thus
be the ‘‘interconnected”’ set of these private cities, and it will necessarily
contain a certain number of crossovers (potential conflicts). These net-
works representing a society probably include repetitions, and there-
fore ‘‘regularities,’”’ that is to say, a ‘’structure.”’ What can we know
about that structure?
There is—sad to say—only one human institution whose structure (in
the sense meant here) is well known, and that is the army (probably the
stablest of all structured human organizations ever invented). The graph
of an army is a ‘‘tree’’ of fourth degree, based on the empirical observa-
tion that a man who commands cannot give an order and supervise its
execution by more than three persons at a time.
This implies that the structural rules, and consequently the weakness
or strength of an army organization, are easy to recognize: we are deal-
ing with a familiar structure. (It is interesting to note that new nations
generally begin their existence by organizing their armies.)
As for the form that structural rules for the mapping of societies may
take, based on the concept of the “private city,’’ | can suggest several
possibilities. We may ask ourselves the following questions:
a. What is the smallest closed circuit? (That is, what number of persons
can constitute the smallest ‘‘group?’’)
b. What is the classification of the degrees for a point in this mapping?
138 Conclusions About Society

Tee STRUCTURAL DIFREREVCES

BETWEEN “THE PRIVATE CITY CONCEPT
Aub HIERARCHICAL CRAGANIZATION
BRING UP IMPORTANT QUESTION
ABouT THE SOUBTY OF THE PUTURE

PRIVATE cTlEs MUTARY ORGANIZATION

WHAT 408 THE HERE THE STRUCTURAL RULES

STRUCTURAL RULES2 ARE ALREADY XNONN:
Bite Coameseatdne
UN EXPLORED
A TRER OFmieDEGREE4.
vy eld DISSEMINATION OF INFOR-
ON WW THE GROUP;
FRAGIUTY.
Figure 60

(That is, how many persons can be in contact with a central individual?)
c. What is the most symmetrical figure for such a graph? (In other words,
what is the simplest legislative code for the group in which there are the
fewest conflicts?) And so on.
It is by trying to answer these questions (something which has never
been done anywhere, as far as | know) that we might discover several
of the fundamental properties of societies.
Thanks to this research, we might also discover what laws concerning
“thresholds’’ govern societies based on communication and societies
based on memory.
When these regularities and other analogous ones become known to
us, it will be easier to settle the question of the Markovian or non-
Markovian nature of a sequence of events that transforms a social or-
ganization of a certain type into one of a different type. To translate
the question into other terms: does the ‘‘mapping” of a society allow
139 Some Superficial Conclusions

for an underlying cycle of different states of that society or not? Of

course, | do not think that the solution is that simple, but | use this
hypothesis simply to show what type of answer | hope to obtain
through these investigations.
As long as these questions, and others of the same kind, are not an-
swered one way or the other (through the discovery of regularities or
of their absence), | believe that present operational methods cannot
lead us very far, and that they will never be ‘‘trustworthy” (likely to
yield certainties). A typical example is ‘‘game theory,” the application
of which to our subject will always be fallacious, since the process we
are studying may very well be cyclical, and since it is therefore impos-
sible to see how this theory, which considers how well strategies bring
the game to its conclusion (win or tie) could be applied to it.
To “conclude this conclusion” | should like to return to the concept
of democratization; that is to say, direct decision regarding operations
by all the persons concerned who stand to lose because of these opera-

THE STRUCTURAL ALUSS

FOR A GRAPH REDAESENTING
A NETWORKoF
Cobls RE OF he Mime

N- Grey ASED ON
! A emt BAee

2) GRQWITS CONTAINING In
CROSSOVERS
3) N-CROvAS Conta m
CROSoV ERS aes
4) le CROSSOVERS CONTAINED
IN TREES STARTING FROM
POINTS OF DEGREE X%
ET
Figure 61
140 Conclusions About Society

INTERPRETAETIONS OF SUBGRAPHS
COMTAINED IN 2 NETWORK :

TREE oF DEGREE Y = HIERARCHICAL ORcANIZAMeV

RAPID DISSEMINATION
GOAT palace ated
NO CONF
SRICHANGEABILTY

we GRGIT + EGAUTARIAN OCRCANIZATION

SS
No CSNTSR
SEMI STABIUTY
NC CoNFPUCT
UNCHANGEABIU
TY

Serene N-GRAPH: IDEAL CacANIZAéTION

“ORGANIZATION WITH No ConPUCS
ban LAqeDd SUBGREV
PUV LEED oF we FRGANIZATION

COMPLETE
Fas N-GRAPA:
taes
AIS IS EXPLOSIVE UTOPIA

EeTnsm2oNe
x(MUM
PERMITS ANY CHANGE

Figure 62

tions. Democracy, in this sense, is an institution based on memory and

not based on communication, since in order to be capable of making a
decision individuals must have a ‘‘memory content” established a priori
(this a priori memory content includes, for example, concepts of per-
sonal interest, self-defense, and so on). In reality, all higher animals live
in a democracy based on memory; we usually call their memory con-
tent “instinct.”
If we wish to fight for a more democratic social organization, we
must use communication in its present state as if it were a tool destined
to obsolesce, whose principal use is the creation and establishment of
“memory contents.’ This means that popularization and instruction
are the most important activities.
The ‘‘Flatwriter,”” a tool which is based on memory and eliminates
information “‘breakdown,” is a good example of how this new society
could be achieved.
One last question remains: do we really want this other society or
not? But that is an emotional question and | can answer it only with
141 Some Superficial Conclusions

reference to myself. The only purpose of my proposition is to demon-

strate how far “‘objective’’ thought can lead us, avoiding the error com-
mon to most predictions, in which thought mistakes its own desires
for reality.
APPENDIX (1972) SOCIETY + ENVIRONMENT

A.1 TERMINOLOGY
A.1.1 SOCIETY CAN BE MAPPED BY A GRAPH
| call ‘““society’’ a set of individuals in which there exists some sort of
“relation” between any two individuals belonging to the set. A person
who has no such “‘relation” to at least one other who in his turn is re-
lated to the others, can be considered a person ‘‘out of society.”
Let us suppose that | would like to sketch the image of a society. |
will first draw all persons belonging to it, then continue by drawing
lines linking any two persons between whom | observe an existing rela-
tion. | will thus represent this society by means of a figure in which
any one person will be linked to any other by at least one ‘‘path” pas-
sing from one to the other by the intermediary of other persons.
If | replace the ‘‘mannikins’’ of this map of the society by points, the
result will be a figure consisting of points and lines, in which there is at
least one path between any pair of points chosen arbitrarily (‘“connec-
ted graph’”’).
Obviously this graph gives an oversimplified image of a society. Using
this figure will require further explanation.
| will thus reexamine the concept of ‘‘relations,’’ which | represented
by lines in the figure. First | will try to find out who drew it. Otherwise
stated, one has to know who the observer of this society is.
| should point out immediately that different observers will see different
“maps’’ for the same society. In most cases it will be difficult to find
two observers who would attribute the same importance to the same rela-
tion; thus the ‘‘importance”’ of relations cannot be observed without er-
ror, and therefore | will not introduce ‘‘importance”’ into my terminology.
143 Terminology

Figure 1

Figure 2

On the other hand, the mere existence of a direct relation between

two individuals belonging to a society can be observed. This means that
the existence or the nonexistence of a “‘line’’ in a graph representing a
society can be agreed on by a large number of potential observers. | will
thus be content to note the existence of a relation.
Another characteristic of such ‘‘relations’’ can be observed and noted:
their ‘“‘direction.’” Let me explain the term.
In observing two persons who ‘‘communicate”’ (and thus a relation
between them) we can generally state that once the communication is
finished, the one or the other (or both) of the two persons changes his
previous behavior. We will say, in such a case, that one of the two per-
sons (or both of them) received an ‘‘influence’’ from the other one.
This influence has a ‘‘direction,”” which points from the person who
exerts the influence to the other one who receives it.
144 Appendix (1972): Society = Environment

Figure 3

| will define “‘influence’’ as a relation between two individuals which

has an observable direction.
In conclusion, a society will thus be represented by a direct connected
graph, that is, by a figure consisting of points and lines, in which there
is no point not linked by at least one path to any other point (when
arrow directions are not considered) and in which every line carries an
arrow representing the direction of the influence in this particular link.
A.1.2 STRUCTURAL CHARACTERISTICS OF SOCIETIES

This representation of a society makes it possible to describe its struc-

tural characteristics by using maps constructed with graphs as explained
earlier. These characteristics do not refer to any measurable size, as |
had to exclude, first of all, the possibility of knowing importance, inten-
sities, and so on, of “‘influences”’ in general. | have thus limited myself
to considering certain topological properties of these graphs, in order
to characterize social organizations. These topological properties indi-
cate characteristics implied by the linkage schemes, the paths, and the
circuits within such maps of a society; they correspond to those of the
influences within a set of individuals.
To get to the description of such important characteristics | shall in-
voke an image: the image of the ‘‘situation”’ of a particular person
within a society. This ‘‘social situation’ will be defined by influences
this particular person receives from and exerts upon the other members
of the society. For example, if he exerts an influence upon four of his
neighbors and he does not receive any influence from anybody, he
could be considered as having more ‘‘power”’ than another person who
also exerts four influences but receives two influences from others.
 145 Terminology

Thus “‘social situation’’ of a particular person will be expressed by the

difference of the sum of influences starting with him and of the sum of
influences having him as their endpoint.
In practice, the social situation of a particular person corresponds to
his balance of influences. But—and this is important—we calculate this
balance without associating any “‘size’’ difference with different arrows,
as we have agreed that ‘size’ of influence is not observable.
We will thus consider, to make an inevitable simplification, any in-
fluence exerted by one person on another one, as having the same im-
portance when observed by anybody not related to any of these two
persons (that is, by an observer not belonging to the same society).
This convention does not mean, obviously, that the two persons linked
by this influence do not attribute to it any value, any importance
whatever.
We can be sure that they do, and even that very probably this impor-
tance will be different for each of the two: the one who exerts the in-
fluence might consider it as being important, and the other, who re-
ceives it, might ignore it; or inversely, one might ignore the influence
he exerts on others and the ones who receive it might appreciate its
importance.
To avoid such ambiguity | had to take as a standard the observation
of someone exterior to the society observed. But if this observer con-
siders all direct influences as equal, he might observe a degradation of
“indirect’’ influences (influences transmitted by several intermediary
persons), degradation through errors, omissions, because of successive
transmissions (degradation resulting from what information theory calls
“noise’’).
We will use, in order to describe this degradation of an indirect influ-
ence because of the necessary transmissions, a simple rule: we wil! sup-
pose that the “‘intensity’’ of an influence will decay in inverse propor-
|
tion to the number of intermediary transmissions necessary to its for-
warding.
We are now ready to define the ‘‘social situation’ of any person within
a society by an observer exterior to this society. It will be expressed by
the difference of the sum of all influences (direct and indirect) exerted
146 Appendix (1972): Society = Environment

Bg | .c

© x

> : *"
Figure4

A B
e e

WwexX

a)
ic 4
SY Oy
AX= BX: CB=DC=4 (qd =4
CX2DB= % ea
Dx*h (A, = 3)

Figure 5
147 Terminology

by this particular person upon all other persons within the society, and
of the sum of all influences starting from all other persons in the society
and received by him.
In order to do this simple calculation it is sufficient to construct the
“path matrix’’ of the graph mapping the society in question. On the
basis of this graph (or of this matrix) we can find both sums necessary
to define the parameter of ‘‘social situation.”’
Using this method, we can obtain not only the social situation of any
person within the society of our example (here containing 7 persons)
that we obtained, but also the real “‘hierarchy”’ established (frequent-
ly in a tacit way) within the society. Obviously, the hierarchy we find
in the figure is only the “‘hierarchy’’ observable by somebody exterior
to this society. It is quite possible that the ‘‘hierarchy”’ as observed by
Mr. A. or by Mr. B. differs sensibly from the ‘‘objective hierarchy.”
Let us suppose, now, that one member of this society decides to leave
it for some reason of his own. Immediately the ‘‘hierarchy” will be
transformed as a consequence of his leaving it. Certain persons remain-
ing within the society will benefit from this defection (their ‘‘social
situation” will be improved); others will be prejudiced. Thus, if we sup-
pose—in order to keep the example simple—that all members of the
society want to get a “‘higher’’ place in the hierarchy, those who bene-
fit by the defection of Mr. X can be considered his ‘‘adversaries,’’ who
are interested in removing him from the society. On the contrary,
those who are prejudiced by the departure of Mr. X. will try to keep
him: they are his ‘‘allies.”’
Using a simple function, which | call that of ‘‘dependence’”’ of a par-
ticular person in the society upon the departure of Mr. X., we can con-
struct a ‘‘table of alliances’’ characteristic of this society.
“Dependence” can be calculated in a simple way: by obtaining the
difference between the social situations of a person belonging to the
society before and after the departure of Mr. X. The first social situa-
tion can be calculated on the graph which maps the society as a whole
(Mr. X. included); the second social situation will be obtained by calcu-
lating on a subgraph of the same graph, the subgraph not containing
Mr. X.
 148 Appendix (1972): Society + Environment

B
A gy siS58
r) oS
~ ec

*d

wi Se
SS eG
®

yrhr | yr i

x ysl Any y?! A.

Figure6

To:
ABCBDEFC Age
aoe A1O 122345 Show! NG
B)2°11234 Se CARER:
C]/o oOCoCOo on A PATH
DI stonliee.3)
E422 3012 (4,,) (S,)
yeh ae 2301 iT
AE) ) & A a he neil
WW THE Sodexy
ThA REAL HieRRdy
IN THE SAME
Socery
Pp
From: 0 loo S50 B2S al 210 so m8 r Ss, = 218-241 = 0,37
B 33 4 eS Spyz 358-40 = -042
(zs © 00° olow APTER TRONGHSION Si tsceahegts 8.85
Di] loS 33 © wD Blas Gide) Sp #3,66-2:83 = 683
E} 25 0 & 0 le B30 a Se +3 08-2, = 0,52
FY 33 ko bo 2 Do lel 4 4 Spt AiG -233 = 1,83
G |33 bee Sv
% ol am Sq = 2,8'-2,28 70,65
241 400 393 229 lp 2 be
Figure 7
149 Terminology

Faron TRE SOCIETY

Figure 8

Es
w
A
NT Yt
a
e

ie
ie GQ
AND [TS SUB- GRAPHS
F NAPPING DEPENDENCIES
A EXcLUbeD © excunes B Excryped Q Excwred

3 te bd ExCLudeD FE EKCWWDED

Figure 9
eee) aay)
150 Appendix (1972): Society = Environment

We obtained thus, by a method sufficiently simple to be used by a

ten-year-old child, a fairly good description of the real structure of a
society. This method can be used for any society imaginable.
A.1.3 “EGALITARIAN” AND “HIERARCHIC”’ SOCIETY
| will try to define, after all these preliminary explanations, two con-
cepts of society, concepts which | consider important: ‘‘egalitarian”’
society and ‘‘hierarchic” society.
| call a society ‘‘egalitarian”’ if all its members have the same social
situation. Put otherwise, in an egalitarian society the difference of the
totality of influences exerted and of the totality of influences received
will be the same for everybody. Such a society thus contains no ‘‘upper
class’’ influencing the others.
An egalitarian society is thus a possible one. There are a large number
of graphs which satisfy this condition (even though certain constraints
restrict this number, constraints we will investigate later on).
Another type of society having great importance because it is a very
common one, is ‘hierarchic’: this type of society can be represented
by a ‘‘tree’”’ (a graph in which between any pair of points there is but
one path). This society is characterized by a ‘‘degressive’’ hierarchy of
social situation, starting from the ‘‘apex” (i.e., the person represented
by the apex is the most powerful, the persons just below him are some-
what less powerful, and so on). Its other characteristic is a ‘‘progressive”’
hierarchy of alliances, progression starting from the same apex. Thus
“dependence,” or change in social situation caused by the departure of

Figure 10
151 Terminology

aes
Taras
sya
Se
Figure 11

somebody from the society, will be less noticeable for persons at the
“bottom” of the society than at its ‘‘top.’’
We saw before that this method of representing a society is based on
the exchange of influence within a set of individuals. The two types of
society sketched above are practically two schemes of exchange dia-
metrically opposed. In egalitarian society exchange can start with any
member of the society and will with certainty reach all other members,
perhaps slowly, but surely. On the other hand, in hierarchic society,
there is but one person who can start an influence which will reach all
members of the society, and it will do so relatively quickly. As for the
probability that this influence will reach everyone, it is very low (be-
cause the noncooperation of only one person somewhere in the society
represents a locus for a certain number of paths).
Obviously enough, these two types of society are not the only ones
possible: | picked them because they are particularly important. As a
matter of fact, all social utopias have as a goal an egalitarian society,
and all technical organizations are based on hierarchic society. Evi-
dently, neither is a perfect model, but there are many organizations
tending toward the one or the other of these models.
A.1.4 SOCIETY CONTAINS BOTH PEOPLE AND OBJECTS
| would like to correct a simplification | had to make (among others)
when | began this paper. | defined society as a set of individuals in
which there exists some sort of relation (influence) between any two
individuals. In reality, we are related not only to men by a system of
152 Appendix (1972): Society + Environment

influences, but also to objects.

| will be thus obliged to introduce a new definition for society, as being
a set of persons and objects linked by a system of influences. Thus a society
can be considered as a mixed mechanism containing men and objects.
In order to be more precise, | have to state what | consider the criterion
of difference between human beings and objects, from the point of view
of this study. This difference results from the fact that men are conscious
of their situation in society, as opposed to objects, which do not care.
This definition makes it possible for us to imagine other social alter-
natives, which could be realized more easily than most social utopias
(even if these alternatives are submitted to the constraints | will talk
about later). One such alternative would be a society egalitarian for
human beings but hierarchic for the objects belonging to it.
A.1.5 SOCIETY AND ENVIRONMENT ARE SYNONYMS
The first important result we get by applying our new terminology
(that of the definition of society and its mapping by graphs) will be the
equivalence of society and environment.
This equivalence is expressed in the definition ‘‘set of persons and
objects.’’ Indeed, the definition generally used for environment, ‘‘the
set of objects influenced by and influencing men,” is inadequate. Why
not ‘‘objects and persons’’? As for the definition of society, it was
discussed in the preceding paragraph.
“Society” and “‘environment” are thus the same set. My own term
for this set is ‘‘others.’’ This term is really the most accurate, because
it refers to the person who uses it. Thus, if Mr. X. is talking about so-
ciety or about his environment and he uses the term ‘‘the others” he
includes Mr. Y., his house, and so on, but will not include himself. On
the other hand, if Mr. Y. uses the same term, he will include Mr. X., a
tree, and so on, and the whole system he is linked to, except himself.
“The others’ are others for every one of us.

A.2 CRITICAL GROUP

A.2.1 “VALENCE” AND DEGRADATION OF INFLUENCE
After this long section about terminology we are still far from the con-
clusion of this study. We have still to make clear ceratin basic concepts,
158 Critical Group

but happily, we have the most difficult ones behind us.

The first concept to be investigated is the one | call ‘‘valence’’: it sig-
nifies a property, observable and biologically determined, belonging to
the human animal. This property defines how many centers of interest
can occupy simultaneously the attention of man. For example, | can
read two books simultaneously (even if with some difficulty), perhaps
even three; but, surely, | could never understand ten different books at
the same time. In this case my valence will be perhaps three, perhaps
more, but by no means as much as ten.
“Valence” will thus limit the number of persons who can influence
(or who can be influenced by) one member of a society during a given
period of reference. Valence will be visualized in the map of a society
(or of an environment) by the ‘‘degree”’ of a given point corresponding
to a person (we call ‘‘degree’’ the number of lines incident to a given
point).
The second key concept, that is, the degradation of an influence by
successive transmissions, was mentioned earlier, when | discussed the
way to calculate real hierarchy (social situation) within a society.
Degradation of influence through successive transmissions also implies
an observable and biologically determined property of the human ani-
mal: indeed, this degradation depends on our mental capacity. | call
the ‘‘channel capacity’”’ of a particular person (or of a species) the capa-
city for transmitting a message with a number of errors, where this
number is characteristic of the ‘‘channe! capacity”’ of this person (or
species or sort of object).
This property is a very limiting one for a society: it implies that an
indirect influence submitted to more intermediary transmissions than
admitted by the ‘channel capacity” will decay completely: it will be
indistinguishable beyond a given number of transmissions.
“\/alence’”’ and ‘‘channel capacity’’ of the human animal represent
two natural thresholds which cannot be crossed without great difficulty.
These two thresholds affect the rhythm of exchange of influences
among human beings (and objects), thus within social organizations
(or environments) which in the end are dependent on the numerical
value of these thresholds.
154 Appendix (1972): Society + Environment

A.2.2 THRESHOLD VALUES

The possibility of practical application of these two thresholds (valence
and channel capacity) for societies or environments comes from the
fact that these thresholds determine the numerical size (quantity) of
“elements’’ (men and objects) that can belong to a society without
having a disturbing effect on its functioning. tmplicitly, the same thres-
holds determine the number of links within a society. Thus, for ex-
ample, it would be impossible to realize an ‘‘egalitarian’”’ society con-
taining n’ humans, m’ objects and w’ links if the respective thresholds
do not admit more than n’ humans, m’ objects, and w’ links, and when
n,m, and w are respectively smaller than’, m’, and w’.
It would be just as impossible to conceive of a ‘“‘hierarchic’’ society
containing n” humans, m”’ objects and w’””’ links if the specific thres-
holds imply numbers less for both ‘‘elements” and “‘links’’ for a “‘hier-
archic’’ society. Expressed simply, a society or an environment having
a determined structure (in the sense used in the first section) cannot
contain more elements and links than admitted by certain established
threshold values.
We will call ‘‘critical groups” the largest set of elements (humans, ob-
jects and links) for which the functioning of a group, characterized by
a determined structure, still can be assured.
The concept of the ‘‘critical group” is perhaps the most important re-

e e
/ ]Ay EXAMPLE:
e VALENCE: Us 4
ge —> — |
5 CHANNEL
CAPACITY: EC <5
Rasa eax STRUCTURE ! BQAUTARIAN
e<——@ IN THIS CASE
THE CRITICAL GROUP
aNVas CONTANS AQ PERSONS

Figure 12
155 Critical Group

sult of this study because a comparison of any organization with

the ‘‘critical group” corresponding to its structural category shows
immediately whether this organization is realizable or not. Most
utopias or projects fail less because of the impossibility of imple-
menting the ideas on which they are based than because they exceed
the size admitted by the “critical group.’’ Very often it is exactly
the success of a group’s ideology that is the tool of its self-destruc-
tion, for this success promotes the group’s expansion and when ex-
pansion passes the critical group size, the group ‘‘explodes.”’
A.2.3 “CRITICAL GROUP SIZE" AS A SPECIES CHARACTERISTIC
“Critical group” is a result of two biological factors (valence and
channel capacity) and one topological factor (the structure of a
society). Its size is thus independent of any ideology, technique,
or knowledge; otherwise expressed, of any “‘artificial’’ factor inven-
ted by man. The three decisive factors depend thus on ‘‘natural
laws,” and the rule of the “‘critical group” is itself a ‘‘natural
law.”
Two of these factors (valence and channel capacity) are biological
factors: thus they differ with every species. The third factor is invar-
iable in the sense that it is the same for any species.
Thus, ‘‘critical group size’’ varies with every species: it is different for
men, for monkeys, lions, herrings, or bees. But for any species, it can
be known, and the numerical size of the critical group could be con-
sidered as a species characteristic.
If we consider, for example, an animal species, let us say, elephants,
we will find that a herd of elephants varies as to the number of indivi-
duals belonging to it, but that this herd never exceeds a given number:
that of the critical group size of elephants.
Alienation of man could thus be a consequence of enormously ex-
ceeding human critical group size: we live with more people than we
can tolerate, and with more objects than we can rely on, and all this
without becoming biologically a different species.
Exceeding critical group size puts an overload on the brain of an indi-
vidual, an overload which he cannot bear.
156 Appendix (1972): Society + Environment

A.3 CRITICAL GROUP AND ENVIRONMENT

A.3.1 THE OVERPRODUCTION OF WASTE
Statements concerning the critical group for living beings (men and ani-
mals) have their analogues for objects, but | will reserve my interest for
the real world, which contains both living beings and objects.
| have said already that environment means ‘’the others.’’ We are not
necessarily in contact with all others, living beings or objects; thus it is
more accurate to reserve the expression “‘the others’ for those with
whom we are linked in an observable manner.
“Environmental crisis’’ results from improvements in our manner of
observation: today there are many known relations which were before
nonobservable. Many other relations, which could have been observable,
were not observed because they did not attract our attention.
One such relation, a very important one, concerns production of waste.
Any living organism, or any organization containing living beings, func-
tions through selection of components useful for its survival. The pro-
portion of useful components to all components in an environment is
very small; thus the organisms or organizations mentioned reject a quan-
tity of components which is far greater than the quantity retained. A
living being is thus a ‘‘waste factory.”
If waste is ‘‘recyclable,’’ it is not in any amount. If it is produced in
an amount above a determinable threshold, pollution (accumulation of
waste) begins. It is a question of a critical amount of waste, this amount
being determined by the structure of the organism or organization
which makes the selection of useful components and thus produces
waste, and by its relation to ‘‘other’’ organisms or organizations.
The key to avoiding overproduction of waste is thus the operation of
selecting what is useful. Today, with our way of practicing selection,
the principal production of humanity is waste: about 70 percent of
human energy is invested in it.
A.3.2 “RUBBISH IS BEAUTIFUL,” OR ADAPTATION TO MAKE WASTE
USEEUE

As the preceding paragraph points out, waste is waste because it was

considered useless during the process of selection, very often for arbi-
trary reasons. We could reduce overproduction of waste by a simple
157 Critical Group and Environment

method: by changing the use criteria for certain objects, thus changing
the process of selection.
To explain this idea, | will use a historical (or even “‘prehistorical’’) ex-
ample. | have in mind the “‘heroic”’ period of the agricultural revolution.
The first thing the farmer did was to cut down and tear out native
vegetation from the land he wanted to sow; then from season to season
he threw away stones that he found on his land. The product of this
deforestation and of this disposal of stones was genuine rubbish to the
farmer: wood and stones. One of the first inventions of the primitive
farmer was the ‘‘recycling’’ of this rubbish in the form of using it to
construct nis shelters.
This recycling did not imply any material transformation of the rub-
bish, not even any intervention of some new technology: it consisted
only of a change in man’s attitude toward some waste which began to
accumulate in quantities over the critical amount.
Prehistoric man thus avoided one kind of pollution by nothing more
complicated than changing his own attitude.
Let us imagine, as a second example. a similar change of attitude for
today: | planned several years ago to propose an international compe-
tition under the title ‘‘rubbish is beautiful.’’ The basic idea was inspired
by the fact that many of today’s artistic movements use rubbish for
works of art. So why not try to transfer heaps of rubbish into monu-
mental art, monuments for our time? A large part of rubbish is not bio-
degradable, and much effort is required to dispose of it, outside the
course of everyday life. Why should we dispose of it? Why should we
not use it for the construction of ‘‘megasculptures,’’ collective ones if
possible? Let us imagine, for example, a pyramid of plastic bottles, or
a large sculpture made of wrecked cars.
The conclusion is that being faced with pollution (that is, we exceed
the critical amount of rubbish) it would be easier to change our attitude
toward waste than to stop waste.
Expressed otherwise, a critical amount of useless components is
characteristic of a ‘‘species of objects’ (exactly as critical group size
was for a living species), and of the organization men impose on these
objects. It is easier to change this organization than the characteristics
158 Appendix (1972): Society + Environment

of the species, once waste exceeds the critical amount.

A.3.3 “CRITICAL GROUP” FOR PRODUCTION
Let us try to correlate the two kinds of thresholds:
1. The critical group size of a living species depends on the ‘‘valence”’
and ‘‘channel capacity” of this species, and of the structure of the
group (‘’structure”’ in its topological meaning).
2. The critical anount of waste depends on the human group organiza-
tion, on its attitude toward waste and on the properties of the transfor-
mation process (selection of useful components).
Although all factors of the critical group are independent of human
will, factors of the critical amount for objects are largely dependent
on human desires.
It is interesting to remark that the critical amount of rubbish is
partly a function of the critical group of the human species, and partly
of the transformation process producing a specific object.
This remark opens the question whether a return to social groups
with fewer members than the ‘‘critical group’’ would not resolve pollu-
tion problems, at least in a large number of cases. Would not such a
return to social groups smaller than the critical group also reduce com-
munication between groups, and thus commerce (the most ancient
type of communication) and overproduction (production of goods for
other use than one’s own)?
| do not pretend to be able to answer these questions; the only answer
| can give is that they seem to me very important ones. But there is one
thing | am sure of: a return to groups smaller than the critical one
would resolve a great many of today’s economic problems: relations
between production, property, and exchange. (| shall develop these
ideas in more detail).
In the following sections | discuss some social organizations and ana-
lyze them according to the points considered in the first three sections.

A.4 REALIZABLE UTOPIAS

A.4.1 THE “NONPATERNALIST"” SCHEME
| have investigated up to this point the ‘‘objective’’ elements of this
study: definitions, a terminology based on graphs (thus on mathema-
159 Realizable Utopias

WAPLENSYTAMey - me. — >

}\ MR xX
eS ToMoRROW
=~ ren ao

ae!
eis 2
CS ee I ToHoRROW
aes TOHORAOW
Figure 13

\ON *+ \M?
ye? Srna,

HR x
TODAY
EC\ISION +
id LM PEM eV 7
SNORROW

Figure 14
160 Appendix (1972): Society + Environment

tics), critical groups (a concept belonging both to biology and to mathe-

matics) characteristic of man and environment. | have made no moral
judgments, and | have mentioned examples only as illustration, without
committing myself by saying whether | consider these examples desir-
able or not.
Now | will change my approach, and | will talk about schemes and or-
ganizations and state my opinion about them. Thus | plan to make
moral judgments, and | will use statements from the three preceding
sections for reference, to be as precise as | can be in this condensed
study.
| will start by explaining what | call ‘‘paternalist’’ and ‘‘nonpaternalist”’
schemes.
| call ‘“‘paternalist’’ any organization in which one or more persons
make decisions, decisions which are supposed to be implemented by
other persons who have to take all risks involved in the decisions, with-
out any risk to the persons who made them. Obviously enough, | don’t
want to suggest that the decision makers, who reach their decisions at
the expense of “‘others,’’ do so by bad intentions. |n this category be-
long all sorts of ‘‘experts’’: people who pretend to know better, what
to do for the well-being of others, than the ‘‘others’’ themselves.
| call ‘‘nenpaternalist’’ any organization in which all decisions are
made by the persons who implement them and who take the whole
risk implied by the decisions, and where there exist no ‘‘others’’ who
risk getting the ‘‘fallout’’ from the decisions.
| personally think that the nonpaternalist scheme is desirable, and
that the paternalist scheme is dangerous, inefficient, and unjust. My
judgment that it is dangerous and inefficient is an ‘objective’ judg-
ment (it can be explained in general terms comprehensible to anybody),
but my judgment about its injustice is a moral judgment.
A.4.2 “OBJECTIVE LANGUAGE” IS A CONDITION SINE QUA NON FOR
NONPATERNALIST SCHEMES

The nonpaternalist scheme is realizable only if any individual involved

in a process is informed about the following points:
1. The possibilities he can choose from in making his decision (‘‘reper-
toire’’)
161 Realizable Utopias

IMPLEMENTATOW

IN PRASTRUCT.
WRX

INDIVIDUAL
WARN ING

TRE OTHERS”
Figure 15

VRX

aie
al
THE OTHERS
162 Appendix (1972): Society + Environment

i pal
es
2. The consequences for him personally; what he can expect from his
decision (warning about consequences to the individual)
3. The rules of the context wherein he will implement his decision
(“infrastructure’’)
4. The consequences of his decision for others (warning about conse-
quences for the community).
Conditions necessary for such information cannot be assured unless
an ‘objective terminology,” a ‘‘code,” is used. (Thus, the section of
this appendix devoted to terminology is the guarantee that this is a
“‘nonpaternalist’’ study.)
This is a necessary condition because without an objective terminolo-
gy neither the ‘‘repertoire”’ nor the ‘‘warnings”’ nor the “‘infrastructure”’
can be described without imprecision; and genuine decisions cannot
be made without knowing these elements. If a particular person does
not know them, he can be taken advantage of by some ‘’expert’’ who
tries to impose upon him his own decisions; for someone not knowing
such a terminology, the expert will become a self-appointed ‘‘trans-
lator’’ who might use the ‘‘others’ ‘’ ignorance.
The “repertoire” of social organizations (obviously oversimplified)
will include the following:
1. Hierarchical society, represented by a ‘‘tree’’ (a graph in which there
is but one path between any two points).
2. Egalitarian society, represented by a strongly connected, well-oriented
graph (a graph in which there are several paths between any pair of
points and in which arrow directions are ordered so that any region is
163 Realizable Utopias

Figure 18

hz a
mingh
a
Figure 19

limited by a closed circuit, with the same arrow direction.

3. “Undetermined” society, represented by a connected, directed
graph without supplementary conditions.
Consequences of the choice of one organization, whether they concern
the individual or the community, can be foreseen by the simple for-
mulas discussed in the section on social situation: warnings will thus
describe both hierarchies and alliances within a particular social organ-
ization.
Evidently, my hyopthesis in the first sections requires that the termi-
nology describe equally well a society or an environment.
Thus we are prepared to implement nonpaternalist schemes.
A.4.3 SOME REALIZABLE EXAMPLES OF EGALITARIAN SOCIETY
Personally, and for moral reasons, | am for egalitarian society, and |
will devote this subsection to its realizability. | would like to enumer-
ate some examples which can be carried out following the nonpater-
nalist scheme.
1. The ‘‘noncompetitive’’ society
We are often victims of a superstition which interprets competition as
164 Appendix (1972): Society + Environment

an imperative of the survival instinct. Competition might be a means to

assure survival of the fittest if it is directed at scarce commodities (like,
for example, food in an arid region).
Once all means of survival (food, water, air, shelter, available space,
and so on) are abundant (naturally or artificially), competition for sur-
vival is no longer necessary. We (humans) invent, then, a ‘‘fictive scar-
city,” a ‘privilege’ (material or immaterial) which can be assured only
in limited quantities. Such a privilege (scarce by definition) becomes
the “premium” inciting to competition. All societies today are based
on an internal struggle for privileges (iuxury, status, and so on) which
are not directly necessary for survival.
Let us imagine a society which would renounce these privileges. It
would be in essence egalitarian, and it could be realized only if the
number of its elements (persons and objects) did not exceed the limit
imposed by the critical group.
2. The ‘‘anonymous” society
We all like to ‘“‘sign’’ things we make: shoes, works of art, omelets, or
poems. Is it true that a shoe, a work of art, an omelet, or a poem makes
me happier if | know who made it?
The persons-objects relation is organized in two ‘‘path sections”: (a)
from the producer to the object (or idea) produced, (b) from the object
(or idea) to the user. Of these two path sections | consider the second
one important. If the ‘‘producer-object’’ section has a reduced impor-
tance, this means that it is not important to know who made a thing;
the important thing is to know for whom it was made.
um
An “anonymous society,”’ in which anybody knows how to make
things which will be considered important, is necessarily an ‘‘egalitarian”’
society, based on nonpaternalist organization: such a society is anti-
elitist.
3. Society with ‘‘weak communication”
When we investigated the idea of critical group, we saw that decisive
factors were “‘valence’’ and ‘’channel capacity.’’ “Critical group”’ was
thus a relatively smal! group with a very reduced amount of communi-
cation, as Most communications could not reach persons situated far
from the source (distance measured by the number of transmissions
165 Realizable Utopias

aaa Ss m SS Pig
OF WEAK
COMMUN (CATION
(VALENCE: 4)

Figure 20

necessary to convey influence).

We considered “‘critical group” as an “‘island’’ with a zone of non-
communication surrounding it. But conditions for the critical group
could be interpreted otherwise: in a communication network contain-
ing a number of persons and objects there can exist a critical group
around each person (or object) in the society, considered as the ‘‘center’’
of this group. Such an interpretation is possible if the links composing
the network are arranged in such a way that the number of links relat-
ing a particular person to the others does not exceed the number ad-
mitted by ‘‘valence.”’
This society, even though it contains a very large number of indivi-
duals, contains a relatively small number of links: if it is based on
“weak communication.”
A.4.4 INDIA AS A POSSIBLE ALTERNATIVE MODEL FOR WESTERN—
Ars SOMES
| find it rather interesting that the only place where | met an observable
example of all three types of egalitarian society was India. Although
noncompetitive society, anonymous society, and the society of weak
communication exist separate/y in western-type countries, in India
| observed all three together.
Indian society is composed of very small groups (smaller than the
critical group). These groups are not linked very tightly by local com-
munication because of technological inefficiency (to their benefit).
166 Appendix (1972): Society + Environment

Indian artisans for many centuries have considered it undesirable to

sign their work: it could be produced by anybody. What is considered
important for a work (or product) is the way the future user will utilize
it, and this way (‘‘code’’) is strictly indicated.
Noncompetitive society does exist in India but in an incomplete form,
since abundance, which would be necessary to its complete develop-
ment, is lacking. But because of the poverty of the country, noncom-
petition characterizes the basic attitude of people: it is appreciated,
contrary to the situation in our countries, where competition is admired
as a positive, desirable feature.
It is not uninteresting to remark that industrial development in India
(a development which is considerably more advanced than many people
would think) is proceeding less by large industrial concentration than
(always for technological reasons) by artisanal decentralization: arti-
sans (workers) work at home for industrial organizations, and they
often make up by inventiveness what they lack in tools.
| know that | cannot avoid being superficial and partial by the picture
| have sketched. My intention here is less to discuss |ndia’s policy than
to use what | observed to support the theoretic concepts stated before,
knowing that the facts might develop in an unforeseeable way and that
the reasoning behind them is complex, conditioned by a long history.
Even so, | was fascinated to find that in this very rational country (the
only one among developing countries where a military dictatorship has
never existed), there exists a model from which we could, perhaps,
learn something.
A.4.5 AN INDUSTRIAL ‘‘ECOSYSTEM"”
The idea of relating the ‘‘critical group’’ of society to that of produc-
tion (concepts linked by similar thresholds) led to the society of ‘‘weak
communications,” social and artisanal.
This idea could lead us even further toward a social and industrial
“ecosystem,’’ which could be successful provided that links are weak
within the system. This system could be thus represented by a map in
which “valence” and ‘‘channel capacity’’ would have very low numeri-
cal values.
| have been talking in the preceding sections about a ‘‘social ecosys-
167 Limited Effects of Technological Utopias

tem” (without using the expression). Let us look now very briefly at
an industrial ecosystem.
It would be, to give it a concrete form, a ‘coexistence’ of small-scale
industrial organizations (smaller than the critical group). These organi-
zations would function in such a way that waste produced by one
would represent raw material for the other (similarly to the example
of materials rejected by farmers becoming raw material for builders).
The waste of western industrial production could represent today,
for example, a very important source of raw material for developing
countries. Plastic bottles, food cans, cardboard wrappings, even news-
papers are materials very intelligently used in India, or in other coun-
tries of an equal stage of development. Indians transform these mater-
ials into objects of beauty or of everyday use, objects which are often
sought on the Western market!
An exchange of waste could be more important than the exchange of
finished products going on today. Such an exchange could appeal to
individual inventiveness within small groups (smaller than the critical
group) and, on the other hand, it could prevent accumulation of indus-
trial waste over the critical amount.
It seems to me quite probable that a social and industrial ecosystem
having but weak internal communication could resolve most of our
social, economic, and political problems.

A.5 LIMITED EFFECT OF TECHNOLOGICAL UTOPIAS

| have referred in this appendix mostly to social (or environmental)
utopias. Why did | deliberately omit technological ones?
| will present here an example to show how small the direct impact of
a technological utopia, even of a very advanced nature, could be. Let
us suppose, for the sake of example only, that we can forecast imple-
mentation of some technological inventions (inventions which in real-
ity are not yet out of laboratories). | choose here three inventions
which could have a very large effect:
1. Advance control of weather
2. Large increase of staple food
3. Transport of energy with no conducting lines.
168 Appendix (1972): Society + Environment

These three inventions could be realized in several ways:

1. By hardware (‘‘things’’ used by man: satellite relay for weather con-
trol, “green revolution’”’ for basic food, energy beams for energy
transport)
2. By software (mutation of man himself: better weather resistance of
the human body, better efficiency of human metabolism).
What would the earth’s landscape become if these inventions could be
implemented ?
It would become (whether by hard or soft technology) a paradise: no
buildings (as weather shelters would no longer be necessary), no towns,
almost no factories. The only manmade objects visible on the landscape
would be meeting places (where people would gather if their presence
were necessary) and paths or roads leading to these areas.
Let us think. This utopia is not new: it existed in prehistoric times,
when the only manmade objects were assembly places and paths lead-
into to them.
The two pictures are identical. The most developed technology we
can imagine leads us to prehistory.
If our species has changed since prehistory, this may be less the con-
sequence of technology than of transformations of human social organ-
izations.
This example is not intended as a proof, only to demonstrate the pri-
mordial importance of relations among individuals, which | consider
to take precedence over technological inventions.

CONCLUSION
This short study led to very important results: the possibilities of reali-
zation of social or environmental utopias.
My first conclusion was that ‘‘society’’ and ‘‘environment’’ mean the
same thing: a set of persons and objects around us. This equivalence of
society and environment signifies, logically, that ‘structural laws’’
valid for the one are valid as well for the other. These laws are charac-
terized by ‘threshold conditions,’’ which means that both society and
environment function differently above and below certain strictly de-
finable dimensions: | called these thresholds the ‘‘critical group size”
169 Conclusion

(for society) and ‘‘critical amount” (for objects in the environment).

Although ‘‘structural laws’’ are the same for both society and environ-
ment my moral attitude—conditioned by the fact that we are men and
not objects—is a different one toward society than toward environment.
As a matter of fact, | find admissible a ‘‘paternalist’’ attitude toward
objects and | consider the same attitude inadmissible toward humans.
| demand a nonpaternalist attitude toward men, which means that no-
body has the right to make decisions for other people who might suffer
from these decisions.
These two conclusions, one structural and the other moral, are not in-
dependent of each other: it is the structural law which determines what
consequences (risks) can be expected from decisions we make, whether
concerning society or environment.
For moral reasons, the most desirable society seems to be, in my
opinion, the one | called ““egalitarian.’”
This type of society is consid-
ered utopian because it has not often been realized. Our new knowledge
about critical groups shows the reasons for the apparent unrealizability
of this type of society, which results from the fact that nearly every
attempt to set up such a society passed limits imposed by the critical
group.
We cannot improve the actual situation (social and environmental)
by technological progress alone. Such progress, at best, only leads us
back to some situation which existed before and which could be imple-
mented as well without technological progress.
We cannot improve our society and our environment except under
the condition that all improvements are decided on by all of us. This
type of decision making is implied by the nonpaternalist scheme de-
scribed earlier. The way to such improvement could be through a social
and economic ecosystem with weak communication.
Critical group size, a fundamental condition of survival, is a charac-
teristic of our species. Our future could depend on our answer to the
question: can we get back to groups satisfying the criteria of critical
group, or are we undergoing a “mutation” (in the biological sense)
into another species characterized by a different critical group?
 - =~; a a 4The

pe ro > op = Git ' @* cern y Sans ave

oa np = : oe a. RPsee gk wets or i a

ES Gipl Rawlive
imine nierwi
. Cae )
ee he ars tte
ee ee en ee,
= w a eit " -

miata tbainehas-paiedc
ate
7 a a de =
 $4.95
ARCHITECTURE

Toward a Scientific Architecture

by Yona Friedman
Today in architecture and city plan-
ning, Friedman (who has lived and
practiced architecture in France for
many years) observes that there are
no strict rules that allow an accurate
prediction of the results of a particu-
lar decision. Instead of the intuitive
rules or ‘tricks of the trade’ used by
both professions, Friedman pro-
poses a complete list of objective
rules and risks involved as choices
are made among several million pos-
sible plans for living space.
“Much of the fascination of this
book lies in Yona Friedman's pro-
vocative ideas: city barometers,
effort maps, intervention permits and
oee
ta
Al
£
so on. But it is also fascinating for
the carefree audacity with which he
expresses his particular form of pos-
itivism. ... He proposes...a
method which would provide each
client with a repertoire of all possi- Do

ble arrangements of space, together

with a technical device—a
‘flatwriter—by which to choose his meer
Ce
aeePe

own particular environment, as if

from an architectural menu, with
inbuilt warnings against socially dis-
advantageous consequences. ’’—The
Architectural Association Quarterly

The MIT Press

_ Massachusetts Institute of
Technology
0
re)
Cambridge, Massachusetts 02142 0 262 560194 FRITP