Copyright 1989 by Northwestern University, School of Law Printed in U.S.A.

Northwestern University Law Review Vol. 83, Nos. 1 & 2

LAW AS A SOCIAL SYSTEM

Niklas Luhmann*

In the classical division of labor between jurisprudence and sociol-

ogy, jurisprudence is concerned with norms, and sociology, in contrast,
with facts. The jurist's task is to interpret norms and apply them. The
sociologist may concern himself only with the existing context of the law,
with its social conditions and consequences. But this classical view was
already out of date, if not anachronistic, even at the time when Hans
Kelsen gave it its most precise formulation. "Social-engineering" ap-
proaches and the jurisprudence of interests had tied the apphcation of
law to facts that had not been taken into account in formulating norms
but instead had to be ascertained subsequent to the formulation of the
legal text. Pragmatism had postulated that all practical apphcation of
the law should consider how different constructions of the law would
affect legal outcomes; it was concerned not only with the impact on fu-
ture decisions within the legal system but also with controlling actual
consequences within social reality.
The resulting dissolution of the sharp demarcation between jurispru-
dence and sociology has given rise, since the beginning of this century, to
the hope that sociology will be able to make a contribution to the admin-
istration ofjustice. From the perspective of the law, however, sociology's
function remains more that of an auxiliary science. Aside from a few
exceptions (the concept of an institution, for example), sociology has had
no influence on legal theory and scarcely any impact on legal doctrine.
Nor is it clear whether a special discipline called "legal sociology" can
provide the law with information, or whether all branches of sociology
would be available to do so. And there is still no adequate sociology of
legal doctrine or legal theory.1
There has not been much movement on any of these questions in the
last two decades. It is clear, however, that the quite optimistic expecta-
tions for a sociological contribution to the administration of justice have

* Professor of Sociology, University of Bielefeld. Visiting Professor of Law, Northwestern Uni-

versity School of Law, 1989-90. Translated by Shierry Weber Nicholsen, Ph.D. Cornell University.
Ms. Nicholsen is a translator of social science, most recently of Jiirgen Habermas' O N THE LOGIC OF
T H E SOCIAL SCIENCES (1988), and his T H E N E W CONSERVATISM: CULTURAL CRITICISM AND T H E
HISTORIANS' D E B A T E (forthcoming).
1
For some recent studies, which do not, however, share a common theoretical basis, see HlS-
TORISCHE SOZIOLOGIE DER RECHTSWISSENSCHAFT, IUS COMMUNE SONDERHEFT 2'6 (E.V. Heyen
ed. 1986) [Historical Sociology of Jurisprudence, special issue on common law].

136

HeinOnline — 83 Nw. U. L. Rev. 136 (1988-1989)

83:136 (1989) Law as a Social System

diminished and become more realistic. At present, moves toward a radi-

cal alteration in the way these questions are posed can be expected
neither from jurisprudence nor from sociology. To the surprise of schol-
ars of both disciplines, they are coming from elsewhere—from research
that is attracting more and more attention under such names as general
systems theory, cybernetics (of the third or fourth generation), multiva-
lent logic, theory of automata, information theory, and, recently, as a
general theory of self-referential "autopoietic" systems,2 with which we
shall be concerned here.
This detour by way of general autopoietic theory is currently pro-
ducing more confusion than clarity and more problems and open ques-
tions than answers. The confusion is closely related to the fact that the
offerings of existing theories have their origins in mathematics, biology,
or neurophysiology, and do not take matters of psychic or social fact into
consideration. As yet there has been no place in this discussion for sys-
tems that conduct their operations with the aid of the medium of "mean-
ing." The new discovery is that biological systems, if not physical
systems in general, are characterized by a circular, recursive, self-referen-
tial mode of operation. The mode of analysis that has emerged from this
discovery has dethroned the "subject" in its claim to be unique in its self-
referentiality. This does not have to mean that psychic and social sys-
tems are now to be interpreted in terms of the model of biological sys-
tems. A mere analogy between them would miss the mark, as would a
merely metaphorical transfer of biological terms to sociology. The chal-
lenge is rather to construct a general theory of autopoietic systems that
can be related to a variety of bases in reality and can register and deal
with experiences deriving from such diverse domains as life, conscious-
ness, and social communication. Current uncertainty is due primarily to
the fact that a general theory of this kind does not exist, and conse-
quently one is frequently working too directly with concepts borrowed
from mathematics or biology, without adequate concern for the appro-
priateness of the transposition.
In the application of the theory of autopoietic systems to the specific
case of the law, there is an additional problem of coordination among
multiple levels. One can conceive of law as a social system only if one
takes into consideration the fact that this system is a subsystem of soci-
ety, and that there are other subsystems as well. To conceive of society
as itself a differentiated social system presupposes a general theory of
social systems that can deal not only with the comprehensive system of
society as a whole but also with other social systems, such as face-to-face
2
An autopoietic system is a system that "produces and reproduces its own elements by the
interaction of its elements." See Teubner, Introduction to Autopoietic Law, in AUTOPOIETIC LAW, A
NEW APPROACH TO LAW AND SOCIETY 1-11, at 3 (G. Teubner ed. 1987). See also, N. LUHMANN,
RECHTSSOZIOLOGIE (2d ed. 1983) (English translation, A Sociological Theory ofLaw, E. King & M.
Albrow trans. 2d ed. 1985).

137

HeinOnline — 83 Nw. U. L. Rev. 137 (1988-1989)

 NORTHWESTERN UNIVERSITY LAW REVIEW
interaction, or organizations. Theoretical decisions must therefore be
distributed across several levels and must be checked to see whether what
is asserted of the law does not hold for society as a whole, or even for
every social system or every autopoietic system as well.
The following reflections focus on the legal system and must there-
fore largely disregard these problems of coordination among multiple
levels.3 In the treatment of a relatively concrete subject, this omission
will produce the appearance of excessive abstraction. The reader should
not let this intimidate him; nor should he see it as proof, in and of itself,
of the scientific character of the treatment. In fact, it is only in this way
that one can confront general theories with the realities of concrete areas
of investigation to see whether the theories are functional and what mod-
ifications they might need.

I.
There are two innovations that especially lend themselves to use in a
theoretically grounded sociology of law: (1) the theory of system differ-
entiation, inspired by general systems theory, which conceives of differ-
entiation as the establishment of system-environment relationships in
systems; and (2) the assumption that such differentiation is possible only
through the establishment of a self-referential closedness in the systems
becoming differentiated. Without such closure, the systems would have
no way of distinguishing their own operations from those of the environ-
ment. With the aid of these two concepts we can achieve an understand-
ing of the social character of law and, at the same time, the legal system's
own reflective accomplishments. In other words, doctrine or legal theory
can be better understood as one formulation of the legal system's self-
referentiality. This understanding does require, however, a much more
precise mode of presentation than has tended to be customary, a presen-
tation that is consistent with systems theory.
Formulations such as the statement that there are "connections be-
tween" law and society (which presupposes that law is something outside
of society) especially must be avoided. The legal system is a differenti-
ated functional system within society. Thus in its own operations, the
legal system is continually engaged in carrying out the self-reproduction
(autopoiesis) of the overall social system as well as its own. In doing so,
it uses forms of communication that, for all their esoteric quality, can
never be so abstract as to be completely removed from normal, compre-
hensible meaning. This means not only that the legal system fulfills a
function for society—that it "serves" society—but also that the legal sys-
tem participates in society's construction of reality, so that in the law, as
everywhere in society, the ordinary meanings of words (of names, num-
3
See N. Luhmann, SOZIALE SYSTEME: GRUNDRISS EINER ALLGEMEINEN THEORIE (1984)
[Social Systems: Outline of a General Theory].

138

HeinOnline — 83 Nw. U. L. Rev. 138 (1988-1989)

83:136 (1989) Law as a Social System
bers, designations for objects and actions, etc.) can, and must, be presup-
posed. In the legal system, then, Mr. Miller is still Mr. Miller. If he is
only claiming to be Mr. Miller, and this question must be examined
within the legal system, then a language that is generally comprehensible
is indispensable for resolution of that question as well.
The legal system, however, is distinct in many ways from law's envi-
ronment within society (and of course from its extra-societal environ-
ment as well). The law is not politics and not the economy, not religion
and not education; it produces no works of art, cures no illnesses, and
disseminates no news, although it could not exist if all of this did not go
on too. Thus, like every autopoietic system, it is and remains to a high
degree dependent on its environment, and the artificiality of the func-
tional differentiation of the social system as a whole only increases this
dependency. And yet, as a closed system, the law is completely autono-
mous at the level of its own operations. Only the law can say what is
lawful and what is unlawful, and in deciding this question it must always
refer to the results of its own operations and to the consequences for the
system's future operations. In each of its own operations it has to
reproduce its own operational capacity. It achieves its structural stability
through this recursivity and not, as one might suppose, through
favorable input or worthy output.
In this conceptualization both the dependence and the independence
of the law are more strongly emphasized than in the customary expres-
sion "relative autonomy." When sociological theory is used to formulate
a theory of the legal system, it reveals many more aspects of dependence
and many more aspects of independence than one tends to notice in the
normal activity of the law, and consequently theory has to abandon the
amorphous formulation "relative autonomy." Differentiation gives rise
to an escalating relationship in which aspects of dependence and aspects
of independence both increase, because differentiation leads to greater
complexity in relationships between the system and the environment.
For this reason, the concept of the autonomy of the legal system cannot
be formulated on the level of (causal) relationships of dependence and
independence. Rather, the concept of autonomy refers only to the sys-
tem's operative closedness, as a condition for its openness.4
A theory of this kind, however, is convincing only if it succeeds in
precisely defining the elements of the closed character of the system and
how those elements determine the system's openness. This can be done
by describing more precisely the components of the particular elemen-
tary operations peculiar to the law (those which occur nowhere else but
in law) and how they are reproduced through reference to one another.

4
See generally 1 E. MORIN, LA METHODE: LA NATURE DE LA NATURE (1977) [Methodol-
ogy: The Nature of Nature]. For a biological perspective see F. VARELA, PRINCIPLES OF BIOLOGI-
CAL AUTONOMY (1979).

139

HeinOnline — 83 Nw. U. L. Rev. 139 (1988-1989)

 NORTHWESTERN UNIVERSITY LAW REVIEW

In a way that no other system does, the law processes normative

expectations that are capable of maintaining themselves in situations of
conflict.5 The law cannot guarantee, of course, that these expectations
will not be disappointed. But it can guarantee that they can be main-
tained, as expectations, even in case of disappointment, and that one can
know this and communicate it in advance. From the sociological point
of view, then, normativity is nothing but counterfactual stability. To for-
mulate this differently: in that it protects expectations, the law frees us
from the demand that we learn from disappointments and adjust to
them. It thereby holds out the prospect of resolving conflicts (and at the
same time makes it possible to seek out and withstand conflicts), for it
contains a preliminary decision (however unclear it may be in the indi-
vidual case) about who has to learn from disappointment and who does
not.
Processing these expectations requires a binary code that contains a
positive value (justice) and a negative value (injustice), and that artifi-
cially excludes both contradictions (justice is injustice, injustice is justice)
and other values (utility, political expediency, and so forth). This coding
is of decisive significance for the differentiation of the legal system, as it
provides the system with its own internally constituted form of contin-
gency. Everything that enters the law's sphere of relevance can be either
lawful or unlawful, and anything that does not fit into this code is of legal
significance only if it is important as a preliminary question in decisions
about justice and injustice.
One could show through more detailed analysis that this coding ful-
fills a dual function. The first function of the code serves to differentiate
the system for the specific task of the law. It simulates the problem of
the disappointment of expectations by providing that either the expecta-
tion or the conduct that disappoints the expectation will elicit either the
positive or the negative evaluation. To this extent, the coding is tied to
the law's function. At the same time, however, the coding also serves the
system's ongoing process of checking for consistency, that is, the actual-
izing of its memory. For memory is nothing but checking for consis-
tency, and to this end it presupposes, presumably even on the
neurophysiological level,6 a binary coding that can ascertain both consis-
tencies and inconsistencies and can link them to further operations.
Thus, the second function serves the autopoietic reproduction of the sys-
tem—the closure of the system's reproduction complex. It makes it pos-
sible to examine all processing of normative expectations in terms of the
key question whether or not the processing is compatible with previous
processing.
Once this dual function, and with it the law's autopoiesis, has been
5
For a more detailed discussion of this point see N. LUHMANN, supra note 1.
6
See Foerster, What is Memory That it may Have Hindsight and Foresight as Well?, in THE
FUTURE OF THE BRAIN SCIENCES 19-64 (S. Bogoch ed. 1969).

140

HeinOnline — 83 Nw. U. L. Rev. 140 (1988-1989)

83:136 (1989) Law as a Social System
assured, the system can develop reflexive processes and, ultimately, self-
reflection. It can regulate its own regulation, and thereby also regulate,
legally, alterations in the law. Further, it can evaluate the system as a
whole from its own perspectives (e.g., in terms of the idea of justice).

II.
The next sections of this essay will deal with some of the conse-
quences of this theoretical point of departure. Especially important here
are aspects in which this theory leads to views that are new or that differ
from ones previously accepted.

A.
An especially important implication of this theory of the law's auto-
poietic character is that the boundaries of the system must be drawn dif-
ferently than has been customary (even in the way sociologically oriented
systems theory has dealt with law). Up to now the law has been treated
either from the perspective of jurisprudence, as a complex of norms, or as
a system of knowledge, in abstraction from real social behavior. Jurists
saw the legal system as a macro-text. Or, as is customary in sociology,
the focus was shifted to institutions that are concerned with law on a full-
time basis, whether those institutions were organizations (primarily the
courts), or the legal profession. This perspective permitted empirical
treatment of such problems as "access to the law." Yet distinguishing
between the legal system and the state as the basis for organizations and
the source of power was difficult. Political influence on the law was con-
ceived as a kind of input (of the law into the law). Alternatively, the
legal system as a whole was even conceived from the standpoint of the
political system, as an "implementation" of politics. For all its ambiva-
lence, this perspective has left a definite mark on jurists' attitudes toward
the relationship between law and politics.7
Assuming that the system has a self-referential, closed character
leads to completely different notions about the boundaries of the system.
They are defined not at the institutional but at the operative level. And,
as is evident to the sociological observer, the system's boundaries are de-
fined by the legal system itself, with the aid of a recursive referral of
operations to the results of (or the prospects for) operations by the same
system. In these terms, every communication that makes a legal asser-
tion or raises a defense against such an assertion is an internal operation
of the legal system, even if it is occasioned by a dispute among neighbors,
a traffic accident, a police action, or any other event. It is sufficient that
7
One significant consequence is that jurists, more than other professionals, have an understand-
ing of the political context of their own practice. See Lange & Luhmann, Juristen—Berufswahl und
Karrieren, 65 VERWALTUNGSARCHIV 113-62, 157 (1974) [Jurists—Choice of Profession and Ca-
reers. Archives of Administration].

141

HeinOnline — 83 Nw. U. L. Rev. 141 (1988-1989)

 NORTHWESTERN UNIVERSITY LAW REVIEW

the communication be assigned a place within the system, and that has
already occurred with the use of the code lawful/unlawful. Of course,
the law can also be observed from the outside, as in a news report in the
press. And within the educational system there is also a didactic treat-
ment of law that only simulates legal cases and thus does not aim at a
decision. Consequently not every reference to the law is an operation
internal to the legal system. But whenever a communication occurs in
the context of the administration of justice, the context of providing for
conflicts within the law, or the context of an alteration of the law—that
is, in the processing of normative legal expectations—we are dealing with
an operation internal to the legal system, and this operation simultane-
ously defines the boundaries between the legal system and the everyday
life context that occasions the posing of a legal question.
These system boundaries are a good place to study thefilteringeffect
of the legal system. One sees clearly, for example, how difficult it can be
in ongoing life relationships (marriages, work relationships, relationships
between neighbors) to resort to the law to give force to one's own views.
The rigidity of the binary code makes the reasons for this difficulty clear:
asserting one's own legal position is tied to designating opposing views as
unlawful. A look at the legal cultures of the Far East also shows that
recourse to the law can be interpreted as an intention to engage in con-
flict, and consequently it is institutionally discouraged.
Clearly there is a connection between the complexity of the law, its
resulting opaqueness, and how high this threshold of discouragement is.
Corruption, which a look at various civilizations will show to be a nor-
mal phenomenon, has an equally discouraging effect on potential users of
the legal system. Corruption in law is a normal phenomenon: it is only
realistic to assume that the law accommodates dominant interests; it
could not conduct itself otherwise and still be accepted. (This does not
mean, however, that corruption is a part of official legal policy or that it
is consciously cultivated). Rather, what is amazing is the degree to
which the law can be purged of corruption in spite of this. With a de-
crease in corruption, the threshold of discouragement is thereby lowered;
people have confidence in a judge who is impartial. Yet, this relief itself
leads to an increase in the complexity of the law. With less corruption to
filter people out of the legal system, the number and diversity of cases
increases, and as a result there is increased need for regulation. With this
increase in complexity, the threshold of discouragement shifts its location
from corruption to complexity. It thereby acquires a form against which
the legal system itself is powerless and which is the subject of recurrent
complaints throughout the history of law.8
If one adopts a self-referential autopoietic theory, it no longer makes
8
An example is the discourspreliminaire [preliminary discourse] in S. LINGUET, 1 THEORIE
DES Loix civiLES, ou PRINCIPES FONDAMENTAUX DE LA SOCIETE (London 1767) [Theory of Civil
Law, or Fundamental Principles of Society].

142

HeinOnline — 83 Nw. U. L. Rev. 142 (1988-1989)

83:136 (1989) Law as a Social System

sense to assume that the structures of the legal system, which themselves
regulate the production of its operations, can be specified as input and
output. The specification of structures always presupposes operations of
the system itself. This does not contradict the assumption of a normal
complicity with dominant interests on the part of the law. Nor does it
exclude the possibility that an outside observer could describe the legal
system with the aid of an input-transformation output model. But such a
description would be compelled to give the transformation function the
form of a "black box," and to take into consideration the fact that the
law adjusts its reactions to its condition at any given time, that it can
change even if external interests do not change, and that it thus does not
function as a "trivial machine." To the degree to which these factors are
taken into consideration, however, it makes sense to move from an input-
output model to the theory of self-referential systems.9 It is better suited
to the existing state of affairs.

B.
The most important advantage of this theory of a closed self-referen-
tial legal system may lie in its close resemblance to the notions of legal
doctrine and legal theory, a closeness which by virtue of its alienation
effect proves surprising and irritating at the same time.10 Sociological
theory attempts to reconstruct not only jurists' actions, but also their
conceptions, or at least the way the legal system generates self-observa-
tions and self-descriptions. We are not concerned here with a "critique
of ideology" or a sociology of knowledge in the classical manner in which
modes of thought are linked to interests or social positions and explained
in those terms.
Our starting point is the thesis that a self-referential system can link
its operations together and reproduce them only through concurrent self-
9
See also the distinction between couplagepar input and couplagepar cloture [coupling through
input; coupling through closure] in Varela, L'auto-organisation: De Vapparence au mecanisme, in
L'AUTO-ORGANISATION: D E LA PHYSIQUE AU POLITIQUE 147-64 (P. Dumouchel & J. Dupuy eds.
1983) [Auto-organization: From Appearance to Mechanism, in Auto-organization: From Physics
to Politics].
10
A parallel problem arises in the relationship of law to theology, and it has been more exten-
sively discussed in the literature. See D. POLLACK, DIE RELIGIONSTHEORIE NIKLAS LUHMANNS
UND IHRE SYSTEMTHEORETISCHEN VORAUSSETZUNGEN, (1984) [Niklas Luhmann's Theory of
Religion and Its Systems-Theoretical Presuppositions]; F. SCHOLZ, FREIHEIT ALS INDIFFERENZ:
ALTEUROPAISCHE PROBLEME MIT DER SYSTEMTHEORIE NIKLAS LUHMANNS (1982) [Freedom as
Neutrality: Old-European Problems with Niklas Luhmann's Systems Theory]; THEOLOGIE UND
FUNKTIONALE SYSTEMTHEORIE: LUHMANNS RELIGIONSSOZIOLOGIE IN DER THEOLOGISCHEN DlS-
KUSSlON (M. Welker ed. 1985) [Theology and Functional Systems Theory: The Reception of Niklas
Luhmann's Systems Theory by Theology]; Moritz, Religion und Gesellschaft in der DDR, 110 THEO-
LOGlSCHE LITERATURZEITUNG 573 (1985) [Religion and Society in the German Democratic Re-
public. Journal of Theological Literature]; Pannenberg, Die Allgemeingultigkeit der Religion, 11
EVANGELISCHE KOMMENTARE 356-57 (1978) [The Universal Validity of Religion. Protestant
Commentary].

143

HeinOnline — 83 Nw. U. L. Rev. 143 (1988-1989)

 NORTHWESTERN UNIVERSITY LAW REVIEW
observation and self-description. To put it very simply, one needs "rea-
sons" in order to be able to deal selectively with the multitude of possible
internal connections, and to check for consistency and inconsistency.
Consequently, all processing of expectations is always accompanied by a
supervisory observation through which the way the world is observed is
itself observed—that is, the way one communicates correctly or incor-
rectly within the system is itself the subject of communication. From the
sociological point of view one would speak here of "redundancy," of a
diminishing of the surprise effect of individual operations.11
Every complex system must balance variety, that is, the number and
diversity of its basic elements, against redundancy. In a complex envi-
ronment it is not possible to operate in a completely rigid fashion, with-
out surprises. Rather, the system must be open to irritations that disrupt
the usual practice. If the law, however, is to provide security, this open-
ness cannot be carried too far. There must be a provision for redundancy
so that knowledge of one or more elements (knowledge of important
court decisions, for example, or knowledge of decisions about laws) can
be relied upon to permit inferences about how the system will behave in
concrete instances.
This issue of the relative degrees of variety and redundancy is
closely connected to the system's relationships with its environment.
One can proceed on the assumption that in interactions between elastic
and rigid systems the elastic systems will adapt to the rigid ones, just as
sand conforms to stone but stone does not conform to sand. A legal
culture of argumentation that produces a high degree of variety, that em-
phasizes the individual nature of each case and is content with vague
general formulas like "proportionality" or "balancing interests," will
tend to open the legal system to adaptation to rigid environmental sys-
tems such as large-scale organizations whose form is set by technology or
capital investment. Whereas a rigid, highly redundant legal system will
be able to maintain itself, whatever the social consequences may be, in
the face of the more elastic systems of its environment and to turn such
highly elastic communications media as money or political power to its
own ends.
This is only one of many examples of the way sociological analysis
produces an "alienation effect"12 through its special understanding of
the way systems observe and describe themselves. Reconstructing argu-
mentation as the management of redundancy does not grasp argumenta-
tion the way it is intended; it understands argumentation not as a search
for convincing rational grounds but as a way of mastering contingency
11
On the reconstruction of argumentation as the management of redundancy see N. LUHMANN,
DIE SOZIOLOGISCHE BEOBACHTUNG DES RECHTS 31-38 (1986) [Sociological Observation of the
Law]. See also G. Lazzaro, Entropia della legge (1985) [Entropy in Law].
12
An example of this effect is found in Brechtian theater, where the spectator is prevented from
identifying with the characters in the play—translator.

144

HeinOnline — 83 Nw. U. L. Rev. 144 (1988-1989)

83:136 (1989) Law as a Social System

and as a condensation of the systemic context. The sociological descrip-

tion of the system's own self-description could not be accommodated
within that self-description (although there is more to be said on that
point). For that reason, in observing the legal system, sociological de-
scription always uses the schema manifest/latent as well, and with the
help of this schema it also sees that the system does not see that it does
not see what it doesn't see.
But in contrast to the aims of a critique of ideology, no unmasking
or enlightening effect is intended here. Rather, this way of seeing things
follows logically from the assumption that every autopoietic system dif-
ferentiates its own operations with the aid of its own distinctions, and
thus, if it wants to preserve this differentiation it is prevented from distin-
guishing itself in turn from these distinctions.
How far into legal doctrine this impossibility extends shall remain
an open question here. Certainly it applies to the code itself. To deal
with the question whether the distinction between justice and injustice is
being used justly or unjustly would lead the system into paradoxes and
block at least the operations based on this question. Observation and
description of the legal system in terms of legal theory must presuppose
the acceptability of the code. It may proceed neither on the basis of a
tautology (justice is what is just) nor on the basis of a paradox (what is
just is what is unjust). It has to "tune out" this possibihty of defining the
unity of the system within the system itself; it has to de-tautologize and
de-paradoxicalize the description of the system and at the same time
make the operations through which this is done invisible.13
If it is important to him to do so, the sociologist can observe, with
the aid of the schema manifest/latent, legal theory's efforts to de-tautolo-
gize and de-paradoxicalize the system; he locates the latent functions of
the manifest intention of the legal discourse, which will be directed else-
where. In doing so, he can make use of general systems theory's distinc-
tion between natural and artificial necessities.14 The operations that
serve to de-tautologize and de-paradoxicalize the system will seem to the
system to be naturally necessary. An observer, in contrast, can recognize
the function of these semantic efforts and speculate about other, func-
tionally equivalent possibilities; to him, every specific semantic solution
to this problem appears historically determined and contingent, depen-

13
See also Y. Barel, Leparadoxe et Ie systeme: essai sur Ie fantastique social (1979) [Paradox and
System: Essay on the Fantastic in Society]; Barel, De la fermeture a I'ouverture en passant par
I'autonomie?, in L'AUTO-ORGANISATION, supra note 9, at 466-75 [From Closure to Openness by
Way of Autonomy?]. As a case study in the history of legal theory see Luhmann, Die Theorie der
Ordnung and die natiirlichen Rechte, 3 RECHTSHISTORISCHES JOURNAL 133 (1984) [Natural Rights
and The Theory of Order. Journal of Legal History].
14
See Lofgren, Some Foundational Views on General Systems and the Hempel Paradox, 4 INT'L
J. GENERAL SYS. 243 (1978).

145

HeinOnline — 83 Nw. U. L. Rev. 145 (1988-1989)

 NORTHWESTERN UNIVERSITY LAW REVIEW
dent on the supply of plausibility in the specific sociohistorical
circumstances.

C.
Finally, with the help of a general theory of self-referential auto-
poietic systems it is possible to connect systems theory to a theory of
evolution more adequately than before. What results is a weakening of
the concept of "adaptation" to the environment, a concept that cannot
adequately explain either the high degree of form constancy in natural
evolution nor the accompanying tempo of innovations. This is true for
the theory of the evolution of living systems, but even more true for the
theory of social evolution.
Special evolutionary paths become possible when the differentiation
of particular autopoietic systems is successful; for as soon as this occurs a
system can vary its structures, insofar as this is compatible with its con-
tinued self-reproduction. In constructing and altering structures, auto-
poietic systems can make use of contingent impulses from the
environment that occur and disappear again, as well as of errors in the
reproduction of their own operations. The possibilities are often re-
stricted more by the demands of internal consistency than by problems of
survival in the environment. In other words, very often a system fails to
make full use of the degrees of freedom the environment permits it and
restricts its own evolution to a greater degree than would be ecologically
necessary.15 Even with this modification to the theoretical apparatus of
classical Darwinism, however, it is still correct to characterize evolution
as an unplanned (not coordinated and in this sense making use of "acci-
dents") differentiation in variation, selection, and restabilization.
Accordingly, a theory of the evolution of law has to clarify two pri-
mary questions: (1) what problem leads to the differentiation of a partic-
ular evolution of law within a general social evolution, and (2) what is
the nature of the autopoiesis of law that allows it to be maintained even
when structural alterations take place? The answer to these questions
must start from the principle of variation, for a specific selection mecha-
nism can be formed only if the pertinent variation manifests specific
peculiarities.
The problem that gives rise to a special evolution of the law must lie
in uncertainty about whether expectations, and which expectations, can
be maintained, or at least be proven to be counterfactually justified, in
the case of conflict. This problem becomes relevant, if it was not so from
the outset, because a segmentary social structure establishes who is to be
on what side, who is to confirm claims, to take oaths, and if necessary, to
fight. The evolution of law then begins with the loosening of the struc-
15
On this point, with regard to a broadly conceived concept of organizations, see K. WEICK,
THE SOCIAL PSYCHOLOGY OF ORGANIZING (2d ed. 1979).

146

HeinOnline — 83 Nw. U. L. Rev. 146 (1988-1989)

83:136 (1989) Law as a Social System
tures of segmentary societies, and especially with the introduction of a
sufficient measure of uncertainty into social conflicts. For it then be-
comes a question of how this uncertainty is to be resolved, and selection
criteria can be developed for that. A certain independence of religious or
tribal political roles from preexisting ties of kinship or proximity was
probably decisive for this development. In any case, whether or not an-
evolution of the law is set in motion does not depend on the prior institu-
tionalization of the competence to make legally binding ("judicial") deci-
sions. Such an arrangement is still inconceivable in fairly well developed
late archaic societies, and presumably even for Mycenean culture. It pre-
supposes a critical mass of already existing, already evolved legal rules
that make it possible to think of this judicial competence as connected
with law. Thus, in theoretical terms, the autopoiesis of law, the produc-
tion of law by law, must already be possible for the central institution of
a court that makes binding decisions, the institution that in turn makes
possible the autopoiesis of law, to be possible. Evolution does not work
directly; it works epigenetically. Only in this way can innovations that
presuppose themselves arise. This is why contemporary observers give a
mythic or religious interpretation to the paradox of the asymmetry in the
origin of this circle: The Areopagus, for example, is instituted through
divine intervention. Or, God puts the law under the bush. Or later, and
in more civilized form, law is created by God in the form of human
nature.
Only when adequate differentiation in variation and selection has
been established, and when every legal claim is no longer both lawful and
unlawful at the same time, depending on the person involved, can criteria
for selection among selection criteria be developed. A long period of
practicing law and observing its transformations over time is required
before possibilities for distinguishing between selection and restabiliza-
tion arise. The legal system that has already evolved develops possibili-
ties for reflection, puts its own "justice" into question, and has recourse
to moral ideas in order to protect a subsistence economy (limiting tax
levies and debt collection, for instance, as in the reforms of the lawgiver
Solon in classical Athens).16 Religion and morality place limits on the
structure of argumentative justification in the law, and thereby also limit
the possibility of giving law the stability of tradition. In addition, due to
peculiarities of its organizational practice, the law can become so com-
plex that legal knowledge can no longer be taken for granted as part of
the normal knowledge of the aristocracy. Thus there arises a need for
special educational arrangements. We are familiar with the result of this
process in the concept of an institution (which originally meant "teach-
ing"). The function of stabilizing the law is transferred to processes of

16
For modern perspectives see J. SCOTT, THE MORAL ECONOMY OF THE PEASANT: REBEL-
LION AND SUBSISTENCE IN SOUTHEAST ASIA (1976); Thompson, The Moral Economy ofthe English
Crowd in the 18th Century, 50 PAST AND PRESENT 76 (1971).

147

HeinOnline — 83 Nw. U. L. Rev. 147 (1988-1989)

 NORTHWESTERN UNIVERSITY LAW REVIEW
doctrinalization and systematization, which then in turn outlast changes
in society by virtue of their own potential for innovation, which is inher-
ent in their concepts.
When the peculiarities of modern positive law are considered
against this background, it becomes obvious that in many respects this
kind of evolution no longer functions. Perhaps it is too slow for our
circumstances. At any rate, the impetus to variation no longer lies in
anticipating conflicts that can be expected; instead, the law regulates
modes of behavior that are themselves provided with the capacity for
conflict. The law itself creates the conflicts that it needs for its own
evolution, and thereby perfects its own autopoiesis. It ordains, for exam-
ple, that only a limited amount of wine is eligible for subsidy, and in
doing so it gives rise to problems that can in turn be fed into the legal
system as legal problems.17 As a consequence, the law evolves—there is
no question of planning here—so rapidly that traditional means of stabi-
lization no longer come into play. The law evades the control of doc-
trine. Nor can it any longer properly be described as a system of norms,
to say nothing of a system of "knowledge." At this point it can only be
described as a social system defined by its own code. Stabilization now
lies only in the positive character of legal validity—in the fact that spe-
cific norms are given force by decisions (whether it is the decision of the
legislator, the judge, or the current opinion of the commentators), and
have not yet been changed. For this reason the stability of the law must
be understood as something completely temporal, and objective questions
come into the picture only from the standpoint of complexity. They
make alterations difficult, and as a result, the law, despite its accelerated
tempo of change, remains by and large the same. Here we could also cite
Odo Marquard's remark about the functional concept of religion and the
possibility of religion being exchanged for any other functional
equivalent: Don't worry, given the complexity of the process, it would
take so long that we would always die before the exchange had been
completed.18

III.
Legal theory has found it difficult (and perhaps it always will) to
grasp this positive quality of the law in the absence of any conception of
an external (especially a moral) justification. The 19th century's attempt
to understand law as a guarantee of freedom (and that means freedom for
17
To show that the phenomenon is not so new, an example from the 18th century: For the sake
of the button makers, the law ordains that buttons may not be made out of the same fabric as the
clothing. See V. de Riqueti, Marquis de Mirabeau, 1'ami des hommes 90 (Paris 1883) [The Marquis
de Mirabeau, Friend of Man].
18
See Marquand, Religion und Skepsis, in Dm RELIGIOSE DIMENSION DER GESELLSCHAFT:
RELIGION UND IHRE THEORIEN 45 (P. Koslowski ed. 1985) [Religion and Skepticism, in The Reli-
gious Dimension of Society: Religion and Its Theories].

148

HeinOnline — 83 Nw. U. L. Rev. 148 (1988-1989)

83:136 (1989) Law as a Social System

irrational and immoral conduct) and thereby to accommodate it to the

disintegration of the traditional unity of reason and morality, did not
succeed. Even Kelsen still needs a fundamental norm, even if it is one
with the ambivalent status of an epistemological hypothesis. And for the
normal jurist, the idea that even good, pertinent arguments lead only to
the confirmation of argumentation itself—to the strengthening of its re-
dundancy—must still be completely unacceptable. In this situation the
theory of autopoietic systems offers at least the possibility of an adequate
description. Whether this description can be introduced into the legal
system itself (Le., used as its self-description) must be left an open ques-
tion (which means, left to evolution). In this situation the theory of auto-
poietic systems can only make use of its own autopoiesis as clearly as
possible.
At this point the question of the basis and the justification for legal
validity leads us to assume an escalating relationship between closedness
and openness in a system. Only as a self-referential closed system can the
legal system develop "responsiveness" to social interests.19 Viewed in
this way, evolution selects (on the level of organisms as well as on the
level of social systems) forms that permit greater complexity in combin-
ing closedness and openness. But that certainly does not mean better
adaptation to the powers that be; it does not mean more efficient
corruption.
A second, related point concerns the creative character of para-
doxes.20 The term "paradox" signifies here a phenomenon of observation
or description—that accepting a description has as its consequence the
acceptance of the opposite description. The observation of paradoxes,
something which occurs, for example, in the application of the code to
itself,21 blocks the system's observation and description, even though at
19
Responsiveness in the sense discussed in P. NONET & P. SELZNICK, LAW AND SOCIETY IN
TRANSITION (1978). See also Teubner, Substantive and Reflexive Elements in Modern Law, 17 LAW
AND SOC. REV. 249-284 (1983).
20
See K. Krippendorff, Paradox and Information, in 5 PROGRESS IN COMMUNICATION SCI-
ENCES 45-71 (B. Dervin & M. Voigt eds. 1984). The assumption of the fruitfulness of paradoxes
corresponds, moreover, to the old rhetorical practice of demonstrating, by turning customary opin-
ions and preferences into paradoxes, that their opposite also admits of truth—an undertaking that
can also render itself paradoxical. See for example the two works by Ortensio Lando, PARADOSSI,
ClOE SENTENTIE FUORI DEL COMMUN PARERE (Vinegia 1545) [Paradoxes, or Statements Be-
yond the Pale of Common Sense], and CONFUTATIONE DEL LIBRO DE PARADOSSI NUOVAMENTE
COMPOSTA, IN TRE ORATIONI DISTINTA (n.p., n.d.) [Refutation of the Recent Book of Paradoxes, in
three separate orations].
21
But it can occur in many other cases as well, where distinctions are used as justifications.
Thus, for example, every justification of the law in terms of transcendental theory must answer the
question whether the distinction between transcendental and empirical is itself transcendental or
empirical. If the distinction is a transcendental one, then the empirical is also something transcen-
dental. But the distinction itself may be an empirical one, since Kant, an empirical individual, intro-
duced the distinction at a specific time in history, namely at the end of the 18th century (in his
CRITIQUE OF PURE REASON).

149

HeinOnline — 83 Nw. U. L. Rev. 149 (1988-1989)

 NORTHWESTERN UNIVERSITY LAW REVIEW
the same time the observer must concede that the system's own auto-
poiesis is not blocked by the paradox. In other words, the system can
simultaneously both be observed and not be observed as a paradoxical
system. The observer must then transform this self-paradoxicalization
into a quality of his object by asking how the system de-paradoxicalizes
itself.22
These reflections hold both for outside observation and for self-ob-
servation. Consequently, they state the problem in such a way that soci-
ology and legal theory could collaborate on it. That would presuppose,
of course, that legal theory reconceived things it had previously taken for
granted and saw them now as functions of de-paradoxicalization, thus
making the transition from natural to artificial necessities.23 And such
reconceptualization will probably become possible only when sociology
can offer a good deal more theoretical certainty for this step into the
unknown, this illumination of what has been latent, than it has hitherto
been able to do.

22
See supra note 11.
23
See supra note 12.

150

HeinOnline — 83 Nw. U. L. Rev. 150 (1988-1989)