Autopoietic Law - A New Approach to Law and Society [Reprint 2011 ed.] 9783110876451, 9783110114591

Citation preview

EUI -

Series A - 8

Teubner (Ed.), Autopoietic Law: A New Approach to Law and Society

European University Institute Institut Universitaire Europeen Europäisches Hochschulinstitut Istituto Universitario Europeo

Series A Law/Droit/Recht/Diritto

8

Badia Fiesolana — Firenze

Autopoietic Law: A N e w Approach to Law and Society

Edited by

Gunther Teubner

W DE 1988 Walter de Gruyter · Berlin · New York

Library of Congress Cataloging-in-Publication

Data

Autopoietic law. (Series A — Law ; 8) Includes index and bibliography. 1. Law — Philosophy. 2. Sociological jurisprudence. 3. Order (Philosophy) I. Teubner, Gunther. II. Series. K235.A97 1987 340M15 87-30301 ISBN 0-89925-3903-0 (U.S.)

CAP- Kur^titelaufnähme

der Deutschen

Bibliothek

Autopoietic law — a new approach to law and society / ed. by Gunther Teubner. — Berlin ; New York : de Gruyter, 1987. (European University Institute : Ser. A, Law ; 8) ISBN 3-11-011459-3 NE: Teubner, Gunther [Hrsg.]; Istituto Universitario Europeo : European University Institute / A

© Copyright 1987 by Walter de Gruyter & Co., Berlin. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced in any form — by photoprint, microfilm, or any other means — nor transmitted nor translated into a machine language without written permission from the publisher. Dust Cover Design: Rudolf Hübler, Berlin. Setting: Arthur Collignon GmbH., Berlin. — Printing: Gerike GmbH., Berlin Binding: Verlagsbuchbinderei Dieter Mikolai, Berlin. Printed in Germany

Preface This book grew out of a conference held at the European University Institute in Florence. Legal theorists and legal sociologists from different countries had been asked to discuss the potential of the theory of autopoiesis for the analysis of law. Contributions that deal with issues of legal theory — legal system, autonomy of law, legal evolution and cognition in law — are presented in this volume. Contributions that deal with law's relation to politics and the economy will be published in a second volume, entitled "State, Law, Economy as Autopoietic Systems" (de Gruyter, Berlin, 1988). I would like to thank Joyce Reese for her precise and thorough editorial assistance, Ian Fraser for his sensitive translation of complex texts, and the Institute's Publications Officer, Brigitte Schwab, for her helpful activities in coordinating the publication. November 1987

Gunther Teubner

Table of Contents Preface

V

Bremen, Firenze Introduction to Autopoietic Law G U N T H E R TEUBNER,

Bielefeld The Unity of the Legal System

1

NIKLAS LUHMANN,

FRANCOIS E W A L D ,

12

Paris

The Law of Law

36

Paris On the Supposed Closure of Normative Systems

51

Bruxelles Between Order and Disorder: The Game of Law

70

Berlin Biological Metaphors in Legal Thought

97

JEAN-PIERRE DUPUY,

F R A N C I S OST,

HUBERT ROTTLEUTHNER,

Hannover The Communicative Autonomy of the Legal System

128

Ann Arbor The Autonomy of Law: Two Visions Compared

152

D A V I D N E L K E N , London Changing Paradigms in the Sociology of Law

191

Bremen, Firenze Evolution of Autopoietic Law

217

Bremen Perspectives on a Post-Modern Theory of Law: A Critique of Niklas Luhmann, 'The Unity of the Legal System'

242

HANS-GEORG DEGGAU,

R I C H A R D LEMPERT,

G U N T H E R TEUBNER,

KARL-HEINZ LADEUR,

Table of Contents

VIII

Stanford Accounting for Law THOMAS HELLER,

PATRICK NERHOT,

283

Montpellier, Firenze

The Fact of Law

312

Bielefeld Closure and Openness: On Reality in the World of Law

335

NIKLAS LUHMANN,

PETER KENNEALY,

Firenze

Talking About Autopoiesis — Order from Noise?

349

Authors' Biographical Sketches

369

Name Index

373

Subject Index

378

Introduction to Autopoietic Law G U N T H E R TEUBNER

Bremen, Firenze Is the practice of legal reasoning bound to end in "strange loops", "tangled hierarchies", and "reflexivity dilemmas" (Hofstadter, 1979: 692; 1985: 70)? Is the legal process nothing but a closed cycle of recurrent legal operations: "computation of computation of computation..." (von Foerster, 1981: 296)? And are the social dynamics of the legal system based upon the "paradoxes of self-reference" (Wormell, 1958; Quine, 1976)? Up to now, the intricate problems of self-referential relations have not been part of the discourse of lawyers; they have been discussed outside the law, in logic, linguistics, cybernetics and general systems theory. Now the theory of legal autopoiesis is importing the logic of self-referentiality into the legal world. Legal autopoiesis breaks a taboo in legal thinking — the taboo of circularity. Legal doctrine, legal theory and legal sociology have all regarded circularity as a subject not to be broached. Circular arguments have been viewed as petitio principii forbidden by the iron law of legal logic. Legal autopoiesis now presumes to invalidate this iron law by transferring circularity from the world of ideas to that of hard facts. The message is that circularity is not a flaw in legal thinking which ought to be avoided (Fletcher, 1985: 1263), but rather that the reality of law consists of a multitude of circular processes. Autopoiesis proposes, as a new and promising research strategy, to identify circular relationships within the legal system and to analyze their internal dynamics and their external interactions. What has been done so far in this field seems to be marginal. Legal methods are only slightly informed about the circular relation between purpose and norm in teleological reasoning (Alexy, 1978: 289 ff.); legal hermeneutics have only begun to study the perplexities of the hermeneutic circle in "Vorverständnis und Methodenwahl" (Esser, 1970); legal sociology has so far granted the luxury of circularity in feedback loops only between law and society. In the autopoietic perspective, these phenomena are viewed as only a few special cases in the encompassing circular self-referential reality of law. The whole legal system is seen as a dynamic cyclical reproduction of legal elements embedded in hypercyclical relations of legal structures and processes (see

2

Gunther Teubner

Teubner, 1987 a). Law, like other autopoietic systems, is nothing but an "endless dance of internal correlations in a closed network of interacting elements" (Maturana, 1982: 28). Circularity suggests closure. Thus, an autopoietic legal system is seen as operationally closed. Apparently, this runs against modern conceptions of an open responsive legal order, adapting to, and at the same time, shaping the social environment. Operational closure of law raises the suspicion of a new legal formalism, of a self-sustaining autarchy of law, of a new ideology for the legal profession — especially in the minds of legal sociologists. However, operational closure is only a half-truth. "L'ouvert s'appuye sur le ferme" (Morin, 1977: 197 ff.). A radical closure of the system — under certain conditions — means its radical openness. This is one of the most challenging theses of autopoietic theory. The more the legal system gains in operational closure and autonomy, the more it gains in openness toward social facts, political demands, social science theories, and human needs. This sounds paradoxical, and many formulations within this new theory seem to enjoy paradoxes. However, if it can be shown that this combination of openness and closure renders the legal system more responsive to social reality, and if concrete mechanisms coupling normative closure with cognitive openness can be identified, then legal autopoiesis has a potential of going beyond the theory of "open systems" which views itself as the legitimate successor of the older concepts of closed systems. This book launches an inquiry into the world of legal autopoiesis — it is a project in interdisciplinary theory building. Our collective effort is to transfer a new and complex theory, the theory of autopoietic systems, from the general context of systems theory to the context of legal theory. This is done with the cooperation of international experts in the fields of both legal theory and systems theory. The book represents the first major endeavor to develop this approach within the field of law. What sense is there for legal theory to make use of the theory of autopoietic systems? This newly developed theory, formulated by biologists (Maturana, 1970; Maturana et al., 1974; Varela, 1979; Zeleny, 1981; Roth and Schwegler, 1981) and transferred to the social sciences (Hejl, 1982 a; 1982 b; Luhmann, 1981; 1984; 1985; Teubner, 1985; 1987 a; 1987 b; Teubner and Willke, 1984) cannot yet claim with authority to be a fruitful paradigm. It is thus used in a more experimental manner as a strictly heuristic device. What follows for the problematic law and society relation if it is reformulated in terms of autopoiesis and self-referentiality? What hypotheses and consequences are implied with regard to doctrinal practice and legal policy? While the relations between law and other spheres of society will be dealt with in a separate volume (Teubner, 1987 c), the idee directrice of this book is to examine the potential of autopoiesis for legal theory and to reformulate fundamental legal concepts in the light of this theory. Although

Introduction to Autopoietic Law

3

it is tempting to start with a critical assessment of autopoiesis as such, it seems to be a more fruitful approach if we first explore the changes in legal theory suggested by the concept, and then follow up with a critical examination of the concept and its legal consequences. Thus, the book is organized as a debate between protagonists and critics of autopoiesis concentrating on the following issues: — — — —

Legal Legal Legal Legal

System Autonomy Change Cognition

1. Legal System What is autopoiesis? An autopoietic system produces and reproduces its own elements by the interaction of its elements. According to Niklas Luhmann, the main representative of social autopoiesis, the decisive innovation in comparison to older theories of self-organization is that certain systems are capable not only of creating an autonomous order but of creating their own elements as well. Autopoietic systems are not only identified in biology as cells and organisms, on the basis of life, but they can also be identified as social systems (interaction, organization, society), on the basis of meaning. The basic element of a social system is communication — not the human being. Communication as the unity of utterance, information and understanding constitutes social systems by recursively reproducing communication. This conceptualization of social systems brings about three important changes in social theory: (1) a radical temporalization of social systems — this means that the elements are not "stable" but are "events" which necessitate the system to recursively use events to produce events; (2) a reformulation of the old maintenance problem: what has to be maintained is the recursively closed organization of an open system; closure as a condition for openness is one of the central tenets of autopoiesis; (3) epistemological consequences: for the theory of autopoiesis, observation is an activity which can be performed by living systems, psychic systems and social systems. Observation is itself an operation of an autopoietic system. If society is an autopoietic system of recursively self-reproducing communication, can one then also identify the law as being an autopoietic system of its own? In other words: Is there a legal autopoiesis, beyond social autopoiesis? According to Luhmann, in his essay, "The Unity of the Legal System", 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

4

Gunther Teubner

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. Circularity replaces extra-legal foundations of law as does Kelsen's fictitious Grundnorm. If the recursive reproduction of legal acts constitutes legal autopoiesis, then operational closure is the main condition for its environmental openness. This interplay of openness and closure is represented in the legal system by the combination of normative closure and cognitive openness. The important consequence is that only a limited scope exists for the combination of normative expectations with cognitive expectations. This, according to Luhmann, is the key to the explanation of the modern crisis phenomena of law. Finally, autopoiesis of law limits, to a considerable degree, the variability of legal functions. The autopoietic closure sets effective limits to the political instrumentalization of law. What is the significance of legal autopoiesis in the history of legal thought and what is its role in modern society? "The Law of Law" is Francois Hwald\ answer to this question by which he attempts a contextual interpretation of legal autopoiesis. He deals with two issues: What role does autopoiesis play in the conflict between legal theory and legal sociology? Does autopoiesis supply an adequate theory for the law of the modern age which Ewald calls the "social law"? In the contemporary dilemmatic schism between formal analytical theories of law and more empirically-based legal sociology, legal autopoiesis appears in Ewald's view as a kind of "Columbus' egg". Legal autopoiesis is at the same time post-Kelsen and post-sociological. Both the internal viewpoint of the Kelsenian tradition which has been unable to account for differences between legal systems, and the critical project that inspired the external viewpoint of legal sociology are in deep crisis today. In this historical situation of legal thought, autopoiesis is emerging as an ambitious attempt to reconcile the antagonistic views. Autopoiesis transcends and conserves at the same time the split between pure theory and sociology. The view of law as a self-reproductive system of communication, the legal act as the element of the legal system, the combination of normative closure and cognitive openness — these features of legal autopoiesis give it a chance, Ewald maintains, to overcome the contemporary crisis in legal theory and sociology. In Ewald's view the significance of legal autopoiesis is not confined to its role in legal thought. What is at stake is an adequate theory of law under conditions of modernity which Ewald characterizes by the destruction of natural law concepts, the loss of universalism, the particularization of political sovereignty and the relativization of science. Under these circumstances law can no longer be founded on principles outside itself. Legal practice in the twentieth century, according to Ewald, has found a solution

Introduction to Autopoietic Law

5

to this need for self-foundation: the formulation of principles of law based exclusively on legal reflexivity. Legal autopoiesis may well serve as a theoretical re-formulation of those factual developments in modern legal practice which Ewald describes as "social law". But is autopoiesis in a position to act as a critique of the legality of social law? Ewald argues that as long as autopoiesis seeks to develop in the traditional forms of theoretical discourse, it is condemned to reformulate empty tautologies. The practice of social law is in need of a self-description and self-critique that makes objectivity of legal judgment possible but avoids at the same time universal statements: objectivity without universality. Ewald asserts that legal autopoiesis must take on a "resolutely atheoretical form" if it is to become the critical legal philosophy of our time. Jean-Pierre Dupuy continues these efforts to locate legal autopoiesis in the context of contemporary social and legal philosophy by confronting it with the normative theories of Robert Nozick and Friedrich von Hayek. He takes up the paradox of closure and openness in autopoietic systems and supplies three different interpretations of how the legal system can be open and closed at the same time: (a) closure/openness refer to different domains of the legal system (normatively closed, but cognitively open); (b) legal closure implies legal openness ("order from noise"); and (c) selftranscendence of a normative order, the distancing of law in relation to itself, what he considers to be the most subtle form of openness. His main thesis is that a unit is self-transcending insofar as the complexity of the behavior of which it is capable is infinitely superior to the control capacities of the unit. In this perspective, Dupuy analyzes Nozick's theory of entitlement with its self-referential character of pure procedural justice and Hayek's theory of law as a "spontaneous" social order. In both theories, he identifies important parallels to the constructs of self-reference and autopoiesis. Dupuy ends with a paradoxical interpretation of legal autopoiesis locating it at the "height of modernity": If, at the same time, one sees the irreducible Grundlosigkeit of law and its very necessity, the "squaring of the circle" is to posit that law is self-legitimizing and self-founding. "The autonomy of the legal system is, finally, undecidably, decided". The dialectics of legal openness and closure are taken up by Francois Ost from a different angle — from the angle of the individual actor in the perspective of game theory. He criticizes the autopoietic paradigm in its almost autistic closure by contrasting it with three positions. (1) Ashby's theorem: A program programming itself is a logical impossibility; changes must be produced from the outside; (2) Atlan's auto-organisation relative·. according to the "order from noise" principle the internally undetermined system transforms external perturbations into an organizational order; (3) Castoriadis' societe autonome·, a society is autonomous when it is constantly able to free itself of the laws, values and traditions which it has imposed on itself. Ost develops a model of le jeu du droit in which closure is

6

Gunther Teubner

represented by regulation and convention while openness and dynamics are introduced into the system by individual actors and their strategies. One of the crucial steps in legal autopoiesis is the transfer of biological models into social and legal spheres. Hubert Rottleuthner analyzes and criticizes the use of "Biological Metaphors in Legal Thought". He asks the question: How are concepts changed when they are transferred from one field to another and how does a biological metaphor allow the solution of major problems in social and legal contexts? He begins with an historical analysis of the conceptual transfer describing how the concepts of "organism" and "evolution" have been used in legal thinking. He then critically assesses the transplantation of autopoiesis into the field of law. He identifies three areas in which the concept of autopoiesis can help solve specific problems: (1) A new concept of law can be explicated which would enable hitherto disparate elements of legal theory to be integrated or seen in a different light. In this context, Rottleuthner examines the concept of autopoiesis by referring to the legal theories of Kant, Kelsen and Hart; (2) Legal regulation processes can be explained by pointing to the limits of law's capacity for social control and identifying information requirements for legal control. According to Rottleuthner, autopoiesis helps identify autonomous social structure and organization. What is missing, however, is a more concrete analysis of regulated systems; (3) Theories of evolution are reinterpreted in terms of social and legal change. Autopoiesis stresses the role of an internal evolution of law. Again, Rottleuthner insists on a more empirically-oriented formulation of the theory of autopoiesis, which would lead to the identification of concrete evolutionary mechanisms.

2. Legal Autonomy Legal autopoiesis is probably most controversial in its insistence on legal autonomy. After all, modern law is shaped by external constraints, social pressures and political decisions. How can one, after the social science revolution in law, after sociological jurisprudence and legal economics, still describe the legal system as autonomous (Friedman, 1985)? Hans-Georg Deggau, in his essay on the "Communicative Autonomy of the Legal System", carefully elaborates the conditions of the possibility of legal autonomy based upon legal communications as elementary units of a legal system. Deggau identifies legal autonomy in the "conditionalized legal normativity" which is capable of producing the necessary surplus for autopoietic closure. Normativity, as the basis of legal autonomy, produces, at the same time, the specific relation of law to other social systems, since, in its normative structures, law has a "structural affinity" towards other social systems. The very function of law is to congruently generalize normative structures in its social environment.

Introduction to Autopoietic Law

7

This is a concept of legal autonomy which stands in sharp contrast to classical socio-legal assumptions about the interrelations between law and society. It is precisely this difference which Richard Lempert focuses on in his critique of legal autopoiesis. He constructs two models of legal autonomy. An "Anglo-American vision" as an empirically-oriented theory about actual operations of law is contrasted with a "Continental vision" which is abstract — developed by logic, analogy and a firm a priori theoretical commitment to a model. While the latter stems from a general concept of autopoietic legal autonomy as a radical operational closure, the AngloAmerican vision is able to distinguish varying degrees of autonomy by defining empirically testable criteria of relative autonomy. From this clearcut juxtaposition it becomes evident that autonomy of the law in the autopoietic sense is either "palpably wrong" or "trivial". This is a harsh statement. Perhaps it tells something about the relationship between "autopoiesis" and "law and society" as the well-known incompatibility between competing pradigms. This is precisely the point where David Nelken picks up the argument. He interprets the controversy between Luhmann and Lempert on legal autonomy as a conflict indicative of a paradigm change. This enables him to point to radical differences in the conception of legal autonomy, in regard to its theoretical framework, its characteristic meaning, its policy concerns, and methods of investigation. His main point, however, is to show that legal autopoiesis represents a full-fledged paradigm change from mainstream law and society. For this purpose, Nelken draws an intellectual map of competing socio-legal approaches placing "law and society" at the intersection of two axes, one dealing with the relationship between law and society (conceived ontologically), the other with that between law and sociology (conceived as epistemologies). Nelken sees the great merit of autopoiesis theory to collapse these axes into each other by showing how the extremes do in fact meet up. Two consequences are important: autopoietic theory shows how, where and why law, society and sociology actually became distinguished from each other as an evolutionary outcome of functional differentiation. At the same time autopoiesis theory resists characterizing law as a mere passive object of the sociological gaze and treats it rather as a subject reproducing its own conceptions of itself and of society.

3. Legal Change If the identity of an autopoietic system is defined by its organizational closure, by the recursive reproduction of its elements, then autopoiesis appears to be resistant to change, learning and evolution. However, if one takes into account the fundamental distinction between autopoietic

8

Gunther Teubner

organization, which is kept constant, and concrete structures of the system, which can change within the boundaries of the autopoietic organization (Maturana, 1982: 282 f.), then it becomes clear that autopoiesis does not exclude evolution but implies a redefinition of evolution. Gunther Teubner, in his essay on "Evolution of Autopoietic Law", attempts to resolve two problems of autopoietic evolution. How does the legal system evolve into autopoietic closure? How does legal evolution operate after an autopoietic closure of the legal system? The first question of pre-autopoietic evolution is answered with reference to the construction of a "hypercycle" (Eigen and Schuster, 1979). It is by mechanisms of "blind" evolution that "socially diffuse law" develops higher forms of autonomy via the cyclical constitution of its system components. It gains autopoietic closure by the hypercyclical connection between them. The answer to the second question of postautopoietic evolution is "internalization". Autopoietic law is separated from the general evolution of its social environment and develops internal legal mechanisms for the evolutionary functions of variation, selection and retention. At the same time, legal development is coupled to broader social developments by specific mechanisms of co-evolution. Karl-Hein£ Ladeur's contribution is also concerned with legal change, focusing on the adaptability of law under modern conditions. Ladeur asks if and to what degree the functions of modern law, understood as an autopoietic system, can vary. He criticizes Luhmann's version of legal autopoiesis as too static in regard to the variability of functions, and develops a three-level-model of the relationship between legal structures and functions. Social functions of law change through the interaction of those three levels: (I) the organized level (structure — subsumption model in the classical sense); (II) the level of the "organizing function" (controlling transformations on level I); (III) the level of "virtualizing" structure and function. On this level potential destabilization leads to "strange loops" intertwining all three levels with the result that fluctuations of level I, being mitigated on level II, are strengthened on level III and lead to a change of function of level II. More concretely, Ladeur observes not only structural but functional changes in modern law which point to an emerging function of law that could consist in securing an abundance of alternatives, openness, flexibility for a new legal rationality of bargaining and compatibilization as opposed to the classical function of law which consisted in congruently generalizing social expectations. While Ladeur and Teubner stress the internal dynamics of legal evolution, Thomas Heller focuses on the co-evolution of different autopoietic systems, more precisely, on the evolutionary dynamics in the interaction between law, economy and culture. Drawing on the Hispano-American biculturalism controversy in the United States, Heller stresses the intersystemic origins of perturbations of the legal system and develops a model of autopoietic co-evolution, based on the assumptions of self-referential reproduction of

Introduction to Autopoietic Law

9

core practices, system closure, instability and locality. The goals of the analysis are: (1) to define precisely the reproductive autopoietic core in the legal system; (2) to describe the interactive effects among co-existent systems, law, economy and culture; (3) to explain the internal dynamics of self-reproduction that create, at the same time, variation under the influence of external perturbations. Heller construes a multitude of cyclically organized system-internal "matching processes". Under the influence of environmental perturbations, the legal system develops a heuristic model of processed information as legal claims and matches it against existing legal practices. In the negative case this leads to the development of "causal models" and even "causal models of causal models" which in turn are used for reworking the system's prefiguration rules. Heller discusses the political relevance of such a cyclical model as a problem of "the politics of autopoiesis". In opposition to both neo-liberal and structuralist theories, autopoiesis is situated in a post-structuralist discourse, emphasizing the complex, the local, the closed and the unstable against global, coherent, artificing and equilibrating mechanisms.

4. Legal Cognition Legal autopoiesis suggests to rethink questions of legal epistemology, especially in its insistence on normative closure and cognitive openness. "The Fact of Law" is one of the most important epistemological questions that Patrick Nerhot takes up. He contrasts traditional realist positions in law, especially positions of legal positivism, with a "radically constructivist" position which is suggested by legal autopoiesis. Legal facts are not imported into the law from the outside, they do not represent an independent reality, but they are constructed within the law by operations of the legal system. This view does not render the distinction between law and fact meaningless, but it reinterprets and connects this distinction to the difference between normative and cognitive operations. Nik/as Luhmann, in his concluding essay "Closure and Openness: On Reality in the World of Law", takes up themes of legal epistemology in order to answer to the objections which have been raised during the proceedings of the conference. The main issues are: How can one dissolve the paradox that law is at the same time an open system and a closed system? How can one insist on legal autonomy and stress at the same time its inter-dependencies with society at large? How is co-evolution of law and society conceivable under the premise that law as well as society are autopoietic systems of their own? How can one account for the apparent contradiction that the legal system is constructing its own reality by

10

Gunther Teubner

recursive legal operations and is at the same time exposed to the constraints of extra-legal reality? Luhmann approaches these interrelated questions first by posing a dilemma and second by proposing a solution. The theory of self-referential systems, he maintains, leads inevitably to the following dilemma: on the one hand, no system is in a position to operate outside its boundaries; no system has a direct access to reality "out there". On the other hand, the rate of structural evolution requires the assumption that a system's environment is non-random and structured and produces effective constraints for the system. An epistemological theory must be in a position to resolve this dilemma. Luhmann's proposal consists in a distinction. He distinguishes between internal information processing and external constraints via a materiality continuum and via simultaneous presence of events. Luhmann insists on the strictly system-specific and exclusively internal character of information processing. A transfer of information from the environment to the system and vice versa is impossible. Via information the environment is inaccessible to the system. However, the environment effects its presence upon the system by imposing drastic constraints which limit the system's operational potential. Luhmann identifies two mechanisms of environmental coupling. One mechanism he calls materiality continuum. It is the material-energetic "basis" of meaning systems which exerts considerable constraints on them. The second mechanism is the simultaneous presence of events in several meaning processing systems. This allows for the interpenetration of autonomously operating mechanisms. With this dualism of information and interpenetration, Luhmann attempts to answer the above-mentioned questions. Legal operations cannot reach beyond the boundaries of law. In this respect law is a closed system. In its cognitive operations law still can operate only within its boundaries, but it opens itself to its environment by constructing a legal reality. Beyond those mechanisms of normative closure and cognitive openness, the law is coupled to its social environment via mechanisms of interpenetration. Law participates in language structures and reality constructions of general social communication. And it is connected to other spheres of society by the simultaneous presence of social events which are selectively processed in law and in its environmental subsystems. This implies a new understanding of legal autonomy. Autonomy is no longer internal causal determination of the system's operations. Legal autonomy is the selective combination of legal operations with legal operations. This autonomy is compatible with political and economic pressures on the legal system. Autonomy is in danger only once the legal coding as such is in danger of being replaced by criteria of economic utility and political expediency.

Introduction to Autopoietic Law

11

Bibliography Theorie der juristischen Argumentation. Die Theorie des rationalen Diskurses als Theorie der juristischen Begründung. Frankfurt a. M.: Suhrkamp. ESSER, JOSEF ( 1 9 7 0 ) Vorverständnis und Methodenwahl in der Rechtsfindung: Rationalitätsgarantien der richterlichen Entscheidungspraxis. Frankfurt a. M.: Athenäum. E I G E N , M A N F R E D and PETER SCHUSTER ( 1 9 7 9 ) The Hypercycle: A Principle of Natural Selforgani^ation. Berlin: Springer. FLETCHER, GEORGE P. ( 1 9 8 5 ) "Paradoxes in Legal Thought", 8 5 Columbia Law Review A L E X Y , ROBERT ( 1 9 7 8 )

1263.

(1981) Observing Systems. Seaside: Intersystems Publications. Total justice. New York: Russell Sage. H E J L , PETER M. (1982 a) "Die Theorie autopoietischer Systeme. Perspektiven für die soziale Systemtheorie", 13 Rechtstheorie 45. — (1982 b) So^ialwissenschaft als Theorie selbstreferentieller Systeme. Frankfurt a. M.: Campus. HOFSTADTER, DOUGLAS R . ( 1 9 7 9 ) Gödel, Escher, Bach: An Eternal Golden Braid. New York: Basic Books. — (1985) "Nomic: A Self-Modifying Game Based on Reflexivity in Law" in D. Hofstadter, Metamagical Themes. New York: Bantam. L U H M A N N , N I K L A S (1981) "Evolution des Rechts" in N . Luhmann, Ausdifferen^ierung des Rechtssystems. Beiträge %ur Rechtsso^iologie und Rechtstheorie. Frankfurt a. M.: Suhrkamp, — (1984) Sociale Systeme. Grundriß einer allgemeinen Theorie. Frankfurt a. M.: Suhrkamp. — (1985) "The Self-Production of Law and its Limits" in G. Teubner (ed.) Dilemmas of Law in the Welfare State. Berlin: de Gruyter. M A T U R A N A , HUMBERTO R . (1970) Biology of Cognition. Urbana. — (1982) Erkennen: Die Organisation und Verkörperung von Wirklichkeit. Braunschweig: Vieweg. M A T U R A N A , HUMBERTO R., F R A N C I S C O E . V A R E L A and R. U R I B E ( 1 9 7 4 ) "Autopoiesis: The Organization of Living Systems, Its Characterisation and a Model", 5 Bio Systems

FOERSTER, HEINZ VON

FRIEDMAN, L A W R E N C E ( 1 9 8 5 )

187.

La methode, Vol. 1 , La nature de la nature. Paris: Seuil. The Ways of Paradox. Cambridge: Harvard University Press. R O T H , G E R H A R D and HELMUT S C H W E G L E R (eds.) ( 1 9 8 1 ) Self-organising Systems. An Interdisciplinary Approach. Frankfurt a. Μ.: Campus. TEUBNER, GUNTHER ( 1 9 8 5 ) "After Legal Instrumentalism? Strategic Models of PostRegulatory Law" in G. Teubner (ed.), Dilemmas of Law in the Welfare State. Berlin: de Gruyter. — (1987 a) "Hypercycle in Law and Organization", European Yearbook for Legal Sociology. — (1987 b) "Social Order from Legislative Noise" in G. Teubner (ed.), State, Law, Economy as Autopoietic Systems. Berlin: de Gruyter. — (1987 c) State, Law, Economy as Autopoietic Systems. Berlin: de Gruyter (forthcoming). TEUBNER, G U N T H E R and H E L M U T W I L L K E ( 1 9 8 4 ) "Kontext und Autonomie. Gesellschaftliche Selbststeuerung durch reflexives Recht", 5 Zeitschrift für Rechtsso^iologie 4. V A R E L A , F R A N C I S C O J. (1979) Principles of Biological Autonomy. New York: Elsevier. WORMELL, C. P. (1958) "On the Paradoxes of Self-Reference", 67 Mind 267. ZELENY, M I L A N (1981) Autopoiesis: A Theory of Living Organisation. New York: North Holland. MORIN, EDGAR ( 1 9 7 7 )

QUINE, W I L L A R D V. ( 1 9 7 6 )

The Unity of the Legal System NIKLAS LUHMANN

Bielefeld

I. Whenever the word "system" is used, the unity of the system is usually tacitly assumed. An older epistemological tradition defined a system explicitly as a construction from a principle. 1 Neo-Kantianism in particular took up this idea and developed it further, defining it as the appropriate way of ordering legality of objects which lay claim not to empirical existence but to validity (see for example Liebert, 1920: 108 ff.). Jurisprudence has remained true to this interpretation, though the unity of knowledge has now been replaced by the unity of state and nation. 2 Coing for example writes that "by a system we mean the ordering of knowledge (!) according to a uniform perspective" (1956: 26). Insofar as the science of law regards itself not simply as jurisprudence but as the particular science of valid objects, it seems an obvious step to combine this claim with the theory of validity and thus to adopt the neo-Kantian definition of system. But, as jurists realize perfectly well there is no equivalent for this in the actuality of the practice of law. 3 In system-theoretical research it is not even men1

2

3

In this context, unity was often understood as "purpose", as in the Critique of Practical Reason. In Grimm's Dictionary (1942: col. 1433) both these elements are included in the definition: "a meaningfully structured whole, the parts of which are connected for a purpose or are ordered together under a higher principle, an idea or a law". But Schleiermacher's interpretation differed considerably from this definition, particularly — and characteristically — with reference to the problem of contingency. "The purpose moves further away from the idea [i. e., the inner unity of the work, N. L.] the more arbitrary its production is" (Schleiermacher, 1977: 175). Cf. Engisch (1936). This position is by no means as extreme and as isolated as it may appear in a political analysis after the event. The interpretation of the legal order as an institution (Dürkheim, Duguit, Santi Romano) also attributed its unity to the extra-legal fact of society, and in German sociology after 1933 the concept of society was simply replaced by that of Volk (people, nation). Accordingly the idea of unity plays a less important part in more recent publications on legal theory, where greater stress is laid on the method and technique of systematization. Cf. for example Pawlowski (1981: 220ff.); Krawietz (1981: 2 9 9 - 3 3 5 ) .

The Unity of the Legal System

13

tioned. The unity of the system is simply assumed and no thought is given to the question of how the system acquires this unity. This is an unsatisfactory state of affairs. It seems to lead to a situation where sociologists leave it to jurists and jurists leave it to sociologists to formulate theories on the unity of law. Jurists basically content themselves with considerations of consistency, i. e., interpret unity as (relative) interdependence of decisions. Unity is posited to the extent that a change in legal practice would involve too many other changes to remain practicable. The sociology of law is interested only in correlations between legal and extra-legal variables and, though it may talk of a legal system, it never clearly perceives the unity of this system.4 Thus many relevant and widely discussed problems remain unclarified: for example the problems of the demarcation (or fusion?) of law and politics, of the implications of social engineering in law, of the relevance of result orientation or of that new critique of law which argues that there is too much law. These new theoretical approaches as well as the traditional approaches are hampered by the fact that no clear concept of the unity of law exists, indeed that one cannot be sure that this is a conceptual problem at all, or to what extent it is one. Nor is it certain that there is one (and only one) concept which could express the unity of an object. Systems theory is certainly not necessarily the only way of tackling the question of the unity of law. But as no other methods can be seriously contemplated at the moment, we will work on the assumption that the unity of law must be realized and we will therefore examine how the unity of the legal system is to be interpreted.

II. New departures in systems theory have sought an answer to precisely this problem. In this connection they use the idea of self-reference or ideas derived from it. What is proposed amounts in fact to a theory of selfreferential systems. Here the traditional linking of the idea of self-reference to consciousness as the basis of operation is abandoned. Thus the theory of the subject-ness of consciousness (in the sense of subiectum, hypokeimenon) and with it the primacy of the epistemological difference between subject and object are rejected. Instead, two kinds of operations are distinguished, i. e., self-reproduction and observation. Observation in this context may mean both self-observation and observation of others. If one connects this still embryonic theoretical development with the old question of unity in multiplicity (to which the theory of the transcendentality of consciousness sought a possible answer) then one can define self4

See for example Friedman (1975); Aubert (1983: 28). Aubert explicitly states that a concept of law and a precise determination of the limits of the legal system are of no sociological interest.

14

Niklas Luhmann

referential systems as systems which themselves produce as unity everything which they use as unity. The resultant widening of the theoretical framework is underlined by the fact that both "produce" and "unity" have manifold meanings. "Produce" has very different meanings depending on whether it refers to chemical, organic, conscious, or communicative processes (i.e., chemical, organic, mental or social systems). 5 "Unity" means not only unity of the system itself but also and above all: unity of the final elements of which the system consists and unity of the processes into which the operations of this system combine these elements. The innovative aspect of this theory is the extension of the concept of unity to cover the constitution of the elements of which the system consists. This theoretical advance is underlined by the increasing use of the term "autopoietic" systems, "autopoiesis" being a term originally coined by Maturana (see Maturana, 1982). It is no longer merely a matter of "existence" or "preservation of existence" 6 nor merely of problems of selforganization which relate only to the structural level. 7 It is a question of the elements of which the system consists. The crucial insight is that elements presuppose enormous complexity in terms of their energy/matter basis but nonetheless function within systems as indissoluble units in terms of the respective system — for example as molecules, cells, ideas, actions. The indissoluble unity of an element for the system can be constituted only by the system; it follows from its capacity to associate which is produced within the system. An autopoietic system therefore constitutes the elements of which it consists through the elements of which it consists. In doing so it sets limits which do not exist in the substructure complexity of the environment of the system. 8 5

6

7

8

The extensibility of the theory of self-referential systems beyond this overall area is a matter of controversy. Whereas it was traditionally limited to consciousness, the tendency today is to limit it to living systems (or even only to cells in their genetic context). See for example Varela (1979; 1981: 3 6 - 4 9 ) . Nor is it an extension of the irreducible doubling of self-preservation and selfconsciousness — which Henrich regards as the basic structure of modern philosophy. See Henrich (1976: 9 7 - 1 4 3 , especially llOff.). On this discussion, which as started in the 1950s and 1960s by speculations on selfprogramming by computers, cf. Yovits and Cameron (1960); Yovits, Jacobi, and Goldstein (1962); von Foerster and Zopf (1962). Even in contemporary literature, self-reference is still seen largely in terms of self-organization. A representative work in this respect is Jansch (1980). Self-organization, however, always means the capacity of systems to change their o w n structures on the basis of their o w n experience with their environment. Accordingly, Maturana formulates as follows: "We maintain that there are systems that are defined as unities as networks of productions of components that: (1) recursively, through their interactions, generate and realize the networks that produces them; and (2) constitute, in the space in which they exist, the boundaries of this network as components that participate in the realization of the network" (Maturana, 1981: 21). However, the formulation does not clarify what I am primarily concerned with in the text above: the production of the unity of the element.

The Unity of the Legal System

15

In the sphere of action systems, Parsons explicitly queried the constitution of the "unit act" but perceived only two possible answers, one subjectivevoluntaristic, the other cognitive-analytical (1937: 43 ff.). The more evident it became in the development of the theory that the personal system can only contribute one component to action, the greater was the compulsion to adopt a "purely analytical" theory of the system of action. The decomposition of action and the conception of the emergent unity of action based on it was an act which could only be performed by an outside observer (analyst). This theory, with its unclarified epistemological foundations, thus amounted to a kind of "arranged marriage" between analysis and reality. It is precisely at this point that the theory of self-referential systems moves into a new and different starting position. Even a system of action consists only of elements which are formed by the operations of the elements of the system itself. It therefore cannot, as Parsons pointed out, decompose its elements itself. But (in contrast to Parsons' position) there is no point in entrusting this task to an analyst because in each decomposition of the elements he would fail to realize that the system produces and reproduces itself.9 In contrast to some previous attempts to introduce the idea of autopoiesis into the sphere of sociology, we regard autopoiesis not only as a new explanation of the individuality of organic systems or systems of consciousness (cf. Hejl, 1982a; 1982b: 45-88). 1 0 If this were the case, the concept would contribute little to the analysis of social systems. In fact, social systems can themselves be regarded as special kinds of autopoietic systems. Only thus can one arrive at an explanation of the unity of a social system (in this case the legal system) by the autopoiesis of the unity of precisely this system. Perhaps the most important result of this movement towards a theory of self-referential systems is a new conception of the closed nature of these systems. This closure does not, however, mean the absence of an environment nor does it mean complete determination by itself.11 Rather, closure consists in the fact that all operations always reproduce the system. For sense-systems this can also be expressed as follows: the system, with every operation, controls its own possibilities of negation and thus performs precisely what Heinz von Foerster describes as computation of computa9

10

11

It is obvious that here not only analytical reductionism but every reduction of selfreferential systems to another level of reality is rejected. Ergo: no reduction of consciousness to life, no reduction of life to chemical processes, etc. Dias' attempt to apply the idea of autopoiesis to the judicial decision-making process presupposes its subjective bases and treats the legal system only as a "consensual domain" of subjective will-formation but not as an autopoietic system in its o w n right (Dias, 1980: 2 5 7 - 8 2 ) . The older system-theoretical tradition was always based solely on this interpretation. See Hall and Fagen (1956: 25 ff.).

16

Niklas Luhmann

tion. 12 In other words, the system must have at its disposal the code difference between "yes" and "no", and must always be able to re-negate the constantly concomitant, constantly implied negations. The law-code of legal and illegal is merely an application of this very general principle and juristic reality is nothing but the correlate of a self-referential mode of operation which reproduces itself according to this code.

III.

What, then, is this elemental unity, this "unit act", this non-decomposable quasi-atom the production of which — in a closed, recursive procedure — constitutes the legal system? In a theory of autopoietic systems in which the unity of the system is simply the recursive closure of the production of the elements of the system by the elements of the system, everything depends on the answer to this question. However, we cannot immediately give an unequivocal answer but must first introduce a clarifying distinction; namely, a distinction between communication and action and, correspondingly, between reproduction and self-observation] self-description of the system. The reason for this difficulty lies in the general theory of social systems (cf. Luhmann, 1982: 366 — 79). Social systems can only reproduce themselves by (always self-referential) communication. It is only by means of understanding (i. e., through a "second mind") that a communication system reproduces itself. Here, because it is a question of the openness of the situation for acceptance or rejection, the smallest still negatable sense unit functions as an element. But at the same time this process necessitates a reduction to action, which determines who can continue to communicate with whom. 13 A powerfully simplifying selfobservation or self-description of the system by itself must therefore be introduced into the self-reproducing social system in order that sufficiently simplified connections (partners, subjects, information, communications, and in all these respects, the corresponding omissions) for the formation of further elements can be pinpointed sufficiently rapidly. This is, as stated, the necessity of constant self-continuation. Without it the system would simply cease to exist. 12

13

Or simply as "computation". See von Foerster (1973: 35—46). Reality according to this view is simply that which makes this process — as self-restriction — possible. The argument is neither logically nor theoretically completely persuasive. As always with function data, other, functionally equivalent reductions cannot be ruled out. But reduction to action has become so widespread and prevalent that even sociology scarcely perceives this as a problem but interprets social systems simply as systems of action. Frequently the error is made — a fatally obvious linguistic error — of interpreting communication simply as a kind of action and thereby including communication in a general theory of action. See for example Moles and Rohmer (1977: 15 ff.).

The Unity of the Legal System

17

If one interprets the legal system as a social system, one finds the same relations here. Law exists only as communication (or, in psychological terms, as the prospect of communication). Communication here means a synthesis of information, communication and comprehension, and not merely the action of communication as such. The differentiating-out of law therefore lies first of all in the thematic control of communicative processes. Every meaning, even the meaning of purely factual events, can become legally relevant, i. e., can enter into the self-reproduction of the legal system. The practice of jurisprudence calls for knowledge of the world. But within this sphere of relevance there exists the core of legally valid action which changes the legal position, engenders legal consequences, and thereby makes possible new normative expectations which would not have acquired a legal quality without this action of engendering. The legal system itself determines what kind of events have this effect. It is impossible to identify these events without a knowledge of the legal system. The constitution of these elements is an autonomous achievement of the legal system which occurs in the process of self-observation and self-description of this system. The related dogmatic questions of a juristic theory of action, of problems of omission, attribution, etc., cannot be dealt with here. We wish only to propose the argument that the differentiation of the legal system requires the universalization of possible relevances and it also requires that the system describe itself as an action system. 14 The main theoretical consequence of this is that events can only maintain the quality of elementary unity of the legal system when they change the legal position. The reason for the attribution of unity is that through it the difference between continuity and discontinuity can be operationalized and it is therefore normally possible to determine easily and rapidly enough what has and what has not changed as a result of a certain legal action. The function of formation of unity — always presupposing differentiation — lies in the autopoiesis of the system, i. e., as in all systems in the both continuing and discontinuing production of ever new elements. 15 This means that the self-reproduction of law takes the form of change of law, of the transference of the quality of normative validity to partially new expectations. Law, therefore, finds itself in constant trivial variation, 14

15

Particularly interesting in this context are the threshold cases of birth, giving birth, and death, which, although they change the legal position, can neither be understood as legal acts nor do they owe their relevance to other acts. It could be said that they owe their legal relevance to a decision by the legislator. (Note the formulation of Para. 1 of the German Civil Code (BGB) which is concerned only with dating.) The option of interpreting law as the codification of over-positive human rights is at least left open. Analyses which concentrate radically on the temporal dimension are rare not only in legal-theoretical but also in sociological literature. However, one example of such an approach is 0sterberg (1976, especially 64 ff.).

18

Niklas L u h m a n n

and the proven major forms such as the contract and the statute are merely differentiated forms of this state. The positivity of law is its autopoiesis and, precisely because of this, divergent reproduction is possible — whether in the form of almost imperceptible evolution or in the form of planned structural change. Structural changes, therefore, do not require any special "legitimating" principle, ius eminent or super-norm. They are in any event only possible in law.

IV. The new perspective in the theory of self-referential systems changes the starting point from which sociology must examine the relationship between society and law. The binding power of law is not merely a symbolic expression of society as a social fact. The unity of law cannot in this direct manner be explained in terms of the unity of society, and normative validity certainly cannot be explained in terms of an evaluated and assessed social fact. On the contrary, when one enquires into the autopoiesis of the comprehensive social system of society, one first encounters a fundamental problem, and only when this has been solved can the question of the unity of the legal system be answered. The social system consists of meaningful communications — only of communications, and of all communications. It forms its elementary units from the synthesis of information, communication, comprehension, i. e., from the synthesis of three selections, which can be partially (but only partially) controlled by the system. As such element formation always presupposes society and always continues society, there is no communication outside society and therefore no communication of society with its environment. No man can communicate (in the sense of achieving communication) without thereby constituting society, but the social system itself (precisely for this reason!) is not capable of communication: it can find no addressees outside itself to which it could communicate anything. To this extent, society is an autopoietic system in the strict sense of the thoery of self-referential systems. It possesses its own unity as autopoiesis and only as autopoiesis. In this respect it is a recursively closed system, which can neither derive its operations from its environment nor pass them on to that environment. It cannot communicate with the environment but it can and must necessarily communicate about the environment. Granted, communication always implies communication about communication and reproduces itself by this means, but it cannot communicate only about communication. This is a very clear, very unequivocal state of affairs, which does not pose any fundamental difficulties in the concept of unity or in the demarcation of system from environment. All social systems which form within society

The Unity of the Legal System

19

are therefore bound to form their final elements as communications. The legal system, too, consists only of communicative actions which engender legal consequences — it does not, for example, consist of physical events nor of isolated individual behavior which no one sees or hears. It consists solely of the thematization of these and other events in a communication which treats them as legally relevant and thereby assigns itself to the legal system. The problem is that no other social system, no partial system which forms within society can also be a unity in this sense, as typology of operation and therefore as communication. For in all other social systems there is communication within the intra-social environment and possibly also communication with the respective environments. All partial systems execute and reproduce society insofar as they are communicative systems — and it is only this that makes them social systems. This alone does not confer on them either unity or self-referential self-containedness, for these attainments are for them only reached through their differentiation from an intra-social environment which also communicates. Mere participation in the autopoiesis of society does not in itself make the partial systems autopoietic systems, nor does it make them masters of their own unity. According to the theory of self-referential systems, "unity" is a nontransferable quality. We are therefore confronted with the questions of if and how a partial system of society, in this case the legal system, can nevertheless organize its own recursive closure. Or in other words how, while participating in social communication, it can still find its own mode of forming basic elements for which there are no equivalents in the partial system environment.

V. The example of the legal system illustrates how the problem of differentiation is solved in social systems — of course no direct conclusions can be drawn from this about other kinds of systems. The clarification of this problem is a step forward in two areas, namely in the theory of the legal system and also in the more general theory of social differentiation. The differentiation of a legal system is fundamentally based on the distinguishability of normative and cognitive expectations. The solution of the problem lies in the utilization of this difference with an orientation towards the function of law (see Section VII, infra. For a detailed treatment see also Luhmann, 1985). Expectations are fixed as normative if when these expectations are created it is decided that they do not need to be changed in the event of being disappointed. The opposite is true of cognitive expectations. It is therefore a matter of the difference between not learning and learning in still

20

Niklas Luhmann

uncertain future situations, of the advance choice of strategy, of the absorption of uncertainty. This contingent relation, i. e., the function which determines the normative or cognitive nature of the expectation, is then removed in the determination of meaning and is expressed only in the quality of the expectation. Its quality of expectation or of knowledge depends on the form of uncertainty absorption on which it is based. In any event this is an artificial difference which certainly does not have to apply to all behavioral expectations but is only used when an advance determination of the strategy of dealing with disappointment in the communication context of society is required. Legal systems use this difference to combine the closure of recursive self-production and the openness of their relation to the environment. In other words law is a normatively closed but cognitively open system. 16 The autopoiesis of the legal system is normatively closed in that only the legal system can bestow legally normative quality on its elements and thereby constitute them as elements. Normativity has no purpose beyond this (in the sense of an intendible end). Its function is continuous making possible of self, from moment to moment, from event to event, from case to case and it is designed precisely to have no end. The system therefore reproduces its elements by its elements by transferring this quality of meaning from moment to moment and thereby always providing new elements with normative validity. In this respect it is closed to the environment. This means that no legally relevant event can derive its normativity from the environment of the system. In this respect it remains dependent on the self-generating connection of legal elements and on the limits of this connection. At the same time, and precisely in relation to this closure the legal system is a cognitively open system. In each of its elements and in their constant reproduction it is dependent on being able to determine whether certain conditions have been met or not. By programming it makes itself dependent on facts and it can also change its programs when the pressure of facts dictates this. Every operation in law, every juristic processing of information therefore uses normative and cognitive orientations simultaneously — simultaneously and necessarily linked but not having the same function. The norm quality serves the autopoiesis of the system, its selfcontinuation in difference to the environment. The cognitive quality serves the coordination of this process with the system's environment. 17 If the 16

17

Formulations of this kind in systems theory can be traced back to the famous definition of a cybernetic system by Ashby (1956: 4). Ashby defined such a system as one "open to energy but closed to information and control". It may be useful for purposes of clarification to examine a parallel case. The economic system is also differentiated as an autopoietic system. It ties all operations to payments and is, in monetary terms, a closed system. Outside the economy there are no

The Unity of the Legal System

21

cooperation of (normative) closure and (cognitive) openness is a condition of constant self-reproduction and if this in turn is nothing but the unity of the system, "self-reference" can never be total self-determination nor can it ever be total (or even adequate) self-observation. It is always only concomitant self-reference, and is to a certain extent the relation of unity which makes possible the self-realization of unity. It is not a "grounding" principle or a regulative idea and it is certainly not an all-justifying metanorm. If one compares theories here, it is clear that "concomitant selfreference" occupies the space which a theological construction of the world assigned to creatio continua, a space which transcendental theory called on certain privileged facts of consciousness to fill. But the theoretical proposal outlined here eschews any kind of asymmetrical superposition of that concomitant principle. 18 Our primary concern here is with the fundamental asymmetry without which recursive (i. e., symmetric) autopoiesis cannot form systems. Autopoietic processes are recursive, i. e., necessarily symmetrically structured — and this is an aspect of this theoretical approach which has significant and wide-ranging implications for legal theory. The norm quality of each element is owed to the norm quality of other elements, to which the same rule applies. There can therefore be no norm hierarchies. 19 In terms of normativity a strict symmetry exists even between the law and the judge's decision. Laws are only regarded as norms because they are intended to be used in decisions, just as these decisions can only function as norms because this is provided for in laws. In terms of normativity, the relationship between the rule and its application is circular (cf. Eckhoff and Sundby, 1975: 123 — 51 [The authors equate this circularity with cybernetic feedback, a view I do not share]; Eckhoff, 1978: 41—51).20 Autopoiesis operates beyond the sphere of

18 19

20

payments, not even as input or output of the economy. Payments serve the exclusive purpose of making other payments possible, i. e., they serve the autopoiesis of the system. But precisely this closure is also the basis of the wide-ranging openness of the system, because every payment requires a motive which is ultimately related to the satisfaction of a demand. Cf. also Luhmann (1983: 153). The tradition is, however, indicated by the use of the word "concomitant". The same applies — here, too, in contrast to current theoretical ideas — to the autopoiesis of the life of organic systems. Cf. Roth (1981: 1 0 6 - 1 2 0 ; 1982: 3 7 - 4 8 ) . Less radical is the theory (first formulated by Adolf Merkl and Hans Merkl) of the graduated structure of the legal order. It states that at every stage the creation and the application of law are in a hybrid relation to each other and that only this hybrid relation is graduated. Cf. Ohlinger (1975, especially 10 f., 38 f.). This is how one has to think if one wishes to construct a hierarchy with more than two levels. (This is the same problem with clean/unclean in the Indian caste system.) It would be taking one step beyond this to interpret the relation of the creation/the application of law as circular at every level, i. e., as self-referential. Then graduated structure would be a decomposition and hierarchization of the fundamental self-reference of the system.

22

Niklas Luhmann

deduction and beyond the sphere of causality. 21 Thus there is no attribution of norms to final principles or authorities in which normativity and cognitivity, validity and being fuse. 22 Normativity as a clinging to expectations despite disappointments is always and everywhere the same. Even Hart's theory of law (1961) has to come to terms with these presuppositions. On a superficial view this theory constitutes the unity of law as a self-compensatory relation. A layer of primary rules, because of their immanent weaknesses (uncertainty, static character, inefficiency), has to be complemented by a layer of secondary rules. But in legal operations the interrelationship of these two spheres can only be recognized and only be practiced if the same norm quality is involved in both spheres. Otherwise compensation for the inadequacies of law would lie outside the law (which is of course possible). Of course there are also asymmetries in the legal system. No system can get by without them, for this would mean existing as pure tautology in total indeterminability. However, all asymmetries must be introduced into the legal system via cognitive orientations, and they thus articulate willingness to learn. The judge is only cognitively, not normatively, dependent on the law: 23 he must find out whether a law has been passed or not, he can err in this respect, in which case he has to correct his error. In principle, the same applies to asymmetries on an equal level in the time dimension: if a judge wishes to make his decision dependent on other decisions, then he must include these decisions as a fact requiring recognition in his information processing method. The medium of normativity in which the system reproduces itself is structured in this manner. But this always presupposes the basic circularity which differentiates the system and compels it into autonomy. The theory of the closed and thereby open system makes it possible to grasp the function of formality in law and to see this in connection with the peculiar susceptibility to disturbance of law on the symbolic level. 21

22

23

Von Foerster would say it operates on the level of a second order cybernetics, on which the reference to reality is replaced by recursivity. This concept is expanded in von Foerster (1974: 2 7 - 4 6 ; also 1979: 5 - 8 ) . The semantics of central fusion have indisputably dominated the tradition of legal theory to such an extent that it has even been proposed to introduce the basic norm as an epistemological hypothesis. From the perspective of the sociology of knowledge, this type of internal semantics of the legal system demonstrates the difficulty of conceptually realizing the complete differentiation of the legal system. In terms of the history of ideas, all central fusion would appear to be secularizations of the idea of God. Perhaps a further elucidation is useful here. Of course, the judge is the addressee of the normative expectations of the legislator. However, this is a completely different problem. The judge is also the addressee of the normative expectations of the public, of the litigants, of his colleagues, etc. The fulfillment of the normative expectations of another does not mean subordination nor does it guarantee the quality of generating or regenerating of normativity in one's decisions. According to the theory proposed here; this can only occur by means of recursive reference reproduction of elements by means of elements.

The Unity of the Legal System

23

Legal forms are valid because they are valid. They are formal precisely by virtue of the fact that they contain their justification within themselves. They serve as symbols of the circular self-reference of the system. The use of form should not be confused either with logical deduction or with conditional programming, because forms are tautologically valid, whereas deduction and conditioning are built precisely on asymmetries. Like the metaphor or the joke, the form functions only when it remains unquestioned. The use of form occurs when it is insisted that valid rules should be applied, that legally valid sentences should be carried out — because otherwise the legal order itself would be called into question. Forms are carried out in a ritualistic manner. Here references to the world are eliminated and replaced by references to the system itself — a typical characteristic of rituals. 24 Forms are therefore all the more susceptible to every kind of symbolic breakdown — and one form of breakdown is that these forms are no longer believed in. 25 This is why no developed legal system can rely entirely on forms. Selfreference is not only practised simply as self-reference. Its symbolization through forms is transformed into a simultaneous practice of self-reference and external reference. This does not mean that forms become superfluous but they can now be related to the fact that the connection between closure and openness must be guaranteed. This is ultimately why formalism in law is equated with conditioning and logical schematization. However, further analyses are required to make this clear.

VI. Our provisional answer to the question of the unity of the legal system is: autopoietic reproduction of elements by elements on the basis of circular, recursive normativity. Like every good theory, of course this raises more questions than it answers. Our first task is to arrive at a more precise idea of how the continuous simultaneous processing of normative and cognitive aspects of meaning is achieved as a system. For this there are two different mechanisms which operate in conjunction with one another, namely conditioning and binary schematization. Conditioning is a quite general technique of forming systems without which no complex systems could develop at all (cf. Ashby, 1962: 109). For 24

25

Cf. the widespread interpretation of ritual as a restricted code — for example in Wallace (I960, especially 233 ff.); Rappaport (1971b: 2 3 - 4 4 ; 1971a: 5 9 - 7 6 ) ; Bloch (1974: 5 5 - 8 1 ) . This statement can also be reversed, which leads to the Romantic idea of form. Romanticism forces itself through doubt to form.

24

Niklas Luhmann

if systems exceed a certain minimum size yet are to be kept mobile, logically possible connections of elements must be inhibited and at the same time kept available under certain conditions. In the legal system such conditionings are deployed with the additional, special function of combining closure and openness, normative and cognitive expectation. They thereby play an indispensable part in the differentiation of the system. Whether the conditions which confer normative quality to a result or a relation are present is something which can and must be ascertained from a cognitive perspective. Paragraph 1 of the German Civil Code, for example, states that a person's legal capacity begins at birth. This condition has to be examined cognitively: that which is not born does not possess legal capacity. Any attempt to introduce a normative expectation here — for example that a person ought to be born — would throw the entire conceptual system on which legal decisions are based into confusion. This is why the structure of the legal system — insofar as it takes the form of decision-making programs — consists of conditional programs (Luhmann, 1985), which establish an "if a/then b" relation between the conditions (which have to be cognitively ascertained) and the conferment of norm quality. Although it is logically conceivable that validity could be made dependent on future events and therefore left uncertain in the present, such a conditioning of success is incompatible with the orientation function of expectation generalizations. Purposive programs as a means of transfer of norm quality are excluded — if not logically then by the function of law. 26 They would destabilize and particularize law to an intolerable extent. 27 Any purpose orientation of the state must always lay claim to political instruments, i.e., to the power to make collectively binding decisions. 28 26

27

28

The discussion of this issue is riddled with misunderstandings. In particular, a vital distinction is not always grasped. Of course, conditional programs can be built into purposive programs as sub-routines: one has to meet the requirements for obtaining a driving license before being allowed to drive. And these requirements are so selected as to ensure that only those capable of driving actually drive. Nonetheless, the fact that some people who are incapable of driving do actually drive does not make the law on driving licenses invalid; and a person's driving license does not become invalid because the holder is not in a position to buy a car. The non-achievement of a purpose has no legal consequences. With characteristic indecisiveness Unger (1976: 86) concedes that modern jurisprudence has "increasingly accepted the notion that the meaning of a rule, and hence the scope of a right, must be determined by a decision about how best to achieve the purposes attributed to the rule. But all such purposive judgments are inherently particularistic and unstable; the most effective means to any given and varies from situation to situation and the purposes themselves are likely to be complex and shifting" (Unger, 1976: 86). See also the following discussion on the deformation of law by the welfare state (Unger, 1976: 194 ff.). This does not conflict with the rule (in constitutional law) that all administration is tied to law, but it obliges lawyers constantly to consider what minimum amount of

T h e Unity o f the Legal System

25

The legal system itself — and this applies equally to public and to private law — can react to the peremptory demand for consequence orientations only by an internal differentiation between doctrine or interpretation and validity. Interpretation can afford to risk the uncertainties of a glance at the future and can operate with extremely simple causal assumptions. But it must then in its results make clear, purpose-free statements of validity. This technique of juristic conditioning can be applied occasionally whenever problems requiring regulation arise. It becomes universalized and therefore compelled when the legal system is encoded through the binary schematism of legal/illegal. This codification leads to a doubling of normative expectation quality into a positive and a negative value. Then it can (and must) be determined in principle for all events and states about which communication occurs in the legal system, in what conditions they are to be treated as legal or as illegal. In this connection a technicization of this schematism develops, in the sense that the two code values are reversible by mere negation. This excludes the possibility of anything being simultaneously legal and illegal 29 and also of law being interpreted as a weapon by parties with different chances of winning or of losing. For operative purposes, differentiation is oriented not simply towards the function of law but towards the antithesis between legal/illegal. 30 The legal system must accordingly be understood as the unity of legality and of illegality, or more precisely as the control by this antithesis of its autopoiesis with the result that every operation of the system must be conditioned by the system itself. The legal/illegal codification cannot be understood if the function of law is regarded as being solely to resolve conflicts. On the contrary, the severity of the win or lose principle often makes the settlement of a dispute more difficult. Or else it merely forms the menacing background against which negotiations on an amicable agreement take place. In the allocation of action to the category of legal or illegal there is always an element of the reproduction of the legal system, a kind of "surplus value" which is

29

30

conditional p r o g r a m m i n g must be demanded as a condition o f the legality of politicaladministrative purposive action. Cf. Ohlinger (1982: 17—49). T h e United States also appears to be wrestling with the problem of increasing legal " e n c h a i n i n g " o f administration and increasing resistance to this problem. The Greek tragedies illustrate that this maxim was not generally accepted in antiquity. Cf. Wolf (1950 to 1956). O n e may even assume that the development o f law was one of the starting points for the development o f the Sophistic technique o f the forcing of antithetical decisions which in each case led to different consequences. Cf. L l o y d (1966, especially 111 ff.). It may also have led to the development o f logic based on ontology. See Plato's Lysis, 215 E , for the metaphor of the love and desire o f the antithesis for one another.

Niklas Luhmann

26

skimmed off for the benefit of the system and which, it may be said, results from the quarrelsomeness of litigants. We cannot at this stage deal with the consequences which result when structural decisions about conditioning and binary schematization are introduced. It must suffice here to point out that important groundwork has been done and points of contact have been established for detailed legaltheoretical development. This has been achieved not by a leading idea of law such as justice or social order, which would only be more or less imperfectly realized in this way. The reasons lie in the conditions of the possibility of attaining closure and openness at the same time, of combining them and thus of bringing the specific area of functioning of law to a level of high, socially adequate complexity. Ultimately it is always a matter of reformulations of system unity — reformulations by which structural restrictions of correspondingly attainable complexity are accepted and built into the system.

VII. The unity of a self-referential system can only be realized as one which determines, and is determined by, itself. Indeterminacy in that which belongs to a system always means incomplete differentiation, or in other words dependence on an overall social basis of operation which is not ordered as the difference between partial system and environment. Our reflections so far have outlined the problems which arise and have to be solved if a social sub-system in society attempts to attain a distinct unity. It must then itself combine closure and openness, i. e., reproduce its unity itself and constantly expose this process to the difference that exists between itself and the social environment. This may come about because the system fulfils a specific function which is not performed anywhere else in society. As a result it becomes possible for the system to treat everything else (however important this may be in itself for society) as environment. It is a matter of controversy whether a general evolutionary trend towards functional specification can be assumed and how such a trend could be explained. 31 At any rate such a tendency, if it prevails, creates optimal conditions for autopoietic autonomy of social sub-systems because the subsystems relate everything which they use as unity to their function and at the same time can assume that there is no equivalent for this in their environment. By means of the functional relation concomitant with all operations, a self-referential closure can be realized which is not to be

31

It is not unusual for this idea to appear in the construction of scales of ethnological development. Cf. Naroll (1956; 1970).

The Unity of the Legal System

27

found in the use of communication as a basic operation alone. The subsystem then copies as it were the autopoiesis of society — not on the general basis of recursive communication but by exclusive orientation to a function. The thesis introduced above that the legal system is differentiated as a normatively closed but cognitively open system is a first indicator for this functional orientation. But it is not in itself sufficient, since normative expectations also exist outside the law — for example in morals and religion, in the norms regulating the functional orientation of systems and their implications for other systems. The legal system can at best claim a specific use of normativity for itself. The theory of the self-referential closure of the system must therefore contain a theory of the function of law.32 The function of law is not simply a purpose or a regulative idea. If it were a purpose it would be realizable in time and law would therefore end. If it were a regulative idea it would be posited as unrealizable from the outset, which would discredit law or at least raise the question of the real function of real law. Instead we interpret law in terms of a principle of structural selection. 33 Law consists of the exploitation of conflict perspectives for the formation and reproduction of congruently (temporally/objectively/socially) generalized behavioral expectations. 34 The prospect of the disappointment of an expectation and thus, if one clings to the expectation, of a conflict, serves as a principle of selection by means of which generalizations can be tested. In the time dimension — and this indicates a certain predominance of temporal problems — generalization is normative, i. e., is dictated by the intention to cling to expectations despite disappointment and in the face of the facts. But not everything which is thus projected is socially testable (see — an evolutionary theory perspective — Luhmann, 1970). The prospect of winning in the case of a conflict adds further restrictions. The permanent interaction of these factors — an interaction which is coordinated not by any "principle" but only by evolutionary means — creates a large number of congruently generalized behavioral

32

33

34

This does not rule out an analytical separation of these questions; in fact it implies the need for such a distinction. In other words, one can accept the conception of the autopoietic system and at the same time define the function of law differently from this essay, and vice-versa. However, the compatibility of these theoretical parts must be considered in each proposed substitution. The ambivalence of the formulation is intentional. What is meant here is selection of the structure by the structure of the selection. Only this definition corresponds to the mode of operation of a self-referential system. Cf. also Luhmann (1985). This does not however bring out clearly enough the operative connection between the prospect of winning (if necessary, by means of physical violence) in the event of expectation. See also Luhmann (1981 a: 92—112).

28

Niklas Luhmann

expectations which can then be systematized by juristic skill, by comparisons of cases, by concepts and by doctrine. 35 The result becomes, and is experienced as, law.

VIII. This definition of the function of law clearly distances law from the intentions of welfare state control — a purpose for which it can, however, also be used. 36 The political system characteristic of welfare states can enlist the service of law because law can control the outcome of conflict. 37 It can create new subjects of law by creating new possibilities of conflict — for example by prohibiting the cultivation, trade and consumption of drugs — and hence it can cause law to come into being. 38 The "natural" order of the genesis of law is reversed. It is not the norm projection which tests itself against its prospects in the event of conflict but conflict is, so to speak, conjured up out of nothingness in the expectation that orientation will form law from it. 39 Moreover the functional orientation towards conflict was also institutionally embedded. Insofar as legal generalizations developed on the basis of sentencing practice they occurred only when preceded by conflicts. As a result, legal concepts from the outset had a certain closeness to reality. It is well known that the establishment of law by judges has not died out and is perhaps becoming even more important. But at the same time there

35

36

37

38

39

The juristic idea of the unity of the legal system discussed at the beginning of this article refers in particular to this. Even today, law is still widely interpreted as a form of "social engineering", i. e., as an instrument for realizing social intentions. Cf. Ziegert (1975). This is a very wideranging survey and an attempt to establish an "expressive" dimension of law in opposition to the "social engineering" interpretation. However, this interpretation proves unsatisfactory as soon as one asks whether and to what extent the welfare state puts excessive strains on law. For the political system, this may be a question of the limits of causal efficiency. See Luhmann (1981b: 94 ff.). But ought not this question — from the standpoint of the unity and the particular function of the legal system — be posed differently? On the "common" use of physical force in the legal system and in the political system — a subject of major importance since the 18th century — see also Luhmann (1981 c: 1 5 4 - 7 2 ) . See the critical study by Duster (1970). Unfortunately, the author confuses the issue in places with moralistic considerations. This perspective has obvious parallels with the questions which Habermas discusses in terms of the difference between rational systematics and life-world. See Habermas (1981). A detailed comparison is difficult to make because all the key ideas (communication, action, system) are differently defined and the expectations connected with a critique of modern society diverge. For an attempt at a synthesis, see Teubner (1982).

The Unity of the Legal System

29

are also the legislative "mystery tours" (Esser) which do not emanate from conflicts which are to be anticipated or about to be fought out but which instead fix certain states through a political act of will — states which ideally are expected not to produce conflict but approval and obedience. It is problematic that such pre-decided conflicts lack the tension and the uncertainty which motivate the generalization of law. Subjects of conflict of this kind are arbitrarily specifiable but are ultimately interpreted merely as artefacts of politically regulated power towards binding decisions — and not as a tried and tested restriction of expectations. The problem has served as an impetus for forced "democratization" of politics. It has brought us the (unfortunately posed) question of the political legitimation of lawmaking, a question which has given rise to a substantial body of literature. The theoretical approach proposed here seeks a fundamentally different solution. Only by interpreting the unity and the characteristic function of the legal system in the context of the functional differentiation of the legal system can one discern why and how law (similarly to money but in a quite different manner) is "exploited" in the political system in a manner quite alien to its function, without any consideration for its specific regeneration capacity. This discussion of the unity and function of the legal system provides an opportunity to examine a striking new development in the study of law, a completely new kind of critique of law which has recently been much debated at congresses and conferences as well as in literature. The critique of law is of course as old as differentiated forms of law (see above all Nörr, 1970). The fact that no law can satisfy all wishes has always been vexing and has kept the search for remedies and redresses going. In the hands of people, lawyers and courts, law is obviously not what it should be. Following the invention of printing in the 16th century, criticism of the complexity of law — which could now so to speak be read by all and was therefore public — also came into its own. This discussion continued into the 18th century and revolved around the open question of whether redress was more likely to come from the authority of the legislator — which then, to be consistent, would have to forbid the quoting of the opinions of jurists in court — or from the systematizing logic of the dogmatism of law.40 This criticism remains embedded in a sceptical-realistic world view and therefore, like the legal system itself, is not fully differentiated. "Giacche liti, giudizi ed avvocati ci han da essere finche durera il mondo ne rimedio e facilmente da sperare a certi inevitabili difetti della giurisprudenza" (Muratori, 1958: 119). 40

See for example Muratori (1742, quoted from the Milan edition, 1958: 111 f.), which proposes an option for reason and against authority. Here it does not even occur to the author to deny that legislation can be a means of clarifying and simplifying law, whereas today the precise opposite is true.

30

Niklas Luhmann

Today a completely different kind of world view and a completely different kind of criticism of law are becoming prevalent, though they have not yet found a valid formulation. On the one hand, they activate human rights movements in the sphere of communicatively concentrated internationality, championing the cause of all structurally underprivileged people without any consideration of legal-technical feasibility. On the other hand, they complain about "too much of a good thing" — in the legal sphere, for example, about the "juridification of life", about excessive use of legistu