In the Introduction to Autopoietic Law: A New Approach to Law and Society (1988) Gunther Teubner writes,
According to Luhmann . . . the basic units of a legal system are neither, as lawyers are used to describing them, legal norms, nor, as sociologists define them, actors and organizations — law is a system of communications. However, legal autopoiesis is brought about only if an emergent element of the legal system is created: the legal act. Legal acts are those communicative events that change legal structures. Here we find the basic circularity that defines the legal system: the circular relationship between legal acts and legal norms. (4)
Systems create their own elements–in this case the legal act; they don’t create outputs. Thus, to speak of systems transforming inputs into outputs is misleading. In “Closure and Openness: On Reality in the World of Law,” Luhmann wrote
[S]tructures of the system can be built up only by operations of the system. This too must take place in such a way as to be compatible with the system’s autopoiesis; in the case of social systems, for instance, with communication. There is accordingly no input and no output of structures or operations of the system, and at this level, there are no exchange relationships with the environment. All structures are operationally self-specified structures of the system, which orient its operation to these structures. In this respect, too, the system is a recursively closed system. As presumably no one who accepts the distinction between system and environment will deny, it cannot operate outside its boundaries, in its environment, as an observer can see them, cannot be actualized as operations of the system.(Teubner (Ed.), Autopoietic Law: A New Approach to Law and Society, 1988. p. 337)
There are no structural inputs and outputs for operationally closed systems. Nor is there an exchange relationship between systems and their environments because events happen within systems. This is seen most clearly in the case of information. There can be no exchange of information between system and environment because information isn’t information until it is constructed as such by a system. Information is not packaged and exchanged; it is constructed. It must make a difference in a system.
Yet continued autopoiesis depends on an environment. Environments stimulate autopoiesis. According to Luhmann,
A system can reproduce itself only in an environment. If it were not continually irritated, stimulated, disturbed and faced with changes in the environment, it would after a short time terminate its own operations, cease its autopoiesis. . . The question remains how the environment impinges on the system, and what relevance this has for the system’s self-reproduction, for the continuation of its own operations.” (335)
The law, for instance, needs an environment that is has nothing to do with the law. Otherwise no more legal decisions would need to be made. The law is perturbed by realities having nothing to do with the legal/illegal distinction. The economy, politics, mass media, science and the other function systems have nothing to do with the legal/illegal distinction. Nothing can be legal or illegal for the economy, politics, mass media, or science. Events can only be legal or illegal for the law.
Where the economy is concerned, all the law can observe is whether something, such as a payment or a business practice, is legal or illegal. It cannot distinguish between smart and dumb business decisions, for instance. The observation of illegal/legal is a cognitive operation on the part of the law. (It is not only minds or brains that engage in cognitive operation.) Cognition creates structures in a system, but cognitive operations cannot change the binary code of system. Cognitive operations ensure the system’s openness to the environment.
[The] cognitive operations guarantee the system’s openness to the environment (not its relations with the environment!); and likewise, that as regards the function and code of the system, cognitively oriented operations are secondary by comparison with the system’s recursively closed autopoiesis. (341)
One event can register in two or more systems. For instance, paying a fine that was imposed by a judge is observable by economy and the law. Such an observation has no temporal duration; a system does not make an observation over time; it only makes a series of observations with no duration. So if a structure is to emerge, observations must be strung together. In the law, one legal decision (observation of legal/illegal) is linked to a previous decision. In the economy, one payment is linked to another payment, and a payment is over as soon as it happens. If the event had duration, it might cause the economy and the law to get stuck together. As Luhmann argues,
That the elementary operations have the character of events can guarantee a high degree of interpenetration [structural coupling] of the various systems, preventing, through the disappearance of the events, the systems from becoming stuck to one another. (342)
The lack of duration of operations protects system autonomy. It doesn’t get stuck to anything in its environment.
The system may produce ideas of its own or communications of its own about its environment, but can never grasp and reprocess thereby everything that it itself presupposes as environment. (Not least for this reason, the environment appears to the system as a “horizon.”)
Epistemological “constructivism” concludes from this that what the system, as the level of its operations, regards are reality is a construct of the system itself. Reality assumptions are structures of the system that uses them. (337)
Unstable elements “acquire duration only by continually replacing disappearing elements by other ones” (342).
Autopoiesis does not mean reproduction despite threat from without or despite natural decay of the elements. Instead, the elements are produced in order to end immediately, the system continually disintegrates itself. And the autopoiesis stops are every moment–unless continued. (342)
The system tests its own reality assumptions. It performs consistency tests to determine if reality agrees with the assumptions. Assumptions are structures of schemata, and the presuppose a memory function. (Memory is a consistency check.) Every system must start with some schemata, some assumptions about reality–some assumptions or expectations that might be confirmed or disappointed, rather than randomness or chaos. These schemata, of course, belong to the system, not the environment.
So a system has no access to a reality beyond itself. A system presupposes a structured (not random) reality or medium. But the system cannot treat its environment as chaos; it must start with some kind of assumption–an assumption that the environment is something other than mere chaos.
Luhmann argues that systems do not transform chaos into order. He refers to the assumption that
autopoietic systems transform chaos into order (order from noise), whereas such chaos is in fact nowhere to be found. Even modifying this into “order from order plus noise”, the statement remains so unspecific that not much can be done with it.
Chaos cannot be found because it would have to be observed and observation requires a distinction. But in chaos there can be no distinctions.
Again, a system cannot begin with the assumption that its environment is chaos.
Taking a lead from the neo-Darwinist objections to Darwin, this objection may also be formulated in terms of evolutionary theory: a system that had to start from purely random relationships to its environment and wait for noise or irritation would take too much time to build up its own order for it to be capable of evolution. The rate as which complex structures are built up cannot be explained in this way. (338)
For post-ontological theory design, it is quite possible to assert that everything that “is” is formed through complexity reduction, and that the autopoietic systems in consequence form everything that acts as a unit for them through their own operations. (339)
A person (as a unit), for example, is formed through complexity reduction. A person is something that acts as a unit for an observer. A person is something that can be addressed; it is an addressee; it is can be an agent that a cause is attributed to–that is to say, it is held responsible for something. The complex conglomeration of systems–cells, immune system, circulatory systems, brain, psychic system, etc.–that might be analyzed from a scientific point of view are irrelevant when a person is observed as a unit (an addressee or agent) for purposes of communication. All that complexity must be reduced for communication to happen.
The law reduces complexity by using only one distinction–legal/illegal. It also uses the common knowledge of society. It uses the normative concept of man or woman, for example.
Should questions such as whether women, etc,. really exist arise, they can be turned aside or referred to philosophy. . .
In this sense, the law participates in society’s already achieved construction of reality, without having to work it out itself. It makes use of language and a more or less consistent use of words inside and outside the legal system.
Nevertheless , the law differentiates out within society as an autopoietic system on its won, by setting up a network of function-specific communication which in part gives words a narrower sense, in part a sense incomprehensible for non-legal communication, in part adding coinages of its own (for instance liability, testament), in order to make the transformations needed by law communicable. (340)
the surprise value of one and the same event differs in different systems. The now notorious practice of economic contributions to political parties was a surprise only for the legal system, while the political system was surprised only by the fact the legal system was surprised thereby and the economy continues to attribute no significant informational values to slight sums. (343)
the legal quality (whether validity or invalidity) of claims and decisions can be derived only from other operations of the same system (for instance, by reference to statutes or precedents or to such dubious recourses as “prevailing opinion”); it cannot be supplied from external sources like religion or politics or the economy and if in the legal system such references to external sources can be found, then these references are in turn already legal norms, which legally legitimate block acceptance of external norms or decisions . . . Any other view would have the problem of explaining how law and morals, rules of sound management, etc. could then be distinguished. (345-46)
As for autopoiesis in general, it can also be said about autonomy that it either exists or does not. It cannot be realized a little bit. Neither can there be relative autopoiesis–and if so, compared to what? All that is gradual is the degree of differentiation of the system according to the quantity and nature of the operations that it can carry out. The struggles of “autonomy” of the law against theological and political tutelage, against the legal exemption of the church, against the linking of legal doctrine to theological appropriateness, against royal interference in legal process, etc., are . . . to be understood as processes of law’s increased differentiation, i.e., as aspects of the historical transition to functional differentiation of the social system, ending ultimately in the universality of the legal system’s functional competence.
. . .
The law’s autonomy is in danger only when the code itself is in danger–for instance when decisions are taken in the legal system itself increasingly according to the difference between beneficial and harmful rather than the difference between legal and illegal (346-47)