Professor Luhmann, in the following I will pose a series of questions on central concepts in your systems theory. I would like to start with the basic idea: Why systems theory?
You have to begin with a good distinction to elaborate a theoretical framework. At the moment at least, the distinction system/environment is, in my opinion, the best one to start with. It is not the only possible. But there is so much literature about this, and so many developments within systems theory, that, as I see it, the best way to integrate scientific discussions is to use the distinction system/environment as a starting difference - and not for example being/not being, or true/false which lead to another kind of methodological approach.
Why universalism?
It is at least one specialty among others to have an interest in universal theories. To be universal is not to be the only correct one, but to claim to be able to explain everything, or to be able to comment or describe whatever happens in the world or in the society.
Why functionalism?
Functionalism in the traditional sense is another term for comparative interest. So, you can compare in a purely quantitative sense, finding that two is less than three and 2 million is more than 1000. For the more qualitative aspects of scientific descriptions, functionalism is to define a problem and to see what kind of solutions you can compare. As I see it, modern society cannot be described starting from one principal idea of humanity or justice or integration or consensus or reason. You have to have a framework which can describe very different things like religion and economy, or politics and family, with the same set of concepts, so that you have a comparative capacity. This makes it possible to describe the unity of modern society by the comparability of very different functional systems. This can best be done with functional methodology.
Why constructivism?
Constructivism is the consequence of some theoretical positions which focus on operational closure. This means that a system can only work within the system and not outside. A system can never operate in its environment. If you have the intention of being consistent in your approach, and if you have a universal interest, then cognition can be made or constructed only within the system. There is no way operationally to cross the boundary of the system. In cognitive science, this idea comes from brain research, as the brain is an operationally closed system. So, if we have to use our brains to make science, how can we get into the environment? So this is not any longer the transcendental and not even the linguistic discussion.
Why operationalism?
The development of systems theory has shown an increased interest in self-referential figures. The first discussion took place in the fifties and the sixties and focused on self-organization. But this term refers only to the structure, as the structure is the product of the system and cannot be imported from the environment. With the term "autopoiesis" and with biological or neuro-physiological research, there was a stronger emphasis on the type of operation. The idea is that an autonomous or operationally closed system reproduces its own operations by the network of its own operations. This has very important consequences for all kinds of theories, partly cognitive constructivism but also evolutionary theory, as you cannot adapt to your environment if you cannot check what is outside. Also the question of time becomes important. The idea of operational autonomy or operational closure has revolutionary effects on other traditional fields of social science.
Why not rationality?
It is partly the consequence of the decay of traditional notions of rationality. Also the Weberian approach to bureaucracy as a command/obey framework has been more or less openly criticized. In the development of the theory of the firm, the concept of bounded rationality arose. You have premises and you have to have a two-step reduction of complexity - first to declare your premises and then to try to optimize if this is possible. In this, you have to take into account the limits of information capacity, and so on. I think that rationality needs a new definition. I do not deny the possibility of rationality, and I do not deny the function of the concept. But I have in mind a more Utopian, a more complex definition of rationality based upon the idea of a re-entry of the system/environment distinction into the system. In a sense you differentiate by throwing the environment out, you have no contact with the environment. But then you re-introduce or copy the difference into the system, so it is a kind of contradiction to your starting point. This seems to give some hope of defining a new concept of rationality.
But the re-entry operation does not mean that the environment enters into the system?
You copy the distinction, not the environment. You do not have a kind of input mechanism, but if you are aware of the boundaries of a system, then you have of course the idea that there must be something on the other side of the boundaries that I cannot see, cannot touch, cannot check with my operations. The system needs a kind of internal distinction of inside/outside.
Why the stress on distinctions and observers?
This is partly due to the Laws of Form of Spencer Brown - the idea that you cannot observe anything without drawing a distinction and that the concept of observing is very broad, including for example actions. You cannot act on something without making a distinction: you do this and not something else, you cannot act without choosing to do so or without defining a goal and an objective, and in that connection you also have to talk about, for example, theory and not about politics. So, in the Laws of Form of Spencer Brown, it was seen as a necessity to construct or to create or to describe the identity of elements, for example of numbers or of variables in algebra. Why is there 1 and 2 and 3 and 4? This is simply presupposed in normal mathematics. In Spencer Brown's calculus you have a first injunction, "draw a distinction", which just differentiates a marked space from the unmarked space. But how to get to a first definition or indication of something? This has to do with the whole structure of this kind of calculus: only at the end and only by exploding the possibilities of mathematical operations can the observer enter the system, so that you can say that the primary distinction is the observer. This kind of move from distinction to the observer coincides with the theory of second order observation by Heinz von Foerster as well as other approaches to developing or enlarging the framework of traditional cybernetics.
Why self-reference and second-order observation?
This is a way of increasing the complexity of the system. I could give many answers to this, partly, of course, because it is the most fascinating development in present days' systems theory, which dissolves the old controversies between say human interest and input/output mechanism. But for me, at the moment at least, the most important thing is that if you have a self-referential loop within the system or if you introduce the output of the system as an input into the system, or if you think of a system as producing the starting point for the next operation, then you have an unresolvable indeterminacy or a self-created uncertainty. This poses the question, not how to control the complexity of the environment, or how to reduce its complexity, or how to find stable identities, but how to cope with your own product - uncertainty.
Why the stress on the time aspect?
This too deals with the problems of rationality and self-reference. If you accept the notion of self-referential systems, which produce their own basis from which to start operations, which produce the network which produces the operations, then there is no possibility of having a kind of fixed point, a kind of reasonable criteria or whatever. The only resource is actually time: that you have a memory, that you cannot change your past, and that the future is distinct from the past. You have to make a decision, which means that time re-enters time because within the decision framework you have the past of the decision, the information you have got, the preparations, and the future of the decision, that is, the way you will handle the goals and risks of the next decisions and so on. So the idea is that time is not only important because you have to do something in time, now, and not yesterday and not tomorrow. It is also important because one can reduce complexity by means of temporal distinctions. Time is a way of getting resources, of getting the framework of "now we are here". As we cannot undo the past, and cannot know the future, we have to use our un-knowledge of the future as a resource of our present decision. So for me it is increasingly important to see temporal differences as a resource, as a condition of the possibility of decision-making.
Why the stress on paradoxes?
I think it is a recent development, which is visible and very different from the years of linguistic analysis and talk about postmodernity. It goes beyond the classical distinction between rhetorical paradoxes, which are just amazing phrases to open the space for further discussion and new arguments, and logical paradoxes which we somehow have to cope with, if we want to have a consistent system - the Russell-Whitehead problem and then the Gödel problem. The basic point is that the observer is a precondition for describing, for indicating, for acting at all. But to see the construction of elements, the construction of something at all, the indication of an object, is itself a paradox. The reason is that you cannot indicate at the same time the object and what is excluded by indicating this object and not another one. There is a kind of invisible space - the observer himself plus the observing. The observer using the distinction cannot observe himself, because then he has to distinguish himself from other observers. He can do it in the next operation, but this repeats the problem.
To have an invisible unmarked area is not the same thing as having a paradox?
No, you could have many paradoxes, in more concrete ways. If you ask what is the distinction, you meet the problem of the unity of the distinction. If we use the distinction, say jurisdiction/legislation and ask for the unity, then we have to go to the legal system. But then what is it? How to indicate the legal system with respect to the society? And what is the unity of the society and the legal system, and so on and so forth. There is a hidden paradox in all kinds of distinctions.
And also a need of de-paradoxation?
I think the foundation of everything is the unfoldment or unfolding of a solution of a paradox by any distinction which is to some extent convincing at the moment. A paradox cannot be solved, but it can be replaced by a distinction which is for the time being convincing. So for example the distinction system/environment. Of course you could ask what is the same in the system and in the environment. You could say: well, the world is inside and outside, but if you want to use system theory you have in a sense to ignore the world and to refrain from asking the question: what is the world? You do not allow yourself to ask ontological questions. That makes the world invisible or makes the observer invisible. Therefore you always find solutions which in a certain historical situation are more convincing than others, using distinctions such as nature and culture or whatever is for a certain time convincing.
How is it possible to combine the idea of observers observing in many different ways with the idea of a familiar world?
A familiar world is always organized in terms of distinctions. We know how to distinguish a tree. Trees and flowers are distinctions. Cities and villages are distinctions, and so is Norway and Denmark. Familiarity is the possibility of repeating the use of certain distinctions and to be able to neglect the other side of a distinction.
But you can observe one phenomenon, for example to be gay, using many different distinctions, each isolating a certain aspect. You have one word and many distinctions? Is it the same phenomenon irrespective of the distinction used?
Steven Holmes talks about "antonym substitution". You can say "I am interested in gays" presupposing the distinction homosexuality/heterosexuality. You may have another interest, for example the political oppression of gays, or how to make the appearance of gays normal and not pathological. To have one term, one object, is to have a switching point for distinctions.
At this time, the participants of the seminar returned from their lunch break, making it impossible to continue the interview. But just as texts and communications never stop, this interview might continue, as a meeting between an author and a reader, in the texts of Niklas Luhmann.
Return to the content of this issue
Return to the Cybernetics and Human Knowing Homepage