BOSTON STUDIES IN THE PHILOSOPHY OF SCIENCE HUMBERTO R. MATURANA and FRANCISCO J.

VARELA

EDITED BY ROBERT S. COHEN AND MARX W. WARTOFSKY

AUTOPOIESIS
AND
COGNITION
The Realization of the Living

With a preface to 'Autopoiesis'

by

Sir Stafford Beer ..

D. REIDEL PUBLISHING COMPANY

DORDRECHT: HOLLAND/ BOSTON: U.S.A.

VOLUME 42 LONDON:ENGLAND
 0977t-jd--
Library of Congress Cataloging in Publication Data

Maturana, Humberto R. 1928-

Autopoiesis and cognition. EDITORIAL PREFACE

(Boston studies in the philosophy of science ; v. 42)

Bibliography: p.
Includes indexes.
This is a bold, brilliant, provocative and puzzling work. It demands a radical
1. Biology-Philosophy. 2. Cognition. 3. Life (Biology)
I. Varela, Francisco J., 1945- joint author. II. Title. III. Series. shift in standpoint, an almost paradoxical posture in which living systems are
Ql 74.B67 vol. 42 [QH331] 501s [574'.0l] 79-24724 described in terms of what lies outside the domain of descriptions. Professor
Humberto Maturana, with his colleague Francisco Varela, have undertaken the
ISBN 90-277-1015-5 construction of a systematic theoretical biology which attempts to define living
ISBN 90-277-1016-3 pbk.
systems not as they are objects of observation and description, nor even as in-
teracting systems, but as self-contained unities whose only reference is to them-
selves. Thus, the standpoint of description of such unities from the 'outside',
i.e., by an observer, already seems to violate the fundamental requirement
Published by D. Reidel Publishing Company, which Maturana and Varela posit for the characterization of such systems -
P.O. Box 17, 3300 AA Dordrecht, Holland
namely, that they are autonomous, self-referring and self-constructing closed
Sold and distributed in the U.S.A. and Canada systems -in short, autopoietic systems in their terms. Yet, on the basis of such
by Kluwer Boston Inc., Lincoln Building, a conceptual method, and such a theoryofliving systems, Maturana goes on to
160 Old Derby Street, Ringham, MA 02043, U.S.A. define cognition as a biological phenomenon; as, in effect, the very nature of
all living systems. And on this basis, to generate the very domains of interac-
In all other countries, sold and distributed
by Kluwer Academic Publishers Group, tion among such systems which constitute language, description and thinking.
P.O. Box 322, 3300 AH Dordrecht, Holland The radical shift in standpoint here requires an imaginative leap and the
abandonment at the outset of the standard characterizations of living systems
D. Reidel Publishing Company is a member of the Kluwer Group in terms of function or purpose, or of organism-environment relations, or of
causal interactions with an external world, or even in terms of information,
coding and transmission. In effect, Maturana and Varela propose a theoretical
biology which is topological, and a topology in which elements and their
relations constitute a closed system, or more radically still, one which from
Autopoiesis: The Organization of the Living was originally published in Chile the 'point of view' of the system itself, is entirely self-referential and has no
under the title De Maquinas y Seres Vivas, © 1972 by
'outside', Leibnizian for our day.
Editorial Universitaria S.A.
The work demands and deserves careful reading. It is technical, formal,
All Rights Reserved difficult, philosophical and boldly imaginative. It is rigorously constructed,
Copyright ©'1980 by D. Reidel Publishing Company, Dordrecht, Holland and insofar as it is a theoretical biology, it remains uncompromisingly abstract
No part of the material protected by this copyright notice may be reproduced or and formal. Yet it smells of the medical laboratory and of the working do-
utilized in any form or by any means, electronic or mechanical,
main of the neurophysiologist. Where the interpretation of the formal theory
including photocopying, recording or by any informational storage and
retrieval system, without written permission from the copyright owner maps it into the domain of the nervous system, the insights and suggestions
for further interpretation are exciting indeed. And we expect nothing less,
Printed in The Netherlands here and to come.
V
vi EDITORIAL PREFACE

Maturana is perhaps best known to the philosopher and the scientific

nonspecialist as a co-author of the classic 1959 paper 'What the Frog's Eye
Tells the Frog's Brain' (with Lettvin, McCulloch and Pitts). Since then, he GENERAL TABLE OF CONTENTS
has worked on the anatomy and neurophysiology of vision, especially on
color vision. He has also been teaching medical students. The problems and
puzzles which emerged in his research and teaching led Maturana to develop
a distinctively alternative theoretical framework in order to answer the ques- EDITORIAL PREFACE V

tions, 'What is a Living System?' and 'What is Cognition?' The consequence

of his investigations, and of his construction of living systems as self-making, GENERAL TABLE OF CONTENTS vii
self-referring autonomous unities, is that he discovered that the two questions
have a common answer. He writes, "Living systems are cognitive systems, and FOREWORD ix
living as a process is a process of cognition."
We are very pleased to introduce this major theoretical work in the Boston INTRODUCTION (by Professor Maturana) xi
Studies in the Philosophy of Science. The integration of biological theory,
formal construction, epistemology (and, further, Maturana's suggestions of
the nature of interacting systems as a kind of biological sociology, and his BIOLOGY OF COGNITION
sketch of the ethical implications of such a construction) - all mark these
two studies as among the most original attempts at a systematic biology in Dedication 2
decades, and as a profo_undly philosophical work. Table of Contents 3
I. Introduction 5
Center for Philosophy and History of Science ROBERTS. COHEN II. The Problem 7
Boston University MARX W. WARTOFSKY III. Cognitive Function in General 8
July, 1979. A. The Observer 8
B. The Living System 9
c. Evolution 11
D. The Cognitive Process 13
IV. Cognitive Function in Particular 15
A. Nerve Cells 15
~· B. Architecture 19
c. Function 21
D. Representation 22
E. Description 26
F. Thinking 29
G. Natural Language 30
H. Memory and Learning 35
I. The Observer 38
V. Problems in the Neurophysiology of Cognition 41
VI. Conclusions 48
VII. Post Scriptum 57
 viii GENERAL TABLE OF CONTENTS

AUTOPOIESIS: THE ORGANIZATION OF THE LIVING

Preface (by Sir Stafford Beer) 63 FOREWORD

Introduction 73
I. On Machines, Living and Otherwise 77
\ ,, 1. Machines 77
',)) 2. Living Machines 78 ; Everything that I say in the introduction is my exclusive responsibility.
II. Dispensability of Teleonomy 85 Francisco Varela has been generous enough to grant me this freedom in what
1. Purposelessness 85 concerns the essay that we wrote together. His views about it he expresses
2. Individuality 87 fully and independently in his book Principles of Biological Autonomy,
III. Embodiments of Autopoiesis 88 published by Elsevier-North Holland, New York, 1979.
1. Descriptive and Causal Notions 88 Also, I wish to acknowledge the subtle debt which the contents of this
2. Molecular Embodiments 90 book owe to the many illuminating conversations about all the topics here
3. Origin 93 contained that I have had over more than fourteen years with my dear friend
IV. Diversity of Autopoiesis 96 Professor Felix Schwartzmann, of the University of Chile.
1. Subordination to the Condition of Unity 96
2. Plasticity of Ontogeny 98
3. Reproduction, a Complication of the Unity 100
4. Evolution, a Historical Network 102
5. Second and Third Order Autopoietic Systems 107
V. Presence of Autopoiesis 112
1. Biological Implications 112
2. Epistemological Implications 115
3. Cognitive Implications 118
Appendix: The Nervous System 124
Glossary 135

BIBLIOGRAPHY 139

INDEX OF NAMES 141

ix
 INTRODUCTION

In this introduction I wish to make some comments in relation to the two

essays that make this book, but in order to do so I must, at least in part, write
about how they came into being.

HISTORY

Since my childhood I have been interested in animals and plants, and I fre-
quently asked myself what made them living. Thus, in 1948, in my first year
as a medical student, I wrote a poem whose first stanza was:

"lQue es la muerte para el que la mira?

lQue es la muerte para el que la siente?
Pesadez ignota, incomprensible,
dolor que el egofsmo trae, para ese;
silencio, paz y nada, para este.
Sin embargo el uno siente
que su orgullo se rebela, que su mente
no soporta que tras la muerte nada quede,
que tras la muerte este la muerte.
El otro, en su paz, en su silencio,
en su majestad inconsciente siente,
nada siente, nada sabe,
porque la muerte es la muerte
y tras la muerte esta la vida
que sin la muerte solo es muerte."
Representation of the cellular autopoietic network.
All arrows that do not cross the boundary of the represented unity indicate production What is death for the beholder? I What is death for the dying? I A weight
relations. The uniformly shaded areas, including the boundary line and the wedges, beyond knowledge or understanding, I A pain for the self-asserting ego, for
together with the names, indicate constitutive relations. The geµeral form of closure with the one; I For the other, silence, peace, and nothingness. II Yet the one feels
respect to production and constitution in a recursive network realized as a concrete his pride in anger I And in his mind he does not accept I That beyond death
unity through the preferential relations of the components within the network, indicate
nothing should arise, I And that beyond death I There should be only death.
order relations and the consequent cleavage of the network as a simple unity from its
medium. The whole represents a closed network of productions, but the arrows across II The other, in his silence, I In his unknowing majesty feels, I He feels noth-
the depicted constitutive boundary of the network indicate the necessary material ing, he knows nothing, I Because death is death I And life without death is
openness of the system as it realizes the physical space. only emptiness. II
xi
xii INTRODUCTION INTRODUCTION xiii

The poem is not a very good one, yet it contains the implicit question: 'What insufficient knowledge. We had to accept that we could recognize living
kind of systems are living systems that they may die, and how come that they systems when we encountered them, but that we could not yet say what they
cognize?' were. I could enumerate features of living systems such as reproduction,
In 1954 I went to study biology in England and the U.S.A., and when I heredity, growth, irritability, and so on; but, how long a list was necessary?;
returned to Chile in 1960, after six years of study and research abroad, I when would the list be completed? In order to know when the list was com-
began to work at the Medical School of the University of Chile in Santiago as pleted I had to know what a living system was, which was, in fact, the question
a research associate in the Department of Biology. There I was involved in that I wanted to answer in the first place by producing such a list. I could
two kinds of activities; I collaborated in teaching a course in general biology speak about adaptation and evolution, about development and differentiation,
for the medical students, and I did research in the fields of neurophysiology and show how all these phenomena were tied together by the phenomenon of
and neuroanatomy. In my teaching my responsibility was to convey to the natural selection; but the question: 'What was the invariant feature of living
medical students some understanding of the organization of living systems as systems around which natural selection operated?', remained unanswered.
autonomous entities, as well as some understanding of their possible origin Every approach that I could attempt and that I did attempt left me at the
on earth. In my research I wanted to apply to the study of form and color starting point.
vision in birds the same approach that J. Y. Lettvin and I had used in the Yet I obviously had some inkling of what was the correct answere, because
study of form vision in the frog. I rejected the unsatisfactory ones. After several years of these various at-
I soon discovered through my teaching that the central question that the tempts I realized that .the difficulty was both epistemological and linguistic,
students would always ask was: 'What is proper to living systems that had its and that both my wife and my old professor, J. Z. Young, were right: ~me can
origin when they originated, and has remained invariant since then in the only say with a given language what the language permits. I had to stop look-
succession of their generations?' At the same time I soon realized in my ing at living systems as open systems defined in an environment, and I needed
research that my central purpose in the study of color vision could not be a language that would permit me to describe an autonomous system in a
the study of a mapping of a colorful world on the nervous system, but rather manner that retained autonomy as a feature of the system or entity specified
that it had to be the understanding of the participation of the retina (or by the description. In other words, any attempt to characterize living systems
nervous system) in the generation of the color space of the observer. with notions of purpose or function was doomed to fail because these notions
As a result of these different activities I entered a situation in which my are intrinsically referential and cannot be operationally used to characterize
academic life was divided, and I oriented myself in search of the answers to ,any system as an autonomous entity. Therefore, notions of purpose, goal, use
two questions that seemed to lead in opposite directions, namely: 'What is or function, had to be rejected, but initially I did not know how. Accordingly,
the organization of the living?' and 'What takes place in the phenomenon of I tried in my lectures several approaches in order to find a way of speaking
perception?' about living systems in a manner that would grasp their autonomy as a phe-
Let me speak about how I faced them. nomenon of their operation as unitary systems. Thus, eventually, I made a
distinction between what I called self-referred and allo-referred systems, a
First Question: What is the organization of the living? distinction that separated systems that could only be characterized with
When this question was first asked by the students, although it was the same reference tc;> themselves, such as living systems, from systems that could only
question that had been lurking in my mind for many years, I could not answer be characterized with reference to a context. I did this in order to emphasize
it. I had prepared myself for this moment, but when tt came and I tried to that whatever took place in living systems as living systems, took place as
answer it in a manner satisfactory for the students I realized that I had to necessarily and constitutively determined in relation to themselves because
think everything anew. I could speak about form and function and astonish their being defined as unities through self-reference was their manner of
my students and myself at the harmony of nature, exalting the fitness of the autonomy; and that whatever took place in other systems took place as
environment and the fitness of the individual. I could claim that the question constitutively determined in relation to the context with respect to which
was a verr difficult one and that it could not yet be answered due to our they were defined as unities. This way of speaking was not fully satisfactory
 .
xiv INTRODUCTION INTRODUCTION xv

but led me to realize that what was indeed needed was the characterization of
a kind of system which, if allowed to operate, would operate in a manner
.l to specify color distinctions. A different approach and a different epistem-
ology were necessary.
indistinguishable from the operation of living systems, and that one should There are many visual configurations, with uniform and variegated spectral
do so using only neighborhood relations realized through the properties of compositions, in simple and complex geometrical forms, that give rise to
the components of the system. It was with such aim that I spoke for the first indistinguishable color experiences. How should one, then, look for the
time in 1969 of living systems as systems defined as unities through the basic invariances in the activity of the nervous system, if any, in relation to the
circularity of their production of their components. perception of color? After we realized that the mapping of the external world
was an inadequate approach, we found that the very formulation of the ques-
Second question: What takes place in the phenomenon of perception?
tion gave us the clue. What if, instead of attempting to correlate the activity
When Jerry Y. Lettvin and I wrote our several articles on frog vision [Lettvin, in the retina with the physical stimuli external to the organism, we did other-
J. Y., H. R. Maturana, W. S. McCulloch and W. H. Pitts 1959; Maturana, wise, and tried to correlate the activity in the retina with the color experience
H. R., J. Y. Lettvin, W. S. McCulloch and W. H. Pitts 1960), we did it with of the subject?
the implicit assumption that we were handling a clearly defined cognitive Such a step entailed two difficulties. On the one hand it required the
situation: there was an.objective (absolute) reality, external to the animal, and definition of a reference which would permit the characterization of the
independent of it (not determined by it), which it could perceive (cognize), activity of the retinal ganglion cells independently of the stimulus as such;
and the animal could use the information obtained in its perception to com- on the other hand it required us to close the nervous system and treat the
pute a behavior adequate to the perceived situation. This assumption of ours report of the color experience as if it represented the state of the nervous
appeared clearly in our language. We described the various kinds of retinal system as a whole. In other words, the new approach required us to treat
ganglion cells as feature detectors, and we spoke about the detection of prey seriously the activity of the nervous system as determined by the nervous
and enemy. We knew that was not the whole neurophysiological story, as system itself, and not by the external world; thus the external world would
was apparent particularly in the discussion of the article called 'Anatomy only have a triggering role in the release of the internally-determined activity
and Physiology of Vision in the Frog (Rana pipiens)'. But even there the of the nervous system. We did this rigorously, and showed that such an
epistemology that guided our thinking and writing was that of an objective approach did indeed permit us to generate the whole color space of the
reality independent of the observer. Thus, when Samy Frenk and I began observer. That was a very fundamental result that we published in a very
to work with pigeons in 1961, first studying form vision, we approached .unknown article [Maturana, H. R., G. Uribe, and S. Frenk, 1968).
that study with the same fundamental view. No problem arose then and But what was still more fundamental was the discovery that one had to
without any difficulty we could characterize many classes of retinal ganglion close off the nervous system to account for its operation, and that perception
cells. Yet, when Gabriela Uribe joined us and we in fact began to study color should not be viewed as a grasping of an external reality, but rather as the
vision in 1964, it soon became apparent to us that that approach leads to specification of one, because no distinction was possible between perception
deep trouble. Neurophysiologically we did not see anything fundamentally and hallucination in the operation of the nervous system as a closed network.
different from what other scholars had already seen. We found the classic Although we arrived at this conclusion through the study of color vision,
types of ganglion cells with separate, concentric or overlapping opponent there are many earlier experimental studies (such as those of Stone on the
spectral preferences. But we also found: (a) that although the geometry rotation of the eye of the salamander in the early 'forties) that could also
of the receptive fields of the ganglion cells with opponent spectral prefer- have led to an understanding of the nervous system as a closed network of
ences had nothing to do with the geometry of the visual object, the geome- interacting neurons. Whether they did or not, I do not know; but if they
try of the visual object had to do with the response of those cells; and (b) did it seems that the implications were not pursued to their ultimate conse-
that we could not account for the manifold chromatic experiences of the quences.
observer by mapping the visible colorful world upon the activity of the Whatever the case, for me this finding had great significance and plunged
nervous system, because the nervous system seemed to use geometric relations me into the study of cognition as a legitimate biological problem. Two
xvi INTRODUCTION INTRODUCTION xvii

inmediate consequences arose from this: the first one was that in my neuro- system - its nervous system included when present - were the same thing.
physiological studies I had to take seriously the in,distingt1-ishabifily in the From this understanding the essay 'Biology of Cognition' arose as an expan-
operation of the nervous system between perceptio1;1 and hallucj11aJiQQ; the sion of my presentation in that symposium.
second one was that I needed a new language to talk about the phenomena of
perception and cognition. The first consequence required that the question:
'How does the organism obtain information about its environment?' be The Word 'Autopoiesis'
changed to: 'How does it happen that the organism has the structure that The second essay included in this book was written in 1972, as an expansion
permits it to operate adequately in the medium in which it exists?' A seman- of the section on 'Living Systems' in the 'Biology of Cognition'. The writing
tic question had to be changed into a structural question. The second ques- of this essay was in fact triggered by a conversation that Francisco Varela and
tion required the actual attempt to describe the phenomena that take place I had in which he said: "If indeed the circular organizaton is sufficient to
in the organism during the occurrence of the phenomena of perception and characterize living systems as unities, then one should be able to put it in
cognition in a language that retained them as phenomena proper to a closed more formal terms". I agreed, but said that a formalization could only come
nervous system. after a complete linguistic description, and we immediately began to work on
the complete description. Yet we were unhappy with the expression 'circular
A Congress in Anthropology organization', and we wanted a word that would by itself convey the central
feature of the organization of the living, which is autonomy. It was in these
Early in May of 1968 the University of Chile entered a state of revolution. circumstances that one day, while talking with a friend (Jose Bulnes) about
The students took over the University in an attempt to reformulate the an essay of his on Don Quixote de la Mancha, in which he analyzed Don
philosophy that had inspired its organization. I joined them. All standard Quixote's dilemma of whether to follow the path of arms (praxis, action) or
academic activities stopped and students and some members of the faculty the path of letters (poiesis, creation, production), and his eventual choice
tried to say something new. It was not easy. Language was a trap, but the of the path of praxis deferring any attempt at poiesis, I understood for the
whole experience was a wonderful school in which one could discover how first time the power of the word 'poiesis' and invented the word that we
mute, deaf and blind one was. It was easy to be caught in one's own ego, but needed: autopoiesis. This was a word without a history, a word that could
if one succeeded in attaining at least some degree of freedom from it, one directly mean what takes place in the dynamics of the autonomy proper to
began to listen and one's language began to change; and then, but only then, living systems. Curiously, but not surprisingly, the invention of this word
new things could be said. This lasted for several months. proved of great value. It simplified enormously the task of talking about the
In September of that year, I accepted an invitation to go to the University organization of the living without falling into the always gaping trap of not
of Illinois at Urbana to the Biological Computer Laboratory of Professor saying anything new because the language does not permit it. We could not
Heinz von Foerster. Furthermore, Professor von Foerster invited me to escape being immersed in a tradition, but with an adequate language we could
participate in a symposium sponsored by the Wenner-Gren Foundation for orient ourselves differently and, perhaps, from the new perspective generate a
Anthropological Research to be held during March 1969 in Chicago, with the new tradition.
purpose of considering the anthropology of cognition. The invitation was Let me now say something about the essays themselves.
to speak on 'The state of the art of the neurophysiology of cognition'. I
accepted this invitation and decided not to speak about neuronal circuits,
nerve impulses or synapses, but rather I decided to consider what should take BIOLOGY OF COGNITION
place in the organism during cognition by considering cognition as a biological When I wrote this essay I did not yet have the word 'autopoiesis', nor had I
phenomenon. In doing this I found that my two apparently contradictory come to the more formal expression of the living organization given in the
academic activities were not contradictory, and that they were, in fact, ad- next essay. Yet, these shortcomings do not detract from what is said because
dressed to the same phenomenon: cognition and the operation of the living the basic relations embodied in the notion of autopoiesis are fully implied,
xviii INTRODUCTION INTRODUCTION xix

although cumbersomely said, by expressions such as 'circular organization' and space. Notions of purpose, function or goal are unnecessary and misleading.
'self-referential systems'. Furthermore, when I wrote the essay I decided not This the essay does, and in this respect no commentary is needed; the essay
to make any concession to existing notions that I considered inadequate or stands by itself. Yet, when we wrote it we were just beginning to realize the
misleading, even if this seemed to make the text particularly obscure. How- fundamental distinction between organization and structure, and we do not
ever, I made a concession which I have always \regretted~ I submitted to the separate the terms with complete rigor. Also, because it was not written
pressure of my friends and\talked about causal relations)when speaking about under the supposition that the reader had read the 'Biology of Cognition', the
the circular organization or" living systems. To do this was both inadequate essay is not fully clear concerning the validity of the statement "Everything
and misleading. It was inadequate because the notion of causality is a notion said is said by an observer" in relation to the distinction between characteri-
that pertains to the domain of descriptions, and as such it is relevant only in zation and description of a system. I shall make some comments about these
the metadomain in which the observer makes his commentaries and cannot themes.
be deemed to be operative in the phenomenal domain, the object of the
description. It was misleading because it obscured the actual appreciation of
Unity, Organization and Stntcture
the sufficiency of the notion of property as defined by _the distinctive opera-
tion performed by the observer when specifying a unity, for the description of Unity. The basic cognitive operation that we perform as observers is the
the phenomenal domains generated by the specified unities. It was misleading operation of distinction. By means of this operation we specify a unity as an
because it obscured the understanding of the dependency of the identity of entity distinct from a background, characterize both unity and background
the unity on the distinctive operation that specified it. It was misleading with the properties with which this operation endows them, and specify their
because it obscured both the understanding of the phenomenal domains as separability. A unity thus specified is a simple unity that defines through its
determined by the properties of the unities that generate them, and the properties the space in which it exists and the phenomenal domain which it
non-intersection of the phenomenal domains generated by the operation of may generate in its interactions with other unities.
a composite unity as a simple unity in a medium and by the operation of its If we recursively apply the operation of distinction to a unity, so that we
components as components. distinguish components in it, we respecify it as a composite unity that exists
There is nothing else that I wish to add as a commentary to this essay. It in the space that its components define because it is through the specified
is a cosmology and as such it is complete. Finally I wish to say that I find it properties of its components that we observers distinguish it. Yet we can
pervading my views and understanding of everything. In a sense it has been always treat a composite unity as a simple unity that does not exist in the
my way to. transcendental experience: to the discovery that matter, met;i space of its components, but which exists in a space that it defines through
phoricall~ speaking'. is the.creation of the spirit (the.~o~e ofexiste~ce ofthj the properties that characterize it as a simple unity. Accordingly, if an auto-
observer m·a domam of discourse), and that the spmt 1s the creat10n of the poietic system is treated as a composite unity, it exists in the space defined
matter it creates. This is not a paradox, but it is the expression of our exi~ by its components, but if it is treated as a simple unity the distinctions that
ence in a domain of cognition in which the content of cognition is cognition specify it as a simple unity characterize its properties as a simple unity, and
itself. Beyond that nothing can be said. defme the space in which it exists as such a simple unity.

Organization and Stntcture. The relations between components that define

AUTOPOIESIS
a composite unity (system) as a composite unity of a particular kind, con-
This article was written as an expansion of two pages of the 'Biology of stitute its organization. In this definition of organization the components are
Cognition', in an attempt to show that autopoiesis is necessary and sufficient viewed only in relation to their participation in the constitution of the unity
to clmacterize the organization of living systems, and that given the proper (whole) that they integrate. This is why nothing is said in it about the proper-
historical contingencies one can derive all the biological phenomenology from ties that the components of a particular unity may have other than those
the characterization of living systems as autopoietic systems in the physical required by the realization of the organization of the unity.
XX INTRODUCTION INTRODUCTION xxi

The actual components (all their properties included) and the actual rela- undergoes structural changes triggered and selected through its recurrent inter-
tions holding between them that concretely realize a system as a particular actions in its medium, that is, its adaptation is conserved, then the outcome
member of the class (kind) of composite unities to which it belongs by its of this history of interactions is the selection, by the recurrent or changing
organization, constitute its structure. Therefore, the organization of a system structural configuration of the medium, of a sequence of structural changes in
as the set of relations between its components that define it as a system of a the composite unity, which results in that the changing structure of the organ-
particular class, is a subset of the relations included in its structure. It follows ism follows the changing structure of the medium through a continued struc-
that any given organization may be realized through many different struc- tural coupling to it. If organization or adaptation are not conserved, then the
tures, and that different subsets of relations included in the structure of a outcome for the composite unity is disintegration. In other words, if a com-
given entity, may be abstracted by an observer (or its operational equivalent) posite unity is structurally plastic its conservation of adaptation results in its
as organizations that define different classes of composite unities. maintained structural coupling to the medium that selects its path of structural
The organization of a system, then, specifies the class identity of a system, change. In this process, the configuration of constitutive relations that remain
and must remain invariant for the class identity of the system to remain invariant in the adapted composite unity, determines the matrix of possible
invariant: if the organization of a system changes, then its identity changes perturbations that the composite unity admits at any instant, and, hence,
and it becomes a unity of another kind. Yet, since a particular organization operates as a reference for the selection of the path of structural changes that
can be realized by systems with otherwise different structures, the identity of take place in it in its history of interactions. Defined in this manner, structural
a system may stay invariant while its stiucture changes within limits deter- coupling (conservation of adaptation) is not peculiar to living systems. It is a
mined by its organization. If these limits are overstepped, that is, if the phenomenon that takes place whenever a plastic composite unity undergoes
structure of the system changes so that its organization cannot any more be recurrent interactions with structural change but without loss of organization,
realized, the system loses its identity and the entity becomes something else, which may follow any changing or recurrent structural configuration of its
a unity defined by another organization. domain of interactions (medium). Therefore, all that is unique with respect
It is apparent that only a composite unity has structure and organization, to adaptation in living systems is that in them the autopoietic organization
a simple unity does not. A simple unity only has the properties with which constitutes the invariant configuration of relations around which the selection
it is endowed by the operations of distinction through which it becomes of their structural changes takes place during their history of interactions.
separated from a background. It is also apparent that as soon as a composite
unity is treated as ·a simple unity, any question about the origin of its proper-
Epistemology
ties becomes inadequate because the properties of a simple unity are given
through its distinction as a simple unity. It is also apparent that although the As soon as a unity is specified, a phenomenal domain is defined. Accordingly,
properties of a composite unity as a simple unity arise from its organization, if a composite unity operates as a siI11ple unity, it operates in a phenomenal
they are realized through the properties of its components. Accordingly, while domain that it defines as a simple unity that is necessarily different from
two simple unities interact through the simple interplay of their properties, the phenomenal domain in which its components operate. Therefore, the
two composite unities interact in a manner determined by their organization emergence of a phenomenal domain as the result of the operational distinc-
and structure through the interplay of the properties of their components. tion of a composite unity as a simple unity, makes phenomenal reductionism
(and, hence, explanatory reductionism) impossible. Furthermore, the dy-
namics of the establishment of unities through operational distinctions that
Stmctural Coupling specify their properties, have the result that all phenomenal domains are
In the history of interactions of a composite unity in its medium, both unity necessarily realized through the operation (interplay) of the properties of
and .medium operate in each interaction as independent systems that, by the unities that generate them, that is, through relations of contiguity. If a
triggering in each other a structural change, select in each other a structural component 'A' through its interaction with a component 'B' triggers an
change. If the organization of a composite unity remains invariant while it interaction of 'B' with 'C' that triggers a reduction in the production of 'D',
xxii INTRODUCTION INTRODUCTION xxiii

an observer may say by considering the whole that 'A' controls the produc- of a unity, either by pointing to its properties if it is a simple unity, or by
tion of 'D'. 'A', 'B', 'C' and 'D', interact through relations of contiguity in pointing to its organization if it is a composite one, from the knowledge
the phenomenal domain that the components define. Relations such as about a unity that he reveals, either by describing its operation as a simple
regulation, control or function, therefore, are not relations of contiguity; unity if it is a simple unity, or by describing both its operation as a simple
they are referential relations specified by the observer who puts himself in a unity and the operation of its components if it is a composite entity. In either
metadomain of descriptions by using his view of the whole as a reference for case, however, the knowledge that an observer claims of the unities that he
his description of the participation of the components that he describes in distinguishes consists in his handling of them in a metadomain of descrip-
the constitution of the composite unity. tions with respect to the domain in which he characterizes them. Or, in other
"Everything said is said by an observer." This I say in the 'Biology of words, an observer characterizes a unity by stating the conditions in which it
Cognition'. The fundamental cognitive operation that an observer performs exists as a distinguishable entity, but he cognizes it only to the extent that
is the operation of distinction. By means of this operation the observer spec- he defines a metadomain in which he can operate with the entity that he
ifies a unity as an entity distinct from a background and a background as the characterized.
domain in which an entity is distinguished. An operation of distinction, how- Thus, autopoiesis in the physical space characterizes living systems because
ever, is also a prescription of a procedure which, if carried out, severs a unity it determines the distinctions that we can perform in our interactions when
from a background, regardless of the procedure of distinction and regardless we specify them, but we know them only as long as we can both operate with
of whether the procedure is carried out by an observer or by another entity. their internal dynamics of states as composite unities and interact with them
Furthermore, the prescriptiveness of an operation of distinction implies a as simple unities in the environment in which we behold them. The fact that
universal phenomenalism of distinctions which, through the specification of the characterization of an entity is also a description made by the observer,
new procedures of distinction or through their recursive application in the and as such also belongs to his descriptive domain ('Biology of Cognition'),
reordering of the distinguished entities, can, in principle, endlessly give rise does not invalidate the operational effectiveness of the distinctions upon
to new simple and composite unities, and, hence, to new non-intersecting distinctions that constitutes the metadomain of descriptions in which the
phenomenal domains. Thus, although a distinction performed by an observer cognitive statements are made. The entity characterized is a cognitive entity,
is a cognitive distinction and, strictly, the unity thus specified exists in his but once it is characterized the characterization is also subject to cognitive
cognitive domain as a description, the observer in his discourse specifies a distinctions valid in the metadomain in which they are made by treating the
metadomain of descriptions from the perspective of which he establishes a characterization as an independent entity subject to contextual descriptions.
reference that allows him to speak as if a unity, simple or composite, existed Therefore, the complementarities system/environment, autonomy/control,
as a separate entity that he can characterize by denoting or connoting the totality/composition, etc., are complementarities only in our cognition of a
operations that must be performed to distinguish it. system that we observe in a context that allows us to establish such relations,
In the perspective of a descriptive metadomain the distinction between the but they are not constitutive features of the referred system because they do
ch11r11~t~rization of a unity and the observer's knowledge. of it that permits not participate in its constitution through the interplay of the properties of
him to describe it in a context, should be clear. In fact, knowledge always its components. Accordingly, that one should not be able to account for or
implies a concrete or conceptual action in some domain, and the recognition deduce all actual biological phenomena from the notion of autopoiesis with-
of knowledge always implies an observer that beholds the action from a meta- out resorting to historical contingencies, is not a shortcoming of such a notion.
domain. Therefore, when an observer claims knowledge of a system, he claims On the contrary, it is to be expected because the notion of autopoiesis only
that he can define a metadomain from the perspective of which he can simul- characterizes living systems as autonomous entities that can be distinguished
taneously behold the system as a simple unity, describing its interactions and as composite unities realized through neighborhood relations.
relations as a simple unity, and its components as components, describing Finally, I would like to add some sociological and ethical comments that
their interactions and relations as components. In these circumstances it is I consider follow from the understanding of the autopoietic organization
legitimate to distinguish between the characterization that an observer makes of living systems. The essay on autopoiesis was supposed to have a second
xxiv INTRODUCTION INTRODUCTION XXV

appendix on social and ethical implications. This appendix, however, was that the relations stated above characterize the organization of a social system
never included because Francisco Varela and I never agreed about its con- as a system, and that all the phenomena proper to social systems arise from
tents. Now I shall use the privilege that I have by writing this introduction in this organization. If one accepts this proposition, and I shall henceforth pro-
order to present the notions that I would have included in that appendix. ceed as if it were accepted, then one has to accept the following implications:
(i) The realization of the autopoiesis of the components of a social sys-
tem is constitutive to the realization of the social system itself. This cannot be
SOCIETY AND ETHICS ignored in any consideration about the operation of a social system without
The central feature of human existence is its occurrence in a linguistic cogni- negating it.
tive domain. This domain is constitutively social. Yet, what is a social system? (ii) A collection of living systems integrating a composite unity through
how is a social system characterized? how do living systems in general, and relations that do not involve their autopoiesis is not a social system, and the
human beings in particular, participate in the constitution of the social system phenomena proper to its operation as such a composite unity are not social
that they integrate? The answers to these questions are central for the under- phenomena. Since the operational coupling of an organism integrating a com-
standing of social dynamics and the process of social change. The following posite unity does not necessarily involve all its properties, an observer may
considerations state the essence of my answers to these questions: see an organism as integrating simultaneously several composite unities that
(I) It is apparent that natural social systems as systems constituted by living may or may not all be social systems.
systems require these for their actual realization. What is not apparent, how- (iii) The structure of a society as a particular social system is deter-
ever, is the extent to which the coupling of living systems in the integration mined both by the structure of its autopoietic components and by the actual
of a social system entails the realization of their autopoiesis.lfr the realization relations that hold between them while they integrate it. Therefore, the· do-
of a social system were to entail the autopoiesis of its comicinents, then the main of social phenomena, defined as the domain of the interactions and the
realization of the autopoiesis of the components of a social system w;~ld be relations that an observer sees taking place between the components of a
intrinsically indispensable and constitutive of its operation as such and not a society, results from the autopoietic operation of the components of the
mere contingenci{Accordingly, any particular feature, cultural or 'otherwise, society while they realize it in the intifrplay of their properties.
of the manner in which the autopoiesis of the components is realized and (iv) In a society, at any instance of observation, the structures of the
their individuality a~d autonomy is restricted, would be a feature of the pecul- components determine the properties of the components, the properties of
.iar social system (society) in question, and not intrinsic to it as a social system. the components realize the structure of the society, and the structure of the
Qr, however, the autopoiesis of the components of a natural social system were society operates as a selector of the structure of its components by being a
not involved in its constitution because the relations that define a system as medium in which they realize their ontogeny.
social do not entail them, then the autopoiesis of the components (and hence (v) An autopoietic system participates in the constitution of a social
their autonomy and individuality) would be intrinsically dispensableJ system only to the extent that it participates in it, that is, only as it realizes
(.
(2) The question, 'What is a social system?' cannot be answered by simply the relations proper to a component of the social system. Accordingly, in
l:' principle, an autopoietic system may enter or leave a social system at any
describing a particular one because we do not know the significant relations
that we must abstract when characterizing its organization. The question I'
Ji
moment by just satisfying or not satisfying the proper relations, and may
participate simultaneously or in succession in many different ones.
must be answered by proposing a system which, if allowed to operate, would
generate a phenomenal domain indistinguishable from the phenomenal
domain proper to a natural social system. Accordingly, I propose that a
It In what follows I shall pursue the consequences of these. notions, and
whenever I speak of a social system or of a society as a social system of a
@ollection of autopoietic systems that, through the realization of their auto- particular kind, I speak of a system defined as a system by the organization
poiesis, interact with each other constituting and integrating a system that
operates as the (or as a) medium in which t!1ey realize their autopoiesis, is
indistinguishable from a natural social systerr:}Or, in other words, I propose
I
l
proposed above.
(3) A society defines the domain in which it is realized as a unity. Such a
domain may or may not include the components of the society itself, and

I
I
xxvi INTRODUCTION INTRODUCTION xxvii

may or may not include other societies, but in any case it constitutes an change of its components, which, thus, become structurally coupled to it. In
operationally independent medium that operates as: (a) a selector of the path our case, we as social beings generate, through our structure-determined
of structural change that the society follows in its individual history, and properties, our societies as the cultural media that select our individual paths
(b ), if stable, a historical stabilizer of the structures that realize the selected of ontogenic change in a manner that leads each one of us to the structure
invariant relations that define the society as a particular social system. The that makes us generate the particular societies to which we belong. A .society,
more varied the medium of selection, the greater the domain of stabilized therefore, operates as a homeostatic system that stabilizes the relations that
relations in the society and, hence, the more fixed the structure of its com- define it as a social system of a particular kind. -
ponents. If the society is a human society this takes place in a language- (7) In general, the domain of states of a system as a composite unity is
ceniered culture, and the stabilization of the structure of the human com- determined by the properties of its components that realize its organization.
ponents is realized through a cultural stabilization of the relations that they If some of these properties change because the structure of some of the com-
must satisfy as social entities. ponents changes, then, while the system either changes its properties without
(4) To the extent that human beings are autopoietic systems, all their change of organization or disintegrates becoming something else, the changed
activities as social organisms must satisfy their autopoiesis. This they do in components either integrate the system in a different manner or uncouple
the social domain through the fulfilment of the basic biological preferences from it. This also applies to social systems, including human societies, because
(states of pleasure) and rejections (states of displeasure) that constitute the it is the actual interplay of the properties of the components that constitutes
inmediate experiential domain in which they, as components of a society, a social system as an actual system in the space in which these exist. In these
necessarily realize their individual worlds and contribute to the determination circumstances, a change in the relations that define a society as a particular
of the individual worlds of others. In man as a social being, therefore, all social system can only take place through a change in the properties of the
actions, however individual as expressions of preferences or rejections, con- components that realize it. It follows that in a human society a social change
stitutively affect the lives of other human beings and, hence, have ethical can only take place if the individual properties, and, hence, conduct, of its
significance. members change.
(5) What determines the constitution of a social system are the recurrent (8) All that matters for the realization of a society is that the component
interactions of the same autopoietic systems. In other words, any biological autopoietic systems should satisfy certain relations regardless of the actual
stabilization of the structures of the intera~ting organisms that results in the structures (internal processes) through which they realize them. Accordingly,
recurrence of their.interactions, may generate a social system. Among human hypocrisy plays an important role in the realization of human societies, per-
beings the basic stabilizing factor in the constitution of a social system is the mitting human beings under stress to feign having certain properties which
phenomenon of l_2_y_e, the seeing of the other as a partner in some or all the they abandon as soon as the stress is removed. This is why in a human society
dimensions of living. In these circumstances, when a human being makes the a social change takes place as a pennanent phenomenon only to the extent
choice of a particular way of living, apparent in his realization of a particular that it is a cultural change: a revolution is a revolution only if it is an ethical
set of social relations, he makes a basic ethical choice in which he validates a revolution.
world for himself and for others that he has explicitly or implicitly accepted (9) Interactions within a society are necessarily confinnatory of the rela-
as partners in living. Accordingly, the fundamental ethical problem that a tions that define it as a particular social system; if not, the organisms that
man faces as an observer-member of a society is the ethical justification of interact do not interact as components of the society which they otherwise
the particular relations of surrender of autonomy and individuality that he integrate. It is only through interactions operationally not defined within
demands from himself and from other members of the society that he gen- the society that a component organism can undergo interactions that lead
erates and validates with his conduct. to the selection, in its ontogeny, of a path of structural change not confirma-
(6) A social system is essentially a conservative system. This is so because tory of the society that it integrates. This is why se>~illl c;!~ittjvity, as the
it is generated through the interactions of structure-detennined autopoietic generation of novel social relations, always entails interactions operationally
systems and operates as a medium that selects the path of ontogenic structural outside the society, and necessarily leads to the generation, by the creative
xxviii INTRODUCTION INTRODUCTION xxix

individuals, of modes of conduct that either change the defining relations of cipate in it as social beings. A humari being that through his interactions with
the society as a particular social system, or separate them from it. S_o_cjal other human beings participates in interactions proper to their social system in
creativity is necessarily anti~Qc::ial in the social domain in which it talces place. a manner that does not involve his autopoiesis as a constitutive feature of it,
(10) In general any organism, and in particular any human being, can be is being used by the social ;ystem but is not one of its members. If the human
simultaneously a member of many social systems, such as a family, a club, being cannot escape from this situation because his life is at stake, he is under
an army, a political party, a religion or a nation, and can operate in one or social abuse.
another without necessarily being in internal contradiction. A human being (13) All kinds of societies are biologically legitimate. Yet not all are
operating as an observer, however, can always define a metadomain from the equally desirable as systems in which an observer human being may wish to
perspective of which he may see his participation in the various social systems live. The capacity that man has as a language-centered social being of becom-
that he integrates, and find it contradictory. Conduct as observer by a human ing an observer, and thus of operating as if he were external to the situation
being implies that he stands operationally as if outside the various social in which he finds himself, allows him, if he has the proper experiences, to
systems that he otherwise integrates, and that he may undergo in this manner contemplate the societies that he integrates and to like or dislike them. If
interactions that do not confirm them. An observer always js po_tentially the observer human being defines a metadomain from the perspective of
antisocial. · which some of the defining relations of the society are undesirable, he dislikes
(11) To grow as a member of a society consists in becoming structurally it, and if he acts acco;dingly he becomes antisocial and may come to validate
coupled to it; to be structurally coupled to a society consists in having the another society with his conduct. A totalitarian society restricts this possi-
structures that lead to the behavioral confirmation of the society. The spon- bility either by specifying the experiences that its components may have, so
taneous course of historical structural transformation of a society as a unity is that they do not operate as observers, or by uncoupling tl:te. dissidents so that
toward its structural coupling to the medium in which it exists, and, there- they may not seduce others to be observers as themselves~owever, there are
fore, toward the stabilization of the mechanisms that generate its defining certain experiences that cannot be fully specified in a human society without
relations through the stabilization of the properties of its components. In the destroying the basic individual structural plasticity needed for the establish-
domain of human societies this means the stabilization of human conduct. ment of consensual domains and the generation of language _imd, hence, for
But, the stabilization of human conduct always entails a restriction of crea- human creativity in general. Love is one of these experiences,]and as long. as
tivity through a restriction of the possible interactions of the individual man has a language he can become an observer through the experience of
human beings outside those prescribed by the society that they integrate. love.
The extreme case of this, of course, takes place in a totalitarian society of any (14) When a human being 'A' encounters another human being 'B' and
kind. Or, in other words, the spontaneous course of the historical transforma- loves him or her, he sees 'B' in a social context and becomes an observer
tion of a human society as a unity is towards totalitarianism; this is so because of the society that 'B' integrates. 'A' may like or may not like what he sees
the relations that undergo historical stabilization are those that have to do in reference to 'B' and act accordingly, becoming antisocial if he does not
with the stability of the society as a unity in a given medium, and not with like what he sees. An absolute totalitarian society must negate love as an
the well-being of its component human beings that may operate as observers. individual experience because love, sooner or later, leads to an ethical evalua-
Any other course requfre.s_iin eJmc::aj choice; it would not be spontaneous, it tion of the society that the loved one integrates.
would be a work of art, a product of human aesthetic design. If human beings (15) A human society in which to see all human beings as equivalent to
were not observers, or capable of being so, the stabilization of their properties oneself, and to love them, is operationally legitimate without demanding
would not appear to matter because they would not be able to desire some- from them a larger surrender of individuality and autonomy than the measure
thing else. that one is willing to accept for oneself while integrating it as an observer, is a
(12) We as human beings exist in a network of social systems and move product of human art, that is, an artificial society that admits change and
from one to another in our daily activities. Yet, not all human beings caught accepts every human being as not dispensable. Such a society is necessarily a
in the mesh of relations generated in this network of social systems parti- non-hierarchical society for which all relations of order are constitutively
XXX INTRODUCTION

BIOLOGY OF COGNITION
transitory and circumstantial to the creation of relations that continuously
negate the institutionalization of human abuse. Such a society is in its essence (1970)
an anarchist society, a society made for and by observers that would not
surrender their condition of observers as their only claim to social freedom
By
and mutual respect.
HUMBERTO R. MATURANA
At this point there is either much more to say, or nothing - therefore, let us
let the reader judge. Thank you.
 TABLE OF CONTENTS

DEDICATION 2

I. INTRODUCTION 5

II. THE PROBLEM 7

Ill. COGNITIVE FUNCTION IN GENERAL 8

A. The Observer 8
B. The Living System 9
c. Evolution 11
D. The Cognitive Process 13

IV. COGNITIVE FUNCTION IN PARTICULAR 15

A. Nerve Cells 15
B. Architecture 19
c. Function 21
D. Representation 22
E. Description 26
F. Thinking 29
G. Natural Language 30
I dedicate this book to my wife, Maria H. Memory and Learning 35
Montanez Luna. I could not have written I. The Observer 39
it if I had not lived as I have lived, and for V. PROBLEMS IN THE NEUROPHYSIOLOGY OF COGNITION 41
the most part of my existence the thread of
VI. CONCLUSIONS 48
my being has been and is braided with hers.
Therefore, strictly, although I wrote this VlI. POST SCRIPTUM 57
book, she is, in every respect, as much its
author as I am. This I wish both to acknowl-
edge and to thank with deep love.
 HUMBERTO R. MATURANA

BIOLOGY OF COGNITION

I. INTRODUCTION

Man knows and his capacity to know depends on his biological integrity;
furthermore, he knows that he knows. As a basic psychological and, hence,
biological function cognition guides his handling of the universe and know-
ledge gives certainty to his acts; objective knowledge seems possible and
through objective knowledge the universe appears systematic and predictable.
Yet knowledge as an experience is something personal and private that cannot
be transferred, and that which one believes to be transferable, objective
knowledge, must always be created by the listener: the listener understands,
and objective knowledge appears transferred, only if he is prepared to under-
stand. Thus cognition as a biological function is such that the answer to the
question, 'What is cognition?' must arise from understanding knowledge and
the knower through the latter's capacity to know.
Such is my endeavor.

Epistemology
The basic claim of science is objectivity: it attempts, through the application
of a well defined methodology, to make statements about the universe. At
the very root of this claim, however, lies its weakness: the a priori assumption
that objective knowledge constitutes a description of that which is known.
Such assumption begs the questions, 'What is it to know?' and 'How do we
know?'.

Biology
(a) The greatest hindrance in the understanding of the living organization lies
in the impossibility of accounting for it by the enumeration of its properties;
it must be understood as a unity. But if the organism is a unity, in what sense
are its component properties it parts? The organismic approach does not an-
swer this question, it merely restates it by insisting that there are elements of
organization that subordinate each part to the whole and make the organism
5
6 HUMBERTO R. MATURANA

a unity [Cf. Bertalanffy, 1960]. The questions 'How does this unity arise?'
and 'To what extent must it be considered a property of the organization
of the organism, as opposed to a property emerging from its mode of life?' II. THE PROBLEM
remain open. A similar difficulty exists for the understanding of the func-
tional organization of the nervous system, particularly if one considers the
higher functions of man. Enumeration of the transfer functions of all nerve
cells would leave us with a list, but not with a system capable of abstract (1) Cognition is a biological phenomenon and can only be understood as
thinking, description, and self-description. Such an approach would beg the such; any epistemological insight into the domain of knowledge
question, 'How does the living organization give rise to cognition in general requires this understanding.
and to self cognition in particular?' (2) If such an insight is to be attained, two questions must be considered:

(b) Organisms are adapted to their environments, and it has appeared ad- What is cognition as a function?
equate to say of them that their organization represents the 'environment'
in which they live, and that through evolution they have accumulated in- What is cognition as a process?
formation about it, coded in their nervous systems. Similarly it has been said
that the sense organs gather information about the 'environment', and through What follows should answer these two questions.
learning this information is coded in the nervous system [Cf. Young, 1967].
Yet this general view begs the questions, 'What does it mean to "gather
information"?' and 'What is coded in the genetic and nervous systems?'.
A successful theory of cognition would answer both the epistemological
and the biological questions. This I propose to do, and the purpose of this
essay is to put forward a theory of cognition that should provide an epis-
temological insight into the phenomenon of cognition, and an adequate view
of the functional organization of the cognizant organism that gives rise to
such phenomena as conceptual thinking, language, and self-consciousness.
In what follows I shall not offer any formal definitions for the various
terms used, such as 'cognition', 'life', or 'interaction', but I shall let their
meaning appear through their usage. This I shall do because I am confident
that the internal consistency of the theory will show that these terms indeed
adequately refer to the phenomena I am trying to account for, and because
I speak as an observer, and the validity of what I say at any moment has its
foundation in the validity of the whole theory, which, I assert, explains why
I can say it. Accordingly, I expect the complete work to give foundation to
each of its parts, which thus appear justified only in tl).e perspective of the
whole.

Note: I shall be speaking of the organism as a unity, but when I wrote this essay I was
not aware that the word unit did not always quite mean unity. Since I cannot now
correct this. I beg the reader to bear this in mind.
7
 COGNITIVE FUNCTION IN GENERAL 9

(7) The observer is a living system and an understanding of cognition as a

biological phenomenon must account for the observer and his role in it.
III. COGNITIVE FUNCTION IN GENERAL
THE LIVING SYSTEM

(1) Living systems are units of interactions; they exist in an ambience. From a
THE OBSERVER purely biological point of view they cannot be understood independently of
that part of the ambience with which they interact: the niche; nor can the
(1) Anything said is said .by an observer. In his discourse the observer speaks niche be defined independently of the living system that specifies it.
to another observer, who could be himself; whatever applies to the one (2) Living systems as they exist on earth today are characterized by
applies to the other as well. The observer is a human being, that is, a living exergonic metabolism, growth and internal molecular replication, all organized
system, and whatever applies to living systems applies also to him. in a closed causal circular process that allows for evolutionary change in the
(2) The observer beholds simultaneously the entity that he considers way the circularity is maintained, but not for the loss of the circularity itself.
(an organism, in our case) and the universe in which it lies (the organism's Exergonic metabolism is required to provide energy for the endergonic
environment). This allows him to interact independently with both and to synthesis of specific polymers (proteins, nucleic acids, lipids, polysaccharides)
have interactions that are necessarily outside the domain of interactions of from the corresponding monomers, that is, for growth and replication; special
the observed entity. replication procedures secure that the polymers synthesized be specific, that
(3) It is an attribute of the observer to be able to interact independently they should have the monomeric sequence proper to their class; specific
with the observed entity and with its relations; for him both are units of polymers (enzymes) are required for the exergonic metabolism and the
interaction (entities). synthesis of specific polymers (proteins, nucleic acids, lipids, polysaccharides)
(4) For the observer an entity is an entity when he can describe it. To [Cf. Commoner, 1965].
describe is to enumerate the actual or potential interactions and relations of This circular organization constitutes a homeostatic system whose function
the described entity. Accordingly, the observer can describe an entity only if is to produce and maintain this very same circular organization by determin-
there is at least one other entity from which he can distinguish it and with ing that the components that specify it be those whose synthesis or
which he can observe it to interact or relate. This second entity that serves maintenance it secures. Furthermore, this circular organization defines a
as a reference for the description can be any entity, but the ultimate reference living system as a unit of interactions and is essential for its maintenance
for any description is the observer himself. i as a unit; that which is not in it is external to it or does not exist. The circular
(5) The set of all interactions into which an entity can enter is its domain
of interactions. The set of all relations (interactions through the observer) in
which an entity can be observed is its domain of relations. This latter domain
l
t
organization in which the components that specify it are those whose syn-
thesis or maintenance it secures in a manner such that the product of their
functioning is the same functioning organization that produces them, is the

I
lies within the cognitive domain of the observer. An entity is an entity if it living organization.
has a domain of interactions, and if this domain includes interactions with the (3) It is the circularity of its organization that makes a living system a unit
observer who can specify for it a domain of relations. The observer can define of interactions, and it is this circularity that it must maintain in order to
an entity by specifying its domain of interactions; thus part of an entity, a l remain a living system and to retain its identity through different interactions.
group of entities, or their relations, can be made units of interactions (entities) i All the peculiar aspects of the different kinds of organisms are superimposed
by the observer. li on this basic circularity and are subservient to it, securing its continuance
(6) The observer can define himself as an entity by specifying his own ! through successive interactions in an always changing environment. A living
domain of interactions; he can always remain an observer of these interact- I system defines through its organization the domain of all interactions into
ions, which he can treat as independent entities. which it can possibly enter without losing its identity, and it maintains
its identity only as long as the basic circularity that defines it as a unit of
8

I
10 HUMBER TO R. MATURANA COGNITIVE FUNCTION IN GENERAL 11

interactions remains unbroken. Strictly, the identity of a unit of interactions there are always parts of the environment that lie beyond any possibility
that otherwise changes continuously is maintained only with respect to the of intersection with the domain of interactions of the organism, and there
observer, for whom its character as a unit of interactions remains unchanged. ,, are parts of the niche that lie beyond any possibility of intersection with

l
(4) Due to the circular nature of its organization a living system has a self- the domain of interactions of the observer. Thus for every living system its
referring domain of interactions (it is a self-referring system), and its condition organization implies a prediction of a niche, and the niche thus predicted as
l.
of being a unit of interactions is maintained because its organization has a domain of classes of interactions constitutes its entire cognitive reality.
functional significance only in. relation to the maintenance of its circularity ' If an organism interacts in a manner not prescribed by its organization, it
and defines its domain of interactions accordingly. does so as something different from the unit of interactions defined by its
(5) Living systems as units of interactions specified by their condition of basic circularity, and this interaction remains outside its cognitive domain,
being living systems cannot enter into interactions that are not specified by although it may well lie within the cognitive domain of the observer.
their organization. T-he circularity of their organization continuously brings (7) Every unit of interactions can participate in interactions relevant to
them back to the same internal state (same with respect to the cyclic process). other, more encompassing units of interactions. If in doing this a living
Each internal state requires that certain conditions (interactions with the system does not lose its identity, its niche may evolve to be contained by the
environment) be satisfied in order to proceed to the next state. Thus, the larger unit of interactions and thus be subservient to it. If this larger unit of
circular organization implies the prediction that an interaction that took interactions is ( or becomes) in turn also a self-referring system in which
place once will take place again. If this does not happen the system dis- its components (themselves self-referring systems) are subservient to its
integrates; if the predicted interaction does take place, the system maintains maintenance as a unit of interactions, then it must itself be ( or .become)
its integrity (identity with respect to the observer) and enters into a new subservient to the maintenance of the circular organization of its components.
prediction. In a continuously changing environment these predictions can Thus, a particular self-referring system may have the circular organization
only be successful if the environement does not change in that which is of a living system or partake functionally of the circular organization of its
predicted. Accordingly, the predictions implied in the organization of the components, or both. The society of bees (the honey producing bees) is an
living system are not -predictions of particular events, but of classes of inter- example of a third order self-referring system of this kind; it has a circular
actions. Every interaction is a particular interaction, but every prediction is a organization superimposed on the second order self-referring systems that
prediction of a class of interactions that is defined by those features of its are the bees, which in turn have a circular organization superimposed on
elements that will allow the living system to retain its circular organization the first order living systems that are the cells; all three systems with their
after the interaction, and thus, to interact again. This makes living systems ; domains of interactions are subordinated both to the maintenance of them-

I
inferential systems, and their domain of interactions a cognitive domain. selves and to the maintenance of the others.
(6) The niche is defined by the classes of interactions into which an
organism can enter. The environment is defined by the classes of interactions
EVOLUTION
into which the observer can enter and which he treats as a context for his
interactions with the observed organism. The observer beholds organism and t (1) Evolutionary change in living systems is the result of that aspect of their
environment simultaneously and he considers as the niche of the organism circular organization which secures the maintenance of their basic circularity,
that part of the environment which he observes to lie in its domain of inter- allowing in each reproductive step for changes in the way this circularity is
actions. Accordingly, as for the observer the niche appears as part of the maintained. Reproduction and evolution are not essential for the living
environment, for the observed organism the niche constitutes its entire organization, but they have been essential for the historical transformation
domain of interactions, and as such it cannot be part of the environment of the cognitive domains of the living systems on earth.
(2) For a change to occur in the domain of interactions of a unit of

Ii
that lies exclusively in the cognitive domain of the observer. Niche and
environment, then, intersect only to the extent that the observer (including interactions without its losing its identity with respect to the observer it must
instruments) and the organism have comparable organizations, but even then suffer an internal change. Conversely, if an internal change occurs in a unit of

l
!
 12 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN GENERAL 13

interactions, without its losing its identity, its domain of interactions changes. THE COGNITIVE PROCESS
A living system suffers an internal change without loss of identity if the
predictions brought forth'.by the internal change are predictions which do not (1) A cognitive system is a system whose organization defines a domain of
interfere with its fundamental circular organization. A system changes only if interactions in which it can act with relevance to the maintenance of itself,
its domain of interactions changes. and the process of cognition is the actual (inductive) acting or behaving in
(3) After reproduction the new unit of interactions has the same domain this domain. Living systems are cognitive systems, and living as a process is a
of interactions as the parental one only if it has the same organization. process of cognition. This statement is valid for all organisms, with and
Conversely, the new unit of interactions has a different domain of interactions without a nervous system.
only if its organizaticn is different, and hence, implies different predictions (2) If a living system enters into a cognitive interaction, its internal state is
about the niche. changed in a manner relevant to its maintenance, and it enters into a new
(4) Predictions about the niche are inferences about classes of interactions. interaction without loss of its identity. In an organism without a nervous
Consequently, particular interactions which are indistinguishable for an system ( or its functional equivalent) its interactions are of a chemical or
organism may be different for an observer if he has a different cognitive physical nature (a molecule is absorbed and an enzymatic process is initiated;
domain and can describe them as different elements of a class defined by the a photon is captured and a step in photosynthesis is carried out). For such an
conduct of the organism. The same applies to interactions that are identical organism the relations holding between the physical events remain outside
for the organism but different for (have different effects) its different internal its domain of interactions. The nervous system enlarges the domain of inter-
parts. Such interactions may result in different modifications of the internal actions of the organism by making its internal states also modifiable in a
states of the organism and, hence, determine different paths of change in its relevant manner by 'pure relations', not only by physical events; the observer
domain of interactions without loss of identity. These changes may bring sees that the sensors of an animal (say, a cat) are modified by light, and that
about the production of offspring having domains of interactions different the animal (the cat) is modified by a visible entity (say, a bird). The sensors
from the parental ones. If this is the case and a new system thus produced change through physical interactions: the absorption of light quanta; the
predicts a niche that cannot be actualized, it disintegrates; otherwise it main- animal is modified through its interactions with the relations that hold
tains its identity and a new cycle begins. between the activated sensors that absorbed the light quanta at the sensory
( 5) What changes from generation to generation in the evolution of living surface. The nervous system expands the cognitive domain of the living
systems are those aspects nf their organization which are subservient to the system by making possible interactions with 'pure relations'; it does not
maintenance of their basic circularity but do not determine it, and which create cognition.
allow them to retain their identity through interactions; that is, what changes (3) Although the nervous system expands the domain of interactions of
is the way in which the basic circularity is maintained, and not this basic the organism by bringing into this domain interactions with 'pure relations',
circularity in itself. The manner in which a living system is compounded as a the function of the nervous system is subservient to the necessary circularity
unit of interactions, whether by a single basic unit, or through the aggregation of the living organization.
of numerous such units (themselves living systems) that together constitute a (4) The nervous system, by expanding the domain of interactions of the
larger one (multicellular organisms), or still through the aggregation of these organism, has transformed the unit of interactions and has subjected acting
compound units that form self-referring systems of even higher order (insect and interacting in the domain of 'pure relations' to the process of evolution.
societies, nations) is of no significance; what evolves is always a unit of inter- As a consequence, there are organisms that include. as a subset of their possible
actions defined by the way in which it maintains its identity. The evolution interactions, interactions with their own internal states (as states resulting
of the living systems is the evolution of the niches of the units of interactions from external and internal interactions) as if these were independent entities,
defined by their self-referring circular organization, hence, the evolution of generating the apparent paradox of including their cognitive domain within
the cognitive domains. their cognitive domain. In us this paradox is resolved by what we call 'abstract
thinking', another expansion of the cognitive domain.
14 HUMBERTO R. MATURANA

(5) Furthermore, the expansion of the cognitive domain into the domain
of 'pure relations' by means of a nervous system allows for non-physical
interactions between organisms such that the interacting organisms orient IV. COGNITIVE FUNCTION IN PARTICULAR
each other toward interactions within their respective cognitive domains.
Herein lies the basis for communication: the orienting behavior becomes a
representation of the interactions toward which it orients, and a unit of
interactions in its own terms. But this very process generates another apparent NERVE CELLS
paradox: there are organisms that generate representations of their own
interactions by specifying entities with which they interact as if these be- (1) The neuron is the anatomical unit of the nervous system because it is a
longed to an independent domain, while as representations they only map cell, and as such it is an independent integrated self-referring metabolic and
their own interactions. In us this paradox is resolved simultaneously in two genetic unit (a living system indeed).
ways: (2) Anatomically and functionally a neuron is formed by a collector area
(dendrites, and in some cases, also the cell body and part of the axon) united
(a) We become observers through recursively generating representations
via a distributive element (the axon, and in some cases, also the cell body
of our interactions, and by interacting with several representations simul-
and main dendrites), capable of conducting propagated spikes to an effector
taneously we generate relations with the representations of which we can
area formed by the terminal branching of the axon. The functional state of
then interact and repeat this process recursively, thus remaining in a domain
the collector area depends on both its internal state (reference state) and on
of interactions always larger than that of the representations.
the state of activity of the effector areas synapsing on it. Correspondingly,
(b) We become selfconscious through self-observation; by making the state of activity of the effector area depends on both the train of impulses
descriptions of ourselves (representations), and by interacting with our generated at the corresponding collector area and on the pre-synaptic and
descriptions we can describe ourselves describing ourselves, in an endless · non-synaptic interactions with distributive elements and other effector areas
recursive process. that may take place in the neuropil and in the immediate vicinity of the next
collector areas. This is true even in the case of amacrine cells, in which the
collector and effector areas may be intermingled. The distributive element
determines where the effector exerts its influence.
(3) Whether one or two branches of a bifurcating axon are invaded by a
nerve impulse propagating along it depends on their relative diameter and on
the state of polarization of their membranes at their origin in the bifurcation
zone. As a result, the pattern of effector activity, that is, the pattern of
branch invasion which a train of impulses determines in the branches of the
distributive element and effector area of a neuron, depends (i) on the spike
interval distribution of the train of impulses, which determines the time that
the axonal membrane at the branching zone has for recovery before the
arrival of the next spike, and (ii) on the non-synaptic influences which, in the
form of local water and ion movements caused by the electrical activity of
neighboring elements, may produce diameter and polarization changes at the
branching zones, and thus modify the invasibility of the branches by the
arriving spikes.
(4) At any moment the state of activity of a nerve cell, as represented by
15
16 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 17

the pattern of impulses travelling along its distributive element, is a function participation in determining the production of nerve impulses, and hence, the
of the spatio-temporal configuration of its input, as determined by the state of activity of the neuron, depends on their relative spatial distribution on
relative activity holding between the afferent neurons, that modulates the the collector area. Inhibition works by shunting off the spreading excitatory
reference state proper· of the collector area. It is known that in many neurons processes; as a result the relative contributions of a point of excitation and a
the recurrence of a given afferent spatio-temporal configuration results in point of inhibition in the generation of a nerve impulse depend on where, on
the recurrence of the same state of activity, independently of the way in the collector, they stand with respect to each other and with respect to the
which such a spatio-temporal configuration is generated [Cf. Maturana and point of emergence of the distributive element. Excitation and inhibition
Frenk, 1963; Morrell, 1967]. [This is so in the understanding that two states must be seen as integral parts in the definition of the spatio-temporal con-
of activity in a given cell are the 'same' (equivalent) if they belong to the figuration of afferent influences, not as indepenclent processes. The shape of
same class, as defined by the pattern of impulses that they generate, and not the collector area (its geometry) determines the class or classes of spatio-
because they are a one-to-one mapping of each other.] Also, the spatio- temporal configurations of afferent influences to which the cell responds.
temporal configuration of the input to a neuron that causes in it the re- (7) The neuropil is the site where the distributive elements and effector
currence of a given state of activity is a class of afferent influences defined areas of many different neurons interrningle with each other and with the
by a pattern in the relations holding between the active afferents and the collector areas of the post-synaptic cells. Here non-synaptic interactions take
collector; a given cl.ass of responses is elicited by a given class of afferent place between neigh boring elements which may cause in each other, as a result
influences. of the local movements of water and ions produced by their independent
(5) For every nerve cell, at any moment, its transfer function at its collec- electrical activity, changes in diameter and polarization at their branching
tor area is a well-defined deterministic process [Cf. Segundo and Perkel, points. Depending on the time constant of these local changes, and on the
1969]. Many neurons have several transfer functions, and different classes of capacity of the axons to homeostatically maintain their diameter at the new
afferent influences change their activity differently, causing them to generate values, the pattern of branch invasion produced in a given effector area by a
different classes of activity in their effector areas. Because every nerve cell given train of impulses may be modified in a more or less permanent manner
participates in the generation of the spatio-temporal configuration of afferent by these non-synaptic interactions. Something similar may happen during
influences on the other nerve cells, all their states of activity must be con- synaptic concomitances at the collector areas if synapses also affect each
sidered as significant for their next states of activity. Thus there are two otl1er non-synaptically, due to their spatial contiguity, causing each other
aspects to consider with respect to the activity of any given neuron: (i) its .more or less permanent changes in size (increase or decrease) and polarization
genesis, which must be considered in reference to the neuron itself and to the (with the corresponding changes in effectiveness) as a result of their inde-
afferen ts to it; (ii) its participation in the generation of activity in other pendent electrical activities. Thus, the neuropil may have to be seen as
neurons for which it is an afferent influence, which must be considered in constituting a plastic system through which acquired self-addressing states of
reference to those other neurous. In both cases the interactions between the activity attain their functional significance as they become specified by the
neurons involved are strictly deterministic, although what is cause in one is non-synaptic and synaptic concomitances generated by the interactions of the
not necessarily cause in another. organism. It is not the repetition of the same state of activity which can cause
(6) The nerve impulses that travel along the distributive element originate neuronal changes of behavioral significance subordinated to the evolving
at the point where this element emerges from the collector area. Each nerve domain of interactions of an organism, but rather it is the occurrence of
impulse is the result of the state of excitation of the collector area at a local concomitant states of activity produced by seemingly unrelated inter-
given moment (as determined by the spatio-temporal configuration of the actions which can cause such subordiz:iated changes in the reactive capacity of
afferent excitatory and inhibitatory influences acting upon it, and on its neurons.
own internal generating mechanisms, if any) that spreads reaching a given (8) It follows that one should expect in a significant number of neurons,
threshold at the point of emergence of the distributor. Excitatory and in- which may vary in different classes of animals according to the organization
hibitory influences, however, do not superimpose linearly; their relative of their different neuropils, a continuous change in their transfer functions
 HUMBERTO R. MATURANA
COGNITIVE FUNCTION IN PARTICULAR 19
18

(from collector to effector area), or in the circumstan~es under which they Thus although at any moment every neuron functions deterministically
witll a definite transfer function, and generates a definite pattern of activity
are activated, as a result of the past history of the organism. However, for_ ~e
understanding of the functional organization of the nervous system it 1s in its effector area, the transfer functions and the patterns of effector activity
necessary to consider that nerve cells respond at any moment with ~efini~e in many of them may change from one moment to another and the organism
transfer functions to classes of afferent spatio-temporal configurat10ns m still will give rise to what the observer would call 'the same behavior'. The
their input, generating definite states of effector activity, and not to particular converse is also the case, and through what the observer would call 'different
behaviors' the organism may satisfy its subordination to tile same aspect of
afferent states. Furthermore:
the maintenance of its basic circularity.
(a) Any interaction is represented in the nervous system by tl~e s~- (9) From these notions it is apparent that the neuron cannot be considered
quence of states of relative neuronal activity leading to the conduct which 1t as the functional unit of the nervous system; no neuron can have a fixed
generates; this conduct should be repeatable to the extent tha: the interaction functional role in the generation of conduct if it must be continuously
(sequence of states of relative activity) is reproducible,_ that 1s, as long as tll_e changing its participation in it. For the same reason a fixed collection of cells
historical transformation of the nervous system (learnmg) does not make it also cannot be considered as a functional unit of the nervous system. Only
impossible. conduct itself can be considered as tile functional unit of the nervous system.
(b) The nervous system always functions in the present, and ~t can o_nly (10) If nerve cells respond to classes of afferent configurations and not to
be understood as a system functioning in the present. The present is.the t~e particular afferent states, they must necessarily treat as equivalent particular
interval necessary for an interaction to take place; past, future and ttme exist afferent configurations that arise through interactions which for the observer
only for the observer. Although many nerve cells may cha~ge continuously, are otherwise unrelated.
their mode of operation and their past history can explam to t.h~ obse~ver
how their present mode of operation was reached, but. not_ how 1t 1s re~1z~d ARCHITECTURE
now, or what their present participation in the determmatton of behaVIor 1s.
(1) In any given nervous system tile great majority (and perhaps the totality)
(c) Any behavior is defined through a sequence of states in the r~cep:or of its neurons can· be assigned to well-defined morphological classes, each
surfaces (external and internal) that satisfy its direct o~ ~direct subor~matlon characterized by a given pattern of distribution of the collector and effector
to the maintenance of the basic circularity of the livmg system. Smee the areas of its elements. As a result, the elements of the same class hold similar
nervous system is continuously changing through experience, ':hat oc~urs relations with each other and with other classes of neurons; the shapes of the
when the observer sees a given behavior reenacted is a sequence of mteract1~ns nerve cells (collector area, distributive element, and effector area) specify
that satisfy this subordination independently of the neuronal process w~ch their connectivity. These shapes are genetically determined and have been
generated them. The more complex tile domain of interactions of an or~an1sm, attained tllrough evolution; the whole architecture of the brain is genetically
the more indirect is this subordination ( an adequate mode of behaVIor sub- determined and has been attained through evolution. The following implica-
ordinated to another), but not the less strict. tions are significant for the understanding of the nervous system:
(d) An organism is a unit to the extent that its conduct results in tile (a) There is a necessary genetic variability in tile shape of nerve cells
maintenance of its basic circularity (and hence identity), and two modes of as well as a variability that results from interactions of the organism with
conduct are equivalent if they satisfy the same class of requirements for th~s independent events during its development. The functional organization of
maintenance. For this reason an organism, as a self-regulated homeostat1c tile nervous system must be such as to tolerate this double variability.
organization, does not require a constant behavior in its deterministic ~om-
ponent elements (in this case, neurons) if their changes be~ome spec1fi_ed (b) Due to the genetic and somatic variability no two nervous systems
through the generation of conduct, and sameness of conduct 1s defined with of animals of tile same species (particularly if they have many cells) are
identical, and they resemble each other only to the extent that tlley are
respect to an observer or a function that must be satisfied.
20 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 21

organized according to the same general pattern. It is the organization defining localization with respect to the possibility of establishing certain functional
the class, and not a,ny particular connectivity, which determines the mode of concomitances. From this it follows that any localized lesion in the ne.rvous
functioning of any given kind of nervous system. system must necessarily interfere in a localized manner with the possibility
of synthesizing some specific conduct (state of neural activity).
(2) The shapes of nerve cells and their packing are such that there is in
general a great overlapping in the collector and effector areas of neurons of
the same class. Also, the spatial distribution and the interconnections between FUNCTION
different classes of neurons is such that any particular part of the_ nervous
system is in general simultaneously related to many other parts; the parts (1) The way the nervous system functions is bound to its anatomical organi-
interconnected, however, differ in different species, and as a result these have zation. The functioning of the nervous system has two aspects: one which
different interacting capabilities. refers to the domain of interactions defined by the nervous system (relations
(3) The organism ends at the boundary that its self-referring organization in general); the other which refers to the particular part of that domain used
defines in the maintenance of its identity. At this boundary there are sensors by a given species (particular classes of relations): Different species interact
(the sensory surfaces) through which the organism interacts in the domain of with different sets of relations (have different niches).
relations and effectors ( the effector surfaces) through which the nervous (2) The nervous system only interacts with relations. However, since the
system modifies the posture of the organism in this domain. The sensory functioning of the nervous system is anatomy bound, these interactions are
surfaces are in general constituted by collections of sensory elements (cells) necessarily mediated by physical interactions; for an animal to discriminate
with similar, though not identical, properties (classes of properties) which in objects visually the receptors in its eyes must absorb light quanta and be
their mode of interaction with the nervous system share the characteristics activated; yet, the objects that the animal sees are determined not by the
of neurons in general. As a result, whenever the organism enters into an quantity of light absorbed, but by the relations holding between the receptor-
interaction within the physical domain of interactions of the sensors, as a induced states of activity within the retina, in a manner determined by the
rule not one but many sensory elements are excited. The effectors are also connectivity of its various types of cells. Therefore, the nervous system
multifarious and differ from each other in the manner in which they change defines through the relative weights of the patterns of interactions of its
the receptor surfaces of the organism during the interactions: action always various components, both innate and acquired through experience, which
leads to a change in _the state of activity of the receptor surfaces. rel_ations will modify it at any given interaction [Cf. Maturana, 1965]. Or,
(4) The architectural organization of the nervous system is subordinated in general, the organization and structure of a living system (its nervous
to the order of the sensory and effector surfaces. This subordination has two system included) define in it a 'point of view', a bias or posture from the
aspects: (i) the receptor and effector surfaces project to the central nervous perspective of which it interacts, determining at any instant the possible
system retaining their proper topological relations; (ii) the topological rela- relations accessible to its nervous system. Moreover, since the domain of
tions specified by the receptor and effector surfaces in their projection interactions of the organism is defined by its structure, and since this struc-
constitute the basis for all the architectural order of the central nervous ture implies a prediction of a niche, the relations with which the nervous
system. As a consequence, this architectural organization constitutes a system system interacts are defined by this prediction and arise in the domain of
that interconnects these surfaces in a manner that permits the occurrence of interactions of the organism.
. certain concomitances of activity and not others in the different neuropils, (3) Due to the properties of neurons, and due to the architecture of the
and thus secures well-defined functional relations between these surfaces, nervous system, interactions within the nervous system give rise to activity
specifying how they modify each other. Truism: the nervous system cannot in aggregates of cells. Also, for the same reasons, any given cell may assume
give rise to a conduct that implies the concomitance of states of activity the same state of activity under many different circumstances of interactions
for which there is no anatomical basis. As a result of its architectural organ- of the organism. Thus, under no circumstances is it possible to associate the
ization every point in the central nervous system constitutes an anatomical activity of any particular cell with any particular interaction of the living
 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR · 23
22

system. When any particular interaction takes place at the level of the sensors, their origin (through concomitant events, their locations, or through the
the relations accessible to the nervous system are given at this level in a consequences of the new interactions which they originate) there is no
certain state of relative activity of the sensing elements and not in the state of possible distinction between internally and externally generated states of
activity of any particular one [Cf. Maturana, Uribe, and Frenk, 1968]. At the nervous activity.
same time, although operational localizations can be established in the (2) The relations with which the nervous system interacts are relations
nervous system [Cf. Geschwind, 1965], these localizations are to be viewed given by the physical interactions of the organism, and, hence, depend on its
in terms of areas where certain modalities of interactions converge, and not anatomical organization. For the observer the organism interacts with a given
as localizations of faculties or functions. As a result of the mode of organiza- entity that he can describe in his cognitive domain. Yet, what modifies the
tion of the nervous system that I have emphasized, localized lesions should nervous system of the observed organism are the changes in activity of the
produce discrete functional deficiencies by impeding the convergence of nerve cells associated with the sensing elements, changes that henceforth
activities necessary for the synthesis of a particular conduct (state of activity). constitute an embodiment of the relations that arise through the interaction.
The anatomical and functional organization of the nervous system secures the These relations are not those that the observer can describe as holding between
synthesis of behavior, not a representation of the world; hence, it is only component properties of the entity in his cognitive domain; they are relations
with the synthesis of behavior that one can interfere. TI1e nervous system is generated in the interaction itself and depend on both the structural organiza-
localized in terms of the organism's surfaces of interaction, but not in terms tion of the organism and the properties of the universe that match the domain
of representations of the interactions it can generate. of interactions that this organization defines. Whenever such a relation recurs
at the sensory surface, the same state of relative activity arises among the
neurons in contact with the sensing elements. Two interactions that produce
REPRESENTATION the same state of relative activity are identical for the nervous system, no
( 1) The fundamental anatomical and functional organization of the nervous matter how different they may be in the cognitive domain of the observer.
system is basically uniform; the same functions and operations ( excitation, (3) Every relation is embodied in a state of relative activity of nerve cells,
inhibition, lateral interaction, recursive inhibition, etc.) are performed in its but also every state of relative activity acts to modify the relative activity of
various parts, although in different contexts, and integrated in different other nerve cells. Thus, relations through their embodiment in states of
manners. A partial destruction of the nervous system does not alter this relative activity become units of internal interactions and generate additional
basic uniformity,'and, although the parts left untouched cannot do the same relations, again embodied in states of relative activity which in turn may also
things that the whole did, they appear in their mode of operations identical become units of internal interactions, and so on, recursively.
to the untouched whole. To the observer, once the boundary of the sensors (4) If an external interaction takes place, the state of activity of the
is passed, the nervous system, as a mode of organization, seems to begin at any nervous system is modified by the change in relative activity of the neurons,
arbitrary point that he may choose to consider; the answer to the question, which in close association with the sensing elements embody the relations
'What is the input to the nervous system?' depends entirely on the chosen given in the interaction. Accordingly, that which the different states of
point of observation. This basic uniformity of organization can best be activity thus generated can be said to represent are the relations given at the
expressed by saying: all that is accessible to the nervous system at any point sensory surfaces by the interaction of the organism, and not an independent
are states of relative activity holding between nerve cells, and all that to medium, least of all a description of an environment necessarily made in
which any given state of relative activity can give rise are further states of terms of entities that lie exclusively in the cognitive domain of the observer.
relative activity in other nerve cells by forming those states of relative activity If an internal interaction takes place, the state of activity of the nervous
to which they respond. The effector neurons are not an exception to this system is modified by one of its own substates of relative activity that em-
since they, by causing an effector activity and generating an interaction, cause bodies one set of relations. However, that which the new state of relative
a change in the state of relative activity of the receptor elements at the activity represents is the relations given in the internal interaction, and not
receptor surfaces. This has a fundamental consequence: unless they imply an independent set of relations or their description, in terms of some kind of
24 HU.MBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 25

entities, such as thoughts, that lie only within the cognitive domain of the actions can only arise through a concomitance of events that indicates the
observer. source (sensory surface or not) of the state of activity caused by them, or
(5) The classes of relations that can be embodied have been defined: (i) through the outcome of new interactions which they initiate. A nervous
through the evolution of the general structural organization of the organism, system that is capable of treating its internally generated states of activity
and particularly, of the sensors, that has defined the classes of relation that as different from its externally generated states, that is, of distinguishing
are accessible to the nervous system; and (ii) through the evolution of a their origin, is capable of abstract thinking.
particular organization of the nervous system that defines for each class of
(b) The nervous system can interact with the representations of its
animals (species) the specific mode of how these relations generate a behavior
interactions (and hence, of the organism) in an endless recursive manner.
relevant to their maintenance.
(6) For any class of relations, the particular relations given as a result of (8) Four comments:
a present interaction are embodied in· a set of particular states of activity
occurring in the present. This is the case independently of the history of the (a) Notions such as embodiment of representation express the corres-
system. However, the relevance of the behavior generated by those states of pondence that the observer sees between relations, or sets of relations, and
activity for the maintenance of the living system is a function of history, and different states of activity of the nervous system, and, as such, lie in his
may depend both on the evolutionary history of the species and on the past cognitive domain. They describe the functional organization of the nervous
experiences of the organism as an individual. In the first case I would speak of system in the cognitive domain of the observer, and point to the ability of
instinctive behavior, and in the second case of learned behavior. The descrip- the nervous system to treat some of its own states as independent entities
tion of learning in ~erms of past and present behavior lies in the cognitive with which it can interact, but they do not characterize the nature of the
domain of the observer; the organism always behaves in the present. The ob- functional subordination of the nervous system to its own states. This sub-
server, however, by interacting with descriptions that he generates can treat ordination is that of a functionally closed, state determined, ultrastable
interactions which do not recur as if they were in the present. This apparent system, modulated by interactions [Cf. Ashby, 1960]. ·
paradox is resolved by generating the notion of time, past, present, and future, (b) The closed nature of the functional organization of the nervous
as a new expansion of the domain of interactions. Whenever an interaction system is a consequence of the self-referring domain of interactions of the
takes place which is an element of a class experienced for the first time, it is living organization; every change of state of the organism must bring forth
sufficient that the state of activity which it generates be followed by the another change of state, and so on, recursively, always maintaining its basic
suppression of a peculiar concomitant internal state of activity (that is circularity. Anatomically and functionally the nervous system is organized
apparent in what the observer calls the emotion of anxiety or uncertainty) for to maintain constant certain relations between the receptor and effector
the organism to experience the recurrence of an interaction of the same class, surfaces of the organism, which can only in that way retain its identity as it
which takes place without such a concomitant state, as not new (in the sense moves through its domain of interactions. Thus all conduct, as controlled
that it can generate an established conduct as is apparent in the absence of through the nervous system, must (necessarily, due to the latter's architectural
anxiety) and, hence known. Any experience without anxiety can be described organization) lead through changes in the effector surfaces to specific changes
as known, and thus serve as a basis for the functional notion of time. in the receptor surfaces that in turn must generate changes in the effector
(7) There is no difference in the nature of the embodiment of the relations surfaces that again ... and so on, recursively. Conduct is thus a functional
generated through either external or internal interactions; both are sets of continuum that gives unity to the life of the organism through its transforma-
states of neuronal activity that can be said to represent the interactions. In a tions in the latter's self-referring domain of interactions. The evolutionary
nervous system capable of interacting with some of its own internal states as subordination of the architecture of the central nervous system to the topo-
if they were independent entities, there are two consequences: logy of the sensory and effector surfaces appears as an obvious necessity.

(a) The distinction between externally and internally generated inter- (c) The ability of the nervous system to interact with its own internal
26 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 27

states, as if these were independent entities, enters these internal states as and functions always in a predictive manner: what happened once will occur
modulating factors in the continuum of behavior. This requires an anatomical again. Its organization, (genetic and otherwise) is conservative and repeats
and functional internal reflection so that the internal organization of the only that which works. For this same reason living systems are historical
nervous system can project itself onto itself retaining its morphological and systems; the relevance of a given conduct or mode of behavior is always
functional topological relations, as the receptor and effector surfaces do in determined in the past. The goal state (in the language of the observer) that
their own projection. This seems to have acquired an autonomous evolutionary controls the development of an organism is, except for mutations, determined
course with the development of the neo-cortex in mammals, which arises as a by the genome of the parent organism. The same is true for behavior in
center of internal anatomical projection, and whose evolution in this line is general; the present state is always specified from the previous state that
accompanied by an increased dependency of the organism on its own states restricts the field of possible modulations by independent concornitances. If
of nervous activity. a given state of relative activity in the nerve cells originates a given behavior,
(d) The closed nature of the functional organization of the nervous a recurrence of the 'same state' of relative activity should gi~e rise to the
system (open only to modulations through interactions) is particularly evi- 'same behavior' no matter how the recurrence originates. The relevance of
dent in systematic observations that explicitly show the subordination of such a behavior is determined by the significance that it has for the main-
conduct to the correlation of activity between the receptor and effector tenance of the living organization, and it is in relation to this relevance that
surfaces [Cf. Held and Hein, 1963]. Experiments such as those of Held and any subsequent behaviors are the same. With the expansion of the cognitive
Hein show that a c~t does not learn to control its environment visually if domain during evolution, the types of behavior have changed as well as how
raised in darkness and carried about only passively, by another cat, when their relevance is implemented; different kinds of behavior are relevant to the
under light. From these observations, it is apparent that the 'visual handling' maintenance of the basic circularity of the living organization through dif-
of an environment is no handling of an environment, but the establishment of ferent domains of interactions, and hence, different fields of causal relations.
a set of correlations between effector (muscular) and receptor (proprioceptor (2) Since the niche of an organism is the set of all classes of interactions
and visual) surfaces, such that a particular state in the receptor surfaces may into which it can enter, and the observer beholds the organism in an environ-
cause a particular state in the effector surfaces that brings forth a new state ment that he defines, for him any one of the organism's behaviors appears
in the receptor surfaces ... and so on. Behavior is like an instrumental flight as an actualization of the niche, that is, as a first order description of the
in which the effectors ( engines, flaps, etc.) vary their state to maintain environment (henceforth denoted by a capital D: Description). This Descrip-
constant, or to change, ·the readings of the sensing instruments according to a tion, however, is a description in terms of behavior (interactions) of the
specified sequence of variations, which either is fixed (specified through observed organism, not of representations of environmental states, and the
evolution) or can be varied during the flight as a result of the state of the flight relation between behavior and niche lies exclusively in the cognitive domain
(learning). The same is apparent in the experiments with innate perception of the observer.
of depth [Cf. Gibson, 1950] that show that there is an innate system of (3) An organism can modify the behavior of another organism in two
correlations between certain states of the receptor and effector surfaces. The basic ways:
reference to a pre-established perception of depth is a description that lies in (a) By interaction with it in a manner that directs both organisms toward
the cognitive domain of the observer,'and as such only alludes to relations, each other in such a way that the ensuing behavior of each of them depends
through the observer, between elements that lie in his cognitive domain; but, strictly on the following behavior of the other, e.g.: courtship and fight. A
as a process, this innate behavior obviously corresponds to one of optimization chain of interlocked behavior can thus be generated by the two organisms.
of sensory states.
(b) By orienting the behavior of the other organism to some part of its
DESCRIPTION domain of interactions different from the present interaction, but comparable
to the orientation of that of the orienting organism. This can take place only
(1) A living system, due to its circular organization, is an inductive system if the domains of interactions of the two organisms are widely coincident; in
28 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 29
this case no interlocked chain of behavior is elicited because the subsequent manner, and the organism becomes an observer: it gen~rates discourse as a
conduct of the two organisms depends on the outcome of independent, domain of interactions with representations of communciative descriptions
although parallel, interactions. ( orienting behaviors).
Furthermore: if such an observer through orienting behavior can orient
In the first case it can be said that the two organisms interact; in the himself towards himself, and then generate communicative descriptions that
second case that they communicate. The second case is the basis for any orient him towards his description of this self-orientation, he can, by doing so
linguistic behavior; the first organism generates (as is apparent for the ob- recursively, describe himself describing himself ... endlessly. Thus discourse
server) a Description of its niche that, in addition to its own significance as a through communicative description originates the apparent paradox of self.
behavior (within the cognitive domain of the first organism, and independently description: self-consciousness, a new domain of interactions.
of it), orients the second organism within its cognitive domain to an inter- (6) A nervous system capable of recursively interacting with its own states
action from which ensues a conduct parallel to that of the first one, but as if these were independent entities can do so regardless of how these states
unrelated to it. The conduct thus elicited by the orienting behavior is denota- are generated, and in principle can repeat these recursive interactions endlessly.
tive: it points to a feature of the environment that the second organism Its only limitation lies in the need that the progressive transformation of its
encounters in its niche and Describes by the appropriate conduct, and that he actual and potential behavior, which in such a system is a necessary con-
can treat as an independent entity. The orienting behavior is, for the observer, comitant to behavior itself, be directly or indirectly subservient to the basic
a second order description (henceforth denoted by italics: description) that circularity of the living organization. The linguistic domain, the observer, and
represents that which he considers it to denote. By contrast, the orienting self-conciousness are each possible because they result as different domains
behavior of the first organism is connotative for the second one, and implies of interactions of the nervous system with its own states in circumstances
for it an interaction within its cognitive domain which, if actualized, originates in which these states represent different modalities of interactions of the
a behavior that Describes a particular aspect of its niche; that which an organism.
orienting behavior connotes is a function of the cognitive domain of the
orientee, not the orienter.
THINKING
(4) In an orienting interaction the behavior of the first organism, as a
communicative descrjption causes in the nervous system of the second one a (1) I consider that in a state-determined nervous system, the neurophysio-
specific state of activity; this state of activity embodies the relations generated logical process that consists in its interacting with some of its own internal
in the interaction and represents the behavior of the second organism (De- states as if these were independent entities corresponds to what we call
scription of its niche) connoted by the orienting behavior of the first one. tl1inking. Such internal states of nervous activity, otherwise similar to other
This representation, as a state of neuronal activity, can in principle be treated states of nervous activity that participate in the specification of behavior, as
by the nervous system as a unit of interactions, and the second organism, if in reflex mechanisms, cause conduct by determining specific changes of state
capable of doing so, can thus interact with representations of its own Descrip- in the nervous system. Thinking thus conceived, and reflex mechanisms, are
tions of its niche as if these were independent entities. This generates yet both neurophysiological processes through which behavior emerges in a
another domain of interactions (and hence, another dimension in the cognitive deterministic manner; they differ, however, in that in a reflex action we can,
domain), the domian of interactions with representations of behavior (inter- in our description, trace a chain of nervous interactions that begins with a
actions), orienting interactions included, as if these representations were specific state of activity at the sensory surfaces; while in thinking, the chain
independent entities within the niche: the linguistic domain. of nervous interactions that leads to a given conduct (change in the effector
(5) If an organism can generate a communicative description and then surfaces) begins with a distinguishable state of activity of the nervous system
interact with its own state of activity that represents this description, generat- itself, whichever way it may have originated. Accordingly, thinking is a mode
ing another such description tl1at orients towards this representation ... , of operation of the nervous system that reflects functionally its internal
the process can in principle be carried on in a potentially infinite recursive anatomical projection (possibly multiply) onto itself.
30 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 31

(2) The process of thinking as characterized above is necessarily inde- becomes apparent that learned orienting interactions embody a function of
pendent of language. That this is so even for what we call 'abstract thinking' non-linguistic origin that, under a selective ·pressure for recursive application,
in man is apparent from the observations of humans with split brains [Cf. can originate through evolution the system of cooperative consensual inter-
Gazzaniga, Bogen and Sperry, 1965]. These observations show that the actions between organisms that is natural language. Particular orienting
inability of the non-speaking hemisphere to speak does not preclude in it interactions, like any other learned conduct, arise from the substitution of
operations that the observer would call abstract thinking, and that the lack one type of interaction for another as a cause for a given behavior, and their
of language only implies that it cannot generate discourse. When we talk origin as a function of the general learning capacity of the nervous system is·
about concepts or ideas we describe our interactions with representations of completely independent of the complexities of the system of cooperative
our descriptions, and we think through our operation in the linguistic domain. interactions to which their recursive application gives rise. Widespread among
The difficulty arises from our considering thinking through our description of animals other than man-orienting interactions are particularly evident in
it in terms or concepts as if it were something peculiar to man, and in some primates, in which it is easy to see how the audible and visible behavior of
way isomorphic with the notions embodied in the descriptions, instead of one individual orients others within their respective cognitive domains [Cf.
attending to the functional process that makes these descriptions possible. Jay, 1968], and in dolphins which seem to have evolved a rich and efficient
system of auditive cooperative interactions [Cf. Lilly, 1967]. In accordance
with all this I maintain that learned orienting interactions, coupled with some
NATURAL LANGUAGE
mode of behavior that allowed for an independent recursive expansion of the
(1) Linguistic behavior is orienting behavior; it orients the orientee within his domain of interactions of the organism, such as social life [Cf. Gardner and
cognitive domain to interactions that are independent of the nature of the Gardner, 1969] and/or tool making and use, must have offered a selective
orienting interactions themselves. To the extent that the part of its cognitive basis for the evolution of the orienting behavior that in hominids led to our
domain toward which the orientee is thus oriented is not genetically deter- present-day languages.
mined and becomes specified through interactions, one organism can in (3) Behavior (function) depends on the anatomical organization (structure)
principle orient another to any part of its cognitive domain by means of of the living system, hence anatomy and conduct cannot legitimately be
arbitrary modes of conduct also specified through interactions. However, separated and the evolution of behavior is the evolution of anatomy and vice
only if the domains of interactions of the two organisms are to some extent versa; anatomy provides the basis for behavior and hence for its variability;
comparable, are such consensual orienting interactions possible and are the behavior provides the ground for the action of natural selection and hence for
two organisms able to develop some conventional, but specific, system of the historical anatomical transformations of the organism. Structure and
communicative descriptions to orient each other to cooperative classes of function are, however, both relative to the perspective of interactions of the
interactions that are relevant for both. system and cannot be considered independently of the conditions that define
(2) The understanding of the evolutionary origin of natural languages it as a unit of interactions, for what is from one perspective a unit of inter-
requires the recognition in them of a basic biological function which, properly actions, from another may only be a component of a larger one, or may be

II
selected, could originate them. So far this understanding has been impossible several independent units. It is the dynamics of this process of individuation,
because language has been considered as a denotative symbolic system for the as an historical process in which every state of a changing system can become
transmission of information. In fact, if such were the. biological function of a unit of interactions if the proper circumstances are given, what makes the
language, its evolutionary origin would demand the pre-existence of the evolution of living systems a deterministic process of necessarily increasing
function of denotation as necessary to develop the symbolic system for the complication. Thus, in the evolution of language, natural selection, by acting
transmission of information, but this function is the very one whose evolu- upon orienting behavior as a function that if enhanced strongly increases the
tionary origin should be explained. Conversely, if it is recognized that language
is connotative and not denotative and that its function is to orient the orientee
within his cognitive domain, and not to point to independent entities, it
I cooperation between social animals, has led to anatomical transformations
which provide the basis for the increased complexity of the orienting conduct
and the diversity of the interactions toward which man can be oriented in his

I.
t
 32 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 33

cognitive domain. The complexity of the orienting conduct has increased because the speaker tacitly assumes the listener to be identical with him and
through an increase in the complexity and variety of motor behavior, parti- hence as having the same cognitive domain which he has (which never is the
cularly through vocalization and tool making. The diversity of the interactions case), marvelling when a 'misunderstanding' arises. Such an approach is valid,
toward which man can be oriented has increased through a concomitant for man created systems of communication where the identity of sender and
expansion of the internal projection of the brain onto itself, by means of new receiver is implicitly or explicitly specified by the designer, and a message,
interconnections between different cortical areas (as compared with other unless disturbed during transmission, necessarily selects at the reception the
primates), between cortical areas and subcortical nuclei [Cf. Geschwind, same set of states that it represents at the emission, but not for natural
1964 J, and possibly also between different cortical layers and cellular sys- languages.
tems within the cortex itself. (5) It behooves the interlocutor to choose where to orient in his cognitive
(4) So long as language is considered to be denotative it will be necessary domain as a result of a linguistic interaction. Since the mechanism of choice,
to look at it as a means for the transmission of information, as if something as in every neuronal process, is state-dependent, the state of activity from
were transmitted from organism to organism, in a manner such that the which the choice (new state of neuronal activity) must arise restricts the
domain of uncertainties of the 'receiver' should be reduced according to the possible choices and constitutes a reference background in the orientee.
specifications of the 'sender'. However, when it is recognized that language
is connotative and not denotative, and that its function is ·to orient the
I The same is valid for the speaker; the state of activity from which his com-
municative description (linguistic utterance) arises constitutes the reference
orientee within his cognitive domain without regard for the cognitive domain
of the orienter, it becomes apparent that there is no transmission of informa-
tion through language. It behooves the orientee, as a result of an independent
I background that specifies his choice. All the interactions that independently
specify the reference background of each interlocutor constitute the context
in which a given linguistic interaction takes place. Every linguistic interaction
internal operation upon his own state, to choose where to orient his cognitive
domain; the choice is caused by the 'message', but the orientation thus
produced is independent of what the 'message' represents for the orienter. In
I is thus necessarily context-dependent, and this dependency is strictly deter-
ministic for both orienter and orientee, notwithstanding the different back-
grounds of the two processes. It is only for the observer that there is any

I
a strict sense then, there is no transfer of thought from the speaker to his ambiguity in a linguistic interaction that he observes; this is because he has no
interlocutor; the listener creates information by reducing his uncertainty access to the context in which it occurs. The sentence, 'They are flying
through his interaction,s in his cognitive domain. Consensus arises only through planes,' is unambiguious for both interlocutors, regardless of the subsequent
cooperative interactions in which the resulting behavior of each organism behavior which it originates in each of them; for the observer, however, who
becomes subservient to the maintenance of both. An observer beholding a wants to predict the course of the ensuing interactions, it is ambiguous.
communicative interaction between two organisms who have already de- (6) If one considers linguistic interactions as orienting interactions it is
veloped a consensual linguistic domain, can describe the interaction as de- apparent that it is not possible to separate, functionally, semantics and syntax,
notative; for him, a message (sign) appears as denotating the object which the however separable they may seem in their description by the observer. This
conduct of the orientee Describes (specifies), and the conduct of the orientee is true for two reasons:
appears determined by the message. However, because the outcome of the
interaction is determined in the cognitive domain of the orientee regardless (a) A sequence of communicative descriptions (words in our case) must
of the significance of the message in the cognitive domain of the orienter be expected to cause in the orientee a sequence of successive orientations in
the denotative function of the message lies only in the cognitive domain of his cognitive domain, each arising from the state left by the previous one.
the observer and not in the operative effectiveness of the communicative 'They are flying planes' clearly illustrates this; each successive word orients
interaction. The cooperative conduct that may develop between the inter- 1 the listener to a particular interaction in his cognitive domain that is relevant
acting organisms from these communicative interactions is a secondary in a particular manner (apparent in the conduct it generates) that depends on
process independent of their operative effectiveness. If it appears acceptable f the previous orientation. The fact that it seems that the observer can more
to talk about transmission of information in ordinary parlance, this is so easily describe the word are (or any word) by referring to its grammatical and

II
I
34 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 35

lexical functions, rather than by specifying the nature of the orientation (9) Orienting behavior in an organism with a nervous system capable of
that it causes (in terms of conduct or interactions), should not obscure the interacting recursively with its own states expands its cognitive domain by
problem. The observer speaks, and any explanation of the word are that he enabling it to interact recursively with descriptions of its interactions. As a
may give lies in the descriptive domain, while the orientation caused by the result:
word itself, as a change of state of the listener, is an internal interaction in his
(a) Natural language has emerged as a new domain of interactions
cognitive domain.
in which the organism is modified by its descriptions of its interactions,
(b) An entire series of communicative descriptions can itself be a as they become embodied in states of activity of its nervous system, sub-
communicative description; the whole sequence once completed may orient jecting its evolution to its interactions in the domains of observation and
the listener from the perspective of the state to which the sequence itself has self-consciousness.
led him. The limit to such complications lies exclusively in the capacity of the
nervous system to discriminate between its own discriminable internal states, (b) Natural language is necessarily generative because it results from
and to interact with them as if with independent entities. the recursive application of the same operation (as a neurophysiological
process) on the results of this application.
(7) Linguistic behavior is an historical process of continuous orientation.
As such, the new state in which the system finds itself after a linguistic (c) New sequences of orienting interactions (new sentences) within the
interaction emerges from the linguistic behavior. The rules of syntax and consensual domain are necessarily understandable by the interlocutor ( orient
generative grammar [Cf. Chomsky, 1968] refer to regularities that the him), because each one of their components has definite orienting functions
observer sees in the linguistic behavior (as he would see in any behavior) as a member of the consensual domain that it contributes to define.
which, arising from the functional organization of the system, specify the
interactions that are possible at any given moment. Such rules, as rules, lie
MEMORY AND LEARNING
exclusively in the cognitive domain of the observer, in the realm of descrip-
tions, because the transitions from state to state as internal processes in (1) Learning as a process consists in the transformation through experience
any system are unrelated to the nature of the intetactions to which they of the behavior of an organism in a manner that is directly or indirectly
give rise. Any correlation between different domains of interactions lies subservient to the maintenance of its basic circularity. Due to the state
exclusively in the cognitive domain of the observer, as relations emerging determined organization of the living system in general, and of the nervous
from his simultaneous interactions with both. system in particular, this transformation is an historical process such that
(8) The coordinated states of neuronal activity which specify a conduct as each mode of behavior constitutes the basis over which a new behavior
a series of effector and receptor states whose significance arises in a consensual develops, either through changes in the possible states that may arise in it as
domain, does not differ in its neurophysiological generation from other a result of an interaction, or through changes in the transition rules from
coordinated states of neuronal activity which specify other conducts of innate state to state. The organism is thus in a continuous process of becoming that
or acquired significance (walking, flying, playing a musical instrument). Thus,. is specified through an endless sequence of interactions with independent
however complex the motor and sensory coordinations of speech may be, the entities that select its changes of state but do not specify them.
peculiarity of linguistic behavior dcies not lie in the complexity or nature of (2) Learning occurs in a manner such that, for the observer, the learned
the series of effector and receptor states that constitute it, but in the relevance behavior of the organism appears justified from the past, through the in-
that such behavior acquires for the maintenance of the basic circularity of the corporation of a representation of the environment that acts, modifying its
interacting organisms through the development of the consensual domain of present behavior by recall; notwithstanding this, the system itself functions
orienting interactions. Speaking, walking, or music-making do not differ in in the present, and for it learning occurs as an atemporal process of trans-
the nature of the coordinated neuronal processes which specify them but in formation. An organism cannot determine in advance when to change and
the sub-domains of interactions in which they acquire their relevance. when not to change during its flow of experience, nor can it determine in
36 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 37

advance which is the optimal functional state that it must reach; both the the reactivity of the system. This anatomical and functional transformation
advantage of any particular behavior and the mode of behavior itself can only of the nervous system must necessarily be occurring continuously as changes
be determined a posteriori, as a result of the actual behaving of the organism that the cells are able to stabilize with a permanency that lasts until the
subservient to the maintenance of its basic circularity. next modification, which can occur in any direction with respect to the
(3) The learning nervous system is a deterministic system with a relativistic previous one, or that subside by themselves after a certain number of inter-
self-regulating organization that defines its domain of interactions in terms of actions, but which are being locally triggered and selected through the actual
the states of neuronal activity that it maintains constant, both internally and concomitances of activity taking place in the neuropil itself.
at its sensory surfaces, and that specifies these states at any moment through (6) All changes in the nervous system during learning must occur without
its functioning, and through the learning (historical transformation) itself. interference with its continued functioning as a self-regulating system; the
Consequently, it must be able to undergo a continuous transformation, both unity that the observer sees in a living system throughout its continuous
in the states it maintains constant, and in the way it attains them, so that transformation is a strictly functional one. Accordingly, what appears constant
every interaction in which new classes of concomitances occur effectively for the observer when he ascertains that the same behavior is reenacted on a
modifies it (learning curves) in one direction or the other. Since this trans- different occasion, is a set of relations that he defines as characterizing it,
formation must occur as a continuous process of becoming without the regardless of any change in the neurophysiological process through which it is
previous specification of an end state, the final specification and optimization attained, or any other unconsidered aspect of the conduct itself. Learning, as
of a new behavior can only arise through the cumulative effect of many a relation between successive different modes of conduct of an organism such
equally directed interactions, each of which selects, from the domain of struc- that the present conduct appears as a transformation of a past conduct arising
tural changes possible to the nervous system in its structural dynamism, that from the recall of a specifiable past event, lies in the cognitive domain of the
which at that moment is congruent with its continued operation subservient observer as a description of his ordered experiences. Likewise, memory as an
to the basic circularity of the organism. Otherwise the organism disintegrates. allusion to a representation in the learning organism of its past experiences,
(4) The analysis of the nervous system made earlier indicated that the is also a description by the observer of his ordered interactions with the
states of neuronal activity that arise in it through each interaction embody observed organism; memory as a storage of representations of the environment
the relations given in the interaction, and not representations of the niche to be used on different occasions in recall does not exist as a neurophysio-
or the environment as the observer would describe them. This analysis also logical function.
indicated that functibnally such embodiments constitute changes in the reac- (7) It is sufficient for a system to change its- state after an interaction
tivity of the nervous system, as a system closed on itself, to the modulating in a manner such that whenever a similar interaction recurs some internally
influences of further interactions. Consequently what the observer calls 'recall' determined concomitant state does not recur, although the same overt
and 'memory' cannot be a process through which the organism confronts behavior is reenacted for it to treat two otherwise equivalent interactions as
each new experience with a stored representation of the niche before making different elements of the same class. Such a peculiar state could be described
a decision, but the expression of a modified system capable of synthesizing a as representing the emotional connotation of uncertainty which, present
new behavior relevant to its present state of activity. whenever a class of interactions is experienced for the first time, is suppressed
(5) It is known that many neurons change their transfer functions as a after such an experience; the absence of such a concomitant state would
result of the different concornitances of activity that occur in the neuropils suffice henceforth to treat differently (as known) all recurrent interactions of
of their collector and effector areas. Although it is not known what these the same class. I maintain that modifications of this sort in the reactivity of
changes are ( development of new synapses or changes in their size, membrane the nervous system constitute the basis for the unidirectional ordering of
changes, or changes in the pattern of spike invasion at the branching points of experiences in a living system through 'recognition' without any storage of
the axons), it can be expected from the relativistic organization of the nervous representations of the niche. First interactions that by error of the system
system that they should result in local morphological and functional changes are not accompanied by the above mentioned concomitant internal state
that do not represent any particular interaction, but which permanently alter (emotional connotation of uncertainty) would be treated as if known, as
38 HUMBERTO R. MATURANA COGNITIVE FUNCTION IN PARTICULAR 39

occurs in the deja vu. Conversely, interference with the suppression of the one of its states can be its input and can modify it as an interacting unit. We
concomitant state of activity corresponding to this emotional connotation can say that every internal interaction changes us because it modifies our
would result in the treatment of any recurrent interaction as if new (loss of internal state, changing our posture or perspective (as a functional state)
recent memory). from which we enter into a new interaction. As a result new relations are
(8) If such a system is capable of discourse, it will generate the temporal necessarily created in each interaction and, embodied in new states of activ-
domain through the ascription of a unidirectional order to its experiences as ity, we interact with them in a process that repeats itself as a historical and
they differ in their emotional connotations, and although it will continue to unlimited transformation.
function in the present as an atemporal system, it will interact through its (3) The observer generates a spoken description of his cognitive domain
descriptions in the temporal domain. Past, present, and future, and time in (which includes his interactions with and through instruments). Whatever
general belong exclusively to the cognitive domain of the observer. description he makes, however, that description corresponds to a set of
permitted states of relative activity in his nervous system embodying the
relations given in his interactions. These permitted states of relative activity
THE OBSERVER and those recursively generated by them are made possible by the anatomical
and functional organization of the nervous system through its capacity to
Epistemological and Ontological Implications interact with its own states. The nervous system in turn has evolved as a
(1) The cognitive domain is the entire domain of interactions of the organism. system structurally and functionally subservient to the basic circularity of the
The cognitive domain can be enlarged if new modes of interactions are living organization, and hence, embodies an inescapable logic: that logic
generated. Instruments enlarge our cognitive domain. which allows for a match between the organization of the living system and
(2) The possibility of enlargement of the cognitive domain is unlimited; the interactions into which it can enter without losing its identity.
it is a historical process. Our brain, the brain of the observer, has specialized (4) The observer can describe a system that gives rise to a system that
during evolution as an instrument for the discrimination of relations both can describe, hence, to an observer. A spoken explanation is a paraphrase, a
internally and externally generated relations, but relations given throu~ and description of the synthesis of that which is to be explained; the observer
by interactions and embodied in the states of relative activity of its neurons. explains the observer. A spoken explanation, however, lies in the domain of
Furthermore; this occurs under circumstances in which the discriminations discourse. Only a full reproduction is a full explanation.
between states of relative activity - that for an observer represent the inter- (5) The domain of discourse is a closed domain_, and it is not possible to
actions of the organism, for the nervous system, that operate as a closed step outside of it through discourse. Because the domain of discourse is a
network - constitute only changes of relations of activity that arise between closed domain it is possible to make the following ontological statement:
its components while it generates the internal and the sensory motor correla- the logic of the description is the logic of the describing (living) system (and
tions that the states of the organism select. This has two aspects: one refers his cognitive domain). ·
to the functional organization of the nerve cells which, with their responses, (6) This logic demands a substratum for the occurrence of the discourse.
discriminate between different states of relative activity impinging upon We cannot talk about this substratum in absolute terms, however, because
them; the other refers to the ability of the nervous system, as a neuronal we would have to describe it, and a description is a set of interactions into
organization, to discriminate between its own states ·as these are distinguished which the describer and the listener can enter, and their discourse about these
and specified by the further states of activity that they generate. From this interactions will be another set of descriptive interactions that will remain in
capacity of the nervous system to interact discriminately with its own states the same domain. Thus, although this substratum is required for epistemologi-
in a continuous process of self-transformation, regardless of how these states cal reasons, nothing can be. said about it other than what is meant in the
are generated, behavior emerges as a continuum of self-referred functional ontological statement above.
transformation. We cannot say in absolute terms what constitutes an input (7) We as observers live in a domain of discourse interacting with descrip·
to our nervous system (the nervous system of the observer), because every tions of our descriptions in a recursive manner, and thus continuously generate
40 HUMBERTO R. MATURANA

new elements of interaction. As living systems, however, we are closed

systems modulated by interactions through which we define independent
entities whose only. reality lies in the interactions that specify them (their V. PROBLEMS IN THE NEUROPHYSIOLOGY OF
Description). COGNITION
(8) For epistemological reasons we can say: there are properties which
are manifold and remain constant through interactions. The invariance of
properties through interactions provides a functional origin to entities or (1) The observer can always remain in a domain of interactions encompassing
units of interactions; since entities are generated through the interactions his own interactions; he has a nervous system capable of interacting with its
that define them (properties), entities with different classes of properties own states, which, by doing so in a functional context that defines these
generate independent domains of interactions: no reductionism is possible. states as representations of the interactions from which they arise, allows him
to interact recursively with representations of his interactions. This is possible
because due to the gerieral mode of organization of the nervous system there
is no intrinsic difference between its internally and externally generated states
of activity, and because each one of its specific states of activity is specifiable
only in reference to other states of activity of the system itself.
(2) An organism with a nervous system capable of interacting with its own
states is capable of descriptions and of being an observer if its states arise
from learned orienting interactions in a consensual domain: it can describe
its describing [Cf. Gardner and Gardner, 1969]. Through describing itself
in a recursive manner, such an organism becomes a self-observing system
that generates the domain of self-consciousness as a domain of self-observa-
tion. Self-consciousness then is not a neurophysiological phenomenon, it
is a consensual phenomenon emerging in an independent domain of inter-
actions from self-orienting behavior and lies entirely in the linguistic domain.
The implications are twofold:

(a) The linguistic domain as a domain of orienting behavior requires at

least two interacting organisms with comparable domains of interactions, so
that a cooperative system of consensual interactions may be developed in
which the emerging conduct of the two organisms is relevant for both. The
specifiability through learning of the orienting interactions allows for a
purely consensual ( cultural) evolution in this domain, without it necessarily
involving any further evolution of the· nervous system; fot this reason the
linguistic domain in general, and the domain of self-consciousness in partic-
ular, are, in principle, independent of the biological substratum that generates
them. However, in the actual becoming of the living system this independence
is incomplete, on the one hand because the anatomical and neurophysiological
organization of the brain, by determining the actual possibilities of confluence
of different states of activity in it, specifies both the domain of possible
41
 PROBLEMS IN THE NEUROPHYSIOLOGY OF COGNITION 43
42 HUMBERTO R. MATURANA

of the nervous system onto itself determine its capacity to single

interactions of the organism with relations and the complexity of the patterns
out some of its own states and interact with them independently?.
of orienting interactions that it can distinguish, and on the other hand l;>ecause
the necessary subservience of the linguistic domain to the maintenance of (3) At any moment each nerve cell responds in a determinis~ic manner,
the basic circularity of the organism through the generation of modes of and according to well defined transfer functions to classes of spatio-temporal
behavior that directly or indirectly satisfy it limits the type of conduct that activity caused at its collector area by the afferent influ~nces imping~ng upo~
the organism can have without an immediate or eventual disintegration, or, of it· this occurs independently of how these afferent influences anse. This
course, reduced rate of reproduction. Consequently, then, although the ru'ode of cellular operation constitutes the basis for an associative process in
purely consensual aspects of the cultural evolution are independent of a which, whenever a given state of activity is produced in the nervous system,
simultaneous evolution of the nervous system, those aspects of the cultural all neurons for which this state generates the proper classes of afferent
evolution which depend on the possibility of establishing new classes of influences enter into activity. Association thus conceived neurophysiologically
concomitances of activity in the nervous system, and generate new relations is an inevitable process that calls into activity all cells that can be activated
between otherwise independent domains, are not thus independent. Accord- at any moment by a given state of the nervous system. No consideration of
ingly, once a cultural domain is established, the subsequent evolution of the meaning enters into such a notion, since meaning, as a description by the
nervous system is necessarily subordinated to it in the measure that it deter- observer, refers to the relevance that a mode of behavior has in the main-
mines the functional validity of the new kinds of concomitances of activity tenance of the basic circularity of the organism as a consequence of self-
that may arise in the nervous system througl1 genetic variability. regulation and not in the mechanisms of the genesis of conduct. Association
in terms ~f representations related by meaning lies in the cognitive domain
(b) Since self-consciousness and the linguistic domain in general are not of the observer exclusively. The nervous system is a system that functions
neurophysiological phenomena, it is impossible to account for them in terms maintaining constant certain states of relative activity, both internally and
of excitation, inhibition, networks, coding, or whatever else is the stuff of at the sensory surfaces, with reference only to some of its other states of
neurophysiology. In fact, the linguistic domain is fully explained only by relative activity. In this context the following considerations about its func-
showing how it emerges from the recursive application of orienting inter-
tional organization are significant:
actions on the results of their applications without being restricted as a do-
main by the neurophysiological substratum; what indeed is the problem is the (a) The nervous system can be described as a system that ~as ev~l~ed
need to account in purely physiological terms, without reference to meaning, to specialize in tlie discrimination between states of neuronal relative act1V1ty
for the synthesis of behavior in general, and for the synthesis of orienting (particularly in man) each of which is defined by the behavior it generates.
behavior in particular. Accordingly, the fundamental quest in this respect This is valid for innate and learned behavior in circumstances in which every
should be to understand and explain behavior is defined either by a set of states of activity maintained constant,
or by their path of variation, both internally and at the sensory surfaces.
(i) how does the nervous system interact with its own states, and
is modified by them as if they were independent entities?; (b) The basic connectivity of the nervous system, and the original
reactive capacity of the nerve cells, with which any animal is endowed by
(ii) how are these states specified neurophysiologically if they development, secures a basic pattern of flow for the nervous activity origi-
are defined by their own effectiveness in bringing forth certain nating at any point in it. Thus, development specifies and determines both an
internal or sensory states in the system?; initial repertoire of behavior over which all new conduct is built in a historical
process of transformation, and an initial structurally specified set of possible
(iii) how is a given effector per!ormance synthesized that is
associations that changes in an integrated manner with the historical trans-
defined by the relative states of activity that it generates in the
formation ofbehavior [Cf. Lorenz, 1966].
sensory surfaces and in the system itself?; and
(c) Any modification of the transfer function of a nerve cell, resulting
(iv) how do the double or triple internal anatomical projections
 PROBLEMS IN THE NEUROPHYSIOLOGY OF COGNITION 45
44 HUMBERTO R. MATURANA

from new concomitances of activity, occurs modifying a preexisting behavior corresponding postsynaptic neurons are activated. Such can occur if the
in a system that operates through maintaining invariant its definitory internai probability of spike invasion at the branching points of the afferent axons in
relations. In fact, any local change-that would lead to the synthesis of a modi- the neuropils is permanently modified in one direction or another by the
fied conduct by the brganism, must be immediately accompanied by other coincident'novel activity in the neighboring structures, which, in the absence
changes arising through the adjustments that this must undergo in the process of synaptic· interactions, cause, through local currents, local processes of
of maintaining constant its internal relations under its changed behavior. This growth or ungrowth in the branching zones of these axons. If this were the
is why it is the immediate relevance of a conduct for the maintenance of the case four things would occur:
organism in the present which at any moment selects the changes that take (i) The state of the nervous system would change, and hence, also its
place during learning, and not the possible value of the conduct for future conduct, according to the new concomitances of activity produced in the
action. neuropils through its different interactions.
(d) It is apparent that the nervous system cannot determine in advance (ii) Each state of activity of the system (as a state of relative neuronal
the concomitances of activity under which it should change in a permanent activity) would be defined by the concomitances of activity in the neuropil
manner; for it to satisfy future needs of the organism, it must operate under that generate it, such that if they recur, it recurs.
non-predictive changes continuously selected by the concomitances of activ-
ity arising in it. For this the nervous system must be capable of successful (iii) Each new functional state of the neuropils would necessarily
operation under the continuous transformation of its capacity to synthesize constitute the basis for their further modification, in such a manner that
behavior, which necessarily results from a continuous change of the neuro- their morphological and functional organization would be under continuous
physiological concomitances that determine the effective spatio-temporal historical transformation.
configuration of activity impinging on the collector areas of its component (iv) These changes in the neuropils would change the participation of
neurons. Accordingly, it would seem of fundamental importance for the the different neurons in the synthesis of behavior, independently of whether
functional transformation of the system that many of its neurons should be or not there are also changes in their transfer functions, by changing the
able to change their relative participation in the synthesis of behavior as circumstances of their activation. Accordingly, if an interaction (as described
elements of different states of relative neuronal activity, independently of by the observer) recurs, no past conduct could be strictly reenacted by the
whether or not this is ?Ccompanied by any change in their transfer functions. ,organism, but this would have to synthesize a new adequate behavior that
In these circumstances the actual problem for the successful operation of the generates, in the context of its present interaction and in a manner that be-
nervous system is the generation at any moment of the optimal configuration came specified through its structural transformation along its history of inter-
of activity necessary to synthesize a given behavior. However, since this actions, the internal and sensory motor correlations that maintain its identity.
continuous transformation of the functional capacity of the nervous system '
necessarily occurs under continuously successful behavior, such optimization (4) Learning is not a process of accumulation of representations of the
requires no other specification than its attainment through the converging environment; it is a continuous process of transformation of behavior through
transformation of behavior itself. continuous change in the capacity of the nervous system to synthesize it.
Recall does not depend on the indefinite retention of a structural invariant
(e) Since the nervous system is an inferential system, that is, since it that represents an entity (an idea, image, or symbol), but on the functional
functions as if any state that occurred once will occur again, a significant ability of the system to create, when certain recurrent conditions are given,
feature of its organization must be its necessary and continuous transformation a behavior that satisfies the recurrent demands or that the observer would
as a function of the new concomitances of activity occurring in it. This class as a reenacting of a previous one. As a consequence, the quest in the
functional requirement could be satisfied, for example, if any new local study of the learning process must answer two _basic questions:
concomitance of activity in the neuropils changes the nerve cells in a deter-
ministic and specific manner which does not represent any entity or event, 'What changes can a neuron undergO' (in any of its component
but which modifies the neurophysiological circumstances under which the parts) which it can maintain constant for a certain time, and
46 HUMBERTO R. MATURANA PROBLEMS IN THE NEUROPHYSIOLOGY OF COGNITION 47

which modify in a definite manner its possible participation in what he thinks it should be. Each component of the nervous system that the
different configurations of relative neuronal activity?'; and observer describes is defined in the domain of interactions of his observations,
and as such is alien to the system which it is supposed to integrate. Every
'What organization of the nervous system would permit continu-
function has a structure which embodies it and makes it possible, but this
ous changes in the relative activity of its anatomical components,
structure is defined by ilie function in the domain of its operation as a set of
as a result of different concomitances in their activity, and still
relations between elements also defined in this domain. Neurons are the
pennit the synthesis of a conduct that is defined only by .the
anatomical units of the nervous system, but are not the structural elements
states of relative neuronal activity that it generates, and not by
of its functioning. The structural elements of the functioning nervous system
the compcnents used?'.
have not yet been defined, and it will probably be apparent when iliey are
(5) The nervous system is a strictly deterministic system whose structure defined that iliey must be expressed in terms of invar!ants of relative activities
specifies tl1e possible modes of conduct that may emerge (be synthesized) between neurons, in some manner embodied in invariants of relations of
from its functioning in a manner that varies according to the species, and ilie interconnections, and not in terms of separate anatomical entities. In man-
reactive perspective from which these modes of conduct may emerge. The made systems iliis conceptual difficulty has not been so apparent because the
reactive perspective, which the observer would call the emotional tone, does system of relations (the theory) that integrates the parts that the describer
not specify a particular conduct, but determines ilie nature (aggressive, fearful, (ilie observer) defines is provided by him, and is specified in his domain of·
timid, etc.) of the course of the interaction [Cf. Kilmer, McCulloch and interactions; as a consequence, these relations appear so obvious to ilie
Blum, 1968]. Changes during development, maturation, hormonal action, observer iliat he treats them as arising from the observation of the parts, and
drugs, or learning, do not modify the deterministic character of this organiza- deludes himself, denying that he provides the unformulated theory that
tion but change the capacity that ilie system has at any moment to synthesize embodies the structure of ilie system which he projects onto them. In a self-
behavior. Furiliermore, although any conduct or functional state always referring system like a living system the situation is different: the observer
arises through a process of historical transformation from pre-existing modes can only make a description of his interactions wiili parts that he defines
of conduct or functional states, the nervous system functions in the present, ilirough interactions, but iliese parts lie in his cognitive domain only. Unless
and past history does not participate as an operant neurophysiological factor he explicitly or implicitly provides a theory iliat embodies the relational
in the synthesis of conduct; nor does meaning, ilie relevance that a particular structure of the system, and conceptually supersedes his description of the
mode of conduct has, participate in it either. Time and meaning are effective components, he can never understand it. Accordingly, the full explanation of
factors in the linguistic domain, but as relational entities do not have neuro- ilie organization of the nervous system (and of the organism) will not arise
physiological correlates in the operation of the nervous system. Nor is the from any particular observation or detailed description and enumeration of
functional unity of the nervous system attained through a specific feature its parts, but rather like any explanation, from the synthesis, conceptual or
of its organization, but emerges from the functioning of its components concrete, of a system that does what the nervous system (or the organism)
(whatever these may be), each one to its own accord, under circumstances does.
that define the ensemble as a unit of interactions in a particular domain [Cf.
Lindauer, 1967, as an example in a social organism], and has no reality
independent of these circumstances. Thus there is no peculiar neurophysio-
logical process iliat could be shown to be responsible for tltis unity and to
explain it. Furiliermore, in a strict sense, although the nervous system has
anatomical components it does not have functional parts since any mutilation
leaves a functioning unit, with different properties as expressed by its possible
interactions, but a unit in the corresponding domain. It appears incomplete
only for the observer who beholds it as an entity from the perspective of
 CONCLUSIONS 49

a living system into parts that he defines, the description of these parts does
not and cannot represent a living system. In principle a part should be de-
VI. CONCLUSIONS finable through its relations within the unit that it contributes to form by its
operation and interactions with other parts; this, however, cannot be attained
because the analysis of a unit into parts by the observer destroys the very
relations that would be significant for their characterization as effective
The aim set forth in the introduction has been accomplished. Through the components of the unit. Furiliermore, these relations cannot be recovered
description of the self-referring circular organization of the living system, and ilirough a description which lies in the cognitive domain of the observer and
through the analysis c.f the domains of interactions that such an organization reflects only his interactions with the new units that he creates ilirough his
specifies, I have shown the emergence of a self-referring system capable of analysis. Accordingly, in a strict sense a unit does not have parts, and a unit
making descriptions and of generating, through orienting interactions with is a unit only to the extent that it has a domain of interactions that defines
other, similar, systems and with itself, both a consensual linguistic domain it as different from that with respect to which it is a unit, and can be referred
and a domain of self-consciousness, that is: I have shown the emergence of to only, as done above with ilie living system, by characterizing its organiza-
the observer. This result alone satisfies the fundamental demand put forth at tion through ilie domain of interactions which specify this distinction. In
the outset: 'T1ze observer is a living system and any understanding of cognition this context, the notion of component is necessary only for epistemological
as a biological phenomenon must account for the observer and his role in it', reasons in order to refer to the genesis of the organization of the unit through
and proves the validity of this analysis. our description, but this use does not reflect the nature of its composition.
Al though the answers to the various questions posed in the introduction
and the fundamental implications of the analysis are to be found in the text (ii) For every living system its particular case of self-referring circular
itself to the extent that the theory adequately founds its whole development, organization specifies a closed domain of interactions that is its cognitive
there are several conclusions that I would like to state explicitly: domain, and no interaction is possible for it which is not prescribed by this
organization. Accordingly, for every living system the process of cognition
(i) The living organization is a circular organization which secures the consists in ilie creation of a field of behavior through its actual conduct in its
production or maintenance of the components that specify it in such a closed domain of interactions, and not in the apprehension or the description
manner that the product of their functioning is the very same organization of an independent universe. Our cognitive process (the cognitive process of
that produces them. Accordingly, a living system is an homeostatic system the observer) differs from the cognitive processes of other organisms only in
whose homeostatic organization has its own organization as tl1e variable the kinds of interactions into which we can enter, such as linguistic inter-
that it maintains constant through the production and functioning of the actions, and not in ilie nature of the cognitive process itself. In this strictly
components that specify it, and is defined as a unit of interactions by this subject-dependent creative process, inductive inference is a necessary function
very organization. It follows iliat living systems are a subclass of the class of (mode of conduct) that emerges as a result of the self-referring circular
circular and homeostatic systems. Also, it is apparent that the components organization which treats every interaction and the internal state that it
referred to above cannot be specified as parts of the living system by ilie generates as if it were to be repeated, and as if an element of a class. Hence,
observer who can only subdivide a system in parts that he defines through his functionally, for a living system every experience is the experience of a
interactions, and which, necessarily, lie exclusively in his cognitive domain general case, and it is the particular case, not the general one, which requires
and are operationally determined by his mode of analysis. Furthermore, the many independent experiences in order that it be specified through the
relations through which the observer claims that these parts constitute a intersection of various classes of interactions. Consequently, although due
unitary system are relations that arise only through him by his simultaneous to the historical transformation they have caused in organisms, or in their
interactions with the parts and the intact system, and, hence, belong ex- nervous systems, past interactions determine tl1e inductive inferences that
clusively to his cognitive domain. Thus, aliliough ilie observer can decompose these make in the present, they do not participate in the inductive process
48
50 HUMBERTO R. MATURANA CONCLUSIONS 51

itself. Inductive inference as a structural property of the living organization self-referring. What occurs in a living system is analogous to what occurs in an
and of the thinking process, is independent of history, or of the relations instrumental flight where the pilot does not have access to the outside world
between past and pre~ent that belong only to the domain of the observer. and must function only as a controller of the values shown in his flight
instruments. His task is to secure a path of variations in the readings of his
(iii) Linguistic interactions orient the listener within his cognitive
instruments, either according to a prescribed plan, or to one that becomes
domain, but do not specify the course of his ensuing conduct. The basic
specified by these readings. When the pilot steps out of the plane he is be-
function of language as a system of orienting behavior is not the transmission
wildered by the congratulations of his friends on account of the perfect flight
of information or the description of an independent universe about which
and landing that he performed in absolute darkness. He is perplexed because
we can talk, but the creation of a consensual domain of behavior between
to his knowledge all that he did at any moment was to maintain the readings
linguistically interacting systems through the development of a cooperative
of his instruments within certain specified limits, a task which is in no way re-
domain of interactions.
presented by the description that his friends ( observers) make of his conduct.
(iv) Through language we interact in a domain of descriptions within In terms of their functional organization living systems do not have inputs
which we necessarily remain even when we make assertions about the universe and outputs, although under perturbations they maintain constant their set
or about our knowledge of it. This domain is both bounded and infinite; states, and it is only in our descriptions, when we include them as parts of
bounded because everything we say is a description, and infinite because larger systems which we define, that we can say that they do. When we adopt
every description constitutes in us the basis for new orienting interactions, this descriptive approach in our analysis of the living organization we cannot
and hence, for new descriptions. From this process of recursive application of but subordinate our understanding of it to notions valid only for man-made
descriptions self-consciousness emerges as a new phenomenon in a domain (allo-referring) systems, where indeed input and output functions are all
of self-description, with no other neurophysiological substratum than the important through the purposeful design of their role in the larger systems
neurophysiological substratum of orienting behavior itself. TI1e domain of in which they are included, and this is misleading. In the organization of the
self-consciousness as a domain of recursive self-descriptions is thus also living systems the role of the effector surfaces is only to maintain constant
bounded and infinite. the set states of the receptor surfaces, not to act upon an environment, no
matter how adequate such a description may seem to be for the analysis
(v) A living system is not a goal-directed system; it is, like the nervous
of adaptation, or other processes; a grasp of this is fundamental for the under-
system, a stable state.-determined and strictly deterministic system closed on
standing of the organization of living systems.
itself and modulated by interactions not specified through its conduct. These
modulations, however, are apparent as modulations only for the observer
(vii) Tiie cognitive domain of the observer is bounded but unlimited·
who beholds the organism or the nervous system externally, from his own
he can in an endless recursive manner interact with representations of hi~
conceptual (descriptive) perspective, as lying in an environment and as
interactions and generate through himself relations between otherwise in-
elements in his domain of interactions. Contrariwise, for the functioning of
dependent domains. These relations are novelties which, arising through the
the self-referring system itself all that there is is the sequence of its own self-
o?server, _have no other (and no less) effectiveness than that given to them by
subservient states. If this distinction is not made, one is liable to fail by
his behav10r. Thus,he both creates (invents) relations and generates (specifies)
including in the explanation of the organism and the nervous system features
~e ~odd ( d?~ain of i~teractions) in which he lives by continuously expand-
of interactions (descriptions) that belong exclusively to the cognitive domain
mg his cogmttve domam through recursive descriptions and representations
of the observer.
of his interactions. The new, then, is a necessary result of the historical
(vi) It is tempting to talk about the nervous system as one would talk organization of the observer that makes of every attained state the starting
about a stable system with input. TI1is I reject because it misses entirely the point for the specification of the next one, which thus cannot be a strict
point by introducing the distortion of our participation as observers into the repetition of any previous state; creativity is the cultural expression of this
explanation of systems whose organization must be understood as entirely unavoidable feature.
52 HUMBERTO R. MATURANA CONCLUSlONS 53

(viii) The logic of the description and, hence, of behavior in general is, linguistic interactions to what we deem are sensory experiences of concrete
necessarily, the logic of the describing system; given behavior as a referential entities, but which have turned out to be, as are thoughts and descriptions,
and detenninistic sequence of states of nervous activity in which each state states of relative activity between neurons that generate new descriptions.
determines the next one within the same frame of reference, no contradiction The question, 'What is the object of knowledge?' becomes meaningless. There
can possibly arise in it as long as the latter remains unchanged by intercurrent is no object of knowledge. To know is to be able to operate adequately in an
interactions. If a change in the frame of reference takes place while a given individual or cooperative situation. We cannot speak about the substratum in
behavior develops, a new one appears, such that the states following the which our cognitive behavior is given, and about that of which we cannot
change are determined with respect to it. If the new sequence of states speak, we must remain silent, as indicated by Wittgenstein. This silence,
(behavior) appears to an observer as contradicting the previous ones, this is however, does not mean that we fall into solipsism or any sort of metaphysical
so because he provides an independent and constant frame of reference in idealism. It means that we recognize that we, as thinking systems, live in a
relation to which the successive sequences of states (behaviors) are contradic- domain of descriptions, as has already been indicated by Berkeley, and that
tory. Such contradiction, however, lies exclusively in the cognitive domain of through descriptions we can indefinitely increase the complexity of our
the observer, or of whatever provides the independent constant frame of cognitive domain. Our view of the universe and of the questions we ask must
reference. Contradictions (inconsistencies) then, do not arise in the generation change accordingly. Furthermore, this re-emergence of reality as a domain
of behavior but pertain to a domain in which the different behaviors acquire of descriptions does not contradict determinism and predictability in the
their significance by confronting an encompassing frame of reference through different domains of interactions; on the contrary, it gives them foundation
the interactions of the organism. Accordingly, thinking and discourse as by showing that they are a necessary consequence of the isomorphism be-
modes of behavior are necessarily logically consistent in their generation, tween the logic of the description and the logic of the describing system. It
and that which the observer calls rational in them because they appear as also shows that determinism and predictability are valid only within the field
concatenations of non-contradictory sequence dependent descriptions, is an of this isomorphism; that is, they are valid only for the interactions that
expression of this necessary logical consistency. It follows that inconsistencies define a domain.
(irrationalities) in thinking and discourse as they appear to the observer arise
from contextual changes in the circumstances that generate them while the (x) The genetic and nervous systems are said to code information
independent frame of reference provided by the observer remains unchanged. about the environment and to represent it in their functional organization.
This is untenable; the genetic and nervous systems code processes that specify
(ix) Due to the nature of the cognitive process and the function series of transformations from intial states, which can be decoded only
of the linguistic interactions, we cannot say anything about that which is througl1 their actual implementation, not descriptions that the observer
independent ofus and with which we cannot interact; to do that would imply makes of an environment which lies exclusively in his cognitive domain [Cf.
a description and a description as a mode of conduct represents only relations Bernal, 1965] . The following is an illustration of the problem:
given in interactions. Because the logic of the description is the same as the Let us suppose that we want to build two houses. For such a purpose we
logic of the describing system we can assert the epistemological need for a hire two groups of thirteen workers each. We name one of the workers of the
substratum for the interactions to occur, but we cannot characterize this first group as the group leader and give him a book which contains all the
substratum in terms of properties independent of the observer. From this it plans of the house showing in a standard way the layout of walls, water pipes,
follows that reality as a universe of independent entities about which we can electric connections, windows, etc., plus several views in perspective of the
talk is, necessarily, a fiction of the purely descriptive domain, and that we finished house. The workers study the plans and under the guidance of the
should in fact apply the notion of reality to this very domain of descriptions leader construct the house, approximating continuously the final state
in which we, the describing system, interact with our descriptions as if with prescribed by the description. In the second group we do not name a leader,
independent entities. TI1is change in the notion of reality must be properly we only arrange the workers in a starting line in the field and give each
understood. We are used to talking about reality orienting each other through of them a book, the same book for all, containing only neighborhood in-
 ' HUMBERTO R. MATURANA CONCLUSIONS 55
54

structions. These instructions do not contain words such as house, pipes, self-specification when witnessed by the observer as ifin their progressive self-
or windows, nor do they contain drawings or plans of the house to be con- decoding into growth and behavior.
structed; they cont~in only instructions of what a worker should do in the
different positions and in the different relations in which he finds himself as (xi) There are different domains of interactions, and these different
his position and relations change. domains cannot explain each other because it is not possible to generate the
Although these books are all identical the workers read and apply different phenomena of one domain with the elements of another; one remains in the
instructions because they start from different positions and follow different same domain. One domain may generate the elements of another domain, but
paths of change. The end result in both cases is the same, namely, a house. not its phenomenology, which in each domain is specified by the interactions
The workers of the first group, however, construct something whose final of its elements, and the elements of a domain become defined only through
appearance they know all the time, while the workers of the second group the domain that they generate. Any nexus between different domains is
have no views of what they are building, nor do they need to have obtained provided by the observer who can interact as if with a single entity with the
them even when they are finished. For the observer both groups are building conjoined states of nervous activity generated in his brain by his concomitant
a house, and he knows it from the start, but the house that the second group interactions in several domains, or with independent descriptions of these
builds lies only in his cognitive domain; the house built by the first group, interactions. Through these concomitant interactions in different domains (or
however, is also in the cognitive domains of the workers. The coding is with several descriptions within the descriptive domain) the observer generates
obviously different in the two cases. In fact, the instructions contained in the relations between different domains (or between different descriptions) as
book given to the first group clearly code the house as the observer would states of neuronal activity that in him lead to definite modes of conduct
describe it, and the decoding task of the workers consists in purposefully (descriptions) that represent these conjoined interactions as singular in-
doing things that will approximate to the construction of the described final dependent entities. The number and kinds of relations the observer can
state; this is why the house must be in their cognitive domain. In the second generate in this manner is potentially infinite due to his recursive interactions
case · the instructions contained in each one of the thirteen identical books with descriptions. Thus, relations, as states of neuronal activity arising from
do ~ot code a house. They code a process that constitutes a path of changing the concurrent interactions of the observer in different domains (physical and
relationships which, if carried through under certain conditions, results in a relational) constitute the elements of a new domain in which the observer
system with a domain of interactions which has no intrinsic relationship with interacts as a thinking system, but do not reduce one phenomenological
the beholding observer. That the observer should call this system a house is domain into another. It is the simultaneous logical isomorphism of the new
a feature of his cognitive domain, not of the system itself. In the first case element (relations) with their source systems through their mode of origin
the coding is isomorphic with a description of the house by the observer, and (class intersection) that gives the new domain thus generated (descriptions)
in fact constitutes a representation of it; in the second case it is not. The first its explanatory capacity. An explanation is always a reproduction, either a
case is typical of the way in which the observer codes the systems that he concrete one through the synthesis of an equivalent physical system, or a
builds; the second corresponds to the way that the genome and nervous conceptual one through a description from which emerges a system logically
system constitute codes for the organism and for behavior, respectively, and isomorphic to the original one, but never a reduction of one phenomenologi-
one would never find in these codes any isomorphism with the description cal domain into another. An adequate understanding of this irreducibility is
that the observer would make of the resultant systems with which he inter- essential for the comprehension of the biological phenomena, the consensual
acts. In what sense could one then say that the genetic and nervous systems domains that living systems generate, and their conjoined evolution.
code information about the environment? The notion of information refers Many conclusions about self-consciousness and knowledge which arise
to the observer's degree of uncertainty in his behavior within a domain of from this mode of analysis have been proposed in one way or another by
alternatives defined by him, hence the notion of information only applies scientists and philosophers from their intuitive understanding, but never, to
within his cognitive domain. Accordingly, what one could at most say is my knowledge, with an adequate biological and epistemological foundation.
that the genetic and nervous systems generate information through their This I have done through the distinction between what pertains to the domain
 HUMBERTO R. MATURANA
56

of the observer, and what pertains to the domain of the organism, and t~rough
carrying to their ultima'te consequences the implicati?ns ?f ~he cucular
self-referring organization of the living systems: the 1mphcat10ns of the POST SCRIPTUM
functionally closed nature of the relativistic organization. of the ner".'ous
system as a system under continuous transformation ~etermm:d b~ relations
of neuronal activity without the system ever stepping outside itself; and
the implications of the non-informative orienting functi?n of linguis~ic No scientific work should be done without recognizing its ethical implications;
interactions. It is only after this has been done that the functional complextty in the present case the following deserve special attention:
of the living and linguistically interacting system can be properly. grasped (i) Man is a deterministic and relativistic self-referring autonomous system
without its being concealed through such magic words as· consciousness, whose life acquires its peculiar dimension through self-consciousness; ethic
symbolization, or information. Most of the detailed work is yet to ~e done, and morality arise as commentaries that he makes on his behavior through
of course but the fundamental first step of defining the perspective from self-observation. He lives in a continuously changing domain of descriptions
which to 'iook has here been taken. As a final remark, one could say what that he generates through recursive interactions within that domain, and
appears to be another paradox, but which points to the conceptual problem: which has no other constant element in its historical transformation than his
maintained identity as an interacting system. That is, man changes and lives in
Living systems in general, and their nervous system in particular, a changing frame of reference in a world continuously created and transformed
are not made to handle a medium, although it has been through by him. Successful interactions directly or indirectly subservient to the
the evolution of their handling of their medium that they have maintenance of his living organization constitute his only fmal source of
become what they are, such that we can say what we can say reference for valid b.ehavior within the domain of descriptions, and, hence,
about them. for truth; but, since living systems are self-referential systems, any final frame
of reference is, necessarily, relative. Accordingly, no absolute system of values
is possible and all truth and falsehood in the cultural domain are necessarily
relative.
(ii) Language does not transmit information and its functional role is the
creation of a cooperative domain of interactions between speakers through
the development of a common frame of reference, although each speaker acts
exclusively within his cognitive domain where all ultimate truth is contingent
to personal experience. Since a frame of reference is defined by the classes of
choices which it specifies, linguistic behavior cannot but be rational, that is,
determined by relations of necessity within the frame of reference within
which it develops. Consequently, no one can ever be rationally convinced
of a truth which he did not have already implicitly in his ultimate body of
beliefs.
(iii) Man is a rational animal that constructs his rational systems as all
rational systems are constructed, that is, based on arbitrarily accepted truths
(premises); being himself a relativistic self-referring deterministic system this
cannot be otherwise. But if only a relative, arbitrarily chosen system of
reference is possible, the unavoidable task of man as a self-conscious animal
that can be an observer of its own cognitive processes is to explicitly choose a
57
58 HUMBERTO R. MATURANA

frame of reference for his system of values. This task he has always avoided
AUTOPOIESIS
by resorting to god as an absolute source of truth, or to self-delusion through
reason, which can be'used to justify anything by confusing the frames of
The Organization of the Living
reference and arguing in one domain with relations valid in another. The
ultimate truth on which a man bases his rational conduct is necessarily 1973
subordinated to his personal experience and appears as an act of choice
expressing a preference that cannot be transferred rationally; accordingly,
the alternative to reason, as a source for a universal system of values, is
by
aesthetic seduction in favor of a frame of reference specifically designed to
comply with his desires (and not his needs) and defining the functions to be HUMBERTO R. MATURANA
satisfied by the world (cultural and material) in which he wants to live.
and

FRANCISCO J. VARELA
 TABLE OF CONTENTS

PREF ACE by Sir Stafford Beer

63
INTRODUCTION
73
I. ON MACHINES, LIVING AND OTHERWISE 77
1. Machines
77
2. Living Machines
78
II. DISPENSABILITY OF TELEONOMY
85
1. Purposelessness
85
2. Individuality
87
III. EMBODIMENTS OF AUTOPOIESIS
88
1. Descriptive and Causal Notions
88
2. Molecular Embodiments
90
3. Origin
93
IV. DIVERSITY OF AUTOPOIESIS
96
1. Subordination to the Condition of Unity
96
2. Plasticity of Ontogeny
98
3. Reproduction, a Complication of the Unity
100
4. Evolution, a Historical Network
102
5. Second and Thb:d Order Autopoietic Systems 107
V. PRESENCE OF AUTOPOIESIS
112
1. Biological Implications
112
2. Epistemological Implications
115
3. Cognitive Implications
118
APPENDIX: The Nervous System
124
GLOSSARY
135
 STAFFORD BEER

PREFACE

This small book is very large: it contains the living universe. It is a privilege
to be asked to write this preface, and a delight to do so. That is because I
recognize here a really important book, both in general and specifically.
Before talking about the specific contents at all, I would like to ex.plain why
this is in general so.

IN GENERAL

We are the inheritors 'of categorized knowledge; therefore we inherit also a

world view that consists of parts strung together, rather than of wholes
regarded through different sets of filters. Historically, synthesis seems to have
been too much for the human mind - where pratical affairs were concerned.
The descent of the synthetic method from Plato through Augustine took
men's perception into literature, art and mysticism. The modern world of
science and technology is bred from Aristotle and Aquinas by analysis. The
categorization that took hold of medieval scholasticism has really lasted it
out. We may see with hindsight that the historic revolts against the scholastics
did not shake free from the shackles of tl1eir reductionism.
The revolt of the rationalists - Descartes, Spinoza, Leibniz - began from
a principle of 'methodical doubt'. But they became lost in mechanism,
dualism, more and more categorization; and they ended in denying relation
altogether. But relation is the stuff of.system. Relation is the essence of
synthesis. The revolt of the empiricists - Locke, Berkeley, Hume - began
from the nature of understanding about the environment. But analysis was
still the method, and categorization still the practical tool of advance. In the
bizarre outcome, whereby it was the empiricists who denied the very existence
of the empirical world, relation survived - but only through the concept of
mental association between mental events. The system 'out there', which we
call nature, had been annihilated in the process.
By the time Kant was devoting his prodigious mind to sorting all this out,
the battle was lost. If the, quoting him, unconscious understanding organizes
sensory experience into schemata, while conscious understanding organizes it
into categories, the notion of identity remains for Kant forever transcendental.
63

Reproduced by kind permission of Sir Stafford Beer

 PREFACE 65
STAFFORD BEER
64
subject; there is no agreement on what it would be like; _and we are rather
Now the individual has vanished, in practical terms; as to the assemblage
short of anyone qualified to do the teaching. Those who resist the whole idea
of individuals called society, that too has vanished into a transcendental
in my view correctly, say that it would endanger the norms of good scholar:
construct. We have no need to legislate through any consensus of actual
ship. There is a deadlock.
people, but only to meet needs that might have arisen from the noumenal
Against this background, let us consider Autopoiesis, and try to answer the
will. question: 'What is it?' The authors say: "Our purpose is to understand the
And what of science itself? Science is ordered knowledge. It began with
organization of living systems in relation to their unitary character". If the
classification. From Galen in the second century through to Linnaeus in the
book .deals. wi~ living systems, then it must be about biology. If it says
eighteenth, analysis and categorization provided the natural instrumentality
anythmg scientific about organization, it must be about cybernetics. If it can
of scientific progress. Ally this fact with the background of philosophical
recognize the nature of unitary character, it must be about epistemology -
thought, and the scene is set for the inexorable development of the world and also (remembering the first author's massive contribution to the under-
view that is so difficult to challenge today. It is a world view in which real
standing of perception) it will be about psychology too. Yes, it is indeed
systems are annihilated in trying to understand them, in which relations are
about all these things. Will you then call this an interdisciplinary study in the
lost because they are not categorized, in which synthesis is relegated to
field of psychocyberbioepistemics? Do so only if you wish to insult the
poetry and mysticism, in which identity is a political inference. We may
authors. Because their topic has not slipped down the hole in the middle.
inspect the result in the structure and organization of the contemporary
Therefore it is not an interdisciplinary study of the kind defmed. It is not
university. abou~ analysis, but synthesis. It does not play the Game of the Categories.
It is an iron maiden, in whose secure embrace scholarship is trapped. For
And it does not interrelate disciplines; it transcends them. If, because of my
many, this is an entirely satisfactory situation, just because the embra~e is remarks about Kant, this seems to say that it annihilates them then we are
secure. A man who can lay claim to knowledge about some categorized bit of
getting somewhere. '
the world, however tiny, which is greater than anyone else's knowledge of
For there resides my belief in the book's general importance. The dissolu-
that bit, is safe for life: reputation grows, paranoia deepens. The number of tion of the deadlock ·within the disciplinary system that I described above has
papers increases exponentially, knowledge grows by infinitesimals, but
got to be metasystemic, not merely interdisciplinary. We are not interested in
understanding of the world actually recedes, because the world really is an
for~ing a league.of dis~ip~in~ry paranoids, but (as Hegel could have told us) in
interacting system. And since the world, in many of its aspects, is changing
a higher synthesis of disciplines. What emerges in this book is not classifiable
at an exponential rate, this kind of scholarship, rooted in the historical search
under. ~e old categories. Therefore it is predictable that no university could
of its own sanctified categories, is in large part unavailing to the needs of
~on~am. it, although all ~niversities can and now do contain interdisciplinary
mankind. mstitut10ns - because, m that very word, suitable obeisance is paid to the
There has been some recognition of this, and inter-disciplinary studies are
ha.llowed categories, and no one cares if the answers slip down the hole in the
by now commonplace in every university. But will this deal with the problem? middle. As to the prediction that universities cannot contain this kind of
Unfortunately, it will not. We still say that a graduate must have his 'basic
work, I have often see it fulfilled. In the present case it is falsified and I offer
discipline', and this he is solemnly taught - as if such a thing had a precise
heartfelt congratulations to the University of Chile. '
environmental correlate, and as if we know that God knew the difference
I say 'heartfelt' for this reason. In the mounting pile of new books printed
between physics and chemistry. He learns also the academic mores, catches
every year that are properly called scientific, one may take hold of one's
the institutional paranoia, and proceeds to propagate the whole business.
candle an~ s:arch like a veritable Diogenes for a single one answering to the
Thus it is that an 'interdisciplinary study' often consists of a group of dis-
honest c~iten~ I have proposed for a metasystemic utterance. There is only a
ciplinarians holding hands in a ring for mutual comfort. The ostensible topic
handful m existence at all, which is not surprising in view of the way both
has slipped down the hole in the middle. Among those who recognize this
knowledge and academia are organized. And yet, as I have also proposed
too, a natural enough debate has ensued on the subject: can an undergraduate
herein lies the world's real need. If we are to understand a newer and still
be taught 'interdisciplinary studies' as his basic subject? But there is no such
66 STAFFORD BEER PREFACE 67
evolving world; if we are to educate people to live in that world; if we are to precisely wrong. He contended that something not being observed goes out of
legislate for that world; if we are to abandon categories and institutions that existence. Autopoiesis says that something that exists may turn out to be
belong to a vanished world, as it is well-nigh desparate that we should; then unrecognizable when you next observe it. This brings us back to reality, for
knowledge must be rewritten. Autopoiesis belongs to the new library. that is surely true.
The second reason why the concept of autopoiesis excites me so much is
IN PARTICULAR that it involves the destruction of teleology. When this notion is fully worked
out and debated, I suspect it will prove to be as important in the history of
The authors first of all say that an autopoietic system is a homeostat. We the philosophy of science as was David Hume's attack on causality. Hume
already know what that is: a device for holding a critical systemic variable considered that causation is a mental construct projected onto changing
within physiological limits. They go on to the definitive point: in the case of events which have, .as we would say today, associated probabilities of mutual
autopoietic homeostasis, the critical variable is the system's own organization. occurrence. I myself have for a long time been convinced that purpose is a
It does not matter, it seems, whether every measurable property of that mental construct imported by the observer to explain what is really an
organizational structure changes utterly in the system's process of continuing equilibrial phenomenon of polystable systems. The arguments in Chapter II
adaptation. It survives. appear to me to justify this view completely, and I leave the reader to engen-
This is a very exciting idea to me for two reasons. In the first place it der his own excitement in the discovery of a 'purposelessness' that nonetheless
solves the problem of identity which two thousand years of philosophy have makes good sense to a human being - just because he is allowed to keep his
succeeded only in further confounding. The search for the 'it' has led farther identity, which alone is his 'purpose'. It is enough.
and farther away from anything that common sense could call reality. The 'it' But that salute to the authors is also self-congratulation, and I turn quickly
of scholasticism is a mythological substance in which anything attested by the aside. If a book is important, if at any time and from any source information
senses or testable by science inheres as a mere accident - its existence is a is received, then something is changed - not merely confirmed. There are two
matter of faith. The 'it' of rationalism is unrealistically schizophrenic, because arguments in this book that have changed me, and one of them effected its
it is uncompromising in its duality - extended substance and thinking sub- change after a profound inward struggle. Perhaps this part of the Preface
stance. The 'it' of empiricism is unrealistically insubstantial and ephemeral at should be printed as an epilogue: if I am not saying enough to be understood
the same time -esse est percipi is by no means the verdict of any experiencing in advance of the reading, then I am sorry. It is too much to hope that the
human being. reader will return.
The 'it' of Kant is the transcendental 'thing-in-itself' - an untestable People who work with systems-theoretic concepts have often drawn atten-
inference, an intellectual gewgaw. As to the 'it' of science and technology in tion to the subjective nature of 'the system'. A system is not something
the twentieth century world of conspicuous consumption ... 'it' seems to be presented to the observer, it is something recognized by him. One of the
no more than the collection of epiphenomena which mark 'it' as consumer consequences of this is that the labelling of connections between the system
or consumed. In this way hardheaded materialism seems to make 'it' as and its environment as either inputs or outputs is a process of arbitrary
insubstantial as subjective idealism made it at the turn of the seventeenth distinction. This is not very satisfatory. For example, a motor car in action is
century. And this, the very latest, the most down-to-earth, interpretation of evidently a system. Suppose that it is recognized as 'a system for going from
'it' the authors explicitly refute. A to B'; then the water in the radiator is evidently an input, and displacement
Their 'it' is notified precisely by its survival in a real world. You cannot is evidently an output. Now consider the. following scenario. Two men
find it by analysis, because its categories may all have changed since you last approach a motor car, and push it towards a second motor car. They then
looked. There is no need to postulate a mystical something which ensures the connect the batteries of the two cars with a pair of leads, and the engine of
preservation of identity despite appearances. The very continuation is 'it'. At the first car fires. They disconnect the leads, and run the engine hard in
least, that is my understanding of the authors' thesis - and I note with some neutral gear. We can guess what they are doing; but how is the objective
glee that this means that Bishop Berkeley got the precisely right argument scientist going to describe that system? Displacement is evidently an input,
68 STAFFORD BEER PREFACE 69

and one output is the rise in temperature of the water in the radiator. In case taken with this remark from the first sentence of Section 3: 'reproduction
my example sounds too transparent, note that Aristotle thought that the ... cannot enter as a defining feature of the organization of living systems'.
brain was a 'human radiator', namely an apparatus for cooling the blood. Finally, in the numbered Paragraph (ii) of Section 3 of Chapter V, the authors
Note also that he was right. say; 'A linguistic domain ... is intrinsically non-informative'. Surely that is
The fact is that we need a theoretical framework for any empirical in- finally absurd?
vestigation. This is the raison d'etre of epistemology, and the authors make All of this is totally alien to what we (most ofus working in cybernetics)
that point. In the trivial example I have just given, we need to know 'all about have believed. Information, including codes and messages and mappings,
motor cars' before we can make sense of the empirical data. But it often was indeed for us the whole story of the viable system. If one thinks of
happens in science that we know nothing at all about our 'motor cars', and sit reproduction, for example, as the process of passing on a DNA code from an
there scratching our heads over data that relate to we know not what. There aging set of tissues to an embryonic set of tissues, then the survival of the
is a prime example of this in current scientific work, which is so embarrasssing code itself is what matters. The tissues of each generation are subject to aging
that scientists in general pretend that it is not there. I am referring to the and finally death, but the code is transmitted. The individual becomes
whole field of parapsychology - to the mass of data which seems to say: insignificant, because the species is in the code. And that is why identity
precognition, telepathy, telekinesis exist. But we flounder among statistical vanishes in an ageless computer program of bits - a program that specifies
artifacts, and lack the theoretical framework for interpretation. This is made the hydrogen-bonded base pairs that link the sugar-phosphate backbones of
clear in the very name of ESP - 'extrasensory perception' which, if one the DNA molecule.
thinks about it, constitutes an internal contradiction of terms. The whole outlook turns out to be wrong, and the book must speak for
Autopoiesis as a concept propounds a theoretical framework within which itself on this score. But it is an extraordinarily condensed book, which is
to cope with the confusion that arises from the subjective recognition of 'the why this preface is inordinately long. I do not know whether the authors'
system' and the arbitrary classification of its inputs and outputs. For the arguments about information led me to understand their concept of auto-
authors explain how we may treat autopoietic systems as if they were not poiesis, or vice versa. What I am now sure about is that they are right. Nature
autopoietic (that is, they are allopoietic) when the boundaries of the system is not about codes: we observers invent the codes in order to codify what
are enlarged. Moreover, autopoietic systems may have allopoietic components. nature is about. These discoveries are very profound. ·
These ideas are immensely helpful, because our recognition of tlie circum- What is less profound but equally important is the political consequence
stances in which a system should be regarded as either auto- or allo-poietic of this crisis about identity. The subordination of the individual to the
enables us to define 'the system' in an appropriate context. That is to say species cannot be supported. "Biology cannot be used any more to justify
that the context is the recursion of systems within which the system we study the dispensability of the individual for the benefit of the species, society or
is embedded, instead of being the cloud of statistical epiphenomena generated mankind, under the pretense that its role is to perpetuate them. " After that,
by our attempt to study it. the world is a different place.
Understanding this changed me. The second change involved the intellec-
tual struggle I mentioned earlier, and it concerns the authors' views on the IN CONTENTION
information flowing within a viable system. In the numbered Paragraph (iv)
of Section I of Chapter III they say: 'The notion of coding is a cognitive The authors know it, and they draw the immediate inference. It is to say that
notion which represents the interactions of the observer, not a phenomenon scientists can no longer claim to be outside the social milieu within which
operative in the physical domain. The same applies to the notion of regula- they operate, invoking objectivity and disinterest; and in truth we .have
tion'. On first reading, this seemed to me plainly wrong. In the numbered known this, or ought to have known it, ever since Hiroshima. But again this
Paragraph (v) of Section 3 of Chapter IV they say: "Notions such as coding, book gives us the theoretical basis for a view that might otherwise shroud
message or information are not applicable to the phenomenon of self-re- something fundamental in a cloak of mere prudence. "No position or view
production". Wrong again, I considered; indeed, outrageous - especially wh~n that has any relevance in the domain of human relations can be deemed free
70 STAFFORD BEER PREFACE 71

from ethical and political implications, nor can a scientist consider himself Consider this argument at whatever level of recursion you please. An
alien to these implications". However, the authors go on to say that they do individual attempting to reform his own life within an autopoietic family
not fully agree betwe,~n themselves on the questions this poses from the cannot fully be his new self because the family insists that he is actually his
vantage point of their own work on autopoiesis - and they refuse to discuss old self. A country attempting to become a socialist state cannot fully become
them further (numbered Paragraph (iv) of Section 2 of Chapter V). socialist; because there exists an international autopoietic capitalism in which
This seems to be because they do not resolve the question (posed a little it is embedded, by which the revolutionary country is deemed allopoietic.
earlier) whether human societies are or are not themselves biological systems. These conclusions derive from entailments of premises which the authors
At this point, then, I ask to be relieved of the tasks of comment and inter- have placed in our hands. I think they are most valuable.
pretation; I ask for permission actively to enter this arena of discussion - Then let me try to answer the obvious question: why do not the authors
where the angels fear to tread. For I am quite sure of the answer: yes, human follow this line of development themselves, and write the second half of the
societies are biological systems. Moreover, I claim that this book conclusively book (as I hope they eventually will) - which would be about the nature and
proves the point. This is a delicate matter, because presumably at least one adaptation of social institutions, and the evolution of society itself? Well, to
of the originators of autopoietic theory disagrees, or is less than sure ... quote their sentence again: "Our purpose is to understand the organization of
Nonetheless, I have read the book many times; and one of those readings was living systems in relation to their unitary character". This formulation of the
exclusively devoted to validating this contention against the authors' own problem begs the question as to what is allowed to be a called a living system,
criteria of autopoiesis at every point. as they themselves admit. "Unless one knows which is the living organization,
The outcome, to which I was admittedly predisposed because of my own one cannot know which organization is living". They quickly reach the
work, says that any cohesive social institution is an autopoietic system - conclusion however (Subsection (b) of Section 2 of Chapter 1) that "auto-
because it survives, because its method of survival answers the autopoietic poiesis is necessary and sufficient to characterize the organization of living
criteria, and because it may well change its entire appearance and its apparent systems". Then they display some unease, quoting the popular belief: " ...
purpose in the process. As examples I list: firms and industries, schools and and no synthetic system is accepted as living".
universities, clinics and hospitals, professional bodies, departments of state, The fact is that if a social institution is autopoietic (and many seem to
and whole countries. answer to the proper criteria) then, on the authors' own showing, it is neces-
If this view is valid, it has extremely important consequences. In the first sarily alive. That certainly sounds odd, but it cannot be helped. It seems to
place it means that every social institution (in several of which any one me that the authors are holding at arms length their own tremendously
individual is embedded at the intersect) is embedded in a larger social institu- important discovery. It does not matter about this mere word 'alive'; what
tion, and so on recursively - and that all of them are autopoietic. This does matter is that the social institution has identity in the biological sense;
immediately explains why the process of change at any level of recursion it is not just the random assemblage of interested parties that it is thought
(from the individual to the state) is not only difficult to accomplish but to be. · ·
actually impossible -in the full sense of the intention: 'I am going completely When it comes to social evolution then, when it comes to political change:
to change myself'. The reason is that the 'I', that self-contained autopoietic we are not dealing with institutions and societies that will be different to-
'it', is a component of another autopoietic system. Now we already know morrow because of the legislation we passed today. The legislation - even
that the first can be considered as allopoietic with respect to the second, and the revolution - with which we confront them does not alter them at all; it
that is what makes the second a viable autopoietic system. But this is in turn proposes a new challenge to their autopoietic adaptation. The behavior they
means that the larger system perceives the embedded system as diminished - exhibit may have to be very different if they are to survive: the point is that
as less than fully autopoietic. That perception will be an illusion; but it does they have not lost their identities.
have consequences for the contained system. For now its own autopoiesis The interesting consequence is, however, that the way an autopoietic
must respond ·to a special kind of constraint: treatment which attempts to system will respond to a gross environmental challenge is highly predictable -
deny its own autopoiesis. once the nature of its autopoiesis is understood. Clever politicians intuit those
72 ST AFFORD BEER HUMBERTO R. MATURANA AND FRANCISCO J. VARELA

adaptations; and they can be helped by good scientists using systems-theoretic

models. Stupid politicians do not understand why social institutions do not
lose their identities overnight when they are presented with perfectly logical So long as ideas of the nature of living things
reasons why they should; and these are helped by bad scientists who devote remain vague and ill-defined, it is clearly impossible,
as a rule, to distinguish between an adaptation of
their effort to developing that irrelevant logic. the organism to the environment and a case of
In an era when rapid institutional change is a prerequisite of peaceful fitness of the environment for life, in the very most
survival in the face of every kind of exponentially rising threat, it seems to general sense. Evidently to answer such questions
me that the architects of change are making the same mistake all over the we must possess clear and precise ideas and defini-
tions of living things. Life must by arbitrary
world. It is that they perceive the system at their own level of recursion to be
process of logic be changed from the varying thing
autopoietic, which is because they identify themselves with that system and which it is into an independent variable or an
know themselves to be so; but they insist on treating the systems their system invariant, shorn of many of its most interesting
contains, and those within which their system is contained, as allopoietic. qualities to be sure, but no longer inviting fallacy
This is allowable in terms of scientific description, when the input and output through our inability to perceive clearly the
surfaces are correctly defined. Nonetheless it is politically blind to react questions involved.
towards the container and contained systems in a way which makes such a
model evident, because at these other levels of recursion the relevant systems Henderson, The Fitness of the Environment
perceive themselves as autopoietic too.
This statement seems to be worth making. I could not have made it so
succinctly without the language developed in this book. I could not have AUTOPOIESIS
formulated it at all without the new concepts that Humberto Maturana and
Francisco Varela have taught me. I thank them both very much, on behalf The Organization of the Living
of everyone.
INTRODUCTION
ST AFFORD BEER
A universe comes into being when a space is severed into two. A unity is
defined. The description, invention and manipulation of unities is at the base
of all scientific inquiry.
In our common experience we enc.ounter living systems as unities that
appear to us as autonomous entities of bewildering diversity endowed with
the capacity to reproduce. In these encounters autonomy apears so obviously
an essential feature of living systems that whenever something is observed
that seems to have it, the naive approach is to deem it alive. Yet, autonomy,
although continuously revealed in the self-asserting capacity of living systems
to maintain their identity through the active compensation of deformations,
seems so far to be the most elusive of their properties.·
Autonomy and diversity, the maintenance of identity and the origin of
variation in the mode in which this identity is maintained, are the basic
challenges presented by the phenomenology of living systems to which men
have for centuries addressed their curiosity about life.
73
74 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA AUTOPOIESIS 75

In the search for an understanding of autonomy classic thought, dominated them as living'; if not a vital force, if not an organizing principle of some
by Aristotle, created vitalism by endowing living systems with a non-material kind, what then? To take only a notable recent example let us mention J.
purposeful driving component that attained expression through the realization Monad's book Le hasard et la necessite. He tries to answer this question but,
of their forms. After Aristotle, and as variations of his fundamental notions, following the emphasis of evolutionary thought, he postulates a teleonomic
the history of biology records many theories which attempt in one way or organization of molecular nature and the subordination of the organization
another to encompass all the phenomenology of living systems under some of the individual to a plan defined by the species, in which the invariance of
peculiar organizing force. However, the more biologists looked for the explicit reproduction is determinant. Yet, teleonomic and evolutionary notions leave
formulation of one or other of these special organizing forces, the more they the question of the nature of the organization of the living unity essentially
were disappointed by finding only what they could find anywhere else in the untouched.
physical world: molecules, potentials and blind material interactions governed Our endeavor is to disclose the nature of the living organization. However,
by aimless physical laws. Thence, under the pressure of unavoidable experience in our approach we make a starting point of the unitary character of a living
and the definite thrust of Cartesian thought a different outlook emerged, system, and maintain that the evolutionary thought through its emphasis on
and mechanicism gradually gained the biological world by insisting that the diversity, reproduction and the species in order to explain the dynamics of
only factors operating in the organization of living systems were physical change has obscured the necessity of looking at the autonomous nature of
factors, and that no non-matieral vital organizing force was necessary. In fact, living unities for the understanding of the biological phenomenology. Also
it seems now apparent that any biological phenomeno~, once properly we think that the maintenance of identity and the invariance of defining
defined, can be described as arising from the interplay of physicochemical relations in the living unities are at the base of all possible ontogenic and
processes whose relations are specified by the context of its definition. evolutionary transformation in biological systems, and this we intend to
Diversity has been removed as a source of bewilderment in the under- explore. Thus, our purpose is: to understand the organization of living
standing of the phenomenology of living systems by Darwinian thought and systems in relation to their unitary character.
particulate genetics which have succeeded in providing an explanation for it Our approach will be mechanistic: no forces or principles will be adduced
and its origin without resorting to any peculiar directing force. Yet, the which are not found in the physical universe. Yet, our problem is the living
influence of these notions through their explanation of evolutionary change, organization· and therefore our interest will not be in properties of com-
has gone beyond the mere accounting for diversity: it has shifted completely ponents, but in processes and relations between processes realized through
the emphasis in the evaluation of the biological phenomenology from the components. This is to be clearly understood. An explanation is always a
individual to the species, from the unity to the origin of its parts, from the reformulation of a phenomenon showing how its components generate it
present organization of living systems to their ancestral determination. through their interactions and relations. Furthermore, an explanation is always
Today the two streams of thought represented by the physicochemical given by us as observers, and it is central to distinguish in it what pertains to
and the evolutionary explanations, are braided together. The molecular the system as constitutive of its phenomenology from what pertains to our
analysis seems to allow for the understanding of reproduction and variation, domain of description, and hence to our interactions with it, its components
the evolutionary analysis seems to account for how these processes might and the context in which it is observed . .Since our descriptive domain arises
have come into being. Apparently we are at a point in the history of biology because we simultaneously behold the unity and its interactions in the domain
where the basic difficulties have been removed. Biologists, however, are of observation, notions arising in the domain of description do not pertain to
uncomfortable when they look at the phenomenology pf living systems as a the constitutive organization of the unity (phenomenon) to be explained.
whole. Many manifest this discomfort by refusing to say what a living system Furthermore, an explanation may take different forms according to the
is. Others attempt to encompass present ideas under comprehensive theories nature of the phenomenon explained. Thus, to explain the movement of a
governed by organizing notions, like cybernetic principles, that require from falling body one resorts to properties of matter, and to laws that describe
the biologists the very understanding that they want to provide. The ever the conduct of material bodies according to these properties (kinetic and
present question is: 'What is common to all living systems that we qualify gravitational laws), while to explain the organization of a control plant one
 CHAPTER I
76 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA

resorts to relations and laws that describe the conduct of relations. In the first
case, the elements used in the explanation are bodies and their properties; in
the second case, they are relations and their relations, independently of ON MACHINES, LIVING AND OTHERWISE
the nature of the bodies that satisfy them. As in this latter case, in our
explanation of the organization of living systems, we shall be dealing with
the relations which the actual physical components must satisfy to constitute
one, not with the identification of these components. It is our assumption 1. MACHINES
that there is an organization that is common to all living systems, whichever
the nature of their components. Since our subject is this organization, not the Machines are usually viewed as concrete hardware systems, defined by the
particular ways in which it rpay be realized, we shall not make distinctions nature of their components and by the purpose that they fulfill in their
between classes or types of living systems. operations as man-made artifacts. This view however is obviously naive
This mode of thinking is not new, and is explicitly related to the very because it says nothing about how they are constituted. That machines are
name of mechanicism. We maintain that living systems are machines and by unities is apparent; that they are made of components that are characterized
doing this we point at several notions which should be made explicit. First, by certain properties capable of satisfying certain relations that determine in
we imply a non-animistic view which it should be unnecessary to discuss any the unity the interactions and transformations of these same components
further. Second, we are emphasizing that a living system is defined by its is also apparent. What is not so apparent is that the actual nature of the
organization and, hence, that it can be explained as any organization is components, and the particular properties that these may possess other than
explained, that is, in terms of relations, not of component properties. Finally, those participating in the interactions and transformations which constitute
we are pointing out from the start the dynamism apparent in living systems the system, are irrelevant and can be any. In fact, the significant properties
and which the word 'machine' connotes. of the components must be taken in terms of relations, as the network of
We are asking, then, a fundamental question: 'What is the organization interactions and transformations into which they can enter in the working
of living systems, what kind of machines are they, and how is their phenome- of the machine which they integrate and constitute as a unity.
nology, including reproduction and evolution, determined by their unitary TI1e relations that define a machine as a unity, and determine the dynamics
of interactions and transformations which it may undergo as such a unity,
organization?'
constitute the organization of the machine. The actual relations which hold
'among the components which integrate a concrete machine in a given space,
constitute its structure. The organization of a machine (or system) does not
specify the properties of the components which realize the machine as a
concrete system, it only specifies the relations which these must generate to
constitute the machine or system as a unity. Therefore, the organizatiori of
a machine is independent of the properties of its components which can be
any, and a given machine can be realized in many different manners by many
different kinds of components. In other words, although a given machine can
be realized by many different structures, for it to constitute a concrete entity
in a given space its actual components must be defined in that space, and have
the properties which allow them to generate the relations which define it.
The use to which a machine can be put by man is not a feature of the or-
ganization of the machine, but of the domain in which the machine operates,
and belongs to our description of the machine in a context wider than the
77
78 HUMBER TO R. MATURANA AND FRANCISCO J. VARELA ON MACHINES, LIVING AND OTHERWISE 79

machine itself. This is a significant notion. Man made machines are all made as a network of processes of production (transformation and destruction) of
with some purpose, practical or not, but with some aim (even ifit is only to components that produces the components which: (i) through their inter-
amuse) that man specifies. This aim usually appears expressed in the product actions and transformations continuously regenerate and realize the network
of the operation of the machine, but not necessarily so. However, we use the of processes (relations) that produced them; and (ii) constitute it (the ma-
notion of purpose when talking of machines because it calls into play the chine) as a concrete unity in the space in which they (the components) exist
imagination of the listener and reduces the explanatory task in the effort of by specifying the topological domain of its realization as such a network. It
conveying to him the organization of a particular machine. In other words, follows that an autopoietic machine continuously generates and specifies its
with the notion of purpose we induce the listener to invent the machine we own organization through its operation as a system of production of its own
are talking about. This, however, should not lead us to believe that purpose, components, and does this in an endless turnover of components under
or aim, or function, are constitutive properties of the machine which we conditions of continuous perturbations and compensation of perturbations.
describe with them; such notions are intrinsic to the domain of observation, Therefore, an autopoietic machine is an homeostatic ( or rather a relations-
and cannot be used to characterize any particular type of machine organiza- static) system which has its own organization ( defining network of relations)
tion. The product of the operations of a machine, however, can be used to as the fundamental variable which it maintains constant. This is to be clearly
this end in a non-trivial manner in the domain of descriptions generated by understood. Every unity has an organization specifiable in terms of static or
the observer. dynamic relations between elements, processes, or both. Among these possible
cases, autopoietic machines are unities whose organization is defined by a
particular network of processes (relations) of production of components,
2. LIVING MACHINES
the autopoietic network, not by the components themselves or their 'static
That living systems are machines cannot be shown by pointing to their relations. Since the relations of production of components are given only as
components. Rather, one must show their organization in a manner such that processes, if the processes stop, the relations of production vanish; as a result,
the way in which all their peculiar properties arise, becomes obvious. In order for a machine to be autopoietic, its defining relations of production must be
to do this, we shall first characterize the kind of machines that living systems continuously regenerated by the components which they produce. Further-
are, and then show how the peculiar properties ofliving systems may arise as more, the network of processes which constitute an autopoietic machine is a
consequences of the organization of this kind of machines. unitary system in the space of the components that it produces and which
generate the network through their interactions. The autopoietic network of
processes, then, differentiates autopoietic machines from any other kind of
a. Autopoietic machines
unit. In fact: (i) in a man-made machine in the physical space, say a car,
There are machines which maintain constant, or within a limited range of there is an organization given in terms of a concatenation of processes, yet,
values, some of their variables. The way this is expressed in the organization these processes are not processes of production of the components which
of these machines must be such as to define the process as occurring com- specify the car as a unity since the components of a car are produced by
pletely within the boundaries of the machine which the very same organization other processes which are independent of the organization of the car and its
specifies. Such machines are homeostatic machines and all feedback is internal operation. Machines of this kind are non-autopoietic dynamic systems. (ii)
to them. If one says that there is a machine M, in which there is a feedback In a natural physical unity like a crystal, the spatial relations among the
loop through the environment so that the effects of its output affect its components specify a lattice organization which defines it as a member of a
input, one is in fact talking about a larger machine M' which includes the class (a crystal of a particular kind), while the kinds of components which
environment and the feedback loop in its defining organization. constitute it specify it as a particular case in that class. Thus, the organization
Autopoietic machines are homeostatic machines. Their peculiarity, how- of a crystal is specified by the spatial relations which define the relative
ever, does not lie in this but in the fundamental variable which they maintain position of its components, while these specify its unity in the space in which
constant. An autopoietic machine is a machine organized (defined as a unity) they exist - the physical space. This is not so with an autopoietic machine. In
80 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA ON MACHINES, LIVING AND OTHER WISE 81

fact, although we find spatial relations among its components whenever' we and is not determined through their operation, because its product is different
actually or conceptually freeze it for an observation, the observed spatial from themselves; allopoietic machines do not have individuality.
relations do not (and cannot) define it as autopoietic. This is so because the
spatial relations between the components of an autopoietic machine are {iii) Autopoietic machines are unities because, and only because, of
specified by the network of processes of production of components which their specific autopoietic organization: their operations specify their own
constitute its organization and they are therefore necessarily in continuous boundaries in the processes of self-production. This is not the case with an
change. A crystal organization then, lies in a different domain than the allopoietic machine whose boundaries are defined by the observer, who by
autopoietic organization: a domain of relations between components, not of specifying its input and output surfaces, specifies what pertains to it in its
relations between pr'.)cesses of production of components; a domain of operations.
processes, not of concatenation of processes. We normally acknowledge this
{iv) Autopoietic machines do not have inputs or outputs. They can be
by saying that crystals are static.
perturbated by independent events and undergo internal structural changes
It is important to realize that we are not using the term organization in the
which compensate these perturbations. If the perturbations are repeated,
definition of an autopoietic machine in a mystical or transcendental sense,
the machine may undergo repeated series of internal changes which may
pretending that it has any explanatory value of its own. We are using it only
or may not be identical. Whichever series of internal changes takes place,
to refer to the specific relations that define an autopoietic system. Thus,
however, they are always subordinated to the maintenance of the machine
autopoietic organization simply means processes interlaced in the specific
organization, condition which is definitory of the autopoietic machines. Thus
form of a network of productions of components which realizing the network
any relation between these changes and the course of perturbations to which
that produced them constitute it as a unity. It is for this reason that we can
we may point to, pertains to the domain in which the machine is observed,
say that every time that this organization is actually realized as a concrete but not to its organization. Thus, although an autopoietic machine can be
system in a given space, the domain of the deformations which this system treated as an allopoietic machine, this treatment does not reveal its organiza-
can withstand without loss of identity while maintaining constant its organiza- tion as an autopoietic machine.
tion, is the domain of changes in which it exists as a unity. It is thus clear that
the fact that autopoietic systems are homeostatic systems which have their An organization may remain constant by being static, by maintaining its
own organization as the variable that they maintain constant, is a necessary components constant, or by maintaining constant certain relations between
consequence of the autopoietic organization. components otherwise in continuous flow or chl:\nge. Autopoietic machines
The consequences of this autopoietic organization are paramount: are organizations of the latter kind: they maintain constant the relations that
(i) Autopoietic machines are autonomous; that is, they subordinate all define them as autopoietic. The actual way in which such an organization may
changes to the maintenance of their own organization, independently of in fact be implemented in the physical space, that is, the physical structure
how profoundly they may otherwise be transformed in the process. Other of the machine, varies according to the' nature (properties) of the physical
machines, henceforth called allopoietic machines, have as the product of their materials which embody it. Therefore there may be many different kinds of
functioning something different from themselves (as in the car example). autopoietic machines in the physical space (physical autopoietic machines);
Since the changes that allopoietic machines may suffer without losing their all of them, however, will be organized in such a manner that any physical
definitory organization are necessarily subordinated to the production of interference with their operation outside their domain of compensations will
something different from themselves, they are not autonomous. result in their disintegration: that is, in the loss of ~utopoiesis. It also follows
that the actual way in which the autopoietic organization is realized in one of
(ii) Autopoietic machines have individuality; that is, by keeping their these machines (its structure) determines the particular perturbations it can
organization as an invariant through its continuous production they actively suffer without disintegration, and hence, the domain of interactions in which
maintain an identity which is independent of their interactions with an it can be observed. These features of the actual concreteness of autopoietic
observer. Allopoietic machines have an identity that depends on the observer machines embodied in physical systems allow us to talk about particular
82 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA
ON MACHINES, LIVING AND OTHERWISE 83
cases, to put them in our domain of manipulation and description, and hence, perfectly predictable. Contrariwise, living systems are a priori frequently
to observe them in the context of a domain of interactions which is external viewed as autonomous, ultimately unpredictable systems, with purposeful
to their organization. This has two kinds of fundamental consequence: behavior similar to ours. If living systems were machines, they could be made
(i) We can describe physical autopoietic machines, and also manipulate by man and, according to the view mentioned above, it seems unbelievable
them, as parts of a larger system that defines the independent events which that man could manufacture a living system. This view can be easily dis-
perturb them. Thus, as noted above, we can view these perturbing independent qualified, because it either implies the belief that living systems cannot be
events as inputs, and the changes of the machine that compensate these understood because they are too complex for our meager intellect and
perturbations as outputs. To do this, however, amounts to treating an auto- will remain so, or that the principles which generate them are intrinsically
poietic machine as an allopoietic one, and to recognize that if the independent unknowable; either implication would have to be accepted a priori without
perturbing events are regular in their nature and occurrence, an autopoietic proper demonstration. There seems to be an intimate fear that the awe with
machine can in fact, be integrated into a larger system as a component respect to life and the living would disappear if a living system could be not
allopoietic machine, without any alteration in its autopoietic organization. only reproduced, but designed by man. This is nonsense. The beauty of life
is not a gift of its inaccessibility to our understanding.
(ii) We can analyze a physical autopoietic machine in its physical parts,
and treat all its partial homeostatic and regulatory mechanisms as allopoietic (ii) To the extent that the nature of the living organization is unknown,
machines (sub-machines) by defining their input and output surfaces. Accord- it is not possible to recognize when one has at hand, either as a concrete
ingly, these sub-machines are not necessarily components of an a:utopoietic synthetic system or as a description, a system that exhibits it. Unless one
machine because the relations that define such a machine need not be those knows which is the living organization, one cannot know which organization
that they generate through the input-output relations that define them. is living. In practice, it is accepted that plants and animals are living but
their characterization as living is done through tl1e enumeration of their
The fact that we can divide physical autopoietic machines into parts
properties. Among these, reproduction and evolution appear as determinant,
does not reveal the nature of the domain of interactions that they define as
and for many observers the condition of living appears subordinated to
concrete entities operating in the physical universe.
the possession of these properties. However, when these properties are
incorporated in a concrete or conceptual man-made system, those who do not
b. Living systems accept emotionally that the nature of life can be understood, immediately
conceive of other properties as relevant, and do not accept any synthetic
If living systems are machines, that they are physical autopoietic machines is
system as living by continuously specifying new requirements.
trivially obvious: they transform matter into themselves in a manner such
that the product of their operation is their own organization. However we (iii) It is very often assumed that observation and experimentation are
deem the converse is also true: a physical system if autopoietic, is living. alone sufficient to reveal the nature of living systems and no theoretical
In other words, we claim that the notion of autopoiesis is necessary and analysis is expected to be necessary and least of all sufficient for a charac-
sufficient to characterize the organization of living systems. This equivalence terization of the living organization. It would be long to state why we depart
may not be apparent for some observers due to several reasons which do not from this radical empiricism. Let us simply say that we believe that epistemo-
pertain to the domain of the organization of autopoietic machines, but logical and historical arguments more than justify the contrary view: every
which are proper within the domain of description and evaluation of the experimentation and observation implies a theoretical perspective, and no
observers who adopt such reasons, and lead them to its a priori negation. The experimentation or observation has significance or can be interpreted outside
following are some of these reasons: the theoretical framework in which it took place.
(i) Machines are generally viewed as human made artifacts with completely
Our aim was to propose the characterization of living systems that explains
known deterministic properties which make them, at least conceptually,
the generation of all the phenomena proper to them. We have done this by
84 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA CHAPTER II

pointing at autopoiesis. in the physical space as a necessary and sufficient

condition for a system to be a living one.
To know that a given aim has been attained, is not always easy. In the case DISPENSABILITY OF TELEONOMY
at hand, the only possible indication that we have attained our aim is the
reader's agreement that all the phenomenology of living systems, including
reproduction and evolution, indeed requires and depends on autopoiesis.
The following chapters are devoted to show this. Teleology and teleonomy are notions employed in discourse, descriptive and
explanatory, about living systems, and although it is claimed that they do not
necessarily enter as causal elements in their functioning, it is asserted that
they are essential definitory features of.their organization. Our present aim
is to show that in the light of the preceding discussion, these notions are
unnecessary for the understanding of the living organization.

1. PURPOSELESSNESS
It is usually maintained that the most remarkable feature of living systems is
a purposeful organization, or what is the same, the possession of an internal
project or program represented and realized in and through their structural
organization. Thus, ontogeny is generally considered as an integrated process
of development towards an adult state, through which certain structures are
attained that allow the organism to perform certain functions according to
the innate project which defines it in relation to the environment. Also,
phylogeny is viewed as the history of adaptive transformations through
reproductive processes aimed at satisfying the project of the species, with
complete subordination of the individual to this end. Furthermore, it is
apparent that there are organisms that may even appear capable of specifying
some purpose in advance (as the authors of this book) and conduct all their
activities towards this attainment (heteropoiesis). This element of apparent
purpose or the possession of a project or program in the organization of living
systems, which has been called teleonomy without implying any vitalistic
connotations, is frequently considered as a necessary, if not as a sufficient,
definitory feature for their characterization. Purpose or aims, however, as we
saw in the first chapter, are not features of the organization of any machine
(allo- or autopoietic); these notions belong to the domain of our discourse
about our actions, that is, they belong to the domain of descriptions, and
when applied to a machine, or any system independent from us, they reflect
our considering the machine or system in some encompasssing context. In
general, the observer puts the machines either conceptually or concretely to
some use, and thus defines a set of circumstances that lead the machine to
85
 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA DISPENSABILITY OF TELEONOMY 87
86
change, following a certain path of variations in its output. The connection 2. INDIVIDUALITY
between these outputs, the corresponding inputs, and their relation with the
context in which the observer includes them, determine what we call the aim The elimination of the notion of teleonomy as a defining feature of living
or purpose of the machine; this aim necessarily lies in the domain of the systems changes the outlook of the problem completely, and forces us to
observer that defines the context and establishes the nexuses. Similarly the consider the organization of the individual as the central question for the
notion of function arises in the description made by the observer of the understanding of the organization of living systems.
components of a machine or system in reference to an encompassing entity, In fact, a living system is specified as an individual, as a unitary element of
which may be the whole machine or part of it, and whose states constitute interactions, by its autopoietic organization which determines that any change
the goal that the changes in the components are to bring about. Here again, in it should take place subordinated to its maintenance, and thus sets the
no matter how direct the causal connections may be between the changes of boundary conditions that specify what pertains to it and what does not
state of the components and the state which they originate in the total pertain to it in the concreteness of its realization. If the subordination of all
system, the implications in terms of design alluded to by the notion of changes in a living system to the maintenance of its autopoietic organization
function are established by the observer and belong exclusively to his domain did not take place ( directly or indirectly), it would lose that aspect of its
of description. Accordingly, since the relations implied in the notion of organization which defines it as a unity, and hence it would disintegrate. Of
function are not constitutive of the organization of an autopoietic system, course it is true for every unity, whichever way it is defined, that the loss of
they cannot be used to explain its operation. its defining organization results in its disintegration; the peculiarity of living
The organization of a machine, be it autopoietic or allopoietic, only systems, however, is that they disintegrate whenever their autopoietic or-
states relations between components and rules for their interactions and ganization is lost, not that they can disintegrate. As a consequence, all change
transformations, in a manner that specifies the conditions of emergence of must occur in each living system without interference with its functioning as
the different states of the machine which, then, arise as a necessary outcome a unity in a history of structural change in which the autopoietic organization
whenever such conditions occur. Thus, the notions of purpose and function remains invariant. Thus ontogeny is both an expression of the individuality of
have no explanatory value in the phenomenological domain which they living systems and the way through which this individuality is realized. As a
pretend to illuminate, because they do not refer to processes indeed operat- process, ontogeny, then, is the expression of the becoming of a system that at
ing in the generation of any of its phenomena. This does not preclude their each moment is the unity in its fullness, and does not constitute a transit
being adequate for the orientation of the listener towards a given domain of from an incomplete (embryonic) state to a more complete or final one (adult).
thought. Accordingly, a prediction of a future state of a machine consists The notion of development arises, like the notion of purpose, in the
only in the accelerated realization in the mind of an observer of its succeeding context of observation, and thus belongs to a different domain other than
states, and any reference to an early state to explain a later one in functional the domain of the autopoietic organization of the living system. Similarly,
or purposeful terms, is an artifice of his description, made in the perspective the conduct of an autopoietic machine that an observer can witness is the
of his simultaneous mental observation of the two states, that induces in the reflection of the paths of changes that it undergoes in the process of main-
mind of the listener an abbreviated realization of the machine. Therefore any taining constant its organization through the control of the variables that can
machine, a part of one or a process that follows a predictable course, can be be displaced by perturbations, and through the specification in this same
described by an observer as endowed with a project, a purpose or a function, process of the values around which these variables are maintained at any
if properly handled by him with respect to an encompassing context. moment. Since the autopoietic machine has no inputs or outputs, any cor-
Accordingly, if living systems are physical autopoietic machines, teleonomy relation between regularly occurring independent events that perturb it, and
becomes only an artifice of their description which does not reveal any feature the state to state transitions that arise from these perturbations, which the
of their organization, but which reveals the consistency in their operation observer may pretend to reveal, pertain to the history of the machine in
within the domain of observation. Living systems, as physical autopoietic the context of the observation, and not to the operation of its autopoietic
machines, are purposeless systems. organization.
 CHAPTER III EMBODIMENTS OF AUTOPOIESIS 89

course depends on the particular way in which the autopoiesis is realized, that
is, on the actual structure of their realization. There are, however, certain
EMB.ODIMENTS OF AUTOPOIESIS general notions which apply to any particular concrete autopoietic system
that we must mention at the outset:
(i) Although indeed energetic and thermodynamic considerations would
necessarily enter in the analysis of how the components are physically con-
The assertion that physical autopoietic systems are living systems requires stituted, and in the description of their proper ties in a specific domain of
the proof that all the phenomenology of a living system can be either reduced interactions, such that they may satisfy the requirements of their participation
or subordinated to its autopoiesis. This proof, obviously, cannot consist in in an autopoietic system, these considerations do not enter in the characteri-
enumerating all biological phenomena and presenting cases of autopoietic zation of the autopoietic organization. If the components can be materialized,
systems that exhibit them; rather it must consist in showing that autopoiesis the organization can be realized; the s11tisfaction of all thermodynamic and
either constitutes or is necessary and sufficient for the occurrence of all energetic relations is implicit. Thus, for example, in the concrete case of the
biological phenomena, if the proper non-determinant contingencies are given. cell, that we shall consider in the next section, energetic relations that make
possible certain reactions with the participation of ATP are not constitutive
1. DESCRIPTIVE AND CAUSAL NOTIONS of the autopoietic organization. However, it is constitutive of the structure
through which the autopoietic organization is realized, that the molecules
An autopoietic system is defined as a unity by its autopoietic organization. which participate in it should have among their properties the property of
The realization of this organization in a physical system requires components entering into the interactions which generate the autopoietic processes and,
which are defined by their role in the autopoiesis and which can only be hence, of holding the required energy relations.
described in relation to this. Furthermore these components can only be
(ii) Notions such as specification and order are referential notions; that is,
realized by material elements which can exhibit the necessary properties
they do not have meaning outside the context in which they are defined.
under the conditions specified by the autopoietic organization, and must be
produced in the proper topological relation within this organization, by the Thus, when we speak about relations of specification we refer to the specifi-
cation of components in the context of that which defines the system as
particular instance (structural realization) of the autopoietic system that they
autopoietic. Any other element of specificity that may enter, however
constitute. Accordingly; an autopoietic organization constitutes a closed
necessary it may be for the factibility [factual characterization] of the
domain of relations specified only with respect to the autopoietic organization
components, but which is not defined through the autopoietic organization,
that these relations constitute, and, thus, it defines a 'space' in which it can
we take for granted. Similarly with the notion of order. Relations of order
be realized as a concrete system; a space whose dimensions are the relations
refer to the establishment of processes that secure the presence of the compo-
of production of the components that realize it:
nents in the concatenation that results in autopoiesis. No other reference is
(i) Relations of constitution that determine that the components produced meant, however conceivable it may be within other perspectives of description.
constitute the topology in which the autopoiesis is realized. (iii) An autopoietic organization acquires topological unity by its em-
(ii) Relations of specificity that determine that the components produced bodiment in a concrete autopoietic system which retains its identity as long
be the specific ones defined by their participation in the autopoiesis. as it remains autopoietic. Furthermore, the space defined by an autopoietic
system is self-contained and cannot be described by using dimensions that
(iii) Relations of order that determine. that the concatenation of the define another space. When we refer to our interactions with a concrete
components in the relations of specification, constitution and order be the autopoietic system, however, we project this system upon the space of our
ones specified by the autopoiesis. manipulations and make a description of this projection. This we can do
How these relations of production are embodied in a physical system of because we interact with the components of the autopoietic system through

88
90 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA EMBODIMENTS OF AUTOPOIESIS 91

the properties of their constituting elements that do not lie in the autopoietic autopoietic organization, and hence its physical boundaries. The production
space, and thus, we modify the structure of the autopoietic system by modi- of constitutive relations through the production of the components that hold
fying its components. Our description, however, follows the ensuing change these relations is one of the defining dimensions of an autopoietic system. In
of the projection of the autopoietic system in the space of our description, the cell such constitutive relations are established through the production of
not in the autopoietic space. molecules (proteins, lipids, carbohydrates and nucleic acids) which determine
(iv) Notions such as coding and transmission of information do not enter the topology of the relations of production in general; that is, molecules
in the realization of a concrete autopoietic system because they do not which determine the relations of physical neighborhood necessary for the
refer to actual processes in it. Thus, the notion of specificity does not imply components to hold the relations that define them. The cell defines its physi-
coding, information or instructions; it only describes certain relations, deter- cal boundaries through its dimension of production of constitutive relations
mined by and dependent on the autopoietic organization, which result in that specify its topology. There is no specification in the cell of what it is not.
the production of the specific components. The proper dimension is that
of relations of specificity. To say that the system or part of it, codes for (ii) Production of Relations of Specifications
specificity, is not only a misnomer but also misleading; this is so, because
such an expression represents a mapping of a process that occurs in the space Relations of specifications are relations that determine the identity (proper-
of autopoiesis onto a process that occurs in the space of human design ties) of the components of the autopoietic organization, and hence, in the
(heteropoiesis), and it is not a reformulation of the phenomenon. The notion case of the cells, its physical factibility. The establishment of relations of
of coding is a cognitive notion which represents the interactions of the specification through the production of components that can hold these
observer, not a phenomenon operative in the observed domain. The same relations is another of the defining· dimensions of an autopoietic system. In
applies to the notion of regulation. This notion is valid in the domain of the cell such relations of specification are produced mainly through the
description of heteropoiesis, and it reflects the simultaneous observation production of nucleic acids and proteins that determine the identity of the
and description made by the designer ( or his equivalent) of interdependent relations of production in general. In the cell this is obviously obtained, on
transitions of the system that occur in a specified order and at specified the one hand, by relations of specificity between DNA, RNA and proteins,
speeds. The corresponding dimension in an autopoietic system is that of and on the other hand, by relations of specificity between enzymes and
relations of production of order, but here again only in the context of the substrates. Such production of relations of specification holds only within the
autopoiesis and not of my particular state of the system as it would appear topological substrate defined by the production of relations of constitution.
projected on our domain of descriptions. The notion of regulation, then, There is no production in the cell as an autopoietic system of relations of
can enter in the description, but does not refer to an actual process in the specification that do not pertain to it.
autopoietic organization.
(iii) Production of Relations of Order
2. MOLECULAR EMBODIMENTS Relations of order are those that determine the dynamics of the autopoietic
That a cell is an autopoietic system is trivially apparent in its life cycle. What organization by determining the concatenation of the production of relations
is not irivial is how the cell is a molecular embodiment of autopoiesis, as it of constitution, specification and order, and hence its actual realization. The
should be apparent in its analysis in terms of the dimensions of its autopoietic establishment of relations of or~er through the production of components
space: that realize the production of relations of constitution, specification and
order, constitute the third dimension of the autopoietic space. In the cell,
(i) Production of Constitutive Relations
relations of order are established mainly by the production of components
(metabolites, nucleic acids and proteins) that control the speed of production
Constitutive relations are relations that determine the topology of the of relations of constitution, specification and order. Relations of order, thus,
92 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA EMBODIMENTS OF AUTOPOIESIS 93

conform a network of parallel and sequential relations of constitution, speci- tions of constitution, specification and order which constitutes autopoiesis.
fication and order that constitute the cell as a system in which the relations In other words, compensation of deformation keeps the autopoietic system in
of production that specify this network as a dynamic physical topological the autopoietic space.
unity, are maintained constant. There is no ordering through the autopoietic · That all the biological features of the cell as a unity are determined by its
organization of the cell of processes that do not belong to it. autopoiesis, is henceforth obvious. In fact, the only thing that defines the cell
If one examines a cell it is apparent that: as a unity (as an individual) is its autopoiesis, and thus the only restriction
DNA participates in the specification of polypeptides, and hence, or proteins, placed on the existence of the cell is the maintenance of autopoiesis. All the
enzymatic and structural, which specifically participate in the production of rest - that is, its structure - can vary: relations of topology, specificity and
proteins, nucleic acids, lipids, glucides and metabolites. Metabolites (which order can vary as long as they constitute a network in an autopoietic space.
include all small molecules, monomers or not, produced in the cell) participate
in the determination of the speed of the various processes and reactions that
3. ORIGIN
constitute the cell, establishing a network of interrelated speeds in parallel
and sequentially interconnected processes, both by gating and by constitutive The production of relations of constitution, specification and order, are not
participation, in a way such that every reaction is a function of the state of exclusive to autopoietic systems. They are inherent to unitary interactions
the transforming network that they integrate. All processes occur bound to a in general, and to molecular interactions in particular; they depend on the
topology determined by their participation in the processes of production of properties of the units or molecules as expressed in the geometric and en-
relations of constitution. ergetic relationships which they may adopt. Thus, the geometric properties of
We as observers can project all cellular processes upon a system of three the molecules determine the relations of constitution, that is, the topology,
orthogonal coordinates, and legitimately say, as valid in the projection, that the physical neighborhoods or spatial relations in which they may enter. The
specification is mostly produced by nucleic acids, constitution by proteins, chemical properties of the molecules determine their possible interactions,
and order (regulation) by metabolites. The autopoietic space, however, is and, hence, the relations of specificity which are a dimension orthogonal
curved and closed in the sense that it is entirely specified by itself, and to relations of constitution. Both together, they determine sequence and
such a projection represents our cognitive relation with it, but does not concatenation of molecular interactions, that is, relations of order. According-
reproduce it. In it, specification takes place at all points where its organization ly, autopoiesis may arise in a molecular system if the relations of production
determines a specific process (protein synthesis, enzymatic action, selective are concatenated in such a way that they produce components that specify
permeability); ordering takes place at all points where two or more processes the system as a unity which exists only while it is actively produced by
meet (changes of speed or sequence, allosteric effects, competitive and non- such concatenation of processes. This is to say that autopoiesis arises in a
competitive inhibition, facilitation, inactivation, etc.) determined by the molecular system only when the relation that concatenates these relations is
structure of the participating components; constitution occurs at all places produced and maintained constant through the production of the molecular
where the structure of the components determines physical neighborhood components that constitute the system through this concatenation. Thus, in
relations (membranes, particles, active site in enzymes). What makes this general, the question of the origin of an autopoietic system is a question
system a unity with identity and individuality is that all the relations of about the conditions that must be satisfied for the establishment of an
production are coordinated in a system describable as an homeostatic system autopoietic space. This problem, then, is not a chemical one, in terms of what
that has its own unitary character as the variable that it maintains constant molecules took or can take part in the process, but a general one of what
through the production of its components. In such a system any deformation relations the molecules or any constitutive units should satisfy. This deserves
at any place is not compensated by bringing the system back to an identical the following considerations:
state of its components as it would be described by projecting it upon a three-
dimensional Cartesian space; rather it is compensated by keeping its organiza- (i) An autopoietic system is defined as a unity by and through its auto-
tion constant as defined by the relation of the relations of production of rela- poietic organization. This unity is, thus, a topological unity in the space in
94 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA EMBODIMENTS OF AUTOPOIESIS 95

which the components have existence as entities that may interact and have the establishment of any system depends on the presence of the components
relations. For living systems such a space is the physical space. Without unity that constitute it, and on the kinds of interactions in which they may enter;
in some space an autopoietic system is not different from the background in thus, given the proper components and the proper concatenation of their
which it is supposed to lie, and, hence, can only be a system in the space of interactions, the system is realized. The concrete question about the factibi-
our description where its unity is conceptually stipulated. Without unity in lity of a molecular autopoietic system is, then, the question of the conditions
the physical space a living system would lack the dynamics of production in which different chemical processes can be concatenated to form topological
relations which constitute it as a concrete entity in that space. unities that constitute relational networks in the autopoietic space. The
(ii) The establishment of an autopoietic system cannot be a gradual second aspect can be stated in the following manner: given the factibility of
process; either a system is an autopoietic system or it is not. In fact, its autopoietic systems, and given the existence of terrestrial autopoietic systems,
establishment cannot be a gradual process because an autopoietic system is . there are natural conditions under which these may be spontaneously gen-
defined as a system; that is, it is defined as a topological unity by its organiza- erated. Concretely the question would be, 'What were or are the natural
tion. T1ms, either a topological unity is formed through its autopoietic conditions under which the components of the autopoietic systems arose
organization, and the autopoietic system is there and remains, or there is no or arise spontaneously on the earth, and concatenate to form them?' This
question cannot be answered independently of the manner in which the
topological unity, or a topological unity is formed in a different manner and
factibility question is answered, particularly in what refers to the factibility
there is no autopoietic system but there is something else. Accordingly, there
of one or several different kinds of molecular autopoietic systems. The
are not and there cannot be intermediate systems. We can describe a system
presence today of one mode of autopoietic organization on the earth (the
and talk about it as if it were a system which, with a little transformation,
nucleic acid protein system), cannot be taken to imply that the factibility
would become an autopoietic system because we can imagine different sys-
question has only one answer.
tems with which we compare it, but such a system would be intermediate
only in our description, and in no organizational sense would it be a transi-
The notions that we have discussed are valid for the origin ( constitution)
tion system.
of autopoietic systems at any level of physical embodiment, molecular or
(iii) Autocatalytic processes do not constitute autopoietic systems because supramolecular. We shall not dwell on the particular circumstances of the
among other things, thr;!y do not determine their topology. Their topology is establishment of any of these embodiments. We shall leave this matter for
determined by a container that is part of the specification of the system, but another inquiry, accepting the existence of living systems as an existential
which is independent of the operation of the autocatalysis. Processes of this proof of the factibility of the spontaneous generation of autopoietic systems.
or similar kind are abundant in the physical space. Coupling of independent We shall consider next the significance of the conditions of topological unity
processes into larger systems is also the rule; these may or may not constitute for the diversity of autopoietic syst~ms.
unities defined by the circumstances of their constitution in a given space,
be this space physical or otherwise. They, however, will not constitute or
participate in the constitution of an autopoietic system unless the system
they conform becomes defined as a topological unity through its embodiment
of an autopoietic organization. A unity is defined by an operation of distinc-
tion; in an autopoietic system its autopoiesis constitutes the operation of
distinction that defines it, and its origin is cocircumstantial with the establish-
ment of this operation.
(iv) The problem of the origin of autopoietic systems has two aspects;
one refers to their factibility, and the other to the possibility of their spon-
taneous occurrence. The first aspect can be stated in the following manner:
I
I!
l
l
 CHAPTER IV DIVERSITY OF AUTOPOIESIS 97

unity determine its phenomenology. In living systems, these conditions are

determined by their autopoietic organization. In fact, autopoiesis implies
DIVERSITY OF AUTOPOIESIS the subordination of all change in the autopoietic system to the maintenance
of its autopoietic organization, and since this organization defines it as
a unity, it implies total subordination of the phenomenology of the system
to the maintenance of its unity. This subordination has the following con-
living systems embody the living organization. living systems are autopoietic sequences:
systems in the physical space. The diversity of living systems is apparent; it is
also apparent that this diversity depends on reproduction and evolution. Yet, (i) The establishment of a unity defines the domain of its phenomenology,
reproduction and evolution do not enter into the characterization of the living but given the way the unity is constituted by its structure defines the kind of
organization, and living systems are defined as unities by their autopoiesis. phenomenology that it generates in that domain. It follows that the particular
This is significant because it makes the phenomenology of living systems form adopted by the phenomenology of each autopoietic (biological) unity
dependent on their being autopoietic unities. In fact, reproduction requires depends on the particular way in which its individual autopoiesis is realized.
It also follows that the domain of ontogenic transformations (including
the existence of a unity to be reproduced, and it is necessarily secondary to
conduct) of each individual is the domain of the homeostatic trajectories
the establishment of such a unity; evolution requires reproduction and the
through which it can maintain its autopoiesis.
possibility of change, through reproduction of that which evolves, and it is
necessarily secondary to the establishment of reproduction. It follows that (ii) ·All the biological phenomenology is necessarily determined and
the proper evaluation of the phenomenology of living systems, including realized through individual autopoietic unities in the physical space, and
reproduction and evolution, requires their proper evaluation as autopoietic consists of all the paths of transformations that they undergo as homeostatic
unities. systems, singly or in groups, in the process of maintaining constant their
defining individual relations. Whether in the process of their interactions the
1. SUBORDINATION TO THE CONDITION OF UNITY autopoietic unities do or do not unite to constitute additional unities, is
irrelevant for the subordination of the biological phenomenology to the
Unity (distinguishability from a background, and, hence, from other unities), maintenance of the identity of the individual unities. If united they produce
is the sole necessary condition for existence in any given domain. In fact, a new unity that is not autopoietic, its phenomenology, that will necessarily
the nature of a unity and the domain in which it exists are specified by the depend on its organization, will be biological or not according to its de-
process of its distinction and determination; this is so regardless of whether pendence on the autopoiesis of its components, and will accordingly depend
this process is conceptual (as when a unity is defined by an observer through or not on the maintenance of these as autopoietic units. If the new unity is
an operation of distinction in his domain of discourse and description), or autopoietic, its phenomenology is directly biological and obviously depends
whether this process is physical (as when a unity becomes established through on the maintenance of its autopoiesis, which in turn may or may not depend
the actual working of its defining properties that assert its distinction from a on the autopoiesis of its components.
background through their actual operation in the physical space). Accordingly,
different kinds of unities necessarily differ in the domain in which they are (iii) The identity of an autopoietic unity is maintained as long as it
established, and having different domains of existence they may or may not remains autopoietic; that is, as long as it, as a unity in the physical space,
interact according to whether these domains do or do not intersect. Unity remains a unity in the autopoietic space, regardless of how much it may
distinction [the distinctiveness and distinguishing of unity], then, is not an otherwise be transformed in the process of maintaining its autopoiesis.
abstract notion of purely conceptual validity for descriptive or analytical
purposes, but it is an operative notion referring to the process through (iv) Only after a unity has been constituted as an autopoietic unity
which a unity becomes asserted or defined: the conditions which specify a (individual) can reproduction take place as a biological phenomenon.
96
98 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA DIVERSITY OF AUTOPOIESIS 99

2. PLASTICITY OF ONTOGENY internally determined, for an observer its ontogeny partly reflects its history
of interactions with an independent ambience. Accordingly, two otherwise
Ontogeny is the history of the structural transformation of a unity. Accord- equivalent autopoietic systems may have different ontogenies.
ingly, the ontogeny of a living system is the history of maintenance of its
(v) An observer beholding an autopoietic system as a unity in a context
identity through continuous autopoiesis in the physical space. From the mere
that he also observes, and which he describes as its environment, may dis-
fact that a physical autopoietic system is a dynamic system, realized through
tinguish in it internally and externally generated perturbations, even thougl1
relations of productions of components that imply concrete physical inter-
these are intrinsically indistinguishable for the autopoietic system itself. The
actions and transformations, it is a necessary consequence of the autopoietic
observer can use these distinctions to make statements about the history of
organization of a living system that its ontogeny should take place in the
the autopoietic system which he observes, and he can use this history to
physical space. There are several comments to this notion of ontogeny:
describe an ambience (which he infers) as the domain in which the system
(i) Since the way an autopoietic system maintains its identity depends on exists. He cannot, however, infer from the observed correspondence between
its particular way of being autopoietic, that is, on its particular structure, the ontogeny of the system and the ambience which this ontogeny describes,
different classes of autopoietic systems have different classes of ontogenies. or from the environment in which he sees it, a constitutive representation
of these in the organization of the autopoietic systems. The continuous
(ii) Since an autopoietic system does not have inputs or outputs, all the
correspondence between conduct and ambience revealed during ontogeny is
changes that it may undergo without loss of identity, and, hence, with
the result of the homeostatic nature of the autopoietic organization, and not
maintenance of its defining relations, are necessarily determined by its
of the existence of any repr~sentation of the ambience in it; nor is it at all
homeostatic organization. Consequently, the phenomenology of an auto-
necessary that the autopoietic system should obtain or develop such a re-
poietic system is necessarily always commensurate with the deformations that
presentation to persist in a changing ambience. To talk about a representation
it suffers without loss of identity, and with the deforming ambience in which
of the ambience, or the environment, in the organization of a living system
it lies. Otherwise it would disintegrate.
may be metaphorically useful, but it is inadequate and misleading to reveal
(iii) As a consequence of the homeostatic nature of the autopoietic the organization of an autopoietic system.
organization, the way the autopoiesis is realized in any given unity may
(vi) The compensatory changes that an autopoietic system may undergo
change during its ontogeny, with the sole restriction that this should take
while retaining its identity, may be of two possible kinds according to how its
place without loss of identity, that is, through uninterrupted autopoiesis.
structure is affected by the perturbations: they may be (a) conservative
(iv) Although the changes that an autopoietic system may undergo without changes in which only the relations between the components change; or
loss of identity while compensating its deformations under interactions are they may be (b) innovative changes in which the components themselves
determined by its organization, the sequence of such changes is determined change. In the first case, the intenial or external interactions causing the·
by the sequence of these deformations. There are two sources of deformations deformations do not lead to any change in the way the autopoiesis is realized,
for an autopoietic system as they appear to be to an observer: one is con- and the system remains in the same point in tlie autopoietic space because its
stituted by the external environment as a source of independent events in the components are invariant; in the second case, on the contrary, the interac-
sense that these are not determined by the organization of the system; the tions lead to a change in the way the autopoiesis is realized and, hence, to a
other is constituted by the system itself as a source of states which arise displacement of the system in the autopoietic space because its components
from compensations of deformations, but which themselves can constitute changed. Accordingly, while the first case implies a conservative ontogeny,
deformations that generate further compensatory changes. In the phenome- the second case implies an ontogeny which is also a process of specification of
nology of the autopoietic organization these two sources of perturbations are a particular autopoiesis that in its determination is, necessarily, a function
indistinguishable, and in each autopoietic system they braid together to form of both the plasticity of the components of the system and the history of its
a single ontogeny. Thus, although in an autopoietic system all changes are interactions.
100 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA DIVERSITY OF AUTOPOIESIS 101

3. REPRODUCTION, A COMPLICATION OF THE UNITY Self-reproduction. Self-reproduction takes place when a unity produces
another one with a similar organization to its own, through a process that
Reproduction requires a unity to be reproduced; this is why reproduction is is coupled to the process of its own production. It is apparent that only
operationally secondary to the establishment of the unity, and it cannot enter autopoietic systems can self-reproduce because only they are realized through
as a defining feature of the organization of living systems. Furthermore, a process of self-production (autopoiesis).
since living systems are characterized by their autopoietic organization,
(ii) For an observer there is reproduction in all these three processes
reproduction must necessarily have arisen as a complication of autopoiesis
because he can recognize in each of them a unita'ry pattern of organization
during autopoiesis, and its origin must be viewed and understood as secondary
which is embodied in successively generated systems through the three well
to, and independent from the origin of the living organization. The depen-
defined mechanisms. The three processes, however, are intrinsically different
dence of reproduction upon the existence of the unity to be reproduced is not
because their dynamics give rise to different phenomenologies which appear
a trivial problem of precedence, but it is an operational problem in the origin
particularly distinct if one considers the network of systems generated under
of the reproduced system and its relations with the reproducing mechanism.
conditions in which change is allowed in the process of reproduction of
Accordingly, in order to understand reproduction and its consequences in
the successively embodied pattern of organization. Thus, in replication and
autopoietic systems we must analyze the operational nature of this process in
copy the mechanism of reproduction is necessarily external to the pattern
relation to autopoiesis.
reproduced, while in self-reproduction it is necessarily identical to it. Further-
more, only in self-copy and self-reproduction can changes in the unities
(i) There are three phenomena that must be distinguished in relation to
produced which embody the pattern reproduced affect the reproducing
the notion of reproduction; these are replication, copy and self-reproduction.
mechanism. The consequences of this will be dealt with in the next section,
Replication. A system which successively generates unities different from
but now it should be clear that the historical interconnections established
itself, but in principle identical to each other, and with an organization which
between independent unities through reproduction varies with the mechanism
the system determines in the process of their production, is a replicating
through which reproduction is achieved.
system. Replication, then, is not different from repetitive production. Any
distinction between these processes arises as a mat"ter of description in the (iii) In living systems presently known on earth autopoiesis and reproduc-
·emphasis that the observer puts on the origin of the equivalent organization tion are directly coupled and, hence, these systems are truly self-reproducing
of the successively produced unities, and on the relevance that this equivalence systems. In fact, in them reproduction is a moment in autopoiesis, and the
has in a domain different from that in which the repetitive production takes same mechanism that constitutes one constitutes the other. The consequences
place. Thus, although all molecules are produced by specific molecular and of such a coupling are paramount: (a) Self-reproduction must take place
atomic.processes that can at least in principle be repeated, only when certain during autopoiesis. Accordingly the network of individuals thus produced is
specific kinds of molecules are produced in relation to the cellular activities necessarily self-contained in the sense that it does not require for its establish-
(proteins and nucleic acids) by certain repeatable molecular concatenations ment a mechanism independent of the autopoietic determination of the
is their production called replication. Such a denomination then, strictly, self-reproducing unities. Such would not be the case if reproduction were
makes reference only to the context in which the identity of the successively attained through external copy or replication. {b) Self-reproduction is a form
produced molecules is deemed necessary, not to a unique feature of that of autopoiesis; therefore, variation and constancy in each reproductive step
particular molecular synthesis. are not independent and both must occur as expressions of autopoiesis. (c)
Copy. Copy takes place whenever a given object or phenomenon is Variation through self-reproduction of the way the autopoiesis is realized
mapped by means of some procedure upon a different system, so that an can only arise as a modification during autopoiesis of a pre-existing function-
isomorphic object or phenomenon is realized in it. In the notion of copy the ing autopoietic structure; consequently, variation through self-reproduction
emphasis is put on the mapping process, regardless of how this is realized, can only arise from perturbations that require further homeostatic com-
even if the mapping operation is performed by the model unit itself. plications to maintain autopoiesis constant. The history of self-reproductively
102 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA DIVERSITY OF AUTOPOIESIS 103

connected autopoietic systems can only be one of continuous complication events in which each event as a state of the network arises in it as a trans-
of autopoiesis. formation of the previous state, it follows that although history cannot
contribute to explain any phenomenon, it can permit an observer to account
(iv) The nature of reproduction depends on the nature of the unity. The
for the origin of.a phenomenon as a present state .in a changing network.
same goes for its origin. Replication takes place independently of autopoiesis.
This he can do because he has observational ( or descriptive) independent
Copy takes place only in heteropoiesis, and can be deemed to take place in
access to the different states of the historical process. It is in this context that
other situations solely as a description. Self-reproduction is exclusively
the phenomenology of autopoietic systems must be considered when viewed
associated to autopoiesis and its origin is bound to it as a historically secon-
in reference to evolution. Biological evolution is a historical phenomenon and
dary phenomenon. The reason for this association will be dealt with in the
as such it must be explained in the present by its reformulation as a historical
next section.
network constituted through the causal interactions of coupled or inde-
(v) Notions such as coding, message or information are not applicable to pendent biological events. Furthermore, biological events depend on the
the phenomenon of self-reproduction; their use in the description of this autopoiesis of living systems; accordingly, our aim here is to understand how
phenomenon constitutes an attempt to represent it in the language of hetero- evolution is defined as a historical process by the autopoiesis of the biological
poietic design. In fact, the notions of coding, message and transmission of unities.
information apply only to the reduction of uncertainties in the communica-
(i) If by evolution we refer to what has taken place in the history of
tive interactions between independent unities under conditions in which
transformation of terrestrial living systems, evolution is the history of change
the messenger acts as• an arbitrary non-participant link. Nucleic acids are
in the realization of an invariant organization embodied in independent
constitutive components in the process of autopoiesis, not arbitrary links
unities sequentially generated through reproductive steps, in which the
between independent entities. Thus, in self-reproduction there is no trans-
particular structural realization of each unity arises as a modification of the
mission of information between independent entities; the reproducing and
preceding one (or ones) which, thus, constitutes both its sequential and
the reproduced unities are topologically independent entities produced
historical antecedent. Consequently, evolution requires sequential reproduc-
through a single process of autopoiesis in which all components have a
tion and change in each reproductive step. Without sequential reproduction as
constitutive participation.
a reproductive process in which the structural realization of each unity in the
sequence constitutes the antecedent for the structural realization of the next
4. EVOLUTION, A HISTORICAL NETWORK one, there is no history; without change in each sequential reproductive step,
there is no evolution. In fact, sequential transformations in a unity without
A historical phenomenon is a process of change in which each state of the
change of identity constitute its ontogeny, that is, its individual history if it is
successive states of a changing system arises as a modification of a previous
an autopoietic unity.
state in a causal transformation, and not de nova as an independent occur-
rence. Accordingly, the notion of history may either be used to refer to the (ii) Reproduction by replication or copy of a single unchanging model
antecedents of a given phenomenon as the succession of events that gave implies an intrinsic uncoupling between the organization of the unities
rise to it, or it may be used to characterize the given phenomenon as a produced and their producing mechanism. As a consequence, any change in
process. Therefore, since an explanation is always given in the present as the realization of the organization embodied in the unities successively
a reformulation of the phenomenon to be explained in the domain of inter- produced by replication or copy from a single model, can only reflect the
actions of its components ( or of isomorphic elements), the history of a ontogenies of the reproducing systems or the independent ontogenies of the
phenomenon as a description of its antecedents cannot contribute to its units themselves. The result is that under no circumstance in these non-
explanation because the antecedents are not components of the phenomenon sequential reproductive cases does a change in the structure of a unity af-
which they precede or generate. Conversely, since history as a phenomenon is fect the structure of the others yet to be produced, and, independently
to be explained in the present as a changing network of sequentially produced of whether they are autopoietic or not, they do not constitute a historical
104 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA DIVERSITY OF AUTOPOIESIS 105

network, and no evolution takes place. The collection of unities thus produced differential realization in a context that specifies the unitary structures
constitutes a collection of independent ontogenies. In sequential reproduc- that can be realized. In a population of autopoietic unities selection is a
tion, as it occurs in self-reproducing systems which attain reproduction process of differential realization of autopoiesis, and, hence, if these are self-
through autopoiesis, or as it occurs in those copying systems in which each reproducing autopoietic unities, of differential self-reproduction. Consequent-
new unity produced constitutes the model for the next one, the converse is ly, if there is sequential reproduction, and the possibility of change in each
true. In these cases, there are aspects of the structural realization of each reproductive step, selection can make the transformation of the reproducible
unity that determine the structure of the next one by their direct coupling structural patterns realized in each successive unity a recursive function
with the reproductive process which is, thus, subordinated to the organization of the domain of interactions which that very same autopoietic unity specifies.
of the reproduced unities. Consequently, changes in these aspects of the If any system that is realized is necessarily adapted in the domain in which it
structure of the unities sequentially generated, that occur either during is realized, and adaptation is the condition of possible realization for any
their own ontogeny or in the process of their generation, necessarily result in system, evolution takes place only if adaptation is conserved by the unities
the production of an historical network in which the unities successively that embody the invariant organization of the evolving lineage. Accordingly,
produced embody an invariant organization in a changing structure as each different evolving systems would differ only in the domain in which they are
unity arises as a modification of the previous one. In general, then, sequential realized and hence in which selection takes place, not in whether they
reproduction with the possibility of change in each reproductive step neces- are ada~tive ~r not.' Thus, evolution in self-reproducing living systems that
sarily leads to evolution, and in particular, in autopoietic systems evolution maintain their identity in the physical space (while the realization of their
.is a consequence of self-reproduction. autopoietic organization is commensurate with the restrictions of the ambience
in which they exist), is necessarily a process of continued adaptation because
(iii) Ontogeny and evolution are completely different phenomena, both in
only those of them whose autopoiesis can be realized reproduce, regardless
their outlook and in their consequences. In ontogeny, as the history of trans-
of how much the way they are autopoietic may otherwise change in each
formation of a unity, the identity of the unity, in whatever space it may
exist, is never interrupted. In evolution, as a process of historical change there reproductive step.
is a succession of identities generated through sequential reproduction which
(v) For evolution to take place as a history of change in the realization
constitute a historical network, and that which changes (evolves), the pattern
of an invariant organization embodied in successively generated unities,
of realization of the successively generated unities exists in a different do-
reproduction must allow for structural change in the sequentially reproduced
main than the unities that embody it. A colle.ction of successive ontogenies
unities. In present living systems reproduction takes place as a modification
in whose structure an observer can see relations of maintained change,
of autopoiesis and is bound to it. This was to be expected. Originally many
but which have not been generated through sequential reproduction, do
kinds of autopoietic unities were pr9bably formed which would mutually
not constitute an evolving system, not even if they reflect the continuous
compete for the precursors. If any class of them had any possibility of self-
transformation ( ontogeny) of the system that produced them. It is inadequate
reproduction,it is evident that it would immediately displace through selection
to talk about evolution in the history of change of a single unity in whatever
the other non-reproducing forms. The onset of the history of self-reproduction
space it may exist; unities only have ontogenies. Thus, it is inadequate to talk
need not have been complex; for example, in a system with distributed
about the evolution of the universe, or the chemical evolution of the earth;
autopoiesis mechanical fragmentation is a form of self-reproduction. Evolution
one should only talk about the ontogeny of the universe or the chemical
through selection would appear with the enhancement of those features of
history of the earth. Also, there is biological evolution only since there is
the autopoietic unities that facilitated their fragmentation (and hence the
sequential reproduction of living systems; if there were non self-reproducing
regularity and frequency of self-reproduction) to the extent of making it
autopoietic systems before that, their different patterns of realization did
independent of external accidental forces. Once the most simple self-re-
not evolve, and there was only the history of their independent ontogenies.
producing process takes place in an autopoietic system, evolution is on its
(iv) Selection, as a process in a population of unities, is a process of course and self-reproduction can enter in a history of change, with the ensuing
 106 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA DIVERSITY OF AUTOPOIESIS 107

total displacement of any co-existing non-self-reproducing autopoietic unities. equivalent pattern of autopoietic realization under historical transforma-
Hence the linkage between autopoiesis and self-reproduction in terrestrial tions. Historically, a species arises when a reproductive network of this kind
living systems. Of course it is not possible to say now what actually took develops an independent reproductive network as a branch which by being an
place in the origin of biological evolution, but this does not seem to offer an independent historical network(reproductively separated) has an independent
insurmountable conceptual difficulty. The fact is that in present day living history. It is said that what evolves is the species and that the individuals in
systems self-reproduction is crucially associated to nucleic acids and their role their historical existence are subordinated to this evolution. In a superficial
in protein specification. We think that this could not have been so if the descriptive sense this is meaningful because a particular species as an existing
nucleic acid-protein association were not a condition virtually constitutive of collection of individuals represents continuously the state of a particular
the original autopoietic process which was secondarily associated to reproduc- historical network in its process of becoming one, and, if described as a
. tion and variation; and we think that this is so because only uncompensated state of a historical network, a species necessarily appears in a process of
changes at the level of the autopoietic process itself can be incorporated transformation. Yet, the species exists as a unity only in the historical domain,
(through sequential reproduction) as reproducible changes of the autopoietic while the individuals that constitute the nodes of the historical network exist
organizations of the next unity in a manner that allows for evolution to take in the physical space. Strictly, a historical network is defined by each and
place. What is not apparent, though, is whether or not there have been other every one of the individuals which constitute its nodes, but it is at any mo-
modes of autopoietic realization, and other sources of variation, than those ment represented historically by the species as the collection of all the simul-
associated with the nucleic acid-protein system, in the history of terrestrial taneously existing nodes of the network; in fact, then, a species does not
living systems. Whichever the case, once self-reproduction appears in auto- · evolve because as a unity in the historical domain it only has a history of
poiesis, any perturbation which modifies the way in which the autopoiesis is change. What evolves is a pattern of autopoietic realization embodied in many
realized, can, in principle, be reproduced in the next generation, and, thus, be particular variations in a collection of transitory individual~ that together
the source of variations if the change affected those processes involved in define a reproductive historical network. Thus, the individuals, though
reproduction. Accordingly the phenomenology of biological evolution and its transitory, are essential, not dispensable, because they constitute a necessary
origin rests on the inception of two processes: self-reproduction and variation. condition for the existence of the historical network which they define.
One refers to i:1ossible forms of complication of the autopoiesis, the other to The species is only an abstract entity in the present, and although it repre-
the introduction of perturbations which irreversibly modify the way the sents a historical phenomenon it does not constitute a generative factor in the
autopoiesis is realized. · Both undergo historical transformations, which, phenomenology of evolution, it is its result.
though coupled, are not equivalent.
(vi) Of the two possible mechanisms that can give rise to sequential 5. SECOND AND THIRD ORDER AUTOPOIETIC SYSTEMS
reproduction, the only one which is accessible to autopoietic systems in the
Whenever the conduct of two or more unities is such that there is a domain in
absence of an independent copying mechanism is self-reproduction, because
which the conduct of each one is a function of the conduct of the others, it is
of the coincidence between the reproducing mechanisms and the reproducing
said that they are coupled in that domain. Coupling arises as a result of the
unity. Sequential reproduction through copy takes place at present only in
mutual modifications that interacting unities undergo in the course of their
relation to the operation of living systems in their domain of interactions,
interactions without loss of identity. If the identity of the interacting unities
particularly in cultural learning; cultural evolution takes place through is lost in the course of their interactions, a new unity may be generated as a
sequential copy of a changing model in the process of·social indoctrination result of it, but no coupling takes place. In general, however, coupling leads
generation after generation.
also to the generation of a new unity that may exist in a different domain
(vii) A species is a population or collection of populations of reproduc- from the domain in which the component-coupled unities retain their identity.
tively interconnected individuals which are thus nodes in a historical network. The way in which this takes place, as well as the domain in which the new
Genetically these individuals share a genetic pool, that is, a fundamentally unity is realized, depend on the properties of the component unities. Coupling
108 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA DIVERSITY OF AUTOPOIESIS 109

in living systems is a frequent occurrence; the following comments are meant therefore, lies in that he sees the system of coupled autopoietic unities as
to show that the nature of the coupling of living systems is determined by a unity in his perceptive domain in terms other than those defined by its
their autopoietic organization. organization. In contrast, a system realized through the coupling of auto-
poietic unities and defined by relations of production of components that
(i) Autopoietic systems can interact with each other without loss of generate these relations and constitute it as a unity in some space, is an
identity as long as their respective paths of autopoiesis constitute reciprocal autopoietic system in that space regardless of whether the components
sources of compensable disturbances. Furthermore, due to their homeostatic produced coincide or not with the unities which generate it through their
organization autopoietic systems can couple and constitute a new unity while coupled autopoiesis. If the autopoietic system thus generated is a unity in the
their individual pathr of autopoiesis become reciprocal sources of specification physical space it is a living system. An autopoietic system whose autopoiesis
of each other's ambience, if their reciprocal deformations do not overstep entails the autopoiesis of the coupled autopoietic unities which realize it, is
their corresponding ranges of tolerance for variation without loss of auto- an autopoietic system of higher order.
poiesis. As a consequence the coupling remains invariant while the coupled In general, the actual recognition of an autopoietic system poses a cognitive
systems undergo structural changes selected through the coupling and, problem that has both to do with the capacity of the observer to recognize
hence, commensurate with it. These considerations also apply to the coupling the relations that define the system as a unity, and with his capacity to
of autopoietic and non-autopoietic unities with obvious modifications in distinguish the boundaries which delimit this unity in the space in which it
relation to the retention of identity of the latter. In general, then, the coupling is realized. Since it is a defining feature of an autopoietic system that it
of autopoietic systems with other unities, autopoietic or not, is realized should specify its own boundaries, a proper recognition of an autopoietic
through their autopoiesis. That coupling may facilitate autopoiesis requires system as a unity requires that the observer performs an operation of distinc-
no further discussion, and that this facilitation may take place through the tion that defines the limits of the system in the same domain in which it
particular way in which the autopoiesis of the coupled unities is realized has specifies them through its autopoiesis. If this is not the case he does not
already been said. It follows that selection for coupling is possible, and that observe the autopoietic system as a unity, even though he may conceive it.
through evolution under a selective pressure for coupling a composite system Thus, presently, the recognition of a cell as a molecular autopoietic unity
can be developed (evolved) in which the .individual autopoiesis of every one offers no serious difficulty because we can identify the autopoietic nature of
of its autopoietic components is subordinated to an ambience defined through its organization, and interact visually, mechanically and chemically, with one
the autopoiesis of all the other autopoietic components of the composite ·of the boundaries (membrane) which its autopoiesis generates as an interface
unity. Such a composite system will necessarily be defined as a unity by the which delimits it as a three dimensional physical unity. In addition tl1e
coupling relations of its component autopoietic systems in a space that the observer may have two kinds of difficulties in the identification of an auto-
nature of the coupling specifies, and will remain as a unity as long as the poietic unity as an actually distinguishable system: on the one hand, he may
component systems retain their autopoiesis which allows them to enter into treat the system as a unity by making an operation of distinction in a space
those coupling relations. different from the space in which it is realized because he has not yet properly
A system generated through the coupling of autopoietic unities may, on a recognized the relations of production of components that constitute it,
first approximation, be seen by an observer as autopoietic to the extent that and, hence, cannot recognize the topological relations which specify its
its realization depends on the autopoiesis of the unities which integrate it. unity in that space; on the other hand, due to his own mode of autopoietic
Yet, if such a system is not defined by relations of production of components organization (and, hence, cognitive structure) he may be unable to interact
that generate these relations and define it as a unity in a given space, but in the space in which the system is realized as a unity, and, hence, he may
by other relations, either between components or processes, it is not an be unable to observe it as a unity because he cannot specify the proper
autopoietic system and the observer is mistaken. The apparent autopoiesis perceptual dimensions. In the first case, the observer makes a unity distinction
of such a system is incidental to the autopoiesis of the coupled unities which which is not commensurate with the autopoietic system, and he thus defines
constitute it, and not intrinsic to its organization; the mistake of the observer, and operates with a different unity;in the second case he makes no distinction
110 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA DIVERSITY OF AUTOPOIESIS 111

at all, and he has no unity with which to operate. In either case the phenome- space. If the higher order autopoietic system undergoes self-reproduction
nology of the autopoietic unity remains unobservable. However, if there is no (through the self-reproduction of one of its component autopoietic unities or
misidentification of the system,. even if its unity is not yet operationally otherwise), an evolutionary process begins in which the evolution of the
observable, its phenomenology can be asserted by the recognition of the manner of realization of the component autopoietic systems is necessarily
organization that constitutes it. subordinated to the evolution of the manner of realization of the composite
unity. Furthermore, it is to be expected that if the proper contingencies are
(ii) An autopoietic system can become a component of another system if
given, higher order autopoietic unities will be formed through selection. In
some aspects of its path of autopoietic change can participate in the realiza-
fact, if coupling arises as a form of satisfying autopoiesis, a second order
tion of this other system. As has been said, this can take place in the present
unity formed from previous autopoietic systems will be more stable, the more
through a coupling that makes use of the homeostatic resorts of the interacting
stable the coupling is. However, the most stable condition for coupling appears
systems, or through evolution by the recursive effect of a maintained selective
if the unity organization is precisely geared to maintain this organization, this
pressure on the course of transformation of a reproductive historical network,
is, if the unity becomes autopoietic. There is then an ever present selective
which results in a subordination of the individual component autopoiesis
pressure for the constitution of higher order autopoietic systems from the
(through historical change in the way these are realized) to the ambience of
coupling of lower order autopoietic unities which on earth is apparent in the
reciprocal perturbations which they specify. Whichever the case, an observer
occurrence of multicellular systems, if not in that of the eucariotic cell itself.
can describe an autopoietic component of a composite system as playing an
It seems that the only limit to the process of constitution of autopoietic
allopoietic role in the realization of the larger system which it contributes to
unities of higher order is that imposed by the circumstances under which a
realize through its autopoiesis. In other words, the autopoietic unity functions
unity. can be specified in a given space.
in the context of the composite system in a manner that the observer would
describe as allopoietic. Yet, the allopoietic function is exclusively a feature of
the description and pertains to a frame of reference defined by the observer.
As we described in Chapter I, there are allopoietic machines whose organiza-
tion is intrinsically different from autopoietic machines, and can be described
(with no reference to function) by pointing out that the product of their
operation is different .from, themselves. Accordingly, when an autopoietic
system is described as having an allopoietic role as a component in a larger
system, the description makes reference only to its participation in the
production of relations that adopt the form proper to an allopoietic system,
but nothing is implied about function which is proper only in the domain of
heteropoietic human design.
(iii) If the autopoiesis of the component unities.of a composite autopoietic
system conforms to allopoietic roles that through the production of relations of
constitution, specification and order define an autopoietic space, the new
system becomes in its own right an autopoietic unity of second order. This
has actually happened on earth with the evolution of the multicellular pattern
of organization. When this occurs, the component (living) autopoietic sys-
tems become necessarily subordinated, in the way they realize their auto-
poiesis, to the maintenance of the autopoiesis of the higher order autopoietic
unity which, through their coupling, they define topologically in the physical
 CHAPTER V PRESENCE OF AUTOPOIESIS 113

the system as a unity in the physical space, biological phenomena are neces-
sarily phenomena of relations between processes which satisfy the autopoiesis
PRESENCE OF AUTOPOIESIS of the participant living systems. Accordingly, under no circumstances is a
biological phenomenon defined by the properties of its component elements,
but it i~ always defined and constituted by a concatenation of processes in
relations subordinated to the autopoiesis of at least one living system. Thus,
Autopoiesis in the physical space is necessary and sufficient to characterize a the accidental collision of two running animals, as a bodily encounter of living
system as a living system. Reproduction and evolution as they occur in the systems, is not a biological phenomenon ( even though it may have biological
known living systems, and all the phenomena derived from them, arise as consequences), but the bodily contact of two animals in courtship is. Strictly,
secondary processes subordinated to their existence and operation as auto- then, although biological and statical phenomena are physical phenomena
poietic unities. Hence, the biological phenomenology is the phenomenology because they are realized through the properties of their physical components,
of autopoietic systems in the physical space, and a phenomenon is a biological they differ because statical phenomena are phenomena of relations between
phenomenon only to the extent that it depends in one way or another on the properties of components (as previously defined), while biological phenomena
autopoiesis of one or more physical autopoietic unities. are phenomena of relations between processes. Therefore, biological phe-
nomena as phenomena of relations between processes are a subclass of the
1. BIOLOGICAL IMPLICATIONS mechanical phenomena which constitute them, and are defined through the
participation of these processes in the realization of at least one autopoietic
A living system is a living system because it is an autopoietic system in the system. The phenomenology of living systems, then, is the mechanical phe-
physical space, and it is a unity in the physical space because it is defined nomenology of physical autopoietic machines.
as a unity in that space by and through its autopoiesis. Accordingly, any
structural transformation that a living system may undergo maintaining its (ii) As the mechanical phenomenology of physical autopoietic machines,
identity must take place in a manner determined by and subordinated to its the biological phenomenology is perfectly defined, and, hence, amenable to
defining autopoiesis; hence, in a living system loss of autopoiesis is disin- theoretical treatment through the theory of autopoiesis. It follows that such
tegration as a unity and: loss of identity, that is, death. a theory as a formal theory will be a theory of the concatenation of processes
of production that constitute autopoietic systems, and not a theory of
(i) The physical space is defined by components that can be determined properties of components of living systems. It also follows that a theoretical
by operations that characterize them in terms of properties such as masses, biology would be possible as a theory of the biological phenomenology, and
forces, accelerations, distances, fields, etc. Furthermore, such properties not as the application of physical or chemical notions, which pertain to
themselves are defined by the interactions of the components that they a different phenomenological domain, to the analysis of the biological phe-
characterize. In the physical space two kinds of phenomenologies can take nomena. In fact, it should be apparent now that any attempt to explain a
place according to the way the components participate in their generation, biological phenomenon in statical or non-autopoietic mechanical terms would
namely, statical and mechanical (machine like). The statical phenomenology be an attempt to reformulate it in terms of relations between properties of
is a phenomenology of relations between properties of components; the components, or relations between processes which do not involve an auto-
mechanical phenomenology is a phenomenology of relations between pro- poietic unity in the physical space, and would fail to reformulate it. Since
cesses realized through the properties of components. What about the bio- ~ ~iological phenomenon takes place through the operation of components,
logical phenomenology, that is, what about the phenomenology of autopoietic It IS always possible to abstract from it component processes that can be
systems, which, as such, takes place in the physical space? Since a living adequately described in statical or non-autopoietic mechanical terms because
system is defined as a system by the concatenation of processes of production as abstracted processes, they in fact correspond to statical or allopoieti~
of components that generate the processes that produce them and constitute mechanical phenomena. In such a case, any connection between the statical
112
 PRESENCE OF AUTOPOIESIS 115
114 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA

2. EPISTEMOLOGICAL IMPLICATIONS
or non-autopoietic mechanical processes and the biological phenomenon ~rom
which the observer abstracts them, is provided by the observer who considers
(i) The basic epistemological question in the domain of the biological pro-
both simultaneously; the biological phenomenon, however, is not and can~ot
blems is that which refers to the validity of the statements made about biolo-
be captured by these explanations which, necessarily, re~ain a ref?rmula.t10n
gical systems. It is presently obvious that scientific statements made about the
of a phenomenon in a non-autopoietical phenomenological domam .•A .biolo-
universe acquire their validity through their operative effectiveness in their
gical explanation must be a reformulation in terms of processes subordina.ted
application in the domain where they pretend validity. Yet any observation,
to autopoiesis, that is, a reformulation in the biological phenomenological
even that one which permits us to recognize the operational validity of a
domain. scientific statement, implies an epistemology, a body of conceptual explicit
(iii) An adequate theory of the biological phenomena should pe:mit the or implicit notions that determines the perspective of the observations and,
analysis of the dynamics of the concrete components of a syste~ in order hence, what can and what cannot be observed, what can and what cannot
to determine whether or not they participate in processes that integrate a be validated by its operative effectiveness, what can and what cannot be
biological phenomenon. In fact, no matter how much we think we understand explained by a given body of theoretical concepts. This has been a fundamen-
biological problems today, it is apparent that withou~ an adequat: t~eory tal problem in the conceptual and experimental handling of the biological
of autopoiesis it will not be possible to answer questions such as: Given a phenomena, as it is apparent in the history of biology, which reveals a con-
dynamic system, what relations should I observe b~t.ween ~ts concrete com- tinuous search for the definition of the biological phenomenology in a manner
ponents to determine whether or not they participate in. processes that such that would permit its complete explanation through well-defined notions,
make it a living system?'; or, 'Given a set of components with well-defined and, accordingly, its complete validation in the observational domain. In this
properties, in what processes of production can they participate so that the respect, evolutionary and genetic notions have been the most successful.
components can be concatenated to form an autopoietic system?' The ans":'e:s Yet these notions alone are insufficient because, although they provide a
to these questions are essential if one wants to solve the problem ~f the ongin mechanism for historical change, they do not adequately define the domain
of living systems on earth. The same questions must be answered tf one wants o{ the biological phenomenology. In fact, evolutionary and genetic notions
to design a living system. In particular, it should be possible to det~rmine from (by emphasizing generational change) treat the species as the source of all
biological theoretical considerations which relations should be sattsfie~ by any biological order, showing that the species evolves while the individuals are
set of components if these are to participate in processes that constltut~ ~n transient components whose organization is subordinated to its historical
autopoietic unity. Whether one may or may not want to make an autop01et1c phenomenology. However, since the species is, concretely at any moment,
system is, of course, a problem that pertains to the ethical domain. However, a collection of individuals capable in principle of interbreeding, it turns out
if our characterization of living systems is adequate it is apparent that they that what would define the organization of individuals is either an abstraction,
could be made at will. What remains to be seen is whether such a system has or something that requires the existence of well-defined individuals to begin
already been made by man, although unwittingly, and with what consequences. with. Where does the organization of the individual come from? Which is the
mechanism for its determination? This difficulty cannot be solved on purely
(iv) The characterization of living systems as physic~ autopoi~ti~ s~stems
evolutionary and genetic arguments, since it is apparent (even for evolutionists
must be understood as having universal value, that is, autop01es1s in the
and geneticists) that any attempt to overcome it by resorting to other notions
physical space must be viewed as defining living systems any~vhere in t~e
of comprehensive nature, is doomed to failure if it does not provide us with a
universe however different they may otherwise be from terrestnal ones. This
mechanism to account for the phenomenology of the individual. Such is the
is not t~ be considered as a limitation of our imagination, nor as a denial that
case when some sort of preformism is introduced by applying informational
there might exist still unimagined complex systems. It is a statement about
notions at the molecular level (nucleic acids or proteins); or when organismic
the nature of the biological phenomenology: the biological phenomenology
notions are used that emphasize the unitary character of living systems but do
is not less and not more than the phenomenology of autopoietic systems in
not provide a mechanism for the definition of the individual. These notions
the physical space.
116 HUMBER TO R. MATURANA AND FRANCISCO J. VARELA PRESENCE OF AUTOPOIESIS 117

fail because they imply the validity of the same notion that they want to applies also to the different phenomenological domains generated through the
explain. operation of autopoietic systems. Accordingly, as an autopoietic system
As is apparent from all that has been said, the key to the understanding cannot be explained through statical or non-autopoietic mechanical relations
of the biological phenomenology is the understanding of the organization in the space in which it exists, but it must be explained through autopoietic
of the individual. We have shown this organization to be the autopoietic mechanical relations in the mechanical domain, the phenomena generated
organization. Furthermore, we have shown that this organization and its through interactions of autopoietic unities must be explained in the domain
origin are fully explainable with purely mechanistic notions which are valid of interactions of the autopoietic unities through the relations that define
for any mechanistic phenomenon in any space, and that once the autopoietic that domain.
organization is established it determines an independent phenomenological
sub domain of the mechanistic phenomenology, the domain of the biological (iii) The development of the Darwinian notion of evolution with its
phenomena. As a result, the biological domain is fully defined and self- emphasis on the species, natural selection and fitness, had an impact in human
contained, no additional notions are necessary, and any adequate biological affairs that went beyond the explanation of diversity and its origin in living
explanation has the same epistemological validity that any mechanistic systems. It had sociological significance because it seemed to offer an ex-
explanation of any mechanistic phenomenon in the physical space has. planation of the social phenomenology in a competitive society, as well as a
scientific justification for the subordination of the destiny of the individuals
(ii) A phenomenological domain is dP,fined by the properties of the unity to the transcendental values supposedly embodied in notions such as mankind,
or unities that constitute it, either singly or collectively through their trans- the state, or society. In fact, the social history of man shows a continuous
formations or interactions. Thus, whenever a unity is defined, or a class or search for values that explain or justify human existence, as well as a con-
classes of unities are established which can undergo transformations or tinuous use of transcendental notions to justify social discrimination, slavery,
interactions, a phenomenological domain is defined. Two phenomenological economical subordination and political submission of the individuals, isolated
domains intersect only to the extent that they have common generative or collectively, to the design or whim of those who pretend to represent
unities, that is, only to the extent that the unities that specify them interact; th~ values contained in those notions. For a society based on economic
otherwise they are completely independent and, obviously, they cannot discrimination, competitive ideas of power and subordination of the citizen
generate each other without transgressing the domains of relations of their to the state, the notions of evolution, natural selection and fitness (with their
respective specifications'. . Conversely, one phenomenological domain can emphasis on the species as the perduring historical entity maintained through
generate unities that define a different phenomenological domain, but such a the dispensability of transient individuals) seemed to provide a biological
domain is specified by the properties of the new different unities, not by (scientific) justification for its economic and social structure. It is true on
the phenomenology that generates them. If this were not the case the new biological grounds that what evolves is mankind as the species Homo sapiens.
unities would not be in fact different unities, but they would be unities of the It is true on biological grounds that competition participates in the specifi-
same class of units that generate the parental phenomenological domain, and cation of evolutionary change even in man. It is true that under the laws of
they would generate a phenomenological domain identical to it. Autopoietic natural selection the individuals most apt in the features which are favorably
systems do generate different phenomenological domains by generating selected survive, or have reproductive advantages over the others, and that
unities whose properties are different from the properties of the unities that those which do not survive or are less successful in the reproductive sense
generate them. These new phenomenological domains are subordinated to the do not contribute or contribute less to the historical destiny of the species.
phenomenology of the autopoietic unities because they depend on these for Thus, from the Darwinian perspective it seemed that the role of the individual
their actual realization, but they are not determined by them; they are only was to contribute to the perpetuation of the species, and that all that one had
determined by the properties of their originating unities regardless of how to do for the well-being of mankind was to let the natural phenomena follow
these were originated. One phenomenological domain cannot be explained by their course. Science, biology, appeared to justify the notion 'anything for
relations which are valid for another domain; this is a general case which the benefit of mankind', whatever the intention or purpose of whoever
118 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA PRESENCE OF AUTOPOIESIS 119

uttered it first. We have shown, however, that these arguments are not valid deformations that it may undergo without loss of autopoiesis. Such a domain
to justify the subordination of the individual to the species, because the is determined for each unity by the particular mode through which its auto-
biological phenomenology is determined by the phenomenology of the poiesis is realized in the space of its components, that is, by its structure. It
individuals, and without individuals there is no biological phenomenology follows that the domain of interactions of an autopoietic unity is necessarily
whatsoever. The organization of the individual is autopoietic and upon this bounded, and that autopoietic unities with different structures have different
fact rests all its significance: it becomes defined through its existing, and its domains of interactions. Furthermore, an observer can consider the way in
existing is autopoietic. Thus, biology cannot be used anymore to justify the which an autopoietic system compensates its deformations as a description of
dispensability of the individuals for the benefit of the species, society or the deforming agent that he sees acting upon it, and the deformation suffered
mankind under the pretense that its role is to perpetuate them. Biologically by the system as a representation of the deforming agent. However, since the
the individuals are not dispensable. domain of interactions of an autopoietic system is bounded, an observer of an
(iv) Biological phenomena depend upon the autopoiesis of the individuals autopoietic system can describe entities external to it (by interacting with
involved; thus, there are biological systems that arise from the coupling of them) which it cannot describe because it cannot interact with them or it
autopoietic unities, some of which may even constitute autopoietic systems cannot compensate the deformations which these cause in it. The domain of
of higher order. What about human societies, are they, as systems of coupled all the interactions in which an autopoietic system can enter without loss of
human beings, also biological systems? Or, in other words, to what extent do identity is its cognitive domain; or, in other words, the cognitive domain of
the relations which characterize a human society as a system constitutively an autopoietic system is the domain of all the descriptions which it can
depend on the autopoiesis of the individuals which integrate it? If human possibly make. Accordingly, for any autopoietic system its particular mode
societies are biological systems the dynamics of a human society would be of autopoiesis determines its cognitive domain and hence its behavioral
determined through the autopoiesis of its components. If human societies diversity, and it follows that the cognitive domain of an autopoietic system
are not biological systems, the social dynamics would depend on laws and changes along its ontogeny only to the extent that its mode of autopoiesis
relations which are independent of the autopoiesis of the individuals which changes.
integrate them. The answer to this question is not trivial and requires con- We shall not explore in this book all the implications that the proper
siderations which in addition to their biological significance have ethical and characterization of the biological phenomenology has within the domain of
political implications. This is obviously the case, because such an answer cognition, but we shall make four remarks in order to show the dependence
requires the characterization of the relations which define a society as a unity of this domain upon the autopoietic organization of the individual.
(a system), and whatever we may say biologically will apply in the domain of (i) For any autopoietic system its cognitive domain is necessarily relative
human interactions directly, either by use or abuse, as we saw it happen to the particular way in which its autopoiesis is realized. Also, if knowledge
with evolutionary notions. In fact no position or view that has any relevance is descriptive conduct, it is relative to the cognitive domain of the knower.
in the domain of human relations can be deemed free from ethical and Therefore, if the way in which the autopoiesis of an organism is realized
political implications nor can a scientist consider himself alien to these changes during its ontogeny, the actual knowledge of the organism (its
implications. This responsibility we are ready to take, yet since we - Maturana conduct repertoire) also changes; knowledge, then, is necessarily always a
and Varela - do not fully agree on an answer to the question posed by the reflection of ontogeny of the knower because ontogeny as a process of
biological character of human societies from the vantage point of this charac- continuous structural change without loss of autopoiesis is a process of
terization of the biological organization, we have decided to postpone this continuous specification of the behavioral capacity of the organism, and,
discussion. hence, of its actual domain of interactions. Intrinsically, then, no absolute
knowledge is possible, and the validation of all possible relative knowledge is
3. COGNITIVE IMPLICATIONS attained through successful autopoiesis.
The domain of interactions of an autopoietic unity is the domain of all the (ii) Autopoietic systems may interact with each other under conditions
120 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA PRESENCE OF AUTOPOIESIS 121

that result in behavioral cpupling. In this coupling, the autopoietic conduct of others a linguistic consensual domain, can treat its own linguistic states as a
an organism A becomes a source of deformation for an organism B, and the source of deformations and thus interact linguistically in a closed linguistic
compensatory behavior of organism B acts, in turn, as a source of deformation domain. Such a system has two remarkable properties:
of organism A, whose compensatory behavior acts again as a source of de-
formation of B, and so on recursively until the coupling is interrupted. In this 1. Through recursive interactions with its linguistically generated states
manner, a chain of interlocked interactions develops such that, although in it can treat some of these states as objects of further interactions, giving rise
each interaction the conduct of each organism is constitutively independent to a metadomain of consensual distinctions that appears to an observer as a
in its generation of the conduct of the other, because it is internally deter- domain of interactions with representations of interactions. When this hap-
mined by the structure of the behaving organism only, it is for the other pens the system operates as an observer. The domain of such recursive inter-
organism, while the chain lasts, a source of compensable deformations which actions is, in principle, infinite because once the system has attained the
can be described as meaningful in the context of the coupled behavior. These mechanism for doing so there is no moment in which it will not be in the
are communicative interactions. In other word·s, if the interacting organisms position of recursively interacting with· its own states, unless autopoiesis is
as dynamic systems have continuously changing structures, and if they lost. Whether an autopoietic system with this capacity does in fact generate
reciprocally select in each other their respective paths of ontogenic structural an endless series of different states during its ontogeny depends, obviously,
changes through their interactions without loss of autopoiesis, then they on whether its history of linguistic interactions in the metadomain of descrip-
generate, as a recursive or expanding domain of communicative interactions, tions has significance for the circumstantial realization of the autopoiesis of
interlocked ontogenies that together constitute a domain of mutually the interacting organisms.
triggering consensual conducts that becomes specified during its generation. 2. A living system capable of being an observer can interact with those
Such a consensual domain of communicative interactions in which the of its own descriptive states which are linguistic descriptions of itself. By
behaviorally coupled organisms orient each other with modes of behavior doing so it generates the domain of self-linguistic descriptions within which
whose internal determination has become specified during their coupled it is an observer of itself as an observer, a process which can be necessarily
ontogenies, is a linguistic domain. In such a consensual domain of interactions repeated in an endless manner. We call this domain the domain of self-
the conduct of each organism may be treated by an observer as constituting a observation and we consider that self-conscious behavior is self-observing
connotative description of the conduct of the other, or, in his domain of behavior, that is, behavior within the domain of self-observation. The observer
description as an obse'rver, as a consensual denotation of it. Communicative as an observer necessarily always remains in a descriptive domain, that is, in
and linguistic interactions are intrinsically not informative; organism A does
a relative cognitive domain. No description of an absolute reality is possible.
not and cannot determine the conduct of organism B because due to the
Such a description would require an interaction with the absolute to be
nature of the autopoietic organization itself every change that an organism
described, but the representation which would arise from such an interaction
undergoes is necessarily and unavoidably determined by its own organization.
would necessarily be determined by the autopoietic organization of the
A linguistic domain, then, as a consensual domain that arises from the coupling
observer, not by the deforming agent; hence, the cognitive reality that it
of the ontogenies of otherwise independent autopoietic systems, is intrin-
would generate would unavoidably be relative to the knower.
sically non-informative, even though an observer, by neglecting the internal
determination of the autopoietic systems which generate it, may describe it In every explanation, be this an actual concrete reproduction, a formal
as if it were so. Phenomenologically the linguistic domain and the domain of representation or a purely rational description, the reformulation of the
autopoiesis are different domains, and although one generates the elements of phenomenon to be explained resorts to the same notions (identity, exclusion,
the other, they do not intersect. succession, etc.). There is, then, a universal logic, valid for all phenomeno-
logical domains, that refers to the relations possible between the unities that
(iii) An autopoietic system capable of interacting with its own states (as generate these domains, and not to the particular properties of the generating
an organism with a nervous system can do), and capable of developing with unities. We have applied this logic (it could not have been otherwise) in this
122 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA PRESENCE OF AUTOPOIESIS 123

book, and the validity of our arguments, as the validity of any rational (eobiogenesis and neobiogenesis), and those which refer to the particular
argument or concrete phenomenological realization, rests on its validity. organization through which recursive descriptive interactions take place in
Furthermore, we have in principle shown through its application that the animals {the nervous system). Autopoietic systems define the world in which
phenomenology of autopoietic systems generates observers, and through they can exist in relation to their autopoiesis, and some interact recursively
them the phenomenology of description within which this logic is also valid. with this world through their descriptions, it being impossible for them to step
For epistemological reasons, in order to say all that we have said about living out of this relative descriptive domain through descriptions. This demands an
systems, we had to assume a space (the physical space) within which the entirely new cognitive outlook: there is a space in which different phenome-
phenomenology of autopoiesis of living systems takes place. To the extent nologies can take place; one of these is autopoiesis; autopoiesis generates a
that we have been successful (free from logical and experiential contradic- phenomenological domain, this is cognition.
tions), we can conclude that such a space is ontologically a space within
which the logic that we have applied in our description is intrinsically valid. If
this were not the case we could not have done what we have done in terms of
characterizing living systems, or of showing how these may generate systems
capable of their own description. We cannot characterize this space in absolute
terms. In linguistic interactions, all that we can do is to describe through
linguistic behavior and construct further descriptions based on these descrip-
tions ~hich always remain in the same domain of operations defined in
relation to the operating system.
A prediction is a statement of a case within a relational matrix; it is a
cognitive statement, and as such it takes place within a descriptive domain.
Thus, unless mistakes are made, if all the relations that define the particular
matrix within which the prediction is made are properly taken, the prediction
is valid. Errors of interpretations may arise only by mis-application, that is,
by pretending that the observer makes a prediction in one matrix when he
is making it in another. 'In particular, predictions in the physical space are
possible, because a description, as an actual behavior, exists in a matrix of
interactions which (by constitution) has a logical matrix necessarily isomor-
phic with the substratum matrix within which it takes place, not because we
have an absolute knowledge of the universe. These cognitive relations are
valid for the possible cognitive phenomenology generated by any closed
system. Living systems are an existential proof; they exist only to the extent
that they can exist. The fantasy of our imagination cannot deny this. Living
systems are concatenations of processes in a mechanistic domain; fantasies are
concatenations of descriptions in a linguistic domain. In the first case, the
concatenated unities are processes; in the second case, they are modes of
linguistic behavior.
Autopoiesis solves the problem of the biological phenomenology in general
by defining it. New problems arise, and old ones appear in a different perspec-
tive; in particular, those which refer to the origin of living systems on earth
 APPENDIX 125

biochemical environment for the autopoiesis of the neurons as well as for

all other cells, and, hence, is a possible source of physical and biochemical
APPENDIX perturbations which may alter the properties of the neurons and thus lead
to (ii) or (iii).

(ii) There are states of the organism (physical and biochemical) which
THE NERVOUS SYSTEM change the state of activity of the nervous system as a whole by acting upon
the receptor surfaces of some of its component neurons, and thus lead to (iii).
The phenomenology of an organism as a unity is the phenomenology of its
(iii) There are states of the nervous system which change the state of the
autopoiesis. The changes that an organism undergoes while maintaining its organism (physical or biochemical) and lead recursively to (i) and (ii).
autopoiesis constitute its conduct. The conduct of an organism is revealed
to an observer by the changes that it causes in the ambience (including the Through this coupling the nervous system participates in the generation of
observer) in which it exists. Accordingly, the conduct which an observer the autopoietic relations which define the organism which it integrates, and,
beholds in any organism, however complex it may seem, is always an ex- accordingly, its structure is subordinated to this participation.
pression of the autopoiesis of the observed organism, and as such, it always
arises through a phenomenology that takes places in the present because
I. The Neuron
history is not a causal component in the mechanism of autopoiesis (see Chap-
ter IV). Yet it appears to us as subjects of self-observation and as observers of Neurons determine their own boundaries through their autopoiesis; therefore
the conduct of other organisms that past experiences determine our and their they are the anatomical units of the nervous system. There are many classes
conduct in the present as if, embodied in modifications of the nervous system, of neurons that can be distinguished by their shapes, but all of them, regard-
they were causal components in the mechanism which generates behavior. It less of the morphological class to which they belong, have branches which put
appears, therefore, as if the operation of the organism as a state-determined them in direct or indirect operational relations with other otherwise separated
system in which time is not a component were determined by temporal neurons. Functionally, that is, viewed as an allopoietic component of the
phenomena, and we speak of learning, memory and recall as embodiments of l)ervous system, a neuron has a collector surface, a conducting element,
the past. We consider th.at this contradiction arises from not distinguishing and an effector surface, whose relative positions, shapes and extensions are
what pertains to the phenomenology of the autopoiesis from what pertains different in different classes of neurons. The collector surface is that part of
to the domain of interactions of the organism as a unity, and, thus, from an the surface of a neuron where it receives afferent influences. (synaptic or not)
inadequate evaluation of the coupling of the structure of the nervous system from the effector surfaces of other neurons or its own. The effector surface
to the ontogeny of the organism. Accordingly, our purpose in this Appendix of a neuron is that part of its surface which either directly (by means of
about the nervous system is to consider its organization as a neuronal network synaptic contacts) or indirectly (through its synaptic or nonsynaptic action
and to evaluate this coupling in which past and present arise as new dimensions on other kinds of cells) affects the collector surface of other neurons or its
from the recursive interactions of the organism with its own states. own. Depending on its kind, a neuron may have its collector and effector
surfaces completely or partly separated by a conducting element (absence
or presence of presynaptic inhibition), or it may have both collector and
A. THE NERVOUS SYSTEM AS A SYSTEM
effector surfaces completely interspaced, with no conducting element be-
The nervous system is a network of interacting neurons coupled in three ways tween them (amacrine cells). The interactions between collector and effector
to the organism of which it is a component: surfaces may be excitatory or inhibitory according to the kinds of neurons
involved. Excitatory afferent influences cause a change in the state of activity
(i) The organism, including the nervous system, provides the physical and of the collector surface of the receiving neuron which may lead to a change in
124
126 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA APPENDIX 127

the state of activity of its effector surface, while the inhibitory influences 2. Organization: The Nervous System As a Closed System
impinging on it shunt off the effect of the afferent influences on its receptor
surface so that this effect does not at all reach its effector surface, or reaches From the descriptive point of view it is possible to say that the properties of
this surface with reduced effectiveness. the neurons, their internal structure, shape and rel.ative position, determine
Operationally the state of activity of a neuron, characterized by the state the connectivity of the nervous system and constitute it as a dynamic network
of activity of its effector surface, is determined by both its internal structure of neuronal interactions. This connectivity, that is, the anatomical and
(membrane properties, relative thickness of branches, and in general all operational relations which hold between the neurons which constitute the
structural relations which determine its possible states) and the afferent nervous system as a network of lateral, parallel, sequential and recursive
influences impinging on its receptor surface. Conversely, the effectiveness of inhibitory and excitatory interactions, determines its domain of possible
a neuron in changing the state of activity of other neurons depends both on dynamic states. Since the properties of the neurons change along the ontogeny
the internal structure of these, and on the relative effectiveness of its action of the organism, both due to their internal determination and as a result of
on their receptor surfaces with respect to the other afferent influences that their interactions as components of the nervous system, the connectivity of
these neurons receive. This is so because excitatory and inhibitory influences the nervous system changes along the ontogeny of the organism in a manner
do not add linearly in the determination of the state of activity of a neuron, recursively selected during this ontogeny. Furthermore, since the ontogeny
but rather have effects which depend on the relative position of their points of the organism is the history of its autopoiesis, the connectivity of the
of action with respect to each other and with respect to the effector surface nervous system, through the neurons which constitute it, is dynamically
of the receiving cell. Furthermore, the internal structure of a neuron changes subordinated to the autopoiesis of the organism which it integrates.
along its life histpry, both as a result of its autonomous genetic determinations Operationally, the nervous system is a closed network of interacting
and as a result of the circumstances of its operations during the ontogeny of neurons such that a change of activity in a neuron always leads to a change of
the organism. Thus, neurons are not static entities whose properties remain activity in other neurons, either directly through synaptic action, or indirectly
invariant. On the contrary, they change. This has three general consequences: through the participation of some physical or chemical intervening element.
Therefore, the organization of the nervous system as a finite neuronal network
(i) There are many configurations of afferent (input) influences on the is defined by relations of closeness in the neuronal interactions generated in
receptor surface of a neuror1 which produce the same configuration of efferent the network. Sensory and effector neurons, as they would be described by an
( output) activity at its effector surface. observer who beholds an organism in an environment, are not an exception to
this because all sensory activity in an organism leads to activity in its effector
(ii) Changes in the internal structure of a neuron (regardless of whether
surfaces, and all effector activity in it leads to changes in its sensory surfaces.
they arise selected by the autonomous transformation of the cell, or by its
That at this point an observer should see environmental elements intervening
history of interactions in the neuronal network) by changing the domain of
between the effector and the sensory surfaces of the organism, is irrelevant
states of activity that the neuron can adopt, change its domain of input-output
because the nervous system is defined as a network of neuronal interactions by
relations, that is, change its transfer function.
the interactions of its component neurons regardless of intervening elements.
(iii) No .single cell or class of cells can alone determine the properties of Therefore, as long as the neural network closes on itself, its phenomenology is
the neural network which it integrates. the phenomenology of a closed system in which neuronal activity always
leads to neuronal activity. This is so even though the ambience can perturb
Generally then, the structure of a neuron and its role in the neuronal the nervous system and change its status by coupling to it as an independent
network which it integrates does not stay invariant, but changes along its agent at any neuronal receptor surface. The changes that the nervous system
ontogeny in a manner subordinated to the ontogeny of the organism which can undergo without disintegration (loss of defining relations as a closed
is both a result and a source of the changes that the neuronal network and the neuronal network) as a result of these or any other perturbation are fully
organism undergo. specified by its connectivity, and the perturbing agent only constitutes a
128 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA APPENDIX 129

historical determinant for the concurrence of these changes. As a closed 4. Architecture

neuronal network the nervous system has no input or output, and there is no
intrinsic feature in its organization which would allow it to discriminate The connectivity of the nervous system is determined by the shapes of
through the dynamics of its changes of state between possible internal or its component neurons. Accordingly, every nervous system has a definite
external causes for these changes of state. This has two fundamental con- architecture determined by the kinds and the numbers of the neurons which
sequences: compose it; therefore, members of the same species have nervous systems
with similar architectures to the extent that they have similar kinds and
(i) The phenomenology of the changes of state of the nervous system is
numbers of neurons. Conversely, members of different species have nervous
exclusively the phenomenology of the changes of state of a closed neuronal systems with different architectures according to their specific differences in
network; that is, for the nervous system as a neuronal network there is no neuronal composition. Therefore, the closed organization of the nervous
inside or outside. system is realized in different species in different manners that have been
(ii) The distinction between internal and external causes in the origin of · determined through evolution; in all cases, however, the following conditions
the changes of state of the nervous system can only be made by an observer are satisfied:
that beholds the organism (the nervous system) as a unity, and defines its (i) Since, due to its constitution as a network oflateral, parallel, sequential
inside and outside by specifying its boundaries. and recursive interactions, the nervous system closes on itself at all levels, the
It follows that it is only with respect to the domain of interactions of the mutilations that it may suffer generally leave a closed neuronal network with
organism as a unity that the changes of state of the nervous system may have a changed architecture. Accordingly, the organization of the nervous system
an internal or an external origin, and, hence, that the history of the causes of is essentially invariant under mutilations, while its domain of possible states,
the changes of state of the nervous system lies in a different phenomenological which depends on its structure, and, hence, on its architecture, is not. Yet,
domain than the changes of state themselves. due to its closed organization, whatever is left of the neural network after
a partial ablation necessarily operates as a different whole with different
properties than the original, but not as a system to which some of its proper-
3. Oiange ties have been selectively subtracted.
/

Any change in the structure of the nervous system arises from a change in the
(ii) There is intrinsically no possibility of operational localization in the
properties of its component neurons. What change in fact takes place, whether nervous system in the sense that no part of it can be deemed responsible for
morphological or biochemical or both, is irrelevant for the present discussion.
its operation as a closed network, or for the properties which an observer can
The significant point is that these changes arise in the coupling of the nervous
detect in its operation as a unity. However, since every nervous system has
system and the organism through their homeostatic operation subordinated
a definite architecture, every localized lesion in it necessarily produces a
to the autopoiesis of the latter. Some of the changes directly affect the
specific disconnection between its parts and, hence, a specific change in its
operation of the nervous system because they take place through its working domain of possible states.
as a closed network; others affect it indirectly because they take place through
the biochemical or genetic coupling of the neurons to the organism and (iii) The architecture of the nervous system is not static, but it becomes
change the properties of the neurons in a manner unrelated to the actual specified along the ontogeny of the organism to which it belongs, and its
working of the network. The results are twofold: on the one hand, all changes determination, although under genetic control, is bound to the morphogenesis
lead to the same thing, that is, changes in the domain of possible states of of the whole organism. This has two implications: (a) the variability in the
the nervous system; on the other hand the nervous system is coupled to the architecture of the nervous system of the members of a species is determined
organism both in its domain of interactions and in its domain of internal by individual differences in genetic constitution and ontogeny; (b) the range
transformations. of permissible individual variations ( compatible with the autopoiesis) is
130 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA APPENDIX 131

determined by the circumstances in which the autopoiesis of the organism is system itself. As a consequence, the nervous system necessarily operates as an
realized. homeostatic system that maintains invariant the relations which define its
participation in the autopoiesis of the organism, and does so by generating
(iv) The architecture of the nervous system and the morphology of the neuronal relations which are historically determined along the ontogeny of
organism as a whole define the domain in which the ambience can possibly the organism through its participation in this ontogeny. This has the following
couple on the organism as a source of its deformations. Thus, as long as the implications:
architecture of the nervous system and the morphology of the organism
(i) The changes that the nervous system undergoes as an homeostatic
remain invariant, or as long as there are aspects of them which remain un-
system while compensating the deformations that it suffers as a result of
changed, there is the possibility of recurrent perturbations as recurrent
the interactions of the organism (itself an homeostatic system), cannot be
configurations of the ambience which couple in the same way on the nervous
localized to any singular point in the nervous system, but must be distributed
system and the organism.
along it in a non-random manner because any localized change is itself a
source of additional deformations that must be compensated with further
5. Referential States changes. This process is potentially endless. As a result, the operation of the
nervous system as a component of the organism is a continuous generation of
There are states of the nervous system which, as referential states, defin'e
significant neuronal relations, and all the transformations that it may undergo
subdomains of the possible states that the nervous system (and the organism) as a closed neuronal network are subordinated to this. If as a result of a
can adopt under perturbations as matrices of possible internal relations. As a perturbation the nervous system fails in the generation of the significant
result when the nervous system is in different referential states it compensates neuronal relations for its participation in the autopoiesis of the organism, the
the same perturbations (characterized as configuration of the ambience) organism disintegrates.
following different characteristic modes of change. Emotions, sleep, wakeful-
ness, are referential states. In the dynamics of the nervous system, referential (ii) Although the organism and nervous system are closed atemporal
states are defined as are all other states of the nervous system, that is, by systems, the fact that the structure of the nervous system is determined
relations of neuronal activity, and as such are generated by change of neuronal through its participation in the ontogeny of the organism makes this structure
activity and generate changes of neuronal activity. What is peculiar to them is a function of the circumstances which determine this ontogeny, that is, of the
that they constitute states on which other states can be inserted as substates history of interactions of the organism as well as of its genetic determination.
in the process of generating the autopoiesis of the organisms. Therefore, their Therefore, the domain of the possible states that the nervous system can
distinction lies in the domain of observation because for the nervous system · adopt as an atemporal system is at any moment a function of this history of
they are part of their dynamic of state to state operations, and in the domain interactions and implies it. The result is the coupling of two constitutively
of observation they constitute independent phenomenological dimensions. different phenomenologies, the phenomenology of the nervous system (and
the organism) as a closed homeostatic system, and the phenomenology of the
ambience (including the organism and the nervous system) as an open non-
B. CONSEQUENCES
homeostatic system which thus braid together in a manner such that the
domain of the possible states of the nervous system continuously becomes
1. Historical Coupling
commensurate with the domain of the possible states of the ambience.
Due to its coupling with the organism the nervous system necessarily partici- Furthermore, since all states of the nervous system are internal states, and the
pates in the generation of the relations which constitute the organism as an nervous system cannot make a distinction in its process of transformation
autopoietic unity. Also due to this coupling the structure of the nervous between its internally and externally generated changes, the nervous system
system is necessarily continuously determined and realized through the is bound to couple its history of transformations to the history of its inter-
generation of neuronal relations internally defined with respect to the nervous nally determined changes of state as much as to the history of its externally
132 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA APPENDIX 133

determined changes of state. Thus the transformations that the nervous system associated to a behavioral change that takes place under maintained auto-
undergoes during its operation are a constitutive part of its ambience. poiesis, occurs due to the continuous dynamic coupling of the state-deter-
mined phenomenology of the nervous system and the state-determined
(iii) The historical coupling of the nervous system to the transformations phenomenology of the ambient. The notions of acquisition of representations
of its ambience, however, is apparent only in the domain of observation, not of the environment or of acquisition of information about the environment in
in the domain of operation of the nervous system which remains a closed relation to learning, do not represent any aspect of the operation of the
homeostatic system in which all states are equivalent to the extent that they nervous system. The same applies to notions such as memory and recall,
all lead to the generation of the relations which define its participation in the which are descriptions made by the observer of phenomena that take place in
autopoiesis of the organism. The observer can see that a given change in the his domain of observation, and not in the domain of operation of the nervous
structure of the nervous system arises as a result of a given interaction of system, and, hence, have validity only in the domain of descriptions, where
the organism, and he can consider this change as a representation of the they are defined as causal components.
circumstances of the interaction. The representation, however, as a phe-
nomenon exists only in the domain of observation and has a validity that
3. Time as a Dimension
applies only in the domain generated by the observer as he maps the environ-
ment on the behaviors of the organism by treating it as an allopoietic system. Any mode of behavioral distinction between otherwise equivalent inter-
The referred change in structure of the nervous system constitutes a change actions, in a domain that has to do with the states of the organism and not
in the domain of its possible states under conditions in which the representa- with the ambience features which define the interaction, gives rise to a
tion of the causing circumstances do not enter as a component. referential dimension as a mode of conduct. This is the case with time. It is
sufficient that as a result of an interaction (defined by an ambience con-
figuration) the nervous system should be modified with respect to the specific
2. Learning as a Phenomenon
referential state (emotion of assuredness, for example) which the recurrence
If the connectivity structure of the nervous system changes as a result of of the interaction (regardless of its nature) may generate for otherwise
some interactions of the organism, the domain of the possible states which it equivalent interactions to cause conducts which distinguish them in a dimen-
(and the organism) can henceforth adopt, changes; as a consequence, when sion associated with their sequence, and, thus, give rise to a mode of behavior
the same or similar conditions of interaction recur, the dynamic states of the which constitutes the definition and characterization of this dimension.
nervous system and, therefore, the way the organisms attains autopoiesis are Therefore, sequence as a dimension is defined in the domain of interactions
necessarily different from what they would have otherwise been. Yet, that of the organism, not in the operation of the nervous system as a closed
the conduct of the organism under the recurrent (or new) conditions of neuronal network. Similarly, the behavioral distinction by the observer of
interaction should be autopoietic and, hence, appear adaptive to an observer, sequential states in his recurrent states of nervous activity, as he recursively
is a necessary outcome of the continuous homeostatic operation of both interacts with them, constitutes the generation of time as a dimension of
the nervous system and the organism. Since this homeostatic operation the descriptive domain. Accordingly, time is a dimension in the domain of
continuously subordinates the nervous system and the organism to the latter's descriptions, not a feature of the ambience.
autopoiesis in an internally determined manner, no change of connectivity
in the nervous system can participate in the generation of behavior as a
C. IMPLICATIONS
representation of the past interactions of the organism: representations
belong to the domain of descriptions. The change in the domain of the Since history as a phenomenon is accessible to the observer only in the
possible states that the nervous system can adopt, which takes place along the domain of descriptions, it is only in this domain that history may participate
ontogeny of the organism as a result of its interactions, constitutes learning. in the generation of the observer's behavior. This, in fact, takes place. Descrip-
Thus, learning as a phenomenon of transformation of the nervous system tions as linguistic behavior constitute a source of deformations of the nervous
134 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA

system and, hence, part of its ambience. Accordingly, the phenomenology of

transformation of the nervous system discussed above also applies to the
interactions of the organism in the domain of descriptions, and the structure GLOSSARY
of the nervous system is also a function of the history of interactions of the
organism in this domain. The implications are obvious. The operation of
the nervous system makes no distinction between its different sources of
deformation, and, accordingly, it makes no difference with respect to this This glossary only contains words that in this work are given a specific
operation whether the deforming agents are physical environmental features meaning or words that are new. All the definitions are direct quotations from
or behavioral interactions with coupled organisms. Therefore, although the the text.
nervous system operates in a state-to-state fashion, time as a mode of behavior
enters in the determination of its states through the descriptive domain as a ALLOPOIETI C MACHINE: machines that have as product of their func-
component in the domain of behavior of the organism. The same occurs with tioning something different from themselves, as in a car.
any other component of the domain of descriptions which even though AUTONOMY: the condition of subordinating all changes to the maintenance
they do not represent states of the nervous system they act, as any behavior, of the organization. Self-asserting capacity of living systems to maintain
as selectors of its path of structural change. This is so even with notions like their identity through the active compensation of deformations.
beauty, freedom and dignity which, as descriptions arise in the domain of AUTOPOIETIC MACHINE: a machine organized (defined as a unity) as a
behavior of the organism through distinctions referred to it as a result of the network of processes of production, transformation and destruction of
coupling of the phenomenology of the nervous system as a closed neuronal components that produces the components which: (i) through their inter-
network and the domain of interactions of the organism. actions and transformations regenerate and realize the network of processes
We have not given a formal description of the nervous system in the (relations) that produced them; and (ii) constitute it as a concrete unity in
language of anatomy or electrophysiology because our purpose was to disclose the space in which they exist by specifying the topological domain of its
the organization of the nervous system as a closed neuronal network, and realization as such a network.
the languages of neurophysiology and anatomy through their references to AUTOPOIETIC SPACE: an autopoietic organization constitutes a closed
function and input and output relations imply the notion of an open system. domain of relations specified only with respect to the autopoietic or-
The distinction is significant because the disclosure of the organization of the ganization that these relations constitute, and thus it defines a space in
nervous system as that of a closed neuronal network leads to a fundamental which it can be realized as a concrete system, a space whose dimensions
notion: are the relations of production of the components that realize it.
The correspondence that the observer sees between the conduct of the BIOLOGICAL EXPLANATION: a reformulation in terms of processes
organism and the environmental conditions with which this conduct appears subordinated to autopoiesis, that is, a reformulation in the biological
to cope, reveals the structural coupling of the organism (nervous system phenomenological domain.
included) to its ambience as this structural coupling is conserved through BIOLOGICAL PHENOMENON: the biological phenomenology is the phe-
philogenic and ontogenic selection. This correspondence, therefore, does not nomenology of autopoietic systems in the physical space and a phenomenon
reveal any particular feature or property of the connectivity of the nervous is a biological phenomenon only to the extent that it depends in one way
system that would permit it to operate with representations of the ambience or another on the autopoiesis of one or more physical autopoietic unities.
in its computation of the adequate conduct of the organism. CODING: A notion which represents the interactions of the observer, not a
phenomenon operative in the .observed domain. A mapping of a process
that occurs in the space of autopoiesis onto a process that occurs in the
space of human design (heteropoiesis) and, thus, not a reformulation of
the phenomenon.
135
136 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA GLOSSARY 137

COGNITIVE DOMAIN: the domain of all the interactions in which an relations between processes realized through the properties of components.
autopoietic system can enter without loss of identity. MECHANICISM: a biological outlook which asserts that the only factors
COMMUNICATIVE DOMAIN: a chain of interlocked interactions such that operating in the organization of living systems are physical factors, and
although the conduct of each organism in each interaction is internally that no non-material vital organizing force is necessary.
determined by its autopoietic organization, this conduct is for the other OBSERVER: a system that through recursive interactions with its own
organism a source of compensable deformations. linguistic states may always linguistically interact with its own states as if
COUPLING (OF UNITIES): whenever the conduct of two or more unities with representations of its interactions.
is such that the conduct of each one is a function of the conduct of the ONTOGENY: the history of the structural transformations of a unity.
others. ORGANIZATION: the relations that define a system as a unity, and deter-
DIVERSITY: variations in the mode in which identity is maintained. mine the dynamics of interaction and transformations which it may
EVOLUTION: history of change in the realization of an invariant organiza- undergo as such a unity, constitute the organization of the system.
tion embodied in independent unities sequentially generated through PHENOMENOLOGICAL DOMAIN: defined by the properties of the unity
reproductive steps, in which the particular structural realization of each or unities that constitute it, either singly or collectively through their
unity arises as a modification of the preceding one (or ones) which, thus, transformations or interactions. Thus whenever a unity is defined or a class
constitutes both its sequential and historical antecedent. of unities is established which can undergo transformations or interactions,
EXPLANATION: a reformulation of a phenomenon in such a way that its a phenomenological domain is defined.
elements appear operationally connected in its generation. PHYSICAL SPACE: the space within which the phenomenology of autopoiesis
FUNCTION: notion that arises in the description made by the observer of of living systems takes place.
the components of a machine or system in reference to an encompassing PURPOSE: the possession of an internal project or program represented and
entity, which may be the whole machine or part of it and whose states realized through the components of a unity.
constitute the goal that the changes in the components are to bring about. REGULATION: a notion valid in the domain of description ofheteropoiesis,
HETEROPOIESIS: the space of human design. that reflects the simultaneous observation and description made by the
HISTORICAL PHENOMENON: a process of change in which each state of designer ( or its equivalent) of interdependent transitions of the system
the successive states of a changing system arises as a modification of a that occur in a specified order and at specified speeds.
previous state in a causal transformation and not de nova as an independent RELATIONS OF CONSTITUTION: determine that the components pro-
occurrence. duced constitute the topology in which the autopoiesis is realized.
HOMEOSTATIC MACHINES: the condition of maintaining constant or RELATIONS OF ORDER: determine that the concatenation of the compo-
within a limited range of values some of their variables. nents in the relations of constitution, specification and order be the ones
INDIVIDUALITY: maintenance of identity by an autopoietic machine specified by the autopoiesis.
independently from its interactions with an observer. RELATIONS OF SPECIFICITY: determine that the components produced
LINGUISTIC DOMAIN: a consensual domain in which the coupled organisms be the specific ones defined by their participation in the autopoiesis.
orient each other in their internally determined behavior through inter- REPRODUCTION: any of the processes of replication, copying or self.
actions that have been specified during their coupled ontogenies. reproduction.
MACHINE: a unity in the physical space, defined by its organization, which SELECTION: a process of differential realization of a production of unities
connotes a non-animistic outlook, and whose dynamisms is apparent. in a context that specifies the unitary organization that can be realized. In
MACHINE, PURPOSE OR AIM OF: the use to which a machine can be put a population of autopoietic unities, selection is a process of differential
by man, sometimes its product. A descriptive device to reduce the task of realization of autopoiesis, and hence, of differential self-production.
conveying to a listener the organization of a particular machine. SELF-CONSCIOUSNESS: the domain of self-observation.
MECHANICAL PHENOMENOLOGY: the phenomenology generated by SELF-REPRODUCTION: when a unity produces another with a similar
138 HUMBERTO R. MATURANA AND FRANCISCO J. VARELA

organization to its own, through a process thatis coupled to the process

of its own specifications. Only autopoietic systems can self-reproduce.
BIBLIOGRAPHY
SPECIES: a population or collection of populations of reproductively
interconnected individuals which, thus, are nodes in a historical network.
STATICAL PHENOMENOLOGY: the phenomenology generated by the
relations between properties of components. Ashby, W. Ross [1960], Design for a Brain: The Origin of Adaptive Behavior (John
STRUCTURE: the actual relations which hold between the components Wiley and Sons, Inc., New York), Second Edition.
wllich integrate a concrete machine in a given space. Berkeley, George [1709] ,An Essay Towrds a New Theory of Vision.
SYSTEM: any definable set of components. Berkeley, George [1710], Treatise Concerning the Principles of Human Understanding.
TELEONOMY: the element of apparent purpose or possession of a project Bernal, J. D. [1965], 'Molecular Matrices for Living Systems', in The Origins of Pre-
biological Systems and of Their Molecular Matrices (Academic Press, New York and
in the organization of living systems, without implying any vitalistic London).
connotations. Frequently considered as a necessary if not sufficient von Bertalanffy, Ludwig [ 1960] , Problems of Life: An Evaluation of Modem Biological
definitory feature of the living organization. and Scientific Thought (Harper Torchbooks, New York), First Edition.
UNITY: that which is distinguishable from a background, the sole condition Chomsky, N. [1968], Language and Mind (Harcourt, Brace and World, New York).
necessary for existence in a given domain. The nature of a unity and the Commoner, B. [1965], 'Biochemical, Biological and Atmospheric Evolution', Proc. of
the National Academy of Science, 53, pp. 1183-1194.
domain in which the unity exists are specified by the process of its distinc- Gardner, R. A. and B. T. Gardner [1969], 'Teaching Sign Language to a Chimpanzee',
tion and determination; this is so regardless of whether this process is Science, 165, pp. 664-672.
conceptual or physical. Gazzaniga, M. S., J. E. Bogen, and R. W. Sperry [1965], 'Observations on Visual Per-
ception After Disconnection of the Cerebral Hemispheres in Man', Brain, 88, Part II,
pp. 221-236.
Geschwind, N. [1964], 'The Development of the Brain and the Evolution of Language',
Monograph Series on Languages and Linguistics, 17, C. I. J. M. Stuart (Georgetown
University Press, Washington), pp. 155-169;reprinted in Selected Papers on Language
and the Brain, Boston Studies in the Philosophy of Science XVI, Robert S. Cohen
and Marx W. Wartofsky, eds., (D. Reidel Publishing Company, Dordrecht and Boston,
1974), pp. 86-104.
Geschwind, N. [1965], 'Disconnexion Syndromes in Animals and Man', Brain, 88.
pp. 237-294 and 585-644; reprinted in Selected Papers on Language and the Brain,
Boston Studies in the Philosophy of Science XVI, Robert S. Cohen and Marx W.
Wartofsky, eds., (D. Reidel Publishing Company, Dordrecht and Boston, 1974),
pp. 105-236.
Gibson, J. J. [1950], The Perception of the Visual World (Allen and Unwin, London).
Held, R. and A. Hein [1963], 'Movement-Produced Stimulation in the Development of
the Visual Guided Behavior', J. Comp. and Phys. Psycho!. 56, No. 5., pp. 872-876.
Henderson, L. J. [1913], The Fitness oftlze Environment (Macmillan, New York).
Jay, P. C. (ed.) [1968], Primates; Studies in Adaptation and Variability (Holt, Rinehart
and Winston, New York).
Kilmer, W. L., W. S. MuCulloch and J. Blum [1968], 'Towards a Theory of Reticular
Formation' in The Mind: Biological Approaches to Its Function, W. C. Corning and
John Balaban, eds., (John Wiley and Sons, Inc., New York).
Lettvin, J. Y., H. R. Maturana, W. S. McCulloch and W. H. Pitts [1959], 'What the
Frog's Eye Tells the Frog's Brain', Proceedings of the IRE 47, No. 11, pp. 1940-
1959.

139
140 BIBLIOGRAPHY

Lilly, J. G. [19671, The Mind of the Dolphin (Doubleday and Company, New York).
Lindauer, M. [1967], Communication Among Social Bees (Harvard University Press,
Cambridge, Massachusetts).
Lorenz, K. (1966], Evolution and Modification of Behavior (Methuen and Co., Ltd., INDEX OF NAMES
London).
Maturana, H. R., J. Y. Lettvin, W. S. McCulloch and W. H. Pitts [1960], 'Anatomy and
Physiology of Vision in the Frog (Rana pipiens)', J. of Gen. Physiol. 43, No. 6 Part
2, pp. 129-175. Aquinas 63 Jay, P. C. 31, 139
Maturana, H. R. and S. Frenk [1963], 'Directional Movement and Horizontal Edge
Aristotle 63, 68, 74
Detectors in the Pigeon Retina', Science 142, pp. 977-979. Arm, D. L. 140 Kant, I. 63, 65, 66
Maturana H. R. (19651, 'Especificidad versus Ambiguedad en la Retina de los Verte- Ashby, W. R. 25, 139 Kilmer, W. L. 46, 139
brados', Biologica 36, pp. 69ff. Augustine 63
Maturana, H. R., G. Uribe and S. Frenk [1968], 'A Biological Theory of Relativistic Leibniz, G. W. 63
Color Coding in the Primate Retina', Arch. Biologia y Med. Exp., Suplemento No. 1 Balaban, J. 139 Lettvin, J. Y. vi, xii, xiv, 139, 140
(Santiago, Chile). Beer, S. 63 Lilly, J. G. 31, 140
Morrell, F. [1967], 'Electrical Signs of Sensory Coding', in 11ze Neurosciences, A Study Berkeley, G. 53, 63, 66, 139 Lindauer, M. 46, 140
Program, Quarton, ed., (The Rockefeller University Press, New York), pp. 452-468. Bernal, J. D. 53, 139 Linnaeus, C. 64
Segundo, J.P. and H. D. Ferkel [1969], 'The Nerve Cells as Analyzers of Spikes', in The Bertalanffy, L. v. 6, 139 Locke, J. 63
Jnterneuron, M.A. B. Brazier, ed., (University of California Press, Berkeley and Los Blum, J. 46, i39 Lorenz, K. 43, 140
Angeles). Bogen,J.E. 30,139
Varela, F. [19791, Principles of Biological Autonomy, (Elsevier-North Holland, New Brazie;;, M. A. B. 140 Maturana, H. v, vi, xiv, xv, 16, 21, 22,
York). Bulnes, J. xvii 72, 118, 139, 140
Wittgenstein, L. (1922], Tractatus Logico-Philosophicus (Routledge and Kegan Paul, McCulloch, W. S. vi, xiv, 46, 139, 140
London). Chomsky, N. 34, 139 Monod, J. 75
Young, J. Z. (1967], 'On the Organization of Living Memory Systems', in Journeys Cohen, R. S. vi, 139 Morrell, F. 16, 140
Jn Science: Small Steps Great Strides, D. L. Arm, ed., (The Twelfth A. F. D. S. R. Commoner, B. 9, 139
Science Seminar). Corning, W. C. 139 Perkel, H. D. 16, 140
Pitts, W. H. vi, xiv, 139, 140
Descartes, R. 63 Plato 63
'Diogenes 65
Quarton 140
Foerster, H. v. xvi
Frenk, S. xiv, xv, 16, 22, 140 Segundo,J.P. 16,140
Sperry, R. W. 30, 139
Galen 64 Spinoza, B. 63
Gardner, B. T. 31, 41, 139 Stone xv
Gardner, R. A. 31, 41, 139 Stuart. G. I. 139
Gazzaniga, M. S. 30, 139
Geschwind, N. 22, 32, 139 Uribe, G. xiv, xv, 22, 140
Gibson, J. J. 26, 139
Varela, F. J. v, ix, xvii, xxiv, 72, 118
Hegel, G. W. F. 65
Hein, A. 26, 139 Wartofsky, M. W. vi, 139
Held, R. 26, 139 Wittgenstein, L. 53, 140
Henderson, L. J. 73, 139
Hume, D. 63, 67 Young, J:Z. xiii, 6, 140

141
23. Marjorie Grene, The Understanding of Nature. Essays in the Philosophy of Biology, 5 8. Gerard Radnitzky and Gunnar Andersson (eds.), Progress and Rationality in Science.
1974. 1978.
24. Don Ihde, Technics and Praxis. A Philosophy of Technology. 1978. 59. Gerard Radnitzky and Gunnar Andersson (eds.), T11e Structure and Development
25. Jaakko Hintikka and Unto Remes, The Method of Analysis. Its Geometrical Origin of Science. 1979.
and its General Significance. 1974. 60. Thomas Nickles, Scientific Discovery: Case Studies. 1980.
26. John Emery Murdoch and Edith Dudley Sylla, The Cultural Context of Medieval 61.Maurice A. Finocchiaro, Galileo and the Art of Reasoning. 1980.
Learning. 1975.
27. Marjorie Grene and Everett Mendelsohn (eds.), Topics in the Philosophy of Biology.
1976.
28. Joseph Agassi, Science in Flux. 197 5.
29. Jerzy J. Wiatr (ed.), Polish Essays in the Methodology of the Social Sciences. 1979.
32. R. S. Cohen, C. A. Hooker, A. C. Michalos, and J. W. van Evra (eds.), PSA 1974:
Proceedings of the 19 74 Biennial Meeting of the Philosophy of Science Association.
1976.
33. Gerald Holton and William Blanpied (eds.), Science and Its Public: The Changing
Re/ationihip. 1976.
34. Mirko Drazen Grmek, Robert S. Cohen and Guido Cimino (eds.), 011 Scientific
Discovery. 1980.
35. Stefan Amsterdamski, Between Experience and Metaphysics. Philosophical Problems
of the Evolution of Science. 1975.
36. Mihailo Markovic and Gajo Petrovic (eds.), Praxis. Yugoslav Essays in the Philosophy
and Methodology of the Social Sciences. 1979.
37. Hermann von Helmholtz: Epistemological Writings. The Paul Hertz/Moritz Schlick
Centenary Edition of 1921 with Notes and Commentary by the Editors. (Newly r
translated by Malcolm F. Lowe. Edited, with an Introduction and Bibliography, t
by Robert S. Cohen and Yehuda Elkana.) 1977. II
38. R. M. Martin, Pragmatics, Truth, and Language. 1979. I

39. R. S. Cohen, P. K. F;_eyerabend, and M. W. Wartofsky (eds.), Essays in Memory

of Imre Lakatos. 1976.,
42. Humberto R. Matur;na and Francisco J. Varela, Autopoiesis and Cognition.
The Realization of the Living. 1979.
43. A. Kasher (ed.), Language in Focus: Foundations, Methods and Systems. Essays
I
Dedicated to Yehoshua Bar-Hille/. 1976.

I
46. Peter Leonidovich Kapitza, Experiment, T1zeory, Practice. 1980.
47. Maria Luisa Dalla Chiara, Italian Studies in the Philosophy of Science. 1980.
48. Marx W. Wartofsky, Models: Representation and the Scientific Understanding.
1979.
50. Yehuda Fried and Joseph Agassi, Paranoia: A Study in Diagnosis. 1976.
51. Kurt H. Wolff, Surrender and Catch: Experience and inquiry Today. 1976.
52. Karel Kosfk, Dialectics of the Concrete. 1976.
5 3. Nelson Goodman, The Structure of Appearance. (Third edition.) 1977.
54. Herbert A. Simon, Models ofDiscovery and Other Topics in the Methods of Science.
1977.
55. Morris Lazerowitz, The Language of Philosophy. Freud and Wittgenstein. 1977.
56. Thomas Nickles, Scientific Discovery, Logic, and Rationality. 1980.
51. Joseph Margolis, Persons and Minds. The Prospects of Nonreductive Materialism.
1977.