From the perspective of second order cybernetics this paper examines in which respects psychology can claim to be a science. It focuses on the limits of mechanistic description in the behavioral sciences. Through the Danish psychologist Iven Reventlow's works, the article analyzes the use of the Galilean concepts of law in psychology. Reventlow attempts to create basic methods and concepts for a Galilean (law determined) psychology in the tradition of Kurt Lewin through work with animal models in the tradition of ethology. His standard experimental model is the male Stickleback guarding its nest - a small fish in its partly self-created world. Reventlow's aim is to describe the "behavioral personality" of the organism keeping description and causal analysis and explanation on the behavioural level. To this end he works with a statistical model which do not hide the individuals characteristics by rolling them into an average. In this process, however, he finds that he cannot make a final separation of the organism and the environment. It is not possible to carry through either the mechanistic or the dualistic point of view. This finding is discussed in the light of von Foerster's and Maturana's second order cybernetic positions on the observer, observation, autopoiesis and the multiverse. The limitations of these theories carries the analysis further. A realistic, non-reductionistic and constructivistic viewpoint is developed from some of N. Luhmann's formulations.
The degree, to which animal and especially human behavior can be described and explained according to scientific methods, has always had a central position in the discussions of psychology's foundation. Is there a science of psychology in the Galilean tradition of objective science 1 which includes the phenomenological aspect (Lewin 1935)? Can this science be based on a materialistic realism and the Galilean idea of objective universal mathematical and deterministic laws? Or is it necessary to take a more constructivistic and organic view of reality which breaks with mechanicism?
These problems are central to the foundation of second order cybernetics especially in the works of von Foerster, Maturana and Varela. The discussion of these topics is continuing. The radical constructivistic positions of von Glasersfeld and Maturanas "bring-forth-ism" have been disputed by many, recently by Johnson (1993). In an earlier paper (Brier 1992) I discussed the limitations of the epistemology and ontology of Maturana's "bring-forth-ism" and the concept of "Multiversa" and suggested that C.S. Peirce's triadic semiotics offered an ontology and epistemology which made it possible to solve some of Maturana's problems. It is my opinion that the philosophical foundation of these fruitful theories of second order cybernetics needs further work. In this paper I will primarily be concerned with Maturana's and von Foerster's theories. I will try to shed light on the subject matter through the presentation and discussion of the work of the Danish psycho-biologist Iven Reventlow, which was done in the 1960s and 1970s. Although Reventlow was not a cybernetician, his work brought him deeply into the central problems of observing autopoietic systems (animals), which to some degree create their own environment.
For me a very important and fruitful feature of second order cybernetics is that it brings epistemology into a dialog with our biological knowledge in a new way (although Maturana won't call it philosophy). Von Foerster sums this up in the following way:
"While in the first quarter of this century physicists and cosmologists were forced to revise the basic notions that govern the natural sciences, in the last quarter of this century biologists will force a revision of the basic notions that govern science itself. After that "first revolution" it was clear that the classical concept of an "ultimate science", that is an objective description of the world in which there are no subjects (a "subjectless universe"), contains contradictions.
To remove these one had to account for an "observer" (that is at least for one subject): ...
After this we are now in the possession of the truism that a description (of the universe) implies one who describes (observes) it. What we need now is the description of the "describer" or, in other words, we need a theory of the observer. Since it is only living organisms which would qualify as being observers, it appears that this task falls to the biologist. But he himself is a living being, which means that in his theory he has not only to account for himself, but also for his writing this theory."
(von Foerster 1984, p. 258)
The Necessity of a Galilean Psychology
The Gestalt psychologist Kurt Lewin's article: "The conflict between Aristotelian and Galilean modes of thought in contemporary psychology" (Lewin 1935) is a very important basis for Reventlow's analysis of complex concepts in psychology. In this article Lewin writes that psychology "... classes together many things with very slight or unimportant relation and separates things that objectively are closely related."
It is well known that Aristotle believed it was only possible to work scientifically on the basis of constant or almost constant (frequent) recurrences. According to Aristotle laws were constituted - in an historical context - by frequent and regular occurrences within a geographical area defined by himself. In this way the "less frequent" and the "rare" become the accidental, and the laws for falling stones distinct from those for the harmony of the spheres. Phenomena were described by the "attributes" that were almost always present. The sum of these characteristics assigned the phenomenon to a "class". These characteristics were then ascribed to each member of the class as "attributes" - the essence of the class that determined its entire behavior. The other characteristic of the classification was that it often worked by pairs of absolute opposites: cold/hot, dry/wet, upward-seeking/downward-seeking. For instance the downward-seeking nature of stones and the upward-seeking nature of fire. Examples of Aristotelian laws in psychology are:
It is this sort of oversimplified classification that Reventlow attempts to break away from. In connection with the application of theoretical probability models in the analysis of behavior, he writes in his doctoral thesis (1970, p. 185):
"The question is ... whether it is possible that these models will enable psychology to make a leap forward comparable to that which occurred in physics when Galileo and Newton started looking for laws rather than just studying phenomena. In this way one law was enough to explain phenomena as diverse as an object's fall to the ground and the movements of the planets. ... It might be that the division of factors into groups such as stimuli, forces of motivation, or learning processes will prove to be far too primitive and will persist only because the effects of these factors have been studied in many experiments. - The results of which have, however, been used solely to explain the conditions of the particular experiment and to create "laws" from this, while they have only rarely been used to extract simple laws from the many experimental results."
This tendency is related to a point that Lewin has identified in Aristotle: a lack of conviction that the entire field under study is determined or governed by laws. The belief in the all embracing character of natural laws is now a cornerstone in the ontology of the exact sciences, but not in the humanities or the social sciences. It is precisely this problem that forms the leitmotif of Reventlow's work. In agreement with Köhler and other gestalt psychologists, Reventlow's attitude is to accept phenomenological data as adequate in the study of man but to avoid using intentional models and explanations in one's theories and models as long as possible.
Konrad Lorenz (1970-71) and Niko Tinbergen (1973) founded the modern European ethology and Reventlow was especially influenced by the latter. By incorporating a gestalt psychological perspective and a respect for statistics rare in an empiricist, he has used the Tinbergen tradition to develop approaches that have infused comparative psychology with new dimensions and depths.
Reventlow's objective is to examine, as precisely as possible the complex phenomena of "everyday life" in situations as natural and simple as possible in order to find meaningful "basic units" of behavior, which can be used as "fixed points" in the analysis of more complex psychobiological phenomena such as human personality. Reventlow's training actually focused on the psychology of perception, but his desire to achieve a greater exactitude and objectivity in psychology has led him to psychobiological behavioral research. This makes his approach interesting for second order cybernetics.
Reventlow's analyses present themselves not only in the purely theoretical area but also in the continuous, almost painfully epistemological deliberations on the behavioral analysis and the construction of a statistical model of motivation on experiments with Sticklebacks, which is found in his thesis from 1970, bearing the symptomatic title: "Studier af komplicerede psykobiologiske fænomener" ("Studies of complicated psychobiological phenomena"). This work has unfortunately not yet been translated into English, but the model is described in Madsen (1974). Other parts of his work and thoughts can be found in Reventlow (1961, 1970, 1972, 1973).
Extending previous attempts to elaborate Reventlow's work (Brier 1980) I will attempt to discover a common foundation for a non-reductionist ethology and phenomenological psychology based on Kurt Lewin's Galilean psychology. I will then attempt to elucidate and elaborate the main theme in Reventlow's pursuit of fundamental functional concepts of psychology that can be used to reveal actual psychological laws. Finally, I will discuss the problems and limits of this paradigm in the view of second order cybernetics.
It appears to me that even in the most stringent presentation this biopsychological
paradigm compels asking some elementary questions: What status do the functional
concepts and laws possess? Are the laws objective, eternal and absolute,
as often claimed in physics? Must one set aside concepts such as free will,
intentionality and attention in order to pursue Galilean psychology? Is
this "mechanization" the connecting link between human and animal
"psyche" - as seems to be the belief in much of cognitive science
(see fx. Gardner 1985). Must psychology thus be coupled with a materialistic
ontology, which considers the use of mental concepts such as intentionality,
a convenient way of talking about phenomena, which in the future will be
reduced to something that can be duplicated by a computer? Or is it necessary
to shift to a phenomenalistic constructivism in a cybernetic theory of
perception, communication and cognition as Maturana and von Foerster do?
My suggestion is a view which is more realistic than much second order
cybernetics. But this realism is not based on physics but on a more complex
and organic worldview.
Reventlows discovery of the limitations to a Galilean Psychology
Reventlow's final choice of experimental animal and situation was male sticklebacks tending their nests. A wild species with its own natural microcosm in the laboratory, which could be studied without disturbing its natural behavior. Still the behavior was simple enough to be modelled in a statistical model. The male constructs, guards and repairs its own nest and waits for a female to lay its eggs in it.
The interesting aspect from a cybernetic point of view is that he insists on studying the individual autopoietic systems behavioral choices in its historical drift within its domain of living. It is obvious that insights into the learning processes as such are extremely important in understanding the emergence of the individual's "behavioral personality". But in all probability it is at first most expedient to work with behavior types in which the learning process does not cast too much of an individual-historical veil upon the motivational structure of behavior. This is significant seen in the light of our general ignorance regarding basic structures of behavior which must somehow determine what is learned. It is well known that we learn only what we are motivated to learn - or need to learn as a matter of survival. Konrad Lorenz (1973) has also argued in favour of the existence of many more types of learning processes than the behaviorists normally work with. He has argued that the characteristics of these types are to a high degree determined by the nature of the basic structures which are their starting points.
If one wishes to include the organism's "behavioral personality" systematically in one's models of the functional organisation of behavior, then learning processes exclude the execution of reproducible experiments with the same organism. A further important methodological point is that I/we roll the results of all the experiments done on different individuals into an average, the individual character's importance to the process disappears, and at the same time the uncertainty as to the value of the results as "general" laws increases. It was for this reason, among others, that Reventlow instead thought it important to start by analysing the functional organisation of instinctive reactions.
Ethology is a biological science that observes animal behavior in nature. It interprets and attempts to explain behavior, i. e. describes the causality underlying it. One method is experimentation with behavior, both in nature and in the laboratory. On the basis of these observations and experiments, and based on a theory which sees behavior as a fairly stable, inherited part of an animal's "survival machinery", one seeks to draw up models of the inner organisation and control of behavior.
The reaction model arrived at for the inner organisation of behavior (a motivation model) is by Reventlow (1970) described in the following way:
"Looked at as a psychological theory, ethology has the great advantage of offering a comprehensive theory which intergrates in a simple fashion the stimulation situation, the inner motivation and the outer observable behavior. The theory can be graphically illustrated as in fig. 1, which visualises the genetically determined connection ("innate releasing mechanism" - usually abbreviated IRM) - which ethologists believe exists between perception and behavior. It is characteristic for this inherited connection, IRM, that only certain parts of the stimulus situation normally found in nature are necessary to release the behavior. These parts are called "sign stimuli" or "Schlüsselreize". Sign stimuli give rise to sense perception which in conjunction with internally motivated conditions works in such a way that the genetically determined behavior is released (...) Ethology sums up in a simple concept the complex interaction between the environment, the individual and behavior.
At times several IRM's can be arranged in a common system as Tinbergen
(1951), for instance, has pointed out in the case of the male stickleback."
(Reventlow 1970, p. 21).
Figure 1: Illustrates ethology's conception of the release mechanism. 1 denotes the outer world, elements of which can be stimuli. 2a the senses, 2 b the perceptive part of the nervous system which conveys information from the outside world. 3 the motivating and coordinating part of the nervous system. 4a the motor center of the nervous system and 4b the movements of the muscles. 5 the total external behavior. 2 + 3 are the release mechanism. 4 the effector section which comprises one or more fixed movements possibly with learned modifications. 2+3+4 shows the congenital nerve connection which in conjunction with certain motivating conditions produces certain effects - without prior learning - resulting in the release of the congenital behavioral dispositions. The difference in length between 1 and 2a illustrates that organisms can never register the entire physical reality, and that precisely the concept of sign stimulus emphasizes the selection by each species of a few but well defined stimuli in complex phenomena.
Reventlow's doctoral thesis (1970) is an attempt to undertake ethological motivation research based on purely behavioral conditions without considering either the physiological or the phenomenological level. The outcome is a probabilistic model of the tendencies of behavior of the male stickleback at its nest:
In this behavioral probabilistic model, P expresses the probability within a certain timespan (t) that a male stickleback will contact its nest or - if it is at the nest - will leave it. A is a constant for all male sticklebacks in all situations at the nest, but w is a parameter characterising the individual in connection with only one sort of experimental manipulation of the nest, i.e., it changes with every new experimental situation.
In this model of the behavioral tendencies of the male stickleback at its nest, Reventlow finishes his structuring of the model with a parameter (A), common to all fish in all situations, here in a simple experiment (e.g. uprooted nest) and lastly a parameter typical of the individual fish in the individual experiment (w)! That is, a type of individual-environmental factor combination (se Madsen (1974) for further details).
Seen from the viewpoint of mechanistic behavioral ontology it should, of course, be possible to split up such a factor into an environment parameter and an individual parameter. (Which says something about individuality and the "inner" motivation-physiological parameters.) But - and this is one of the crucial points in Reventlow's thesis - he is forced by his results to conclude that in the nature of things such a split is not possible.
More penetrating studies would perhaps make it possible to "peel" a number of parametric layers from the individual-environment factor, but never to split the factor! As Reventlow (1970) writes (my translation):
"This is perhaps a necessary cognitive consequence of the fact that we are not able to observe an individual without observing its behavior in a specific environment. In as much as psychology has not yet discovered its ultimate concepts nor a way of measuring them, this seems very important."
To a certain extent even the lowest organisms seem to define their surroundings (Umwelt) in their own ways. There is no final description of the "true" world in itself. As for example Maturana and Varela (1980) point out, the organism "brings forth" a world. But according to Reventlow only to a certain extend. He is not an idealist.
Reventlow seems aware that he has reached a final limit in the Galilean (mechanistic) way of theorising about psychological phenomena in behavioral analysis. This calls for an analysis of the fundamental epistemological and ontological ideas.
The late Danish philosopher of science Johannes Witt-Hansen stresses in his latest book (1980) the "situations of powerlessness" arising in science, the "postulates of impotence" they foster, and he cites a number of examples in mathematics and physics. A classic example is the powerlessness of the Pythagoreans, faced with the realization that the side and the diagonal of a square are incommensurable, because it is impossible to measure the length of the diagonal of a square in units of the side, no matter how small these units are made. Today this relation is called the square root of two. The Pythagoreans were forced to give up the claim of their philosophy that the universe could be described as governed by harmonic relations between natural numbers. After a delay of two thousand years, this "powerlessness" was resolved by a fundamental extension of the concept of numbers to include irrational numbers. In physics, the realization of the impossibility of constructing a perpetual motion machine has had great importance to the development of thermodynamical concepts of energy and entropy. In quantum mechanics, the recognition of our "powerlessness" to measure simultaneously a subatomic particle's position and momentum with an arbitrary precision led to Heisenberg's uncertainty principle and to Bohr's theory of complementarity. With such examples, Witt-Hansen emphasises the fact that situations rooted in powerlessness almost invariably lead to scientific revolution.
In the light of Descartes' philosophy, the works of Galileo and Newton led to an understanding of nature as governed by mathematical-mechanical laws. These applied likewise to animals, which were regarded as insensitive machines. According to Descartes, it was only the human body that was controlled by an immortal soul. Further the soul could not be described through the mechanical-mathematical sciences.
The dualism of Descartes was replaced by a pure materialistic-mechanistic world view by Laplace. The Galilean science and concept of law was considered omnipotent. Galileo's belief, that nature was governed entirely by a number of mathematical laws, was the basis of his science. Nature could be likened to a great book written in the language of mathematics and in Laplace's philosophy this concept of nature includes the animal and human mind. This belief is apparently shared by Reventlow and Lewin. Reventlow has in fact gone so far as to declare that he does not consider it possible to believe in free will and, at the same time, work honestly as a scientific (Galilean) psychologist. This also goes for those who think that we only have a free "won't".
On the other hand, in our discussions on these topics Reventlow has admitted that from an ethological point of view even a stickleback must have intentionality and experiences (Lorenz 1973 also discusses this crucial point). The reaction is not mechanical. The "stimuli" often has to be presented several times before a stickleback "reacts". Further it has to be in a certain motivational state. This implies that motivation is not a simple physiological concept e.g. Hinde (1970). It cannot be explained on the physiological level although it has physiological aspects (see Brier 1992 for a further analysis).
Thus there appears to be a fundamental antithesis between the Galilean concept of law and the concept of intentionality in Reventlow's paradigm. Somehow an epistemology and an ontology in which both law, intentionality and free will can exist side by side has to be described. This can probably not be done without changing the concepts of law, reality, cognition and mind from that of a mechanistic philosophy be it monistic or dualistic. But even a physicalistic evolutionary theory seems to be insufficient (see Brier 1992 for further details).
Reventlow's individual-environment factor leaves him faced with a situation which spells powerlessness for a Galilean - Cartesian paradigm of science in psychology. The reason is that we are here dealing with a factor that for fundamental epistemological reasons - derived from the measuring problem (just as the case is with quantum mechanics) - refuses to be broken down to allow a final description of causal relations between separate elements.
Reventlow sums up his methodological problems in the model shown in figure 2 below, which again can be viewed as an elaboration on some of Maturana's work (Maturana 1983, 1988, Maturana & Varela 1980).
In his further analysis of the problem of the scientific observer Reventlow points out that the main problem is not only how to determine the relation between 1 (exterior world) and 5 (exterior behavior), but also to determine the relation between 2a (the animal's perception of the world) and 6a (the observer's perception of the world), and the relation of this relation to the relation between 5 and 6b (observer's perception of the behavior of the animal).
One of the problems is that we do not have any final knowledge about 1 and 2a. According to Reventlow's methodological results we will never have such a final knowledge in causal deterministic terms. A certain kind of creative construction seems to be going on within certain limits.
Figure 2: Figure 2 illustrates the conditions for observation of an individual organism. 1 represents the exterior world, which stimulates/perturbates both the observed individual as well as the observer. 2, 3 and 4 represent mechanisms in the observed individual (compare with fig. 1), which cause that 5, which represents the total exterior behavior, is brought forth. 6a and b represent the sense organs of the observer, and 6 c the other perceptual parts of the nervous system and what further determines 7, which is his experience of the observed behavior. 8 represents the description of the observations which the observer gives, and which becomes the scientific datum, that is the foundation for the further scientific analysis (...)
When 6a is not situated symmetrical with 2a it is because that some times animals are most likely to react on stimuli, whose physical properties we do not know (...), while we (e.g. through physical measurement apparatus) can get knowledge about appearences of the physical world, that are without significance for the perceptions of animals. In the same way 6b is smaller than 4b and 5, because the animals have behaviors which we do not know, and even some that we cannot perceive or measure yet." (Reventlow 1970, p. 32)
This result reflects back on the relation between 1 and 6a. Within certain limits we also construct what we see. We know that this phenomenon is partly built into our perceptual systems and partly learned through childhood and part of it is caused by scientific training (paradigms and so forth). With great (social) effort we can become aware of and control a part of it. But so far only the mystical tradition promises to clean our senses so completely that we will see the world as it is.
From a philosophical point of view these basic considerations on the border between science and epistemology seem to set limits for objective scientific knowledge. These limits are difficult to determine precisely and objectively. They seem to be like the fractality of a coastline. Seen from far away it is a line, but the nearer you get the more complicated it appears. It was after many years of brabling with these problems that I through Batesons work was guided into the fundamental change of view called second order cybernetics. I now turn to the role second order cybernetics has to play in dealing with these problems and deepening our understanding.
The cybernetic turn
Reventlow's analysis of cause brings us into the heart of von Foerster's and Maturana's analysis of "observing systems". Their attempts to create a new framing of the problem led to the modern version of second order cybernetics, the theory of autopoiesis, of non-trivial machines and observing systems.
Von Foerster (1984) points out that the organism can be modeled as a machine and yet not be trivial (i.e. there is no deterministic mathematical description of its behavior). With their concept of autopoiesis Maturana and Varela (1980) show one of the reasons why. The system organizes itself and produces its own parts. The self-organizing ability and the historical dimension of the living are important reasons why organisms are not trivial machines. They are closed self-referencing systems. But this actually only makes the whole problem more difficult. If information is not transfered from the environment to a scientifically describable system, which kind of dynamics is then going on?
Von Foerster and Maturana answer the question on information and dynamics as follows. The organism reacts to disturbances/perturbation in its system by means of a self-referential dynamics (so as to conserve the sort of system it seeks to be). The word "outside" is not used, because according to these theories the concept "outside" or (objective) reality has no significant objective meaning. As von Foerster expresses it:
"... I see the notion of an observer-independant "Out There", of "The Reality" fading away very much..."
(von Foerster 1984, Preface)
Von Foerster (1979) recognized the necessity of admitting the reality of the observer, at least one more observer to communicate with through signs to establish a world of language and of society to be able to produce knowledge. In the central quotation below he develops his point of view from that of Maturana. But there is no doubt in my mind that Maturana holds similar views:
"Here is Maturana's proposition, which I shall now baptize "Humberto Maturana's Theorem Number One":
'Anything said is said by an observer.'
. . .
I would like to add to Maturana's Theorem a corollary which, in all modesty, I shall call "Heinz von Foerster's Collary Number One":
'Anything said is said to an observer.'
With these two propositions a nontrivial connection between three concepts has been established. First, that of an observer who is characterized by being able to make descriptions. This is because of Theorem 1. Of course, what an observer says is a description. The second concept is that of language. Theorem 1 and Corollary 1 connect two observers through language. But, in turn, by this connection we have established the third concept I wish to consider this evening, namely that of society: the two observers constitute the elementary nucleus for a society. Let me repeat the three concepts that are in a triadic fashion connected to each other. They are: first, the observers; second, the language they use; and third, the society they form by the use of their language. This interrelationship can be compared, perhaps, with the interrelationship between the chicken, and the egg, and the rooster. You cannot say who was first and you cannot say who was last. You need all three in order to have all three. In order to appreciate what I am going to say it might be advantageous to keep this closed triadic relation in mind."
(von Foerster 1979, p. 5-6)
But we are not in a position a priori to claim that reality is material and entirely determined by non-teleological laws, because it will also be of the mind, spontaneous and teleological, and we can never claim to be totally separated from it.
Maturana says he refuses to do philosophy and metaphysics. He is a scientist dealing with epistemological problems from a biological point of view. His conclusion is that the world is as we "see" it in our praxis of living. It is a Multiverse. He never really deals with the problem of the resistance of reality (as it is often called in social constructivism). Why can't we choose to see the world as we like? He might answer that it is because the world would not be consistant with our existence as living beings - our biology. But is our biology an objective independant fact? Can't we choose to see it another way if this point of view is consistent?
Although the theories and concepts of von Foerster and Maturana have led us to a much better grasp of this basic situation of observing and cognition, they seem in their radicality to have removed too much, when they even skip the Kantian "Ding and sich". The problem seems to be that they have attempted to find a final scientific solution to a basic philosophical problem.
Reventlow - a psychologist interested in biology but educated with phenomenological methods and open to epistemological reflections - stops in front of "the Gordien knot " of "information" and the "phenomenon of perception" and "cognition" which are so vital to cognitive science and AI. His response is to point to the limitations of science - an obvious problem for most psychologists - and towards some basic conceptual problems with the words "free will" "determinism", "probability" and "chance".
On the other hand both von Foerster's second order cybernetics and Maturana's "bring-forth-ism" seem to be right in bringing to our attention the creative processes in perception and cognition. As also Reventlow's analysis shows, something unforeseen or even unexpected is brought forth in perception and action.
As I have attempted to show in Brier (1992) one cannot solve the problem of mind and intentionality in an evolutionary philosophy either through mechanical materialism or through physical indeterminism. Neither do I believe that it can be done through pure phenomenalistic idealism or subjective constructivism or mentalism which underestimates the importance of the relative stability of the "outside" world for the possibility of knowledge, communication and truth (see Brier 1993 for further details).
So far this situation has been mostly accepted in man through the phenomenological philosophy of Heidegger. The concept of "dasein" which underlines the "thrownness" of man in the world. In the discussion of differences and similarities in cognition and problem solving in people and computers, Heidegger's concepts have been used by the Dreyfuss brothers (1985) and Winograd & Flores (1988). They use it to show that a person's relationship to the world is fundamentally different from that of the digital computer. Winograd & Flores use Maturanas theory of autopoiesis and the closure of the nervous system to show that this basic condition is common to both people and animals. The basic situation towards the environment is not objective and separated. The "domain of living" - which is a basic concept of Maturana - is rather an integrated part of the structure of the system before any cognitive separation between self and non-self. This seems to be a biological argumentation for Heideggers thrownness.
This epistemological foundation of second order cybernetics connects it with important points in Heidegger's phenomenology. The important point from Heidegger is that as an observer we are always already a part of the world when we start to describe it. When we start to describe it we to a certain degree, separate ourselves from the wholeness of the world of our living praxis. As Qvortrup (1993) through Luhmann and Kierkegaard points out both the psychic and social systems are closed systems (but in my view only to a certain extent). This is the prerequisite for their function and their limitation for getting knowledge. I think figure 2 demonstrates this very clearly.
Niklas Luhmann (1990 p. 3) sums up nicely how cybernetics and the concept of autopoiesis in Maturana's definition gives the new way of looking at things, which seems to give a solution to Reventlow's problem of the w factor, and, at the same time, he holds on to a sophisticated realism which I judge to be fruitfull for second order cybernetics:
"To use ipsissima verba: autopoietic systems "are systems that are defined as unities as networks of productions of components that recursively, through their interactions, generate and realize the network that produces them and constitute, in the space in which they exist, the boundaries of the network as components that participate in the realization of the network." 2 Autopoietic systems, then are not only self-organizing systems, they not only produce and eventually change their own structures; their self-reference applies to the production of other components as well. This is the decisive conceptual innovation. It adds a turbocharger to the already powerful engine of self-referential machines. Even elements, that is, last components (in-dividuals) which are, at least for the system itself, undecomposable, are produced by the system itself. Thus, everything that is used as a unit by the system is produced as a unit by the system itself. This applies to elements, processes, boundaries, and other structures and, last but not least, to the unity of the system itself. Autopoietic systems, then, are sovereign with respect to the constitution of identities and differences. They, of course, do not create a material world of their own. They presuppose other levels of reality, as for example human life presupposes the small span of temperature in which water is liquid. But whatever they use as identities and as differences is of their own making. In other words, they cannot import identities and differences from the outer world; these are forms about which they have to decide themselves."
(Luhmann 1990 p. 3)
We then need a more sophisticated theory of how these identities and differencies are developed, than the usual materialistic mechanism or the eliminativ materialistic theories and even the functionalistic theories of mind.
This model is not sufficient when we want to understand the perception of objects and the communication of knowledge. The precondition for the function of signs in any language is the existence of "something", which is referred to, which is an item used in interpreting language, and which other language users use in society to communicate. C.S. Peirce speaks in his triadic semiotics of: 1. The primary sign (the representamen), 2. The aspect of interest of the domain of living (the object) and 3. The interpretant, usually based in a society of languaging observers (Buchler 1955).
Von Foerster (1980) uses the mathematical idea of "eigenvalues" (also used in quantum physics) to give a model of how objects ("of reaction") are manifested in a living sensing autopoietic system. Eigenvalues are all those values of a function which, when operated on, produce themselves. It is a kind of circular causality. From an example with the function square roots he concludes:
"Thus, independent of primary values, the values for A and B created by applying recursively the square root converge to the single equilibrial value 1, the eigenvalue of the 'square root'.
The eigenvalue 'test' states that an operation applied to its eigenvalue must yield this eigenvalue. And indeed:
This insight (or 'solution') helps us understand the organism which recursively readjusts its behaviour (operates on its motor activity) in accordance with limiting 'objections' until a stable behavior is obtained.
An observer watching this entire process, who has no access to the organism's sensations of the constraints on its movements, will report that the organism has learned to manipulate a particular object successfully. The organism itself may believe that it now understands (or has mastered the manipulation of) this object. However, since through its nervous activity, it has knowledge of its behaviour only, strictly speaking these 'objects' are tokens for the organism's various 'eigenbehavior'. This suggests that objects are not primary entities, but subject-dependent skills which must be learned and hence may even be altered by the cultural context as well. A similar train of thought has been developed by Jean Piaget and his colleagues in Switzerland, who have painstakingly studied the emergence of equilibrial behavior in infants which is indicative of an infant's grasp - note the reference to motor activity - of 'object constancy'."
(von Foerster 1980, p. 23 & 26)
These eigenvalues and objects are what Maturana and Luhmann call structural couplings between the environment and the autopoietic system. Some of these seems to be what the ethologist calls "Sign Stimuli".
To go deeper into an understanding of the process one must analyse the whole process of signmaking, as C.S. Peirce does in his semiotics, and discuss the functionality of meaning which is an important aspect of Luhmann's theories. The problem in my view has been that the ideas of "closeness" of the living and psychic systems has been pushed to far making it difficult to understand what perception and communication are about. My suggestion will be that if we operate with a "partial closure" it will be possible to unite second order cybernetics with the semiotics of Peirce and his idea of "semiotic realism". I have addressed some aspects of it in Brier (1993) but I will have to return to these and other aspects in other papers. My concern here has mainly been the function of "outside reality" in analysis of the behaviors of autopoietic or "observing systems". My point has been that although one rightly has given up notion of "objective reality" in second order cybernetics, one should not give up the notion of a partly independant "outside reality". There is something lacking in the phenomenalistic or idealistic constructivist position which is not corrected by repeatedly referring to "experienced reality". We cannot avoid ontological considerations, but they must, of course, be under a constant development through our critical epistemological discussions and analysis. What we need to develop is a more refined and complex understanding of the role of the concepts of realities to our understanding of our own process of knowing.
Since we cannot, anyhow, avoid speaking of the nature of aspects of reality as a prerequisite for our various scientific paradigms, I would suggest it to be more fruitful to regard it not just as complex, but hyper-complex. By this I mean that reality both in its entirety and its local manifestations cannot be reduced to something simple, deterministic or random, material or spiritual, that can be contained in a linguistic or mathematical formulation. Furthermore, the spontaneous, intentional, anticipatory mind is an irreducible part of that same reality. So we will never be able to make a complete separation of subject and object, neither for our own science nor for the intentional systems we study. Here I refer both to the analysis illustrated by figur 2 and to the analysis in Brier (1992 p. 84-91 and 1993).
For at least 200 years science has recognized that living beings are an intrinsic part of the physical and chemical reality. For more than 100 years it has been recognized that we are an intrinsic part of the biological aspect of reality. Where the physical and chemical aspects have been considered basic for the univers it is only within the last 30 years it is realized how deeply connected to the whole development of the universe our biological aspect is. My point is that we are now on the brink of discovering how the mind aspect penetrates to what we consider the basic levels of our reality as Bateson, Bohm and Peirce have pointed out.
The deep but hypercomplex connection between the observer and its reality also provides an answer to those thoughts about the nature of probability that characterize large sections of Reventlow's dissertation. Because of reality's hyper-complexity no matter what we do in all measurement there will always be "noise", which will affect our results in an unpredictable way, because we always in a somewhat arbitrary way make a "cut" between ourselves and the observed system and between the observed system and its "environment", as we defined it through our own experiences and in our attempts to explain the "reactions" of the observed system(s).
Galilean science has dominated us for over 300 years. It has shown us that reality has aspects amenable to exact mathematical analysis. This has been an enormously productive insight. We must admit that even mind has its "sluggish" sides, especially when reflected in a primitive nervous system, that may partly be described in terms of functional laws. However, this does not mean that the content of all behavior and language could be transferred to computers, as some eliminative materialists and functionalists seem to believe. There is a hyper-complex "background problem" of individual and historical origin. In both physics and psychology (especially the latter) that which can be describ-ed formally has its background in that which is not thus describable: the hyper-complex pheno-mena, which besides the predictable, the regular, comprise among other things the spontaneous, unpredictable (chaotic), and intentional.
In an evolutionary philosophy - which does not deny that reality can
possess "deep" but formal-ly indescribable absolute features
- we may see the development of even more complex and selectively unstable
individual-environment systems far from equilibrium: Maturana and Varela's
autopoietic system is one example of nature's ability to reflect in ever
increasing degrees the spontaneous, unpredictable, and intentional sides
of reality through the ability of self-reflection. This ability finally
allows these systems to be centers of their own and to draw a line between
themselves as systems and the environment. Through the use of language
in society (a system of observing systems) systems can finally represent
themselves socially and by such means establish an individual conscious
point of view with which to reflect on knowledge, existences and meaning.
Thanks to Hugh Gash, St. Patrick's College of Education, Dublin, Ireland
and Axel Randrup, Center for Interdisciplinary Research, Svogerslev, Denmark
and Heinz von Foerster, Pescadero, California, USA for their fruitful critique
of content and language of an earlier version of the present paper. A special
thanks to Jørgen Brier for language revision.
Brier, S. (1980): Der ønskes analyseret (evt. v.h.a. egne undersøgelser), om hierarki- og sandsynlighedsbetragtninger i beskrivelsen af adfærd kan anvendes i - og udbygge - een eller flere motivationspsykologiske teorier eller modeller. (Prize Essay in psychology about the fruitfullness of hierarchy - and probability - deliberations in constructing models of motivation from behavioral analysis. Awarded with the Gold Medal of Copenhagen University).
Brier, S. (1992): "Information and consciousness: A critique of the mechanistic concept of information", in Vol.1, no. 2/3 pp. 71-94 of "Cybernetics & Human Knowing". Aalborg, Denmark.
Brier, S. (1993): "Cyber-Semiotics: Second-order cybernetics and the semiotics of C.S. Peirce." Proceedings from the Second European Congress on Systemic Science, Volume II, pp. 427-436.
Buchler, (ed)(1955): Philosophical writings of Peirce, Dover Publication, inc., New York.
Dreyfuss, H.L. and Dreyfuss, S.E.(1985): Mind over Machine. Macmillan, New York.
Foerster, Heinz von (1979): "Cybernetics of Cybernetics" in Krippendorff (1979), pp. 5-7.
Foerster, Heinz von (1980): "Epistemology of Communication" in Woodward (1980), pp. 19-27.
Foerster, H. von (1984): Observing systems, Intersystems Publication, California, USA.
Gardner, H. (1985): The Mind's New Science, Basic Books, New York.
Hinde, R. (1970): Animal Behaviour: A synthesis of Ethology and Comparative behaviour, McGraw-Hill, Tokyo (International Student Edition).
Johnson, D.K. (1993): "The Metaphysics of Constructivism" Cybernetics & Human Knowing, vol. 1, no. 4, pp. 27-41, Aalborg, Denmark.
Krippendorff, K. (ed)(1979) : Communication and control in society, Gordon and Breach Science Publishers, New York.
Lewin, K. (1935): Dynamic theory of personality. Mcgraw-Hill, N.Y.& London.
Lorenz, K. (1970-71): Studies in animal and human behaviour I and II. Cambridge, Mass. Harvard Univ. Press, USA.
Lorenz, K. (1973): Die Rückseite des Spiegels. Pieper and Co.
Luhmann, N. (1990): Essays on self-reference, Columbia University Press, New York.
Madsen, K.B (1974): Modern theories of motivation, Copenhagen: Munksgaard
Maturana, H. R. (1983): "What is it to see?" Arch. Biol. Med. Exp. 16, pp. 255-269.
Maturana, H.R. (1988): "Ontology of observing: The Biological Foundation of Self Consciousness and the Physical Domain of Existence" The Irish Journal of Psychology, Vol. 9, no. 1, pp. 25-82.
Maturana, H.R. & Varela, F.J. (1980): Autopoiesis and Cognition: The Realization of the Living, Boston, Reidel.
Qvortrup, L. (1993): "The Controversy of the Concept of Information: An overview and a Selected and Annotated Bibliography". Cybernetics & Human Knowing, Vol. I, no. 4. Aalborg.
Reventlow. I (1961): "Ethopsychopharmacological research in Denmark", Bulletin de L'association Internationale de Psychologie Applique, vol. 10, 1961, pp. 118-125.
Reventlow, I. (1970): Studier af komplicerede psykobiologiske fænomener; Munksgaard, Copenhagen. Doctoral thesis. University of Copenhagen.
Reventlow, I. (1972): "Symbols and Sign Stimuli", Danish Medical Bulletin, vol. 19, pp. 325f
Reventlow, I. (1973): "Konfliktforschung im Tierexperiment" im Reinert, G. (Ed)(1973): Bericht über den 27. Kongress der Deutschen Gesellschaft für Psychologie in Kiel 1970. Göttingen: Hogrefe.
Tinbergen, N. (1973): The animal in its world. George Allan & Unwin, London.
Winograd, T. & Flores, F. (1986): Understanding Computers and Cognition. Ablex Publishing Corporation, New Jersey.
Witt-Hansen, J. (1980): Filosofi: Videnskabernes historie; i det 20. århundrede, Gyldendal, Copenhagen.
Woodward, K. (ed)(1980): The Myths of Information: Technology and Postindustrial Culture. Routledge & Kegan Paul. London.
Return to the content of this issue
Return to the Cybernetics and Human Knowing Homepage