Autopoiesis and Configuration Theory:New Approaches to Societal Steering

Autopoiesis and Configuration Theory: New Approaches to Societal Steering

edited by

ROELAND J. IN 'T VELD Department of Public Administration, Universities of Rotterdam and Leiden, and Netherlands School ofGovernment, The Hague

LINZE SCHAAP

CATRIEN J.A.M. TERMEER

and

MARKJ.W. VANTWIST Department of Public Administration, Erasmus University Rotterdam

.. SPRINGER SCIENCE+BUSINESS MEDIA, B.V.

Library of Congress Cataloging-in-Publication Data

Autapaiesis and canfigu,atian theory new appraaches ta societal steerlng I edited by R.J. in 't Veld [et al.].

p. cm. Includes bibliographical references and index. ISBN 978-94-010-5558-1 ISBN 978-94-011-3522-1 (eBook) DOI 10.1007/978-94-011-3522-1 1. PoliCy sciences--Congresses. 2. Social systeNs--Congresses.

3. Autopaiesis--Congresses. 4. Configuration space--Congresses. 1. Veld. R. J. in ·t. II. Title: Conflguration theorl,l. H97.A76 1992 303--dc20

ISBN 978-94-010-5558-1

Printed an acid-free paper

AII Rights Reserved © 1991 Springer Science+Business Media Dordrecht Originally published by Kluwer Academic Publishers in 1991 Softcover reprint of the hardcover 1 st edition 1991

91-33965

No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the copyright owner.

Contents

Foreword

I. Introduction

1.

2.

3.

R. 1. in 't Veld, Autopoiesis, configuration and steering: impossibilitytheorem or dynamic steering theoryM.I.W. van Twist, C.I.A.M. Termeer, Introduction ro configurationapproach: a process theory for societal steeringM.I.W. van Twist, L. Schaap, Introduction ro auropoiesis theory andauropoietic steering

3

19

31

II. Configuration Theory

4. H.I. van Dongen, Some notions on social integration and steering 475. A.I.I.A. Maas, Offering new possibilities of organizing in social situations:

towards a posture of decentering and centering 556. A.A. Voogt, Managing of social cognitive configurations in a multiple

rontat 677. B. Kee, Knowledge, reijication and organization; some critical comments 798. N.L. Rade, The feasibility of developing a political steering theory based on

process-oriented configuration theories 879. E.H. Klijn, G.R. Teisman, Effective policy making in a multi-acror setting:

networks and steering 9910. G. Broekstra, Consistency, configuration, closure and change 113

ill. Autopoiesis Theory

11. G.C.M. Teubner, Autopoieses and steering: how politics profit from thenormative surplus of capital 127

12. I.J. Koppen, Environmental mediation: an aample of applied autopoiesis? 14313. I.Th.M. Snellen, Autopoiesis and steering: the role of identity 15114. I.A. de Bruijn, E.F. ten Heuvelhof, Policy instruments for steering

autopoietic aerors 16115. I.F.M. Koppenjan, I.A.M. Hufen, Auropoiesis, learning and governmental

steering 17116. A.R. Edwards, Communication and steering: what the science of public

administration could learn from Habermas and Luhmann 18317. W.J.M. Kickert, Applicability of autopoiesis to administration science 19318. I.H.G. Klabbers, Simulation: learning environments for (self-)steering

in social systems 207

IV. Conference bnpressions

19. H. Mastik, A stranger in paradise? Autopoiesis, configuration andsocietal steering 225

Bibliography

Subject Index

Authors Index

About the Authors

vi

233

247

251

253

Foreword

The interest in autopoiesis theory and configuration theory is a recent development inthe social sciences. An important question concerns the implications of these theories forsocietal steering. This was the central theme at an international conference organized bythe Department ofPublic Administration ofthe Erasmus University Rotterdam in November1990. Many new and stimulating ideas were offered.

The participants believe that both theories are very promising and can be beneficiallyapplied in public administration and policy sciences. Because a more thorough explorationof the potential relevance of these approaches for steering problems was needed, the participantswere requested to develop their ideas furtheron. This book is thus the result of stimulatingdiscussions and careful analysis. The different disciplinary backgroundsofthe authors increasethe value of the explorations. Scholars from sociology, philosophy, organizational studies,law, public administration and policy sciences have tackled the difficult theoretical issuesand have carefully analyzed the relative advantages and disadvantages for societal steering.

For the first time autopoiesis and configuration theory are brought together in onevolume and for the first time their potential for descriptive and prescriptive approachesof societal steering are examined. It is the editors' wish that this book constitutes a contributiontowards the development of new approaches in public administration and policy science.

R.I. in 't VeldL. SchaapC.I.A.M. TermeerM.I.W. van Twist

Rotterdam, August 1991

Part I

INTRODUCfION

Chapter 1

AUTOPOIESIS, CONFIGURATION AND STEERING:

IMPOSSmILITY THEOREM OR DYNAMIC STEERING THEORY

R.J. in 't Veld

1. Introduction

Public administration badly needs a firm underlying theoretical framework. Such aframework should enable the scholars who concentrate on problems concerning publicpolicies, on questions related to the connections between public authorities andsociety at large, on questions related to effective regulation or steering - the centralprocess in political life that consists of efforts to influence the behavior of citizens andorganizations in a purposeful manner - and the like, to base their considerations andrecommendations on more solid ground.The present state of the art of theoretical concepts for regulation, which we will

refer to as steering theory, is not satisfactory. Either the simplistic, or at least slightlyartificial, assumptions of social choice theory or the mechanistic characteristics ofclassical systems theory in one shape or the other prevail. Moreover the predominantparadigms are embedded up till now in a pattern of comparative statics. While all thedevelopments in the real world take place in a time- dimension- oriented frame, and,therefore, are all of them more or less irreversible, we can hardly rely upon a firmdynamic theoretical background for our scholarly work. Such a dynamic approachshould shed light on the developmental principles of social systems, on the 'laws' ofevolution and transformation, on the nature of social change itself.Some specialists in public administration will immediately criticize the fore-going

by stating that such fundamental paradigms or theoretical backgrounds or bases orframeworks should be looked for in the 'mother disciplines' such as sociology,psychology, economics and political science, and not in the more applied field ofpublic administration or policy analysis, where art and science are amalgamated.Needless to say that we should utilize all potential sources from those disciplines, butthe question of steering itself is central to all who study public administration andthat question is not covered altogether by any of the mother disciplines. Moreover thedeficiencies in the steering capacity of advanced states and other public authorities ­both as to legitimacy and effectiveness- are by many judged to be so painful andappalling that a relevant future evolution of public administration cannot be imaginedwithout a more satisfactory dynamic steering theory.

3

Consequently we may justify the thorough study of important theoretical findings of adynamic nature in the social sciences in general. Both autopoiesis and configurationtheory can be considered as important contributions. Both have to do withorganizational dynamics. It may seem paradoxical however to assume that these twoapproaches bear relevance to steering theory, because on first sight both might beinterpreted as being more closely related to anti-steering, to the denial of extrinsicsteering, in the meaning of purposeful influencing from outside the boundaries of acertain social system. Some will argue, that both approaches are based on the notionof the closed character of social systems. Hence such an approach would be apt toexplain the deficiencies of public policies, rather than to serve as an option forsuccessful steering.We are convinced however, that we need sophisticated analyses of societal

dynamics in order to detect the causes of failures in steering operations. We thinkthat we can develop some fruitful new insights into the possibilities for satisfactorysteering devices, once we could convert the results of applications of autopoiesis andconfiguration theory into steering problems in practical recommendations. Therefore,it is even necessary to be able to explain why public policies have so many tendenciesto ineffectiveness, before we may attempt again to formulate recommendations forsuccessful policy-making. So we deal with the following two questions in this chapter:

Is it possible to define conceptual limits to the possibilities for social steering, once weaccept some of the central characteristics of autopoiesis and configuration theory?Can we, having recognized those limits, derive fruitful insights and consequentlyrecommendations for effective and legitimate steering?

2. Steering problems

2.1 Reflexivity as the central notion in social systems

Reflexivity is a human ability. We refer to it in order to recognize the capacity ofindividual human beings and social systems to reflect upon their environment(s),events, threats, norms, history, expectations, etcetera, and to change their behavior,their tastes or preferences, and their values or norms subsequently according to theirwill. We assume that reflexivity is the essential characteristic of human existence, thatit is omni-present although its intensity, speed and outcomes may differ considerably.Reflections may concern any regularity, observed or detected earlier as to individualbehavior, and therefore as to social relations. Therefore, the outcomes of reflectionmay transcend the earlier observed regularity. As a consequence the regularity maypotentially lose its validity, because the underlying assumptions concerning humanbehavior have undergone change. Elsewhere we have described the relevance of thisfact for social science itself (In 't Veld, 1989).

If we define the social aspect of our world as the domain that is determined byhuman behavior, the foregoing means that social systems do not bear a deterministiccharacter. Any regularity or law concerning social systems therefore, once formulated,in principle has only a temporary validity.

4

However, it takes time to reflect. Moreover, it may depend upon the specificcharacteristics of the environment whether and how soon the results will show inhuman behavior. If the environment for instance is of a very repressive nature, thesocietal generation of change based on reflection will take a relatively long period oftime. Furthermore the behavioral reactions will not be independent of the characterof the environment. So in a very repressive environment the results of reflection mayemerge as explosions, revolutions or other forms of severe social unrest, while alreadyvery moderate reactions are sufficient to bring about the required change in arelatively tolerant environment.The reflexive character of social systems bears relevance too to situations in

which regulation or steering prevails. Both the regulators and the 'regulated', those atwhom steering is directed, will be reflexive. But reflection is tied to specific roles andinterests so it is not at all certain whether the results of both processes of reflectionwill converge. Of course also different or even contradictory outcomes of reflection bydifferent groups might emerge within each of the just-distinguished categories. So it isnot natural or necessary to assume, that the outcomes of reflection will be in any wayconvergent. Consensus may evolve, but also conflicts may deepen. Anyway, theevolution of steering practices will be determined by mutual and interactive reflectionof both parties concerned.

2.2 Law of diminishing effectiveness of steering

In most literature on policy analyses the effectiveness of steering is considered as aconcept that concerns the relations between objectives and effects of a certain policy.According to this concept a sensible choice of tools or instruments will furthereffectiveness, and in general it is unrelated to time. Here we argue - on the basis ofthe existence of reflexivity of the social system under consideration - that graduallythe organizations and the individuals at whom the steering is directed will react to thesteering measures in a reflexive manner. That is to say, at first they will experienceany potential sanction, any attempt to enforce, however indirect, that is caused by thesteering measures themselves, as an incentive to change their behavior in a directionthat corresponds to the intentions of the regulating body. That change at the sametime, in many instances, contains a deviation from their own most preferred alter­native, and therefore is experienced as a disagreeable event. If this deviation wouldnot be brought about by regulation itself, it would be redundant. So in due course thereflecting actors will utilize their learning capacity to avoid the disagreeable effects ofthe policy concerned. And in general they will gradually succeed better and better astime goes by. Avoidance, sabotage, disobedience, resistance, and any other kind ofcreative activity to restore the original level of satisfaction can be observed.The consequence of this first-order pattern of reactions however is that the

objectives of the regulating body will gradually be behavior less and less. We namethis tendency, this regularity the 'law of diminishing effectiveness of steering'. As weshall see later on, this law is not omnivalent. Its validity is restricted by reactions onbehalf of the regulated actors of a higher order in the longer run. But short termreactions upon steering will be generally characterized by the just-mentioned law.Hence ceteris paribus one might expect a gradual decrease of effectiveness. However,reflexivity is not restricted to the regulated organizations and individuals. Regulators,policy-makers also have the capacity to reflect.

5

2.3 Law ofpolicy accumulation

One would expect that the above-mentioned law is only of minor importance becausethe policy makers, the regulators have many resources to their disposal to correcttheir policies in reaction to the observation of the decrease of effectiveness of theirpolicies. However, reality is not as simple as that. Policy-makers tend to have investedheavily in present policies. A policy measure is not an isolated case, but it is mostlyembedded in wider and larger comprehensive sets of policies. Usually such a set con­sists of regulations, of planning and budgeting procedures, of implementationprocesses and last but not least of procedures for evaluation and control. Institutionaland organizational arrangements have been grouped around the nucleus of presentpolicies and often costly information systems have been designed and implemented inorder to serve the execution of present policies. So it would both be a major andabove all very costly operation to remove an existing set of policies, which we proposeto call from now on a policy system, and replace it by another one. A policy systemthus is an internally coherent and consistent set of policies. It is operated throughnetworks of partially independent actors, with related instruments. The typicalcharacteristics of the combination of the 'Rechtsstaat' and the welfare state or any ofits post modernist variations in most cases dominate the policy system's architecture.In particular the extensive care systems in our society, economically referred to asmerit goods, are characterized by a historical evolutionary process as a consequenceof which allocative and distributive objectives are approached through very large,complex policy systems, both very coherent as well as very hard to change.Resistance to fundamental change also characterizes less complex policy arrange­

ments. So the costs of a more than marginal change are quite considerable, andtherefore reactions to decreasing effectiveness of policies are generally given shape aspiecemeal adaptations within the existing architectural framework of policies. Theseadaptations attempt to correct or to compensate the unfavorable effects of thelearning processes on behalf of citizens and organizations subject to the policiesconcerned. So in most cases the wide scope of these adaptive piecemeal policychanges will be to refine the norms underlying the present policies and to focus stillbetter the tools through which the policies are implemented, as well as to tightencontrol, and thus to restore the effectiveness of the policies at an earlier level. Ingeneral norms will be refined, rules will be formulated in greater detail, the numberof digressions into particular cases will increase, etcetera. Rudely stated, the reactionto ineffectiveness is 'more of the same'. We call this the 'law of policy accumulation'.The regular reaction to diminishing effectiveness of policies is to intensify or toaccumulate these policies.Both above-formulated laws are operational at the same time, so we might expect

that after a correcting accumulative policy measure a new period of diminishingeffectiveness will start. A seemingly infinite series of interactive actions could takeplace. In a next paragraph we will observe however that gradually the correctiveadaptations will decrease in impact, that as a consequence effectiveness will decreasemore and more until eventually crisis will emerge. Before we investigate thisdeterioration process, let us have a closer look at the different possible categories oflearning experiences that are feasible, apart from the one we described in the lastparagraph.

6

2.4 First- and higher-order learning

Many authors (Van der Knaap, 1991) have pointed out that learning is a multi-layerconcept. The adaptation of behavior under the assumption of unchanged preferencesand norms is only the most direct and superficial reaction to any disagreeable changein the environment of human being. Gradually however other changes might takeplace in the minds of citizens and in the decision-centers of organizations which canalso be called learning processes. Citizens might change their attitudes towards thepresent policies by a gradual shift of their insight into - as well as support for - theideological, normative patterns underlying those policies. Goal displacement (Wild­avsky, 1979) could very well be described as a more fundamental reaction to theobservation of the policy effects and their dynamics, among which decreasingeffectiveness. We call these reactions second-order learning processes. They lead us toyet another type of policy dynamics. If the support for present policies erodes becausecitizens gradually shift their normative attitudes towards the present policies, thesupport for necessary adaptations - necessary as a consequence of the functioning ofthe law of diminishing effectiveness- might also drift away. This way the arrival of aneffectiveness- as well as a legitimacy- crisis will be hastened.A still more fundamental category of higher order learning processes is to be

detected in the 'constitutional' area. We assume, with many other public choiceanalysts (Arrow, 1963), that our public institutions and our procedures for socialchoice are based upon relative consensus as to the basic principles of operation ofthese institutions and procedures. This consensus relates to fundamental demandssuch as equality, liberty, and sovereignty of citizens in so far as social choice isconcerned. Based upon those fundamental demands more specific values concerningthe actual shape of and relations between specific institutions are constructed.Citizens have emotionally and normatively invested both in their ideological positionsand in their attitudes towards institutions - see for instance the royal family orparliamentary democracy or socialism (Hinich, 1991, In 't Veld, 1991). Once peoplestart to feel very strongly about the deficiencies of certain policies, they might start towonder in more abstract terms, whether the present political system or a present setof policy-making institutions in a certain policy arena is at all still fit to produce anyeffective policies in the long run. As a consequence constitutional values mightchange, with of course very explosive potential consequences. This type of learning isrelated to the earlier mentioned categories of first- and second- order learning as a'meta-category', because it concerns the organizational aspects of the steering entity,the steering of steering, so two logical steps away from the subject matter itself.We might distinguish still another type of learning, called by earlier authors

deutero-learning (Bateson, 1972). The essential characteristic of this type is situatedin the improvement of the ability to learn, the learning capacity. To learn how tolearn is the nucleus of the concept. In organizational terms this type of learningrelates to developments like establishing citizen's movements or consumer'sorganizations in order to protect their rights against involuntary public 'deliveries',like the improvement of the memory of an organisation, like the introduction of athink-tank or like the installment of antennae to observe any relevant change in thesocietal environment, or like the regular repetition of exercises in simulation in orderto get a better ex ante grip on potential future effects of feasible policy measures.Hence both the steering actors and the regulated citizens might profit from their own

7

mode of deutero-learning. Although this category of learning processes is or afascinating nature, we will not deal with it further on in this contribution, because it isnot closely related to our key issue, policy dynamics.A fairly general observation as to the comparison between the different types of

learning is, that the higher the order of the learning process is, the longer it will takeand the more fundamental its character will be.

2.5 Crisis

First- and second-order learning processes will often reinforce each other. Theinteractive relation of the functioning of both the law of diminishing effectiveness andthe law of policy accumulation will already lead to a problematic situation. As policyaccumulation increases the complexity of the policy system, the historical aggregate ofmeasures, and therefore increases the costs of operation of the system, efficiency willgradually decrease. Moreover the original level of effectiveness will generally not becompletely restored by policy accumulation. So in due course both the effectivenessand the efficiency of the policy system at hand will decrease further and further. Aneffectiveness-crisis will emerge from this development more or less suddenly. Themost beautiful example of such a crisis is still to be found in the 'Laffer-curve' thatillustrates a situation in which both an increase and a decrease of the marginalincome tax rate will cause a decrease in total tax revenues. Once this curve representsreality. it is crystalclear that no change within the framework of the existing arrange­ment is going to provide a reasonable solution any longer. An effectiveness-crisis maycoincide with or even be preceded by a legitimacy crisis that might emerge fromsecond-order and/or constitutional learning processes concerning the normativeattitudes towards the policies under consideration.As soon as policy-makers become aware of the existence of a crisis, they will

realize that the only available method to restore normal circumstances, or better, toregain a future perspective on sustainable relations, is to change the existing frame­work, either by creating a new steering system, or to abstain from any regulation inthis particular case in the future. Now they could utilize the existence of the crisis ­assumed that the citizens and organizations concerned are also aware of its existence­to build the momentum, necessary to overcome the difficulties of a change of frame.As we already explained, a change of frame demands sacrifices, both emotional andmonetary ones. The crisis however will modify the outcomes of cost-benefit- analyses,performed by citizens, in a direction favorable to a change of frame; gradually itbecomes clear to them that the status quo is no feasible alternative any longer.So, according to the words of a famous Dutch lawyer: "gouverner, ce n'est pas

prevoir, it is looking back, and waiting, endless waiting until the momentum is thereto perform the changes you knew unavoidable and/or necessary for so long already".In this respect the openness of a crisis is a blessing in disguise, because it willstimulate awareness among citizens. Policy-makers might even stimulate the opencharacter of a crisis in order to accelerate the growth of awareness of citizens. So farwe have dealt with so-called first-order-crises, during which a certain policy system isin jeopardy. In the course of such a crisis second-order-learning might lead to majorshifts in political preferences in a more general way. Such shifts might bring about achange of government within the existing institutional framework of the society underconsideration.

8

It has become clear by now that higher-order-learning-processes of a constitutionalnature might lead to more fundamental institutional crises because citizens havebegun to wonder whether the existing institutions could function any longer in asatisfactory manner at all. Nothing is as contagious as failure, one might say. Ofcourse it depends on the degree of firmness- referred to by opponents as stubborn­ness- of the authorities or the relative repression applied in that society, whether ashift in citizens attitudes towards their political institutions will easily lead to any realchange. However, in general, crises will aggravate as a consequence of an evolution tohigher- order learning processes.

2.6 Internalization

Is it the fate of policy systems to gradually deteriorate in the above-describedmanner? No, some avenues remain unexplored. There may be another evolutionarypath: the nature of second-order learning processes might be such that normativeattitudes of citizens move into a policy-conform direction. Citizens may observe thatthe present, partially effective policies do improve the quality of their own or otherslives, that the normative assumptions underlying the present policies hold a strongerand more positive intrinsic value than earlier assumed. As a consequence they may'internalize' these assumptions and gradually start to consider them as their ownnorms. They may produce spontaneous behavior that corresponds perfectly with theobjectives of the implemented policies. The extreme case is one in which the policiesbecome completely superfluous because without any regulation all citizens wouldalways behave norm-conform. One could formulate a law of increasing effectivenessof policies for this case; such a 'law' would have some validity in a context in whichthis type of second-order learning prevails. It is almost unnecessary to add that suchlearning takes a lot of time; one should think in terms of decades rather than in termsof years. Merit goods provision, for instance, was accompanied in its first phase ofdevelopment by many regulations enforcing citizens to consume those goods; thejustification of enforcement was found in the supposed latent character of citizensneeds: parents would underestimate the value of their children's education; thereforeit was legitimate for the state to enforce parents to send their children to school.Later, this regulation was internalized by almost all, so it became almost redundant.Our concluding observation is that internalization in general leads to increasing

effectiveness of the policy system, and in the extreme case to its redundancy.

2. 7 Life-cycle ofpolicy systems

So far we have defined the characteristics of the landscape that determines the life­cycle of policy systems. The engine of the dynamics formulated above is the humanpotential for reflexivity: both policy-makers and citizens, public bodies and privateorganizations experience learning processes of different orders. The most general caseis the one designed by the dynamic interaction of two 'laws', the law of diminishingeffectiveness and the law of policy accumulation: eventually they will bring aboutcrisis as well as the death of the policy system concerned; put otherwise, in a moreconstructive mode, the crisis at hand opens the window for a change of frame. Theexceptional case is one, in which the gradual redundancy of the policy system iscaused by such intense second- order learning processes in a policy-conform direction

9

that no regulation is needed any longer because spontaneous behavior would producetotally policy- conform effects.Of course we should not forget that the different types of learning might be

operational at the same time, but also simultaneously in opposite directions. Slowchange of normative patterns of citizens in a policy-norm-conform direction might beovershadowed by faster erosion of effectiveness of the policy system as a consequenceof first-order-learning. The net result of different types of learning processes mayeven appear paradoxical; contradictory phenomena can be observed: for instance, thesame people, who supported the policies wholeheartedly at the start of implementa­tion, show first-order behavioral reactions that undermine the effectiveness of thosepolicies. Another confusing observation may be that the public support for undoubt­edly ineffective policies seems to increase, because citizens have gradually internal­ized the norms underlying those policies originally, but not yet to such a strong degreeas to spontaneously change their behavior accordingly.Thorough analysis of the different learning processes operational at the same

time that might function in opposite directions is a necessary condition for rationalbehavior on behalf of the policy-makers.

2.8 Inter-active learning

It is clear that policy-makers may utilize specific tools or instruments in order tofurther learning processes in society at large of such a nature that their objectives willbe met. It is even possible to anticipate future policies by making attempts to bringabout second-order learning processes beforehand. As Reich (1990, pp.152-153)points out, public managers might "stimulate public deliberation over what was goodfor society rather than to decide specific policy." A public manager might feel "thatpublic learning was at least as important a part of his job as policy making, becausethe public had to understand and decide for itself what value it was to place oncertain issues lying within the manager's domain. (... ) Rather than view debate andcontroversy as managerial failures that made policy making and implementation moredifficult, these managers saw them as natural and desirable aspects of the formationof public values, contributing to society's self- understanding." Reich refers to suchpublic managers as the US secretary of education Bennett, who saw it as his business"to raise the level of debate, focussing the public's attention where it hasn't beenfocussed before." We consider a public manager synonymous with a policy-maker.Although it may be clear, that management by speech and other tools could have

an impact on second-order learning, these effects might be overshadowed by lessfavorable short-term side-effects. The policy-maker who enters the normative debateexplicitly runs the risk of being confronted with very hostile attitudes, because citizensmight feel very uncomfortable with his approach. They easily feel manipulated bypolicy-makers that assume the role of a preacher. This hostility can destroy thesupport he needs in the short term to be at all effective. Apart from the risk ofnegative side-effects there is one more major disadvantage of second-order-learningoriented approaches: the time-horizon of many policy-makers is so nearby, that anyfeasible result of second-order learning can only be expected to go far beyond thathorizon.In practice of course not only individual policy-makers and individual citizens are

involved in learning but also many organizations. They may also consider it their

10

mission to influence other learning processes. They may attempt to anticIpate flaws insocietal learning, or to correct them. As a consequence we observe quite complexpatterns of interactive learning in today's society, but the outcome is uncertain inprinciple.

2.9 Remaining questions

probabilistic character of the future

One may wonder what is the relevance of the dynamic steering concept formulated sofar as it seems to suggest that human reflexivity will transcend and therefore poten­tially destroy all presently consolidated components of social systems in general andsteering systems in particular. We suggest however that the formulation of the con­cept of a life-cycle of policy systems as described above has deepened the insight intothe causes of success and failure of public policies, and in the potentially soundreactions on behalf of policy-makers. Can we derive recommendations from thisconcept as to future policy-making?The concept should make us very cautious to accept any broadly formulated

recommendations concerning sets of tools or methods to be applied in future policies.The unpredictability of the rapidity, the relative intensity and even the direction ofeach of the prevailing learning processes as well as of their interactions and netresults create an extremely contingent environment for the policy-maker. It is usefulhowever to show that thorough analysis concerning the existing interactive learningprocesses should be performed by the policy-maker in order to estimate the mostprobable life cycle of a future policy system. If a reasonable guess would be that theeffectiveness of a future policy system will be undermined within a relatively shortperiod of time, and if large investments would be necessary to install that system,then the policy maker acts wisely by dropping the whole idea. So the probabilisticcharacter of the future does not sentence the policy-maker to random behavior. Hecan cope with complexity by accurate analysis, followed by careful decisions. Theconcept developed above leads him to discover which issues should be analyzed:learning processes in particular.

lead-times

In dealing with the time-dimension of policy-making, we neglected so far the specificproblems connected to the lead-times of the development and implementation ofpolicy systems in reality. These lead-times are considerable, presumably longer than adecade. During such a period - while politicians look for sufficient support, and lateron civil servants prepare implementation- the sequential order of the law of diminis­hing effectiveness and the law of policy accumulation may be reversed. A very convin­cing example is to be found in recent tax reforms in the Netherlands. Income tax inthis country too reached a crisis situation during the eighties, leading to a Laffer­curve-alike situation; widespread consensus emerged as to the necessity of change inthe political realm; a commission was appointed to propose a completely new system,based upon the idea that regulations should be far less detailed. When thecommission's proposals were discussed in parliament, one amendment after the otherwas accepted restoring greater detail and refinement: so policy accumulation started

11

even before the proposals were implemented. As a consequence two years lateranother commission was appointed, etcetera, etcetera.The example illustrates, that lead-times of decision-making and implementation

on one hand and learning processes on the other may interfere and contribute stillmore to complexity.

impenetrability

We have described policy-makers here as actors who are subject to the effects ofsocial regularities as any other entity in society. It is more usual still to look atgovernment as a sovereign body which as a rational actor superposes policy actionsupon society. We do not feel, however, that this stereotype is adequate any longer inan advanced society that has a polycentric power structure and a parliamentarydemocracy as its major public infra-structure. A critical comment to our approachmay be that we seem to consider social systems as impenetrable. Our approachimplies that it is impossible to collect knowledge, that will possibly last and maintainits validity forever, on the evolution of social systems. And indeed, the ultimateconclusion resulting from the recognition of the predominance of reflexivity as theengine of social dynamics is that any regularity or law or other component ofknowledge found and defined as scientific knowledge may be transcended and/ordenied by the functioning of reflexivity later on. As soon as human beings becomeaware of the existence of such knowledge they might take initiatives that destroy itsvalidity. For they are able to produce behavior that deviates from the one "forecasted"by the contents of the body of knowledge. This will be relevant in particular in apower-related context, where one might expect that other, competing or enforcingactors will base their strategies upon the same knowledge.So there is no objective truth to be discovered about social systems like the kind

of knowledge collected on stars or stones. Social systems and their evolution cannotbe penetrated by research, for they potentially change as a consequence of observa­tion and of the process of data and knowledge collection itself. Reflexivity allows ananalogy between social systems and black holes: black holes cannot be observedbecause their concentration of energy is already so immense that they attract and 'eat'all surrounding light available. Social systems digest all knowledge available aboutthemselves, and actors utilize it to their own purposes in a reflexive manner therebypotentially terminating the validity of that same knowledge.This ultimately means that acceptance of the conclusion that the wish for durable

knowledge, for permanent understanding of human society cannot be fulfilled. Onemay enter the cave on his way to understanding, but during the detour it will appear,that behind each space there is another one so that the inner center of the caveremains impenetrable.

Intermezzo

In the real world institutions exist, the exercise of power is normal and accepted, andpeople and organizations are caught in manifold relationships. People grow up, learn,..... and die. Organizations react to any changes along the pathways of dynamicconservatism; they become very active in order to survive. Their learning capacityenables them to do so. Society is not situated in a voluntaristic universe. One may

12

utilize heuristic concepts concerning the process of birth of organizations in an emptyuniverse of course, but at the same time no application to real world situations isallowed without taking into account that there is more between heaven and earththan in the philosophy of voluntarist theories.According to autopoiesis theory as developed by Luhmann and Teubner social

systems are characterized by communication. Individuals are conceptualized aspsychic systems. Their central characteristic is consciousness. Individuals are structur­ally coupled to social systems but unequal to them. Selfreproduction of communica­tion thus reproduces the social system. But the central characteristic of autopoieticsystems is, that they reproduce their own mode of reproduction. Social systems maybe coupled to other ones. They re-interpret, however, outside influence within theirinner logic. They are operationally closed. Their potential for reflection may differ.Configuration theory relates to the process of 'organizing'. Organizing according

to Van Dongen and his 'School' is a process of creating a new reality by the forma­tion of social-cognitive configurations. Individuals are only partially included in such aconfiguration. So each individual is multiple included. Complete inclusion leads toisolation and to a totally closed organization without any outside contacts. Such anorganization cannot adapt to change, and therefore will die.Both approaches are carefully described in chapters 2 and 3 below. We will deal

here, however, with the relations between both theories and the dynamics of steeringas described above.

3. Steering within autopoiesis theory

3.1 Autopoiesis and steering

Autopoiesis theory has to do with the concept of more or less closed organizations.They may be autopoietic in different layers (orders). They do not follow technical orhierarchical imperatives from outside their boundaries in an obedient or mechanisticmanner (Van Twist and Schaap, in this volume). They reproduce themselves, with orwithout reflection. We would like to suggest that the typology of different layers oflearning processes could be very relevant to enrich the categorization of autopoieticorganizations. Each social system will interpret and subsequently re-interpret thesignificance of a steering signal and develop a strategy in accordance of that interpre­tation against the background of its objectives, functions and perceived interests.The recognition of first- and second-order learning processes could enlighten the

sequence of re-interpretations of steering signals observed in autopoietic literature.But still more important remains our earlier conclusion, that there are no limits tolearning, that it may unfold like the principle of infinite regress, in an increasinglyfundamental direction. As time goes by, the outcomes may become increasinglysuccessful in the eye of the beholder, the learning organization. Up till now, however,this has been a discussion of an isolated case, while the essential characteristics of anevolutionary path of a certain type of steering, earlier referred to as a policy systemare to be found specifically in interaction. Therefore, we have to move to thephenomenon of interaction in autopoietic theory: structural coupling.

13

3.2 Structural coupling

We will not try here to explain the central concepts of autopoiesis theory as far asrelations between systems are concerned. Other contributions in this volume aredevoted to that. The whole set of notions used in autopoiesis theory is very complica­ted. Within this approach reflexivity of systems is defined as a characteristic where thedistinction between the system itself and its environment is subject to reflection(Luhmann, 1984; Van Twist, 1991). The relation between a system and its envi­ronment can be characterized, according to representatives of the approach, asacausal parallel processing (Teubner, 1991).We restrict ourselves to a discussion of some issues related to steering itself. A

policy system as described above certainly brings about a longlasting relation, calledby Luhmann and Teubner structural coupling, between the entities that implementand maintain the system on one side and the regulated organizations and citizens onthe other. Teubner defines it as follows: "a system is structurally coupled to itsenvironment when it uses events in the environment as perturbations in order to buildup its own structure." The relation to the steering center not only satisfies thiscondition, but bears a very specific character. One might even agree that the policysystem is the linkage between the steering center and the social systems underobservation. In the terminology of autopoiesis theory linkage mechanisms areconditions for structural coupling. The same theory distinguishes a number ofdifferent types of linkages and coupling.Now Teubner seems to hold the belief that there may evolve a specific type of

structural coupling, to be described as ultra-cyclical processes, that could harmonizethe relations concerned. In this case Teubner even speaks about ecological cycles andabout ecological recursiveness. In the fore-going digression upon steering we havemade an attempt to show that discontinuity of a policy system may very well beunderstood as a result of mutual and inter-active learning. There is nothing wrongwith a society in which a process of creative destruction of policy systems exists sothat they may be replaced by other more relevant ones. The law of decreasingeffectiveness and the law of policy accumulation illustrate the nature of linkages inthe case of steering generally. But it is their interaction that creates a cyclical process.However, this is not of a deterministic character. For the outcomes of internalizationmay oppose the effects of lower-order learning.

3.3 Democracy, steering and autopoiesis

We have serious doubts whether the adherents of autopoiesis theory are developing asustainable vision on the most essential characteristics of parliamentary democracyand tht>refore on today's politics. As we tried to clarify above, the linkage of aregulated individual or social system to the state is at least threefold:it is subject to steering impulses, and it reacts to those in a reflexive manner;it will change its attitude towards the norms underlying the present policies, afterreflection on its effects;it might change its constitutional values related to the structure and functioning ofpolicy-making institutions, as it experiences and reflects upon the whole Gestaltof the policy system.

14

Moreover the three linkages are interrelated. Now let us look a little bit closer at theprocess of change within the steering entity, the state itself. The inner logic of a statein an advanced democracy is to aggregate and to weigh individual values andpreferences and to act in accordance with those. In terms of evolution and change thismust mean that, although different actors within public organizations will act uponself-interested impulses, the net outcome of the actions of public authorities shouldremain in accordance with citizens' values and preferences. Controversially the state,although it acts to create order and aims at the destruction of societal variety bysteering, and therefore seems to be the most sovereign and independent actor insociety, is in fact the most dependent actor. Its inner logic commands it to follow thedynamics of the outcomes of the reflexive learning processes of the regulated actors.Cyclical processes in human society could, according to Teubner, evolve into

hyper- or even ultra-cycles. He uses these concepts to describe the relationshipsbetween economy and law. The existence of a highest order-cycle as suggested abovecan only consist, to our understanding, of constitutional learning processes, as theresult of the most basic category of social learning. Of course we have not paidsufficient attention to the numerous institutions in between the individual citizen andthe state. Institutions that are bi-lingual or even multi-lingual in autopoietic terminol­ogy. Institutions that thus mediate between sub-systems in order to produce linkages.Linkages that enable structural coupling. But taking this into account does notnecessarily change the essence of our argumentation. Politics or democracy have todo with the process of revealing citizens' preferences concerning the very shape ofsociety, its public institutions included. So this process is fundamentally different fromall others in society. The constitutional process is hierarchically superposed upon allothers in society. The self-realization of the only psychic system in society, the humanindividual, is the essence of that process.

4. Steering within configuration theory

4.1 Play and game

The observer is struck by the fact that the debate within configuration theory isconcentrated on the micro-level of society. Therefore doubts might be raised as to itsrelevance on the macro-level. Following the insights of Weick and Mead, the conceptof social integration is developed pointing at the distinction between 'play' and 'game'(Voogt, in this volume). 'Play' creates 'ethical space' on the basis of un-values that arethe building stones of social integration. That space is available for the 'game', thatconsists of the process of organizing. A potential translation of these concepts inpublic policy-making would be that we define the constitutional realm as 'play' andthe arena of alternative public policies as the 'game' under consideration. Constitu­tional learning, i.c.learning the rules, then refers to changes of 'play'.The acquired insights in configuration theory concerning the interaction between

game and play could thus be applied to the arena's of public policy-making. Someadherents of configuration theory, in particular Bolk (Van Twist and Termeer, in thisvolume) have developed an interventionist approach, which they refer to as reframing,that bears a constitutional relation to organizing.

15

4.2 Configuration and congelation

In our own consideration of steering we have paid a lot of attention to the generaldelay of the rate of change that exists in the real world. This retardation is the conse­quence of institutionalization that illustrates the congealing phenomena in real worldenvironments. Regulation, common law, policy systems, they are all visible expressionsof coagulation. Change is only possible after a process of unfreezing; that process istime-consuming and costly. In configuration theory a normative underlying assumptionappears to be, that access to a certain configuration should be free, that 'the other'should be allowed to enter. Without the wish to disagree, we simply observe that allprocesses of social coagulation cause a deviation from this assumption. At the sametime one must realize that exactly that type of coagulation is a necessary but notsufficient condition for all implementation of public policies.

4.3 Configuration and steering

The whole structure of configuration theory offers fruitful leads to a steering concept.The main components of the theory are:the social process of constructing reality;the cognitive contents of this process;the composition of individuals included in the social-cognitive configuration.

Hence each of these components can be imagined as a potential subject of steering:the ethical space - play - may be subjected to constitutional change;play is the context for game; the process of organizing may be influenced fromoutside by other constraints than play such as budgetary limitations;the contents of the process of organizing in semi-dependent configurations maybe manipulated by outside hierarchical commands;outside influence could also be exercised by interventions that point at a changeof the composition of the individuals included.

The partial and multiple character of each individual inclusion in organizing relates tothe nature of learning processes. People confront their various experiences and collectknowledge in each of the configurations in which they are included. The particularindividual combination of specific inclusions may be a decisive factor as to theoutcomes of learning. A striking weakness in configuration theory appears to be thatno attention is paid to the interactions between the learning processes of internal andexternal (intervening) actors. The strength of the configuration approach is related toits precise analysis of the psychological aspects of organizing. Steering-relatedarguments on relevant learning processes and their interrelations, developed here,might enrich applications of configuration theory to public policy-making.

S. Final remark

In this chapter an initial approach of autopoiesis and configuration theory in relationto a number of relevant arguments for dynamic steering in the public realm of societywas offered. We argued that interactive multi-layer learning processes form the

16

nucleus of dynamic steering. We have formulated a first series of potential linksbetween dynamic steering concepts and the two theoretical approaches that hold key­positions in this volume. We have attempted to demonstrate the value added by theamalgamation of the three. This theme will be elaborated upon further in most of thesubsequent chapters.

17

Chapter 2

INfRODUCTION TO CONFIGURATION APPROACH:

A PROCESS THEORY FOR SOCIETAL SfEERING

M.J.W. van TwistC.J.A.M. Termeer

1. Introduction

A major goal of discussions on societal steering within the field of PublicAdministration is to diminish steering pretentions. Unrealistic expectations concerningthe possibilities for societal steering dissipate in these discussions. The monocentricperspective on steering is corrected by pointing out that the government is not theonly, nor the central steering authority (Van Vught, 1986). Government operates in apolicy arena of mutually dependent actors who are trying to realize their own privategoals and who each have their own means of influencing other actors (Ringeling andHufen, 1990). This perspective on steering, defined as the polycentric perspective,helps to explain the disappointing results of steering by government. The possibilitiesfor making recommendations regarding the improvement of steering results based onresearch conducted from a polycentric perspective are however limited. Often one hasto confine oneself to the recommendation to lower expectations concerning steeringresults. At times, researchers even resort to using the monocentric perspective whenformulating recommendations.

In this chapter the configuration approach will be discussed. This approach isrelevant to the discussion of societal steering in public administration for a number ofreasons. One is that it illustrates the limitations of the monocentric approach. In thisrespect the configuration approach does not differ from the polycentric perspective.The configuration approach does however provide a more complete conglomerate oftheoretical concepts. These concepts offer a framework for analyzing the dynamics ofsocial interaction-processes, as well as for making recommendations for improving thepossibilities of influencing these processes (=steering). The configuration-approachintegrates important concepts, such as interaction (Van Vught, 1986), problemdefinitions (Hoppe, 1989), steering as constituting common images (Snellen, 1987),promotion of second-order learning processes (In 't Veld, 1989) and the selectiveactivation and deactivation of relations (Scharpf, 1978).

Inspired by Weick's social-psychological theory on organizing, the OrganizationSciences section of the Department of Business Administration at the Erasmus

19

University Rotterdam is working on a theory of social integration. The project is ledby Van Dongen. In affiliation with the project, dissertations by De Laat (1983), Vander Meer (1983), Maas (1988), Van Dijk (1989), Bolk (1989) and Voogt (1990) haverecently been published. Social-cognitive configuration is a central concept in thesedissertations, in some it is even the central notion. The term 'configurationapproach I will be used as a descriptor for this group of authors. The question centralto our discussion of the configuration approach is:

What insights does the configuration approach offer for the analysis of complexsocietal processes and what are the options for making recommendations withregard to the influencing of these societal processes?

In the configuration approach extensive attention is given to scientific philosophicaldebate. That discussion will not be reiterated here. Some aspects of the debate willbe reviewed, but only insofar as they are of importance in placing the configurationapproach in a more concrete context. Subsequently, the configuration approach itselfwill be dealt with. The review of the theory will be somewhat eclectic. Only aspectsrelevant to the central question in this chapter will be dealt with. A number of ideason steering that have been developed within the configuration approach will beelaborated on. Included in the discussion are the normative grounds for steering, theconceptualization of steering processes and the intervention perspectives. The chapterwill be concluded with a short summary.

2. Scientific philosophical aspects

Within the scientific philosophical debate on the development of, or search for,common knowledge at least three different positions can be identified (Voogt, 1990,pp.1O-24):1. Reality can be considered as an objective reality which cannot be changed and

is waiting for discovery. Reality is assumed to have structural characteristics thatcan be described in terms of laws. Knowledge develops by means of areductionist approach through 'closed' models for conceptualization andresearch. This position is based on the assumption of a reality which existsoutside a social context and independent of the subject as observer.

2. Reality can be viewed as the result of an active process of selective perception.Reality is assumed to exist only insofar as it is part of our observations.Knowledge is constructed by individuals through observation. An observergenerates a definition of the observed, and new observations may change thisdefinition. In other words, it is assumed that subject and object areinterdependent.

3. Reality can also be seen as the result of interaction. In this position it isassumed that reality is constructed in social processes. People exchange their'individual' realities through interaction. In attempting to achieve commonknowledge, a common reality is socially agreed upon. In this case, theassumption is made that objects do not exist outside a social context, but areagreed upon by subjects during ongoing interaction.

20

The configuration approach seeks to approach the latter posItIOn and radicalize it.Social dynamics are stated to be of central importance. Every reality can be reducedto a social-interactive situation; the social and the cognitive are inextricablyintertwined. No reality exists in a social vacuum. Moreover, reality is not onlyconstructed, but also continually reconstructed and changed in processes of ongoinginteraction (Maas, 1988, p.38-46; Voogt, 1990, p.24). There is no objective reality,only social realities which unfold in interaction. This unfolding is a continuous processthat cannot be encircled. Precoding the contents of reality is impossible within theconfiguration approach.

The question of what reality is dominates current research. A good deal ofscientific effort is aimed at encircling 'the' reality, so that it can subsequently beexplained. The question of how reality is constructed acts as a starting-point (Maas,1988, pp.21-38, p.50). If one assumes, however, that reality is the result of a process ofongoing interaction, then the question 'what' cannot be considered separate from thequestion 'how'. If reality is assumed to be an agreed upon reality, then attention mustbe given to studying the process by which reality is constructed and reconstructed.Only after doing this can the question of what reality is, be properly addressed.

3. Central concepts

3.1 The cognitive dimension: definitions of reality

The insight that individual realities differ is not new. In the literature on publicadministration there are several theoretical concepts which point that out. Examplesare perspectives, definitions, schemes, causal maps, frames, constructions, andinterpretations. The list can easily be lengthened by choosing different prefixes, e.g.,reality (constructions), problem (interpretations), situation (definitions). In this paperthe term 'definition of reality' has been singled out (Bolk, 1989).

Definitions of reality refer to subjects which are presented in a social situationand to the specific interpretations which one applies to these subjects (Bolk, 1989,p.30). Examples of subjects that can be central in definitions of reality are the actorhimself, his environment, his strategy and the interdependency relations betweenactors. In this context a problem is a formulation of a part of the definition of reality.Problems are special interpretations that are based on definitions of a presentsituation and a future situation which are connected by a definition of instrumentalrelations (Van Dijk, 1989, pp.66ff.). Each of these elements can become reified.

Reifications are interpretations that constitute a seemingly unchangeable worldof objective facts which appear to dictate the rules of observing and knowing. Oncereified, reality is interpreted as a concrete existing reality with thing-likecharacteristics (De Laat, 1984, p.16). Examples of reification are 'the' organization,'the' government, 'the' policy. Reification can be problematic if the assignment ofthing-like characteristics leads to a compulsive search for that 'thing' (Bolk, 1989,p.18). In this way the process of reconstruction might be hampered or evenobliterated. The search for redefinition possibilities and alternative definitions ofreality is ruled out. (Maas, 1989, p.58). In other words, the irreflexive character ofreification diminishes the possibilities for learning and progress. The normativestarting point within the configuration approach is meant to prevent reification.

21

3.2 The social dimension: ongoing interaction

Definitions of reality are constructed, reconstructed and changed in interaction.Interaction is therefore a central notion in the configuration approach. Interactionrefers to the relations between actors (Van Dijk, 1989, p.17). By entering into socialrelations it is possible for actors to coordinate their behavior. This can be done bymeans of verbal and non-verbal communication (Voogt, 1990, p.19). Behavior as wellas language constitute the basis for the construction of common definitions of reality.

The double interact is considered the basic unit for the analysis of socialprocesses using the configuration approach (De Laat, 1984, pp.62-69; Weick, 1979,pp. 1\Off.). The double interact can be defined by analytically distinguishing threesteps: act, interact and double interact. An act by person A (act) leads to a reactionby person B (interact), and this triggers a reaction by person A in response to thereaction of person B (double interact). The essence of the double interact is that itmakes it possible to (re)evaluate earlier behavior. As a process this can be seen as aredefinition of the acts and interacts that have already taken place. Qualifying one'sown behavior is only possible on the basis of a double interact. For example: Johnhits Paul. Paul starts crying. Only after Paul's reaction is it possible for John to qualifyhis original action and consider on the basis of that interpretation his next reaction(stop hitting). If the interact (action of A and consequent reaction of B) is the centralnotion, the individual behavior is in the foreground, not the relationship between theparticipants. The possibilities for reinterpretation are insufficiently considered. Thedevelopment of common knowledge, and the subsequent learning process, is out ofthe question (Voogt, 1990, p.50). An example can help to clarify this. Steering isusually considered as a series of interacts. The administrator undertakes action, theadministrated reacts and with that the interaction is ended and a new interact starts.Not taking into account the reaction of the administrated to the reaction of theadministrator diminishes the possibility of understanding social processes (ln't Veld,1990). (Re)evaluating original behavior is a precondition to learning (De Laat, 1984,p.67); a bureaucrat in a department develops a policy measure. Implementing andevaluating the measure is left to other people. This impedes the possibilities for thebureaucrat to reinterpret his original behavior and to learn from it.

The limitation of the double interact to two people (dyad) has been criticized byMaas (1988, p.IO?). In the double interact the entrance of third parties is denied.That is why the double interact is vulnerable to reification. Limiting the doubleinteract to two persons can lead to circular processes in which reality is frozen as'facticity'. Because of the denial of entrance to third parties, there is a chance that'the' reality is agreed upon as being definitively this or that. Third parties areessential to safeguarding reflection and social change. With these considerations inmind, the double interact is redefined by Maas as an ongoing interaction between twopersons which is always open to third parties. Third parties refer to anything outsideof the parties of interaction. A third party can be a third person, but also a thirdmeaning or third (inter)action. By redefining the double interact as an ongoinginteraction, each situation can be understood as an invitation to 'more', to 'different',to construction and reconstruction (Maas, 1988, p.215).

This conceptualization of ongoing interaction can be interpreted as descriptive,but it can also be seen as normative. In everyday reality reification is a very commonphenomenon. Often enough there is no room for third parties. Important parts of

22

reality are fixed and excluded from interaction by remarks such as 'we don't need her(third person) for that', or 'don't start twaddling about that (third meaning), it isdifficult enough as it is'. In the opinion of the authors we reified as the configurationapproach these sort of remarks limit the possibilities for change and reflection if theyblock ongoing interaction.

3.3 Social-cognitive configurations: the instantaneous record

In processes of ongoing interaction actors exchange definitions of reality, developing acommon social reality which is the starting-point for further actions. To describe thissocial organizing of realities, the notion of social-symbolic (Maas, 1988) or social­cognitive configuration (Bolk, 1989) is introduced.2

A social-cognitive configuration can be seen as an instantaneous record, asnapshot of the social organizing of reality. Social-cognitive configurations can becharacterized by a matching of relatively stable interaction patterns ('who') andshared cognitions (what') (Bolk, 1989, p.36). In other words, the distinguishing featureof a social-cognitive configuration is an ongoing and intensive interaction betweenactors who also share definitions of reality. If definitions of reality are shared withoutthere being a relative 'narrowing down' in the interaction pattern, Bolk (1989, p.36)speaks of an aggregate. Social-cognitive configurations unfold as a spiral ofintertwined social-structural and cognitive-structural aspects. Over time they developas continuing mutual influencing of definitions of reality (cognitive) and ongoinginteractions (social). Depending on the definitions of reality that they possess, actorsengage in interactions. Those interactions subsequently influence the definitions ofreality that are being constructed. That is why social structures are simultaneously theresult of cognitive structures and the basis for their further development. Descriptionof this social organizing of realities means scrutinizing for who=what=who, and so onto detect who-what-who-patterns. The connections between social and cognitiveaspects can be visualized by means of the double helix metaphor (a model developedto characterize DNA structures); spirals of 'whos' and 'whats' that encircle each otherendlessly.

Contrary to the reified entities frequently used in public administration, social­cognitive configurations are assumed not to exist independently of the researcher.Social-cognitive configurations are not predefined. They are deduced by means ofempirical research. Consequently, they do not have to correspond with (precoded)divisions, the way functions, departments, organizations or networks do (Maas, 1989,pp.109-110).

3.4 Multiple inclusions: the actor as the focal point

-Inclusions typify the extent to which an actor thinks and acts in terms of definitions ofreality in a social-cognitive configuration. Inclusions correspond to the level ofcommitment to the interaction processes which accompany social-cognitiveconfiguration (Bolk, 1989, p.47). Actors are always involved in more than oneconfiguration. Inclusions in a certain configuration can be more or less peripheral butare never complete (Bolk, 1989, p.30). Weick (1979, p.95) introduced the term partialinclusion to describe this partial and incomplete commitment. This term denotes thatactors always invest only a part of their personality in a configuration.

23

The term multiple inclusion is preferred within the configuration-approach (Maas,1989, p.102). In contrast to partial inclusion, the term multiple inclusion means thatactors invest their total personality and not only a part of it. In patterns of ongoinginteraction actors are always included in a plurality of social contexts. Other contextsare always present in the background. Actors can introduce definitions of realitydeveloped in one configuration into other configurations they are included in. Theprocess of constructing and reconstructing definitions of reality is influenced by themultiple inclusions of actors. The existence of multiple inclusions is an importantsource of social change. The important condition that non-determined 'others' exist isrealized. In this way a potential variety of knowledge and acts is guaranteed (Voogt,1990, p.129).

3.5 Social change

Investigations by Van der Meer (1983, 1987) and Bolk (1989) have yielded starting­points for analyzing the dynamics of social change. Understanding these dynamics isvery important in developing concepts for steering. In contrast to other methods foranalyzing social change, the configuration approach does not focus on the system orthe individual actor as the central variable but on ongoing interactions. The centralpoint in analyzing the processes of change is the connection between social andcognitive aspects of interaction. Most other methods of analysis virtually ignore thesecognitive aspects. If no attention is paid to the contents of interaction (the change ofdefinitions of reality), a description of social change would go no further than ananalysis of a 'narrowing-down' and of changes in interaction patterns (Bolk, 1989, p.36). In the configuration approach, social change is seen as a process of developing,stabilizing and destabilizing interaction patterns and definitions of reality (Van derMeer, 1983, p.57).

The concept of social-eognitive configuration is considered basic to theunderstanding of social processes in terms of the changing connections between socialand cognitive dimensions in ongoing interactions (Bolk, 1989, p.47). Using thesenotions, different types of social change can be distinguished (Van der Meer, 1983,pp.57-58). Elements of definitions of reality, interaction-patterns and inclusions canstabilize or destabilize. Owing to this, social-cognitive configurations can change ornew configurations can arise. Definitions of reality can be applied to borderinginteraction situations or can be differentiated within a configuration. Changingdefinitions of reality and divergation or convergation of interaction-patterns can resultin changes within configurations and in changes of configurations, such as atightening-up or a blurring of demarcation lines.

The central concepts, introduced in the other sections, can be used to typifycauses of social change. Ongoing interaction creates social change because theinclusion is never total (Van der Meer, 1985, p.65). There is the possibility ofdiscongruence between the definitions of reality of a configuration and those of anactor, the limitation of a configuration's domain, the simultaneous existence ofconfigurations and the resistance of some actors to a given configuration. Secondly,conditions for social change can be specified (Van der Meer 1983, 1987). Thefollowing examples of conditions are given: 3

24

1. Ongoing interaction between actors included at a lower level in a certainconfiguration. This can lead to the development of a new configuration.

2. Ongoing interactions between actors included at a higher level in oneconfiguration and actors included at a higher level in another configuration.This might result in a tighter demarcation of configurations or the developmentof a new configuration.

3. Ongoing interaction between actors which are all included at a higher level indifferent configurations. In cases of conflict between definitions of reality, thismight result in changed inclusions.

4. Actors can experience discongruence between the definitions of reality of aconfiguration in the development stage with those of the configuration in whichthe actor is already included. This can lead to changes within the configurationor even to the development of new configurations.

Bolk (1989, p.39) added a new condition for social change to these examples. Bolkdeclared that social change can also appear in the case of continuing interactionsbetween actors included at a higher level in one and the same configuration. Thiscould lead to the development of new parallel configurations with other, notnecessarily conflicting, definitions of reality. Bolk partly follows and partly criticizesVan der Meer. Bolk argues that high or low level inclusions are not a prerequisite 'inthemselves' for the understanding of social change. Social change always arises fromconfrontation. Consequently, social change is incomprehensible without anintervention by 'the other' (or the third). The other can be understood in terms ofanother definition of reality, a higher level inclusion or a new interaction partnerfrom 'outside' the configuration. It is 'the other' who brings about the confrontationnecessary for social change (Bolk, 1989, p.37). Following these arguments, Bolkintroduced two more primary categories of social change (Bolk, 1989, pAO):1. strengthening of configurations in terms of the narrowing down of social

structures and greater cognitive uniformity, as a result of 'the other' being morehighly included (interaction on the basis of inclusions in one configuration);

2. corroding of configurations in terms of revisions of cognitive contents andfragmentation of the social-structural components, as a result of 'the other' from'outside'.

4. Steering

4.1 Norms for steering

Steering must be based on reasons. In order to decide on intervention in socialprocesses, norms are needed. Most theories consist of norms related to the content.After 'the reality' has been established, the decision can be made to directly influencesocial processes. In these theories reality is not seen as a result of interaction but asan objective fact. 'The' functioning of 'the' organization, for example, is judged withnorms like 'the' efficiency. Or 'the' problems of 'the' society are seen as basic reasonsfor the activities of 'the' government.

In the configuration approach no norms related to content are used. Steeringbased on a predefined reality is neither possible nor acceptable (Voogt, 1990, p.235).

25

It is not possible because every reality results from processes of social interaction.Little success is expected from a prescription of definitions of reality (Voogt, 1990,p.22l). The development of definitions can hardly be influenced from the outside (DeLaat, 1985, p.116). This is not acceptable because processes of ongoing interactioncan be frozen when definitions of reality are prescribed. In a way, by following thisapproach, intervention is halted rather than stimulated (De Laat, 1985, p.171).

However, this approach does not just supply a normless analytical framework(Bolk, 1989, p.41). Some ideas about norms are developed from which reasons forsteering can be derived. In order to intervene, it is not necessary to make a choicebetween definitions or to answer the 'what is' question (Bolk, 1989, p.44). Norms inthe configuration approach are not related to contents but to processes, and arealways formulated as un-values (Van Dongen, 1982). This means that the norms forsteering do not specify 'what ought to be done' (as a value does), but only 'what oughtnot to be done'. Un-values are always formulated as negative preconditions. The mostimportant examples are not-precoding, not-amputating, not-excluding and not­dominating ongoing interaction (Maas, 1989, p.lll). Un-values specify the 'ethicalspace'. According to the configuration approach, this ethical space should not beviolated. If it is, steering is to be called upon. The authors have developed proceduresfor diagnosing blockages in ongoing interaction. Criteria can be the variety ofdefinitions of reality or the repertoire of inclusions.

Steering activities are not directed towards influencing the acts of actors buttowards influencing the interaction processes. Managing, developing and changingsocial-cognitive configurations in a multiple context are the most important aspects ofsteering (Voogt, 1990, p.134). We distinguish two perspectives on steering:management and intervention. Management refers to providing conditions forcontinuing interaction. When an interaction is blocked, intervention is needed (Voogt,1990, p.235).

4.2 Steering as management

Central in the management perspective is the proVISIOn of conditions for continuinginteractions. Management is not aimed at submission, but at maintenance anddevelopment. Management is considered to be instrumental in facilitating learningprocesses. VOO}t developed a managing method which includes several activities(Voogt, 1990). The clarifying of language is one of these. Multiple inclusions cometogether in ongoing interactions. Managing means finding a language related to thedifferent inclusions in order to make it possible for an actor to communicate with hisown inclusions without becoming isolated. Another activity is watching over the'unhealthiness' of the ongoing interaction. An interaction is healthy when the ongoinginteraction is not disturbed and the process of confrontation, interpretation andreinterpretations of definitions of reality can continue unhampered. Safeguarding andactivating the entering of 'the other' (actors and definitions of reality) is also animportant activity. Constant attention must be paid to indications that definitions orinclusions are not taken into account. The assembling of couplings between actorsand definitions of reality is important. Social assembling means that actors withshared elements in their relations are brought together. Afterwards couplings can beorganized based on cognitive criteria (cognitive assembling).

26

The occurrence of fixation is an important indication of disturbance in the ongoinginteraction. Voogt distinguishes cogmlive and social fixation. In a situation ofcognitive fixation, actors included in the interactions have already dealt with the samedefinition of reality for some time and do not want to change it. Social fixation meansthat mutual relations are no longer the subject of reflection. This can result in rigidritual behavior. The entrance of 'the other' is blocked in both situations. Whenongoing interaction is totally blocked and entrance of 'the other' is totally closed, thedecision can be taken to temporarily cancel management activities and to startintervention.

4.3 Steering as intervention

The intervention perspective has been developed by several authors. Voogt developedthe concept of context variation, Bolk the concept of reframing, Van Dijk theconsultancy perspectives and Maas is working on the concept of strategic decentra­tion. Maas (1988) concluded in his dissertation that further development of theconcept of strategic decentration could prove useful. This strategy can be used insituations where the diagnosis is that people have fixated on the (definitions of the)problems they want to solve. Strategic decentration is an intervention procedure inwhich people are stimulated to see their problems in another way by decentring theseproblems. To use this strategy, careful diagnosing is needed. Maas will elaborate onboth intervention and diagnostic methods in his chapter in this book.

Bolk (1989) developed an intervention approach which he called reframing. Thisintervention perspective focusses on the cognitive component of social processes. Bolkdefined reframing as all intervention techniques attempting to change the actor'sdefinitions of reality. By using these techniques, actors are forced to go beyond theircurrent frame of reference. This results in a change of fixed perspectives, relationsand behavior (see also Levy and Merry, 1986, p.96). In terms of the configurationapproach, reframing is a direct way of organizing contacts and guaranteeing acontinuous confrontation with alternative definitions. The desired effect of reframingis dynamizing the cognitive and social development of configurations.

Van Dijk (1989) developed a methodological strategy for changing the problemsolving processes within organizations. His classification of intervention classes isinteresting for our thinking regarding intervening in the cognitive components ofsocial processes. His diagnoses are based upon his theory that, in a problem solvingprocess elements of 'the problem' get reificated (compare to fixation of the cognitiveaspects). He views intervention in the problem solving process as a context variation.He argues that interventions are most effective if they act upon the element of aproblem (a part of the definition of reality) that shows the greatest variation. Adistinction between four classes of intervention is made (Van Dijk, 1989, pp.78-82):1. Transposition: the translation of a problem or element of a problem into

another medium, for example, simulation;2. Homogenization: the problem or element of the problem is brought into

alignment with the reificated problem or element;3. Heterogenization: a significant difference is created between the reificated

problem or element of the problem and the context;4. Equalization of relationship and communicated content. The cogmlive and

social components of interaction are brought in alignment with each other.

27

Voogt (1990, pp.153-154) developed procedures for determining the causes of blockedinteraction. He distinguishes between cognitive and social fixations. His interventionperspectives are related to these diagnoses. The basic principle is context variation. Inthe case of cognitive fixation it is necessary for actors to bring up for discussion theirdefinitions of reality. Because the cognitive component is fixed and not open tovariation, the choice to create variety must be found in the social component. Thedirected discussion can be facilitated by introducing a 'third person'. This may bringabout a change in the social relationship, this being the context of 'reality'. This canresult in a change or dereification of the fixed definition of reality. In the case ofsocial fixation actors do not permit variety in terms of new persons. The socialcomponent is fixed; variety must be introduced in the cognitive component. Interven­tion may introduce third definitions of reality which may partly overlap with thedefinitions of the participating actors. This change in the context of the socialstructure provides an opportunity for change or dereification in the social structure.Voogt will elaborate on this concept in his chapter in this book.

s. Conclusions

In this chapter a dynamic method of analyzing steering processes is described. In linewith the most central concept, this approach is called the configuration approach. Theapproach provides an integration of well-known and lesser known concepts in publicadministration. The concepts for description and prescription are closely related toeach other. The central points of this approach are summarized below:1. The starting point is not the question 'what is 'the' reality' but the question 'how

is reality constructed and reconstructed'.2. The close relation between cognitive ('what', definition of reality) and social

('who' , ongoing interaction) aspects is considered to be basic for understandingthis 'how-question'. A social-cognitive configuration is characterized by amatching between a relatively stable interaction pattern ('who') and sharedcognitions ('what'). It is an instantaneous record of the social process ofconstructing realities. Inclusions is a term for understanding the same processfrom the actor's perspective.

3. Social change can be understood as a developing, stabilizing and destabilizing ofsocial-cognitive configurations. Confrontation with 'the other' (a person, adefinition of reality) is seen as the essential condition for change.

4. Norms related to the content are not used for judging processes, instead normsconcerning procedures are used. Steering is not directed at influencing the actsof persons. Steering is directed at the ongoing interaction processes.

5. Two perspectives on steering have been developed: management and interventi­on. Management concerns facilitating the conditions for ongoing interaction. Ifongoing interaction is blocked, intervention is permitted. Intervention concernsthe creation of a confrontation with 'the other'. Depending on the diagnosis, themain focus for intervention will be the cognitive or the social aspects.

28

Notes

1. The concept of configuration is based on work done by organizational scientists such as Mintzbergand Miller and Friesen. These authors see organizations as complex entities tending towardscertain 'quatum states' or 'configurations'. Configurations are seen as a description of thecharacteristic status quo in an organization and are defined as clusters of, or relations between,variables that display an underlying cohesion. This means that the presence of a certain featuresuggests the probable presence of the other features (Maas, 1988, p.72). The authors who adhereto the configuration approach as presented here, unlike Mintzberg or Miller and Friesen, do notdefine configurations on the basis of a predefined content. To these authors a configuration alwayssuggests content in a social context. A social reality is always constructed in patterns of ongoinginteraction. Consequently, configurations are seen as gestalts with a social context (see also section3.3).

2. Not all of the authors use the term configuration. Weick (1979) and De Laat (1984) speak of'group' and Van der Meer (1983) uses the term 'play'. To achieve some consistency in theconcepts we used in this chapter, the authors replaced those concepts with the term configuration(see also Bolk, 1989).

3. Van der Meer (1983) presents these types of social change as categories. In our view, however,these are neither complete nor mutually exclusive (o.a. Bolk, 1989). Therefore, they are presentedin another format.

4. Voogt specifies five functions in his discussion of management in multiple social contexts. In ourview, however, these are not mutually exclusive. Consequently, they are presented in anotherformat.

29

Chapter 3

INfRODUCTION TO AUTOPOIESIS THEORY

AND AUTOPOIETIC STEERING

M.J.W. van TwistL. Schaap

1. Introduction

A new scientific theory that has been attracting attention in recent years, and will inall likelihood have a considerable effect on several fields of research, is the autopoi­esis theory. Stafford Beer (1980) refers to it as one of the most important innovationsin systems theory. Boulding (1981) predicted that the autopoietic concept is likely tohave a considerable impact on several disciplines, from thermodynamics to sociology.And in a way it already has. Autopoiesis is a dominant concept in a sociologicaltheory developed by Luhmann (1984). In other social sciences, though its value is notundisputed, the attention given to the concept is growing. Morgan (1986) attemptedto adapt autopoiesis to organizational science, while Teubner (1989a) has done thesame in the field of law. By applying autopoiesis theory in the field of publicadministration and policy sciences, new insights can be developed regarding the(im)possibilities of societal steering.

In this chapter the theory of autopoietic systems will be discussed. First thedevelopment of the theory in biology and implications of the theory for the conceptu­alization of living systems will be described. This will be followed by a discussion ofthe application of the theory in the social sciences; Luhmann's theory of autopoieticsocial systems will be given extensive attention. After these introductory remarks thefocus will shift to the field of public administration and policy sciences. Somecommon diagnoses of steering problems will be presented. These diagnoses areincomplete, as they do not take into account that in modern societies severaldependent and independent autopoietic subsystems have developed, resulting fromthe differentiation of society (Luhmann, 1982). Steering problems from theperspective of the theory of autopoietic systems will be discussed with an emphasis onthe political subsystem, explaining why this is just one subsystem among others andcannot be the central steering system.

31

2. The theory of autopoietic systems

The term 'autopoiesis' was coined by two neurobiologists, Maturana and Varela(1972/1980). They were trying to find a term to characterize 'living systems'. Theywanted to develop a new theory about life, as they found the traditional biologicaldefinition lacking. In conventional biological research, living systems (cells, organisms,plants and animals) were classified by their properties, but the question of what lifewas, had never been properly answered. In Maturana and Varela's opinion, lifecannot be defined by the properties of living systems. They countered that thedifference between living and non-living systems is determined by the systems' specificorganization of production relationships. In living systems this organization ischaracterized by a continuous reproduction of itself (autos = self, poiein = to make,produce, remake, conceptualize). The relations of production in a living system areorganized in such a way that they are continuously regenerated in a recursiveproduction process. A definition of life must therefore always take into account thisspecific organization of the production relationships between the components of asystem, whatever its properties.

In Maturana and Varela's theory, as presented here, living systems are continu­ously self-reproducing systems. This theoretical stand has several repercussions, one ofthese being the conceptualization of the relationship between system andenvironment. In its self-reproduction, the autopoietic system is independent of itsenvironment. In the process of self-reproduction, living systems refer only tothemselves. There is no need for input from outside the system, from theenvironment. 1 That is why living systems are characterized by autonomy regardingthe environment. Consequently, self-reproducing systems cannot be understood asopen systems which transform inputs into outputs to create the conditions necessaryfor survival in a continuously changing environment. Autopoiesis demands a newconceptualization of the systems' notion. The concept of open systems has to bereplaced by one that allows for the system being in continuous interaction with theenvironment, but emphasizes that this interaction is always determined by anorganizationally closed system of production relationships.

Autopoietic systems are organizationally closed. They are not, however, isolatedfrom the environment (which contains other (autopoietic) systems). External environ­mental influences are perceived by an autopoietic system as perturbations. Interactionbetween system and environment can best be conceptualized as reciprocal perturba­tion. This means that the structure of the environment can (not has to) be the causeof changes in the structure of a system, but that the environment can never determineor direct these changes. The same holds true for the system's influence on itsenvironment: interaction between system and environment can (not has to) be thecause of structural changes in the environment, but the final result can never bedetermined by the system. The result will be a string of recurrent mutually congruentstructural changes continuing as long as the autopoietic system and its surroundingenvironment do not disintegrate. Maturana and Varela call this structural coupling(1987, p.75). Structural coupling between living systems can facilitate the realizationof autopoiesis. By means of structural coupling, composite unities can develop. Such acomposite unity is called a living system if it reproduces itself by the autopoiesis ofthe coupled unities that constitute it. Maturana and Varela call these second-orderautopoietic systems (1972/1980, p.l09).

32

Second-order autopOietIc systems can also interact with each other, potentiallyresulting in structural congruence. This is called third-order coupling. On this subjectMaturana and Varela's terminology is not very consistent. Sometimes they avoid theterm autopoiesis (1987, p.193). And when they do call third-order coupling 'autopoie­sis',they fail to explain what it means (1972/1980, pp.107-111). This is an unfortunateomission as the coupling of second-order autopoietic systems is an essential conceptfor the application of autopoiesis theory in social sciences. The uncertainty created byMaturana and Varela has led to many misunderstandings, as it is not clear in theirdefinition whether a social system (as a composite unity of second-order autopoieticsystems) is an autopoietic system.

Maybe this can best be illustrated by the empirical phenomenon the theory wasoriginally meant for, cell reproduction. In cell reproduction, not only are the cellsreproducing themselves, but they are also reproducing their own capacity to repro­duce. In other words, not only are the components reproduced, but the relations ofproduction as well. For that reason, cells are autopoietic systems of the first-order.Common to all metacellulars is that they include cells as components of theirstructure. They reproduce themselves through the coupled unities (cells) that they arecomposed of. Consequently, metacellulars are second-order autopoietic systems (1987,p.87). Ongoing interaction between metacellulars (individuals) can result in structuralcongruence. However, here the authors almost never speak of structural coupling, andnever of third-order autopoiesis. The explanation is simple. Firms, schools, footballteams, or other social systems are not continuously reproduced by the autopoiesis ofthe individuals that make them up. Consequently, autopoiesis of social systems isunthinkable to Maturana and Varela.

3. Autopoiesis in the social sciences

Despite Maturana and Varela's resistance, the theory has been applied in the socialsciences. Social scientists are, however, not unanimous concerning the usefulness ofthe theory in their field (this is hardly surprising). The theory was, after all, designedby biologists to provide a characterization of life and is, in fact, a very abstract theory.We are of the opinion that it can be useful in the field of public administration andpolicy sciences. In the first place there is a societal reason: in autopoiesis theory somefundamental questions are asked concerning the possibilities of political steering andthe accomplishment of political goals (Scharpf, 1988). By analyzing policy processesfrom an autopoietic perspective, descriptions and explanations can be given of somedifficulties encountered in political steering and prescriptions so that improvementscan be made. The second reason is a scientific one; with this theory it is possible toanalyze self-referential, self-maintaining and self-organizing tendencies of socialevents, which are disregarded in other theories (Teubner, 1989a).

In the public administration literature very little reference is made to autopoie­sis. In other scientific areas there has been some discussion about applying the theoryof autopoietic systems in the social sciences. Maturana and Varela refused tocharacterize social systems as autopoietic. They speak only of structural couplingbetween living (=autopoietic) systems. This resistance has led to the accusation byother authors of "an unusual degree of parochialism, defensiveness, and quasi­technological dogmatism" (Jantsch, 1980). Jantsch tried to extend the theory, or at

33

least some of its notions, to make it more suitable to the social sciences (1980). This,however, weakens the theory. Another possibility for applying autopoiesis in the socialsciences is to use the theory metaphorically to characterize social organizations andparticularly their tendencies toward self-maintenance (Morgan, 1986). It is alsopossible to formulate a new autopoiesis theory for the social sciences, as Luhmann(1984) and other German authors (Teubner, WiIlke) do. This new theory will beelaborated on, as this approach seems the most relevant to the discussion of societalsteering in public administration and policy sciences.

4. A theory of autopoietic social systems

Luhmann developed his own version of a theory of autopoietic systems. In developinghis theory Luhmann frequently refers to the insights presented in the biologicallyinspired work of Maturana and Varela. Nevertheless, Luhmann's theory isfundamentally different from that of Maturana and Varela and he does not agreewith Maturana and Varela on the issue of applying autopoiesis in the social sciences.He states that there are a number of general principles of autopoietic organizationand that there are different modes of autopoietic reproduction. The general principlesof autopoietic organization not only manifest themselves in life forms, but in othermodes of circularity and self-reproduction as well. Luhmann defines autopoiesis as ageneral form of system-building using self-referential closure. This opened the way fora more general theory of autopoietic systems in which all references to living systemsare carefully avoided. Luhmann makes a sharp and essential distinction betweenmeaning and life as different kinds of autopoietic organization. Meaning-using systemsare distinguished according to whether they use consciousness or communication asmodes of meaning-based reproduction. Luhmann designated the former as psychicsystems and the latter as social systems.

aUopoietic

machines

autopoietic

living systems:autopoiesis: life

non-living systems:autopoiesis: meaning

psychic systems:consciousnesssocial systems:communication

Figure 1 Autopoietic and allopoietic systems (revised from: Luhmann 1984)

Luhmann assumes, in accordance with Maturana and Varela, that the minds ofpeople are operationally closed. All contact of the mind with the environment takesplace in terms of the mind's own specific internal operations. The internal operationspecific to the mind is thought, being the processing of consciousness by a successionof thoughts. In the mind every new thought builds upon previous ones. That is whyLuhmann (1984a, pp.346-377) characterizes the mind as an autopoietic system.

34

Psychic systems (minds) are autopoietic systems that reproduce consciousness throughconsciousness, thought for thought. In reproducing its consciousness the psychicsystem does not import any consciousness from the outside, nor is any consciousnessever exported outside the psychic system. This thesis is essential for theconceptualization of communication. Communication is often represented as thetransmission of information. That evokes an image of a sender giving something awaythat subsequently is collected by a receiver. In Luhmann's opinion, the ontologyimplied in the conceptualization of communication as the transmission of informationis insufficient for an adequate understanding of communication. The impression holdsthat the information transmitted is the same for sender and receiver. That is, ofcourse, possible, but is never guaranteed by the qualities of the content of theinformation. The semantics of communication, therefore, should be renewed.

Just as thought is the specific internal operation of the psychic system, so iscommunication the specific internal operation of the social system. Communicationalso builds on previous communication. In this way communication reproduces itself.This reproduction of communication takes place in what Luhmann calls a socialsystem. Social systems use communication as their particular mode of reproduction.Communication is, according to Luhmann (1984a, pp.203ff), a synthesis of threeselections. These selections involve:

Information. Someone trying to communicate selects certain information. By thisselection, which excludes other possible information, meaning is determined.Utteraru:e. The information selected is presented as an utterance. This utteranceis a selection out of a repertoire of possible actions.Understanding ('Verstehen'). Communication is purely social. This implies thatthe focal point for communication is not the one trying to communicate, but theone who is supposed to understand the information that is in the message. Thatis why understanding is constitutive for communication. Communication canonly become a unity after the selection of the an interpretation by a listener.

Information, utterance and understanding are aspects which cannot exist independent­ly of a social system; they are co-created within the process of communication. Thecommunicative synthesis of information, utterance and understanding is possible onlyas an elementary unit of an ongoing social system.

Rather surprisingly, in Luhmann's opinion (1986, pp.I77-178), only communica­tion, and not people (listeners) or (communicative) action, is the elementary unit ofthe basic self-referential process of social systems. Only communication is necessarilyand inherently social. Action is not, because it does not include the understanding ofthe listener. The understanding of the listener is never part of the action of the onewho is uttering a message. Communication always contains a far richer meaning thanthe utterance of messages alone. The relation between communication and action hasto be reversed. Social systems are not made up of actions of a special kind; they arenot composed of communicative actions. Social systems are made up of communica­tions. Communication of course requires action, but action is only one of the constitu­ting selections of communication, and not the only one.

People, or more precisely their psychic systems, are not the elementary unitswhich social systems consist of. According to Luhmann, the individual is not therelevant unit for the analysis of social systems. Mayntz (1987, p.l(0) and Scharpf(1988, p.65) have criticized Luhmann on this point. Communication, they propose, is

35

not possible without people, and therefore people cannot be decentred that easily.Luhmann does not disagree with them; communication without people is indeed notpossible. Communication always assumes someone trying to utter a message, andsomeone else understanding this utterance of information. But Luhmann maintainsthose people are part of the environment to their communication. Communication isconveyed by people, but takes on a life of its own. It is communication that buildsupon previous communication and selects further communication. One communica­tion draws upon another. That way communication reproduces itself, relativelyindependent of the people communicating. According to Luhmann, communicationbecomes more and more independent. It cannot be reduced to (the consciousness of)the people communicating. People (psychic systems) and social systems constituteenvironments for each other in Luhmann's theory. Of course, without consciousnessno communication is possible. Even so, Luhmann maintains, both operations shouldbe attributed to different operationally closed systems. In other words, consciousnesscan never be part of communication and communication can never be part ofconsciousness. An ongoing self-referential and internal reconstruction in the auto­poietic system (psychic or social) is a precondition for their contact.

Social systems consist solely of communications. Meaning is attributed on thebasis of communication. The environment of a social system (consisting of, amongother things, of other social systems) is internally reconstructed by means ofcommunication. Only in communication can meaningful elements come into existence.Events in the environment that are not selected and are not being reconstructed incommunication do not play a role in the system's development. This selection isdetermined by an ongoing referral to previous communication. By means of self­reference it is determined which events in the environment are relevant, which arenot, which meaning is to be attributed to that event and which is not. Because of theongoing self-reference, the social system continuously proves the relevance of its owninterpretation.

5. The emergence of functional subsystems2

Complexity and differentiation are the key words in Luhmann's societal theory. Thecentral assumption is that the world is too complex to be comprehensible. For thatreason selection is necessary to comprehend part of the world. Differentiation is amedium to do this. Luhmann distinguishes between three forms of systems differentia­tion: segmentation, stratification and functional differentiation, the latter being themore recent development in sociocultural evolution. Functional differentiationorganizes communication processes around specialized (symmetric and non-hierar­chic) functions which have to be performed at the societal level, resulting in theemergence of functional subsystems. Differentiation starts as a differentiation of roles.Functional differentiation gets acquires significance when at least two roles formaround a specialized function with complementary expectations (for example,politicians/public, preachers/laymen). These roles transform irrelevance into rele­vance, into that which is relevant to the function. Subsequently, functional subsystemsform, directing their operations towards their own environment, their own reality,their own autopoiesis. Examples of functional subsystems are the legal system, theeconomic system, politics, education and health care. The differentiation is supported

36

by many professional skills, education, codes, increased use of professionals, andspecific semantics. Some analysts even speak of different rationalities which cannottolerate each other and therefore try to exclude one another. 3

Functional differentiation does not stop when society divides into a number ofsocietal subsystems as mentioned above. It continues within the subsystems. Law, forexample, (European-continental as well as Anglo-Saxon law) consists of a number ofsubdivisions, such as civil law, criminal law and constitutional law. Several areas ofspecific selections have also evolved within these subdivisions.

Subsystems are not just parts of society, they are the whole society. They repro­duce the distinction between system and environment. Every subsystem reconstructsthe whole system and is a specialized version of the identity of society. Therefore,society is not just the sum of subsystems, it is more. The juridical subsystem, forexample, is the whole society as far as it observes juridical aspects. Society consists ofall communications and makes the distinction between communications and non­communications. Subsystems consist of specific communications, such as juridical andeconomic communications.

By way of functional differentiation, functional relations in society become morespecific. The subsystem's openness to the internal environment to which the subsys­tem performs its function increases. Subsystems can ignore everything except certainspecific aspects of their environment. As a result there is an increasing amount ofindependence (Willke, 1983,1987). Differentiated subsystems become closed andindifferent towards aspects beyond their function. They just select those events andthose aspects of events which are relevant to them. This relevance is determined bythe subsystems themselves. Reality is reduced to what is important to the subsystem.In interpreting and reacting to developments in their environment, they take theirown functional concerns as points of reference. Hence they are self-referential.Despite their self-reference and independence, functional subsystems cannot survivewithout other subsystems. Separate functional subsystems deperui on other subsystemsfor the fulfillment of their functions. In return, the internal operations of onesubsystem affect other subsystems. The political subsystem needs support from thejuridical subsystem on order to influence the economic subsystem. The costs of thehealth subsystem, resulting from its internal operations, are a severe problem for theeconomic subsystem. Yet, this economic problem can probably only be solved if thepolitical subsystem takes measures. So the emergence of functional subsystems notonly increases the interdependence, but it also decreases interdependence because theindifference can become greater.4 Subsystems observe fewer aspects from thesocietal environment, but they observe 'more deeply' and more specifically. Becausedifferentiated subsystems limit their operations to specific aspects and comprehendthese better, society can understand more complexity.

6. The state as a self-description of politics

Systems can observe and describe themselves (Luhmann, 1984b). Observing meansutilizing distinctions, especially the distinction between system and environment. It isnot a physical act, nor does it have to do with consciousness in a social context. Socialsystems observe through communications. Observations can be used to reconstitutethe system by re-entering the distinction between system and environment into the

37

system. This is called self-description. Self-description can never be as total or ascomplex as the system itself and is thus a self-simplification. The system needsspecific semantics to describe itself. With these semantics the system can givemeaning to its acts. Systems use self-description for their development. At this pointLuhmann's distinction between politics and the state should be mentioned. Thepolitical subsystem's function is the production of collectively binding decisions (1981,1984b). Like all systems, the political subsystem needs a simplified definition of itself.The political subsystem conceptualizes itself as 'state', gives itself meaning bydescribing itself as the state. That there is a difference between politics and the statecan be illustrated by non-decisions. Non-decisions have a political meaning and arepolitically relevant. But a non-decision is not a binding act by the state. The state isnot involved.

With the help of the state semantics the political system can give its acts, itsdecisions, a hierarchically higher place than the acts of other systems. It goes beyondthe symmetry of societal functions. As stated before (see section 5) all functions insociety have to be fulfilled and the political function is just one of them. But when thepolitical subsystem describes itself as the state, it ignores functional symmetry. Itshould be noted that the hierarchy only exists the political subsystem's self­description. To put it another way, the special position of the political subsystem doesnot result from functional differentiation but from the self-description of the politicalsubsystem. The consequence of this is that the hierarchy between subsystems in theself-description of the political subsystem can cause some problems if politics reallywants to fulfill its function of making collectively binding decisions, if it wants tomake its decisions binding. This difficulty can be added to the previously knownpolicy making problems.

7. Steering Problems in Dominant Analyses

The political system is constantly being confronted with policy making problems.Policy goals are often not achieved and policies sometimes generate negativeoutcomes. Scientists have developed several models for analyzing policy processes(Allison, 1972; Linstone, 1984). The two dominant perspectives are the monocentricand polycentric models. Each stresses certain (causes of) problems and solutions(Klijn and Teisman, in this volume).

The monocentric perspective has a state-centered point of view. Society is per­ceived as being governed by the state. Goals are set by elected politicians. Policymakers have full information and choose the more effective instruments forimplementing policies. Citizens and organizations are supposed to react to theseinstruments in the way government wishes. Problems, preferences, wishes, as well asthe manner in which problems are solved are objective facts. Government possesses a'toolkit' containing policy instruments or "tools of government" (Hood, 1983).Instruments have intrinsic aspects. The outcomes of instrument application depend onthe instruments' characteristics and are uniform whatever the policy area, time andsetting. For most monocentric analysts the character of the policy field or of theactors involved is not important. (Although some analysts do take into accountaspects of the policy field.)

38

The main cause of policy failure is perceived to be the lack of cooperation andcoordination between the actors (individuals, groups and organizations) involved.Many actors are involved in preparing as well as implementing policies. Individualsact on behalf of a number of different organizations, national departments, subnatio­nal governments and bureaucracies or private organizations. To make policiessuccessful, actors have to cooperate, communication among them has to be adequate,and policies should not be fragmented. This is difficult to achieve especially if actorsacting on behalf of several governmental levels are involved. A number ofinstruments have been developed to solve these problems, one of them beingcoordination.5

In monocentric analyses the central government is seen as the one (rational)actor. The government initiates policies, determines policy goals, chooses theinstruments for attaining these goals and decides which actors should perform whichacts. This is in contrast to the other dominant perspective, the polycentric (or network)model, in which it is assumed that coordination cannot solve every policy makingproblem because actors have different interests and different opinions with regard tosocietal problems. They differ in their policy preferences, goals and choice ofinstruments. Each actor has some necessary but insufficient means to reach his goals.From the polycentric perspective, no actor can attain his goal on his own. Every actordepends, to some degree, on other actors' resources. If the government (or govern­mental actors) wants to attain policy goals it has to activate other actors (Scharpf,1978).

The prescriptive function of the polycentric perspective for policy makingprocesses is not yet very clear. In the literature there are two different solutions(Hufen and Ringeling, 1990). The first is to increase the number of actors involved.The other solution is to utilize the interdependence between actors. Actors mustexchange resources and should bargain about it. 6 Actors who want to steer cancreate new arenas. Klijn and Teisman (in this volume) accept the policy freedom ofthe actors required for a policy. Moreover, they see it as a condition for successfulsteering. There is no general rule for improving policy making processes andincreasing the effectiveness of policies. The authors recommend the development oftailor-made makeshifts. These makeshifts should take into account the interdependen­ce between the actors required for effective policy making.

8. Steering problems in the autopoietic perspective

Although monocentric and polycentric analyses may be valid, it is questionablewhether they sufficiently take into account the consequences of the historicaldevelopment (ontogenesis) of the modern welfare state, in particular the resultingoperational closure or independence of subsystems, and the indifference of subsystemsto the environment.7 Polycentric or network analysts seem to concede that society isnot ruled by a monolithic government. They see many intertwining relationshipsbetween actors. They even recognize the importance of the perceptions (or images) ofother actors and their acts. Actors know other actors' reinterpretations and reactionsto some extent, because of previous contacts with the same actors and because of thesupposed (bounded) rationality of actors. 'Stakes and stands' are to some degreeknown, and so are reactions to steering signals. Reality is (to some extent) predictable

39

and actors are rational. Problems are real problems although somewhat colored by anactor's position in a network.

But should not some other aspects be taken into consideration? The polycentricscholar focuses on actors, their interests, preferences, freedoms and strategies. Heseems to acknowledge that these do not have any significance of their own. Inpolycentrist analysis they acquire significance in interaction. But the polycentristoverlooks that policy goals and outcomes are not only influenced by organizationalpositions ("where you stand depends on where you sit"; Allison 1972), but also byprocesses of giving meaning to features. Statements, goals and signals from otheractors are reinterpreted not only with respect to their origin, but also in the specificcommunications of other organizations and functional subsystems. Organizations oractors not only have images of other organizations or actors, but also of themselves.Signals are reinterpreted within the subsystems, even if the consensus between actorsrequired for the policy making process is present. How is it determined which actorsare required for a policy? Can this be done objectively, or is the requirement alsoperceived?

Autopoietic analysis alters this view. Autopoiesis theory presupposes thatsystems determine their own boundaries. Problems and possible solutions do notcome to systems from an outside world, but are generated within the system itself.The relevance of problems and solutions is determined by the system. A problem isnot a problem if it is not considered to be a problem. The search for problems andsolutions is not influenced by the organizational positions of actors, but by theinternal operations of systems, by the specific communication processes in whichsignals, perturbations, acquire their meaning.

In the second place, a system's variations and reactions are not facts, nor canthey be predicted as such. There is no objective knowledge. Knowledge is determinedinternally. It is a reflection of the ontogenesis of the system. Consequentially, systemscannot be objectively aware of another system's reaction to present steering signals, orof its reaction to a previous signal. Knowledge of another system's reactions is nomore than an interpretation and a reflection of the system's own internal operation.Acts are reinterpreted and given meaning in communication, in the 'language' of thesystem. It must be concluded that propositions about causal and final relations aresimply interpretations by a single subsystem, for example, the political subsystem.Scientific research cannot change this; results of scientific analyses are also justinterpretations in scientific language and these results are themselves reinterpreted inother subsystems. Political knowledge is not the same as scientific knowledge.

In the third place, things which occur beyond the boundaries of the system areobserved as perturbations which systems react to by varying their internal operations.Public policies and steering signals are examples of such perturbations. The variationof the internal operations is, however, not determined by the policy signal, but by thesystem itself. The properties of the policy instruments applied are of no importance.Instruments and their signals are not observed as they are meant in politicalcommunication, but are interpreted in the language of the affected system. Systemsgive meaning to signals. If this meaning corresponds with the meaning of theinstrument in political communication, steering is effective. Reactions to instrumentscan change with time or stay the same. The only source of the reaction and changesin reactions is the internal operation of the system. The relevance of information isdetermined internally, not externally by the environment (Luhmann, 1989).

40

9. Steering from the autopoietic perspective

Does analysis from an autopoietic perspective prove that steering is impossible? If asubsystem's operations can only be varied and changed by the subsystem itself, whatcan other subsystems do? Is autopoiesis an anti-steering theory, emphasizing theimpossibility of steering? Is it impossible for the political system to function, i.e., tomake collectively binding decisions (Luhmann, 1981)? The answer would, at firstsight, seem to be yes. Changes only take place through self-observation, self­description, self-reflection and self-steering. Steering from the outside would seems tobe impossible.

This is not just a political problem, but a theoretical problem as well. Luhmannstates that functional systems depend on each other for an adequate fulfillment of allfunctions. Systems cannot replace each other, cannot fulfill other systems' functions,but have to trust that the other systems are doing their job.8 The juridical, economicand other systems have to trust that the political system will fulfill its function.

This theoretical imperative forces autopoiesis theorists to develop steeringpossibilities. In Luhmann's opinion steering problems do not result from actorsresistance but from steering theories. To put it another way: steering problems areonly problematic if one presupposes that steering efforts will lead to the effectsdesired in politics. From an autopoietic point of view, it is clear that a system'soperations do not determine the environment and other system's operations. Steeringis just an internal operation of one system, even when it is directed towards thesystem's environment, as is often the case in the political subsystem.

Political steering, in general, means the political subsystem observes anothersystem by means of its own political distinctions, that is, under the code of power.The political system makes a distinction between, for example, present and futureunemployment and acts on behalf of this distinction. But these distinctions and theobservations are internally generated and so are expectations for the effects ofpolitical acts. Political acts do not direct changes in internal operations of othersystems. Actual changes can only be reached by systems themselves. But in politicalcommunications these changes can still be seen as resulting from political steeringefforts. Steering efforts can be perceived as input for other systems. This constructionis just a component of the distinction the system tries to minimize. "In der AuBenweltgibt es weder Inputs noch Outputs, weder Informationen noch M6glichkeitsbereiche,aus denen Informationen ausgewahlt werden. Die AuBenwelt ist, wie sie ist: stur,m6glichkeitslos und unbekannt". 9

Luhmann's analysis does not indicate how steering can be done; the prescriptivechallenge remains. Some autopoiesis theorists, especially Teubner and Willke, tried tomeet this challenge (Teubner, 1989a; Willke, 1983). We will limit our overview oftheir work to some central concepts. In Teubner's opinion, systems can co-evolve andco-vary, although they cannot make real-life contact with each other. They canstimulate each other by making 'noise', by perturbing the environment. Every socialsystem is a communication system; they use the same 'source', viz., meaning. Everycommunication in a system is communication in the surrounding societal system atthe same time. But what remains is that systems can only be stimulated, not directed,to change themselves. For example, the juridical subsystem cannot change theeconomic subsystem, it can only stimulate the economic subsystem to change itself:"Gesetzgebung als Fremdregulierung durch Selbstregulierung", as Teubner stated. to

41

Systems can perturb the environment and can influence other system's internaloperations by means of interference (Teubner 1989b), attractors (Teubner, 1990b),perturbations (Maturana, Varela, 1972/1980), ultra cycle and tangential response(Teubner, 199Oc, in this volume). In this chapter we will only elaborate on interfer­ence, this being the basis of steering possibilities. Teubner in his chapter will developthe ultra cycle and tangential response.

Through interference the autonomous self-reproducing systems can achievesome sort of contact. Interference can take place in several forms, be it interferenceof events, structures or persons. All social systems use communications, consisting ofinformation, utterance and understanding ('Verstehen'). Communications are givenmeaning within every system, and the meaning depends on the internal operations (orreference frame). Communication in one system is communication in the societalsystem at the same time and can be communication in another system. If it is, thecoupling of systems takes place. But even then, information is generated again in eachsystem, and thus given specific meaning.

Willke used some other ways to conceptualize steering possibilities from anautopoietic point of view (Willke, 1987). His central hypothesis is that, in order torealize their potentiality, western societies have to institutionalize mediating behavior,intermediary institutions and intermedial steering mechanisms. In organizations andsociety three modes of steering can be found:1. reflex and ritual, as the classical form in which a system reacts instinctively to

signals of other systems;2. representation, in which a system tries to get a picture of its environment and the

system's reactions to signals is based on this image;3. reflection, in which a system not only tries to get a picture of its environment

from its own point of view, but is subject to self-consciousness. It can enter intorelations with itself, can understand its identity as one amongst others. Reacti­ons to signals does not only take into account the picture of the system'senvironment, but the self-made external picture as well.

Reflection means that a system's options come under pressure of selection. Ittherefore is a form of steering, that is, self-steering, for which the system needs aninformation processing intermediary net between input and output. Informationprocessing is a qualitative process of compressing and strengthening cognitive imagesmeant to build an adequate internal model of the system's environment (Willke, 1987,p.30). Unfortunately, Willke did not elaborate on this intermediary net.

The type of action steering corresponding with this steering mode is relationingprogramming ('Relationierungsprogramm') 11. The different goals and rationalities ofthe systems involved are not nivellated, but their complexity is maintained. They areinterrelated and their premisses and consequences are coordinated and even related,through the intermediary net. By using this type of steering the state enters intosocietal domains from which it is traditionally excluded, domains beyond its tradition­al function. It becomes possible to use the specialized and decentralized problemsolving capacities of subsystems for goals of steering the whole system. Consequently,the state trivializes its sovereignty allowing other societal actors into its traditionalfields. The state becomes primus inter pares. In return other societal subsystems getresponsibility for the total system and therefore have to take the whole into consider­ation. In this way they are forced to reflection, theoretically and rhetorically.

42

10. Conclusion

The purpose of this chapter was to provide a discussion of the autopoiesis theory. Anattempt was made to summarize some elements of the autopoiesis theory as originallydeveloped by Maturana and Varela to characterize living systems. A brief overview ofpossible applications in the social sciences was given, which focussed in particular onLuhmann's approach. Luhmann developed an autopoiesis theory for the socialsciences in which social systems are self-referential and self-reproducing.

Using Luhmann's theoretical perspective, it was possible to illustrate thedevelopment of the modem welfare state. During this development several societalsubsystems formed. This resulted in a growing number of interdependencies betweensystems and independencies of systems. Society consists of interdependent and relativeautonomous subsystems.

It was stated earlier that because of this development, steering has become avery difficult task. A number of dominant perspectives for analyzing policy makingand steering problems were presented, Le., the monocentric and polycentric (network)perspectives. It was argued that these models were adequate to some extent butoverlooked the consequences of the ontogenesis of the modem welfare state, inparticular the growing autonomy of functional systems. Autopoietic analysis showsdifferent causes of problems, different problems and different solutions. Problems andsolutions come from within systems. Reactions to signals from the environmentcannot be predicted. Signals are interpreted by the system and reactions aregenerated in the internal operations of the system.

Influencing or steering autopoietic systems from the outside seems to beimpossible. Autopoietic theorists must, however, find ways to change systems' internaloperations. Teubner and Willke in particular have tried to, developing the concept ofinterference (Teubner) and reflection (Willke). But a good deal of theoretical develop­ment and research has yet to be done before it is clear how steering is possible in anautopoietic environment.

Notes

1. Of course for their energetic (physical/chemical) processes living systems require material inputfrom the environment (metabolism), but not for the process of self-reproduction.

2. Luhmann 1982a.

3. Snellen 1987. See for autopoiesis and rationality Willke, in Haferkamp/Schmid, 1987.

4. See for increasing interdependence Elias (1969/1982), De Swaan (1988); and for increasing anddecreasing interdependence Wiltke (1987).

5. They were summarized by Pressman (1975). See also Klijn and Teisman in this volume.

6. Pressman used the term 'bargaining arena', see Pressman, op.cit.

43

7. The polycentric perspective, in a way, copes with the illterdependence of subsystems.

8. Luhmann 1986, p.13. Compare Luhmann's legitimation theory, see Heidorn 1982.

9. "In the world outside there are no inputs and outputs, no informations and areas of possibilities,from which informations are chosen. The world outside is like it is: dumb, without possibilitiesand unknown", see Luhmann, 1989b, p.334.

10. "Legislation as regulation from outside by self-regulation", Teubner, 1989b.

II. Willke, 1987, p.54. Other types are: conditional programming: the classical juridical rule: 'if­then'; and goal programming: an elaborated rule is meant as goal oriented instrumental acting.

44

Partn

CONFIGURATION THEORY

Chapter 4

SOME NOTIONS ON SOCIAL INfEGRATION AND STEERING

H.I. van Dongen

1. Introduction

A lot has changed since we started our research program on social integration and Ideveloped the first tentative theoretical formulations. Several books have been publishedand, as was foreseeable, a certain diversity on the basic concepts has developed overtime. In this sense it is of importance to note that the development of the theory is stillunderway and has not yet been finalized. This paper is to be considered as part of theongoing debate on social integration.

In 1940 Gaston Bachelard published La Philosophie du Non. After discussing thetraditional concepts of science and proposing a non-Aristotelian logic, he discussed twophilosophical theorems. The one position defends the superposition of theories, whichmeans that if two theories are constructed, there is the possibility to construct a theorywhich unifies the two. The other position defends the juxtaposition of theories: theorieswhich are opposed might be coordinated by a philosophie du non. Although Bachelardbased this insight mainly on what was going on in physics, it still seems highly relevantfor social theory in the nineties. Apart from the semantic jungle in social theory, we canidentify several mutually exclusive theories. An example can be the fundamentaldifference between analytical theory versus what I call social constructivism. The firsttheory accepts an external universe 'out there'. This universe has invariant universal pro­perties and the goal of social theory is to clarify these properties and their operation (seeGiddens, 1987). Very different is the position of constructivists who are of the opinionthat reality is a social construct. To exist means 'to be constructed' as the mathematicianHeyting (1966) once formulated. The different, often implicit, ontological perspectivesof different theorists determine to a large extent the state of affairs in social theory,although this is often not recognized or acknowledged as such. The consequences are farreaching affecting the definitions of basic concepts, for example, as well as situations inwhich the rules of logical construction are involved. We will elaborate on this issue, asfar as possible, within the context of this paper and discuss the relationship betweenAristotelian logic and behavioral logic as advocated by Herbst and Spencer Brown.

47

2. Logic and Thirdness

In 1980 Rescher and Brandom published The Logic of Inconsistency, a study in non­standard possible-world semantics and ontology. They introduced the word ontologybecause they wanted to make it clear that Aristotelian logic, which is guiding westernscientific development, is based on the ontological perspective of Parmenides, a traditionwhich holds that the world is a consistent system and that knowledge must be coherentand eschew contradictions. Reality does not contradict itself. One of the classical Aristo­telian principles is the law of excluded middle: given a world w, than for any propositionP, either P obtains in W or its negation (- P) obtains in w. There is no other possibility:Tertium non datur. The other ontological position goes back to Heraclitos, a traditionwhich insists that the world is not a consistent system. Rescher & Brandom: "Such aworld would have to have the feature that in some perfectly definite way something bothis and is not so" (1980, p.2). The point here is not epistemological inconsistency but hardexistential inconsistency. Rescher and Brandom argue that an outright rejection ofontological inconsistency is certainly not necessary and perhaps not even desirable in thesystematic scheme of things. For them, accepting a logic of inconsistency, accepting thenotion of a non-standard world, means that in a way the totalitarian epistemic concernfor consistency may be by-passed. The so-called ontological basis for the standard worldhas essentially more to do with ideologies or specific belief-systems. A more severeposition against Aristotelian logic has been given by the intuitionist Brouwer who alsoquestioned, in a formal way, the law of excluded middle and who showed that there areundecidable problems in mathematics. Herbst (1976) and Spencer Brown (1972, 1972)developed a foundation for behavior logic. In his book, Alternatives to Hierarchies (1976),Herbst describes totalitarian logic in terms of a set of assumptions which are taken tobe self-evident and beyond doubt. The rational and logical consequences are then alsobeyond doubt. Herbst calls it: 'mytho-logical' systems. Making a joke on what he calls thistotalitarian logic, he gives the following characteristics of basic axioms:

1. Quality Axiom: Persons are good or evil. They cannot be both.2. Consistent Attribution Axiom: A good person can only have good characteristicsand an evil person can only have evil characteristics.

Subsidiary axioms are:3. Pennanence axiom: Personal characteristics are permanent and not subject tochange.

4. Inconsistence resolution axiom: If a pattern consists of inconsistent attributions(axiom 2), that is, both negative and positive characteristics, then either thepositive characteristics are pseudo-positive or the negative characteristic arepseudo-negative.

Spencer Brown and Herbst show that the making of the primary distinction by itself issufficient to generate the conceptual and logical structure in terms of which Herbst latercomes to apprehend the world in which he finds himself. "These structures therefore maynot need to be assumed to pre-exist either in the external world nor internally, since thedistinction between the internal and external is already by itself sufficient for thestructure of logic". (1976, p.87). This is in contrast to the ideas of Piaget who assumesthat the infant learns a correspondence between the emergent structure of his operationson the one hand and on the other an independent system of logic, a logic out-there. No,what the infant learns is a consequence of a primary distinction he makes between

48

himself and his environment. One might say with the production of environment (auto­poiesis), with the production of organization (post-modernists), the infant also generatesa logical system. This understanding of the dependence of logical and conceptualstructures on the primary distinction between self and environment generates thepossibility of transcending a specific logical and conceptual structuring as the only modeof knowledge. Herbst gives a representation of the primary distinction as a triadic set ofelements, a triad of co-genetic components. Co-genetic then means that the threecomponents come into existence simultaneously. The following shows some possibletriadic sets.

Internal Boundary External

Finite Boundary Infinite

is Boundary is not

(x) (-x)

It is easy to see that the components are co-genetic. If one of the components isremoved, then the other two disappear as well. So, for example, if the boundary betweenoutside and inside is removed, the distinction between outside and inside is no longerpossible. If in the examples given one of the components is removed, we then return towhat Herbst calls the original state, a state free of any distinguishing characteristic. Weprefer to call this original state the ongoing condition of un-differentiatedness. Un­differentiatedness is beyond the possibility of making distinctions and cannot be definedin whatever 'mytho-ontological' perspective. We will in time give a more detailedargument for the re-framing of the original state into the condition of un­differentiatedness. We can from this perspective understand whyWittgenstein wanted tobe silent. We can also understand why Levinas wants to defend a humble position ofresponsibility face to face with the vulnerable other who refers to, but not defines theOther. From this perspective we also can understand that Aristotelian logic has forcedmany philosophers and scientists to attempt to define the undefinable, an ontology tograsp the whole of reality in a coherent, consistent model. This circularity of 'ontologicallogic' to 'mytho-logical ontology' generates serious problems for reflection on socialintegration and ethicality as we will see later on. What becomes clear in Herbst'sargument is that we should understand elements as products of a dynamic process of dis­tinguishing, of drawing boundaries and, in that sense, characteristics of elements areessentially to be understood in terms of triadic relations. Peirce, in his The principles ofPhenomenology (1940, p.91) in which he discusses Thirdness, also shows that meaningdepends on two premises. The first is that every genuine triadic relation involves mean­ing; the second one is that a triadic relation is inexpressible by means of dyadic relationalone. So: "Thus your right hand is that hand which is towards the east, when you facethe north with your head towards the zenith. Three things, east, west and up are requiredto define the difference between right and left." It is therefore necessary, says Peirce, totrain yourself in the analysis of relations. He also shows that every relation which istetradic, pentadic or any greater number of correlates is nothing but a compound of

49

triadic relations. The basic triangularity is an important and very different assumptionin comparison to Aristotelian logic. It is also of importance to note that within severaldomains of social theory and philosophy, this triangularity has been recognized in avariety of ways. In this paper we will not elaborate on that. Still for our reflections onsocial integration we need to take another step.

3. The boundary, the actor, socio-cognitive configurations

So here again we assume that the basic 'building block' of human interaction is a triad.Let us for a moment reflect on the example of Peirce concerning east, west and thezenith. One might say that the cleavage in space, which Spencer Brown talked aboutwhen he discussed the boundary, could be compared with the head in the zenith, inPeirce's example. Two things should be remembered here: the first is Herbst andSpencer Brown's insight that the observer is always part of the logic employed and,secondly, that every distinction is to be understood as a co-genetic process of a triad ofcomponents. The zenith in the example can now be understood in different ways. Firstly,the zenith as a cleavage in space comes into being as a consequence of the distinctionproduced: east - west. Secondly, the zenith can be understood as a definition of the actorwhich comes into being as a consequence of that same distinction. Let us try togeneralize this example. Suppose we have the following 'distinction':

Internal Boundary Apple

As my wife would say: "This is a child's world, in which everything is all the same." It isbasically a condition which refers to un-differentiatedness. We cannot make sense of theboundary, we cannot place ourselves between internal and apple, nor can we understandthe unity of the two poles, namely, internal and apple. We would certainly need a fancyhumorous imagination. This means that the boundary itself separates and unites, whichmeans that the boundary not only is to be understood as a cleavage in space but also asthe processual act through which the poles stick together. So the boundary separates andunites simultaneously and, as such, illustrates the dynamic characteristic of everydistinction. The second understanding of the boundary is in terms of the actor who unitesand separates. Goguen and Varela (1979, p.32) make the following observation: "Theestablishment of system boundaries is inescapably associated with what we call acognitive point of view, that is, a particular set of presuppositions and attitudes, aperspective, or a frame in the sense of Bateson or Goffman; in particular, it is associatedwith some notion or value, or interest. It is also linked up with the cognitive capacities(sensory capabilities, knowledge background) of the distinctor. Conversely, the distinc­tions made reveal the cognitive capabilities of the distinctor. It is in this way thatbiological and social structures exhibit their coherence, and make us aware that they are'conscious' in some degree." The consequence of this argument is that there is nodistinction without a distinctor, so the boundary is to be understood in terms of the actoras well to overcome the static characteristic of a distinction. The distinction is not an'out-there' distinction. The actor as well as the poles of a distinction are defined in termsof one another. This essentially refers to the triad of co-genetic components. The com­ponents by themselves only can refer to the condition of un-differentiatedness and thatgoes for the actor as well. This argument has a series of consequences. What is of

50

importance is to clarify that every distinction can be understood in terms of the distinctorand, secondly, that the distinctor can be understood in terms of the distinction produced.This two-sided understanding of triangularity represents, in our view, the differencebetween socio-structural and the cognitive-structural aspect of any distinction, of possiblerealities, of worlds of meaning. Cognitively speaking, the distinctor is defined in termsof the distinction made, which means that he is included in the cognitive structure in thesense of internalizing the difference between the poles of the distinction as well as interms of the relatedness of the poles. Apart of the distinction produced as a componenton its 'own', the distinctor refers to the condition of un-differentiatedness, which wecannot define in any way. The world of distinctions together with the condition of un­differentiatedness are referred to here as symbolic realities.We use this term to escapethe ontological temptation and leave that as it is without attempting to label it with anywords. We will argue that the basic notion of reflexivity has to do with the ability tochange either the distinctor or the distinction. However, before entering this debate, weneed a further elaboration of our concepts.

4. Multiple Inclusions

More has to be said on the understanding of the actor. First it is of importance tounderstand that the actor emerges in many ways in continuous processes of makingdistinctions. If we, like Weick (1979, p.97), take seriously the idea that meaning isretrospective, then over time any distinction refers to others with whom we learned anew meaning. Consequently, we can reflect on the social-structural aspect of any meaningwe develop over time. Goguen and Varela invoke the idea of an 'observer', one or morepersons, who embodies the cognitive point of view which erected the system in questionand from whose perspective it is subsequently described. So, understanding themultiplicity of distinctions, being reflexive on our sense-making, also requires a historicanalysis of the social contexts within which we learned and developed and gave meaning.Thus, specific cognitive structures, definitions of situations, refer to social-structures, tofaces with whom we learned a specific content. Normally we by-pass this socio-structuralselectivity of our sense-making process. This means that we are often not aware that themultiplicity of socio-cognitive configurations we are part of over time does not need toconverge to one homogeneous definition of reality. We are multiple included in a varietyof possible divergent, different configurations of sense-making processes. It is essentiallythis characteristic of multiple inclusion which offers the possibility to learn and developnew insights. We refer here to the insight of Moscovici and others who point at socio­cognitive conflict as a basic characteristic of learning processes. We tum here to the post­modernists who show that dissensus and not consensus is the motor of change. So, theuniqueness of 'the' individual, of 'the' group, of 'the' organization, of 'the' State is not tobe understood as a relative static entity nor as a relatively homogeneous set ofcharacteristics. The uniqueness of different inclusions in various 'worlds' gives a farbetter insight into the processual characteristic of so-called 'entities' as multiplicities ofcontinuously unfolding socio-cognitive configurations. This understanding also offers aninsight into the process of unfolding of configurations as a continuous negotiation withinas well as between 'individuals'. Convergence and difference simultaneously makeinterlocking behavior a possibility.

51

S. Thirdness, metaphorization and reflexivity

Thus far we have elaborated on overcoming the dyadic structure of rules of constructionin favour of the insight of triangularity. We also advocated the idea that any meaningrefers to an actor or set of actors who possibly share specific distinctions or meanings.Distinctions not only refer to the distinctions themselves, but also to others with whomwe share similar distinctions. By the same token, however, these distinctions also referto possible others who hold a different perspective on reality. This means that thirdnesscan be understood in a cognitive-structural as well as social-structural way. Thirdness atfirst sight refers to the content level, to the cognitive structure, which after furtherscrutiny reveals itself as a social construct of which we, amongst others, are part of. Thethird refers to the possible other actors with whom we share or who produce differentdefinitions of reality. The distinction between thirdness and third is of importance for theprocess of metaphorization. Introducing a possible third as a strategy of metaphorizationmay illuminate a new meaning, which offers the possibility of reflection on one's ownbeliefsystem, one's own set of meanings. Introducing thirdness gives the possibility tounderstand and identify other actors, other social configurations we were not consciousof. This last strategy of metaphorization offers the possibility of reflection on one's ownpossibly idiosyncratic social configuration. Fundamental to this strategy is the 'crossingover' from the socio-structural to the cognitive-structural perspective and visa versa. Inboth cases the purpose is to understand the selective 'Gestalt' of the socio-cognitiveconfiguration as a whole. The ongoing process of metaphorization is to be understoodas the basic idea of reflexivity and essential for social integration and ethicality.Essentially this is what Maas (1988 and in this volume) and Voogt (1990) refer to whenthey discuss the concept of game versus play. Game then in this context refers to theconsensually validated definitions of reality (to wink at Weick) whereas play refers to theongoing methaphorization of possible other 'reality games' as fundamental for changeand development. One might say that ir-reflexive socio-cognitive configurations will arguethat facticity refers to how things really are and not to possibilities. This is what Herbstwas talking about when he discussed totalitarian logic. One might say: a game of logicwithout a logical play, a game in which, for example, the ontological assumption ofconsistency is beyond any doubt.

6. On social integration

Katz and Kahn (1978), referring to Durkheim, observed that in primitive society socialintegration was based on a common value system, a mechanical solidarity based on acollective conscience or a common morality manifested by all societal members andappropriate to all situations. The assumption then is that in highly differentiated societiesthere is less of a universal moral code. Social integration is supposed to depend on nor­mative practices and role interdependencies. Blake and Davis (1968, p.468) remark that:"Presumably, in anthropological and sociological thought since Durkheim andMalinowski, the values must have a function in society; but the question is obscurebecause, for many social scientists at least, what is taken as a value may be dysfunctional,and what is functional may not be valued." They point to the fact that research often usesnorms and values as explanatory principles for concrete behavior, a tendency to explainthe known by the unknown, the specific by the unspecific. An often implicit assumption

52

is to understand values as the cause of norms. Blake and Davis are of the opinion thatit is better not to talk about cause and effect. No, a norm exemplifies a value, which doesnot mean that the norm is caused by the value it exemplifies, or that the value is themotive and the norm simply the expression of this motive with respect to behavior. Theyprefer a more satisfactory use of 'values' by abandoning them as causal agents andrecognize them frankly as sheer constructs by which we attempt to fill in the subjectivelinkages in the analysis of social causation.

This type of argument illustrates elegantly that the discussion on social integrationis often defined in terms of an effort to define a specific cognitive structure, a worldview, as a leading principle for social integration. In the foregoing, however, we arguedin favor of the notion of multiple realities. We also showed that triangularity ofinteraction and sense-making processes are a basic characteristics in the development ofpossible 'worlds'. The heterogeneity of reality can also be seen as fundamental forlearning and development. The problem of wanting to define a specific homogeneouscognitive structure as the ultimate valuesystem is that, as we illustrated, it must be of aselective character. This now means that these kinds of efforts could be regarded asfundamentally of an ideological character. Without elaboration we now point to the factthat, for example, Nozick's theory on the minimal State and Rawls' theory on justice asfairness both can be regarded as based on ontological assumptions, which as a matter offact can be qualified as basically of ideological character. Our approach to socialintegration, which we will indicate here, is far more of socio-structural character. Theassumption, as we saw above, is that symbolic reality is characterized by ongoingprocesses of construction and reconstruction. A next assumption is that we areconfronted with undecidable problems as far as 'truth' of reality is concerned. So wherethere is no perspective in defining a total concept on basic values, it is far morepromising to understand social integration as a social structural characteristic of the waysin which we produce meaning.

In that sense we can interpret the research findings of Kohlberg, Colby and others.What they show is the moral development basically in terms of injustice instead of tryingto identify basic concept of justice. It is similar to the ideas of Bachelard on thejuxtaposition of theories and the 'theory du non' in the following way. It might be so thatpeople have different worldviews which also lead to divergent opinions on justice, socialintegration and so on. This diversity of opinion is as such of great importance for theprocess of social integration. However, this does not mean that there is in the concretesetting of everyday life no convergence on negative values, that is, on what is notacceptable without defining what is morally just. In our opinion, the Kohlberg cross-cul­tural studies illustrate this conclusion. Our concept, as we formulated it earlier on un­values, has to be understood in terms of a dynamic ethical procedure and not as anethical system which is based on a pre-defined concept of justice. The theory of un-valuesas basic for social integration defends the position that the social continuity of sense­making as well as the condition of un-differentiatedness might be disturbed and that assuch is a basic ethical problem. So, if we define the concept of ethical space as basic tosocial integration, we mean by that the continuous explication on the behavioral andattitudinal level of that which is morally not acceptable.

Operationally the question becomes how to define the ethical procedure whichprotects the ethical space. This is where the concept of reflexivity fits in, in the sense thatthe ethical procedure is directed at the question to what extent in processes ofconstruction and re-construction the third or thirdness is violated. Above we argued the

53

double-sided strategy of metaphorization, namely, the socio-structural strategy ofmetaphorization and on the other hand the cognitive-structural strategy of metaphori­zation. We argued that for both strategies the purPOse was to understand the wholenessof a specific socio-cognitive configuration as a possibility. Fundamental to this strategyfor reflexivity is the introduction of thirdness or a third. Irreflexive realities forbid inessence this strategy because of, for example, an absolutistic ontological perspective.Reality is, reality is as it is and not a possibility amongst others. This position does notaccept the social construction of realities, which means that thirds or thirdness isrepressed. Thirdness and thirds only are recognized and accepted if there is a fit with thesocio-cognitive configurations which gives an ontological status to selective definitions ofreality. This now is fundamental for totalitarian constructs of reality which forbid thirdsand thirdness. This is the understanding, for example, of the third, the other as being themost vulnerable, the other who suffers injustice (see also Rade in this volume). Toformulate the link to Levinas, we could say that the development of social moralconscience is co-genetic with the actual concrete safeguarding of ethical space and ethicalspace can be understood as safeguarding the difference, the diversity of other socio­cognitive structures. To become a little more operational, we could say that the on-goingprocess of metaphorization can be seen as the continuous elaboration and explication ofpossible thirds and thirdness. Ethics then becomes processual in the sense of questioningthe openness of socio-cognitive configurations.

7. Steering negatively and positively

Above we showed that socio-cognitive conflict is essential for learning and development.The multiplicity of world views in that sense is of importance in that learning process.Where we can understand the diversity of reality constructions, we also can understandthe social construction of futures (Van Dongen, 1988), which can be understood as anegotiation process on possible realities. Managing can be seen as a process in which twoelements have to be taken into account. The first being the maintenance of ethical space,the safeguarding of the continuity of vulnerable thirds, as a basic condition for socialintegration. We call this negative steering because of the fact that here it is not the issueto develop a concept of justice; it is on clarifying in-justice and taking the consequences.The other element is the process of positive steering. Here the issue is the process ofnegotiation on possible realities and futures. On State level, this is where the politicalprocess fits in. What we can observe in practice is that negative steering which isfundamentally ethical in character becomes increasingly mixed up with ideologicaldebates on possible futures. Essentially, these two aspects of steering differ considerably.In the political process of decision-making it is of great importance to make explicit whatkind of issues belong to the safeguarding of ethical space and which belong to thedomain of negotiable order. When these kinds of distinctions are not made explicit, onemight expect a continuous displacement of scope which in tum may lead toinstitutionalization of equivocality which afterwards could be diagnosed as thedisintegration of the political decision-making process.

54

Chapter 5

OFFERING NEW POSSIBILITIES OF ORGANIZING

IN SOCIAL SITUATIONS: TOWARDS A POSTURE

OF DECENTERING AND CENTERING}

A.J.J.A. Maas

1. Introduction

The other day I observed (and was part of) the following scene. Both my sons hadbeen told by my wife (and their mother) to take a cup of hot chocolate from thekitchen. They ran into the kitchen and, in front of the kitchen sink, they apparentlyhesitated a moment: on the counter stood a cup and a mug of hot chocolate. Whichone would each take? Quick as lightning my elder son took the cup and cheerfullysneaked out of the kitchen. At exactly that moment the younger three-year-old burstout crying, his usual means of expressing a variety of messages and feelings at once.There I was. What could I do? Run after my older son and punish him? (Why?)Console my younger son? (Why?) Walk away? Call their mother? Ignore the wholescene? I was struck silent: many possibilities kept running through my mind and allwere immediately dismissed. Intervene? Stay out of it? Why not both? Finally, Idecided to do both and to invite the boys for a short story-telling session. They lovedAndersen's 'the emperor's new clothes' and I hoped that telling the story would cheerup both my sons (my wife and myself). Within minutes, all of us were sitting cozily onthe couch, concentrating on Andersen's story and leaving behind the incident for amoment.

Later that day my wife and I discussed the incident and were overjoyed todiscover that we had found an alternative means to define the incident, withoutdenying the sociality of the incident. Andersen's story kept us all distant from theincident as well as involved, because we had taught our children to look at the kingsimultaneously as a man of importance, a man of greed and, of course, as a nakedman. Moreover, the story generated a means to cope with the variety of interpretati­ons of the situation all of us had. In the story we found some metaphors which helpedus to discuss the incident with our two children. In other words, it stimulated ourcapacity to understand the incident from several perspectives simultaneously.

In many domains of our life we encounter situations like this, from very simpleto manifold layered or complex, in which we try to organize realities. This kitchenincident illustrates and calls attention to an important issue. This aspect is rarely

55

discussed in the literature on organization theory and public administration, but iscrucial in processes of development, learning, change and intervention: the process ofdecentration and centration. In this chapter I want to initiate a discussion of analternative way of intervention and critically examine the recent literature. In myopinion, there are some missing links in the literature. Using the methodological baseof the Rotterdam theory of social integrating or configuring, I will develop elementsof a radical procedure of decentering and centering. Some aspects of the procedurewill be illustrated with an example of a process of intervening in the complex field ofinforming and advising (I&A) small and medium-scale enterprises (SMEs).

2. Centering and decentering: a philosophical point of view

In a recent publication, Van Keulen (1989) examined the game situation from aphilosophical point of view. He stated that Western philosophers in their analysis ofthe game-situation made interesting changes of perspective. Essentially his analysis ofthe concept could be understood as a ongoing process of centering, from whichDerrida tries to decenter. Below is a brief summary of his method of analysis.

Van Keulen starts by saying that Kant and Schiller made the remark that ineach game the player is the initiator, the central source of the game. The playerprepares himself for the game and initiates it: for these authors the player could bethe performer as well as the spectator of the game. Hegel amends this position byposing that the artist or player interprets and actualizes historic material: the player(a social aspect) and his interpretation of historic knowledge (a cognitive aspectembedded in social historic contexts) are the center of his attention. As a criticalcomment, I would add that in both interpretations a player organizes reality, but isnot related to other players and their possible other definitions of reality. In thatsense, the perspective of the philosopher is anchored. Van Keulen points to Gadamerto illustrate his first change of perspective: for Gadamer not the player, but the gameitself is the central source. According to Gadamer, the player is not offside: he isinvited, caught up and led away by a bewitching game he never initiated. The playeris always partially included: if he understands the game, his understanding is continu­ally different. Gadamer seems to offer the philosopher the possibility to shift hisposition from player to game, but here as well, the position of the philosopheroriginates at a center and is thus a static one.

According to Van Keulen, Derrida offers another critical change of perspective.In his view, the history of Western philosophy shows that the game definition isanchored in mutual substituting centers. Recent developments in philosophy show aclean break, a definitive process of decentration. Derrida accentuates this decenteringprocess by arguing that the central source of the game does not exist, is not containa­ble and is principally absent. In his words: "l'origine a joue." The game is without acenter and has no totality, but is a playing field of infinite substitutes, which he says isdynamics. Not difference (which assumes a difference from an absolute foundation),but differance (Derrida's redefinition) is essential. It is a playing happening, active aswell as passive, not definable in terms of subject and object. Cooper (1989, p.489)refers to this differance as a continuous absence, as a force that is continually beyondour grasp and therefore never properly present. Derrida also refers to this process asa process of metaphorization.2

56

I want to elaborate on Van Keulen's analysis by relating this to the discussions ongame and play by Voogt, Van Dongen and myself (their chapters in this volume;Maas, 1988). Doing so, I want to sharpen Derrida's position of decentering. In thetheory of social integrating, meaning is principally related to an actor, and vice versa.A cognitive distinction we as actors make not only refers to the distinction itself, butalso to other actors with whom we share similar definitions, and to possible otherswho hold a different perspective, logic or paradigm. 'Game' in this context refers tothe situation in which we share consensually validated definitions of reality withothers in an interaction process. To understand development, learning and change ina social context, (social) reflexivity is an essential element. Basic to reflexivity ininteraction processes is that we are open for possible others, possible other definitionsor possible other interactions. Reflexivity means being reflexive of our social contextand relating our context(s) to others, accepting a process of social reflection based onits own rules or framework of the game. This accepting of a third (definition, actor orinteraction), or better thirdness (as Van Dongen calls it), could be basically on thecognitive as well as social level. This reflexive situation is distinct from the 'game­situation' and is understood as 'play'. 'Play' could be redefined (in Derrida's terms) asthe ongoing process of deliberate metaphorization of possible other 'games of reality'.We could refer to 'play' as fundamental to change and development. Organizingrealities from this point of view constantly requires playing and gaming, which couldbe understood as basic for decentering and centering.

3. Centering and decentering: a social-psychological point of view

In another domain of knowledge, social psychology, we could also detect insights intocentering and decentering. It was the kitchen incident, described above, which drewmy attention to the following Piagetian experiment described by Doise and Mugny: "achild agrees that two identical glasses contain the same amount of fruit juice. If theexperimenter then pours the contents of one glass into a tall thin container and thecontents of the other into a wide flat container the child thinks that there is more todrink in the thin glass" (1984, p.28). In their analysis, Doise and Mugny focus mainlyon the child who first concentrated on the level of liquid in the glasses and failed tointegrate the differences in height with differences in width. But, how would a childbecome aware of the opposite centration as well? In the experiment it was concludedthat it was the interaction with the physical environment which made the childappreciative of the opposite centration. I want to amend this position. In my opinion,in this experiment the social relation between child and experimenter is crucial tounderstanding the cognitive development of the child. By questioning, and thus byoffering the child another definition and therefore an alternative social context, theexperimenter can challenge the child to understand the other centration as well. Thechild must accept this social situation as well. In a later stage of development, thechild (in relation to the experimenter) could be challenged (e.g., by offering metap­hors, such as those in Andersen's story?) to link the different meanings and decidethat a 'third meaning' ("it is the content, e.g., the chocolate, that I like") is a moreappropriate one. In this situation, interaction with another person who offers a thirdmeaning disrupts the child's present understanding which reveals conflict among hisown responses. In this last case cognitive development or change meant: a process of

57

social decentration conditioned the partial or complete integration of the twodifferent centrations.

To reach a deeper understanding of the mechanisms and the content ofcognitive development, the researchers introduced social interaction into the experi­mental situation. They designed experiments in which interactions between severalindividuals became their object of study (1984, pp.77ff). These experimental situationsshowed that interindividual encounters lead to cognitive progress insofar as socio­cognitive conflict occurs during the interaction in the experimental situation. Socialand cognitive poles are inseparable, because the way to resolve a task must be amatter of conflict between social partners. However, in these experiments as well, thesocial interaction between the testees (with their logics and cognitive distinctions) andthe experimenter (with his logics) never was the subject of critical research anddiscussion. Exactly that analysis could have revealed more about the ways bothdistinctions or logics became related. This analysis could also have offered us deeperinsights into those aspects or metaphors which helped the experimenter gain access tothe logic of both testees. In short, such an analysis could have provided us with insightinto possible stages of social decentering, the process of metaphorization andthirdness. Further research on this issue is required (see Bogenrieder, in press).

4. Centering and decentering: an organization theoretical point of view

Most of the books on change and transformations I know are written from the pointof view of a consultant, a possible third person. In fact, this observation means thatthe relation between consultant and clients rarely is the subject of analysis anddiscussion. However, Levy and Merry have presented a rich and extended overview ofthe literature on development, learning, change, transformation and procedures whichcould be used as an incentive to reflexivity. They identify 'reframing' as a pivotalelement in the creative process, the problem solving process and communicationprocesses, by which they mean: "using techniques that either force or enable partici­pants to go beyond their current frame of reference (or paradigm), to have newperspectives on the current situation, and to choose a new perspective to adhere to.In reframing, there is a radical, sudden change in perception and behavior, a jump toa different logical level, while the situation itself (the 'facts') may remain quiteunchanged -indeed, even unchangeable. What turns out to be changed as a result ofreframing is, first of all, the meaning attributed to the situation, and therefore itsconsequences, but not its concrete facts." (1986, p.103). Bolk specifies for organizati­on-consultancy the techniques referred to in this definition as "the organization ofcontacts, the development of alternative interaction frameworks, investigating and,consequently, strengthening 'weak signals', explicitly investigating and producingalternative definitions." (1989, p.45). A common characteristic of these reframingmethods is that they offer a third person the possibility of linking distinctions made inone social context to distinctions made in another social context, and so on: thirdnessis organized. Conditions for selecting which procedures to use in which contexts are,however, not specified.

To overcome this, Van Dijk offers an important argument. He exploresreframing from another perspective and directs attention to the process in whichactors reconstruct repetitive reifications (homogeneous meanings, or fixations). For

58

him that is the moment to perform interventions which are aimed at producing avariation of contexts for actors with these fixations. In his analysis he distinguishedfour forms of variation of context: transposing, homogenizing, heterogenizing andtreating social and content levels as equal (1988, pp.78ff). Consequently, his mainargument for reframing is a negative one, namely, if fixation or repetitive reificationis the main principle of organizing.

I want to elaborate both approaches and thereby specify some elements of aposture of decentering and centering. Fixation in essence means not being able tomaintain distance, to lock onto and indulge in staring at your navel. This distance canbe defined as social distance and cognitive distance. The fixations therefore also haveto be understood as fixations at the cognitive and at the social level. In this respectthe approaches developed by Levy and Merry, Bolk as well as Van Dijk have to beunderstood: some of Bolk's techniques are social-structurally oriented, others arecontent oriented. Likewise, Van Dijk's intervention scheme in fact can take eightforms, each from a cognitive and from a social perspective. Moreover, the schemecould be extended to other forms of contexts variation as well, such as planning,forecasting, qualifying, diagnosing, strategy making, managing change, learning,developing, negotiating, communicating, conflict handling, and so on, and so on.

In this fixation situation, a posture of decentering offers the possibility ofstepping aside, maintaining distance and being detached and relating to other possibleways of organizing. In general, a social fixation requires decentering from the socialaspects to a cognitive approach (of reframing), and a cognitive fixation requiresdecentering from the cognitive aspects to a social approach (of reframing). Thismeans that reframing or offering variations of contexts can be understood as beco­ming reflexive at the cognitive level and as becoming reflective at the social level. Forthis paper the first situation is especially important. It presupposes the absence offixations at the social level and, consequently, is a reframing within a social contextwhere distinctions are shared and can be linked to distinctions in other contexts. Inthis situation social connectedness is a prerequisite; a posture of decentering heremeans an ongoing process of cognitive metaphorization aimed at relating, or linkingvarious other possible reality games in order to understand other definitions of'centers of reality', in addition to actor's cognitive fixation(s). In other words, astrategy like that offers the possibility of seeing reality through someone else's eyes,from a perspective actors are initially not part of. Using possibilities of answering thequestion 'how is an incident or problem viewed through someone else's eyes?' makesone aware of a (social) integrative view on a particular incident or problem. Thisprocess of metaphorization or 'play', this ongoing process of exploring for linking textsor sameness, is followed by a posture of centering or 'game' which refers to theprocess of consensually validating definitions of reality. The posture of decenteringand centering requires an ongoing process of play and game, game and play, and soon. Most authors studying development, learning, change and transformation takethese processes for granted. In my opinion, the distinctions made here are crucial tofacilitating learning and negotiating organizing.

s. Rotterdam theory of social integration in a nutshell

The posture of decentering must be based on a methodological principle, a method ofdereifying, in which a minimum number of definitions and criteria is presumed in

59

advance. An illustration of such an approach is the Rotterdam theory of socialintegrating, which is essentially an empty theory (Van der Meer, 1983, p.44; Pavel,1990). The developers of this theory have produced a method to analyze socialprocesses along which reality unfolds, is 'completed and not', as it can be described interms of constructing and reconstructing realities. The theory of social integratingshould be conceived as a type of relational thinking, which means the ongoing processof exploring 'thirds'. Or, as Van Dongen puts it in this volume: "meaning refers to anactor, and vice versa. Distinctions made not only refer to the distinction itself, butalso to others with whom we share similar distinctions. However, these distinctionsalso refer to possible others who hold a different perspective on reality." The maininterest is to relate distinctions produced by actors in various contexts, socially as wellas symbolically. An alternative definition of this theory is that in a particular situationone is especially interested in the extent of (social) organizing (of realities). Centralto the process of organizing realities is the arranging and rearranging of ongoinginteraction into patterns. These processes of ongoing interaction will at least alwaysbe open to 'a third'. 'A third' refers to anything (person(s), meaning(s), (inter)actionor multiple interactions) outside the parties to an interaction.

Defining this way of understanding realities as 'normative' would be an undeser­ved reproach. The relational modes in which people in an ongoing interactive processproduce and reproduce their organizational contexts are the only selective activity.The central idea of the theory is that understanding and defining realities is funda­mentally a social process. A cognitive definition does not exist in a social vacuum. Acognitive map of the world refers in most situations to a specific social context. Henceit is possible to describe most situations in terms of a spiral relationship between thesocial structure and the cognitive or definitional structure. This double helix refers tothe phenomenon that in most situations cognitive structures do not exist in a socialvacuum and, reciprocally, social structures do not exist without a cognitive structure.In the process of maturation actors belong to more than one social context: they aremultiply included. These socio-cognitive contexts could be referred to as socio­cognitive configurations. Learning and developing might be understood as a result ofthe socio-cognitive conflict between different socio-cognitive configurations. As VanDongen stated (1988), one may recognize that if people within and between configu­rations define reality in a similar way, then the consequences of the rules of construc­tion are expected to be real.

I want to sharpen the methodological framework here. Looking to socialdynamics we can detect two extremes as basic for understanding social processes: aconfiguration as well as an aggregate. Both social situations have to be interpreted asstudies of an instantaneous record, which refer to a process. Thus configuration has tobe understood as configuring, and aggregate as aggregating. Both concepts portray theextent to which people share patterns of ongoing interaction, whether solid or not.Studied as an instantaneous record, social-cognitive configurations are characterizedby a matching between relatively stable ongoing interaction patterns and sharedcognitions (Bolk, 1989). In actual situations they could be characterized by one of thefollowing qualities:

high degree of organizing,solid and intensive communication between members (especially on the strate­gist's territory),configurations are nodal points of information and consultation,

60

formal consultations (records of meetings are present),configurations are recognizable and open to others outside the configuration,configurations are junctions or links, not necessarily restricted to the boundariesof a department, division or organization,configurations are strongly embedded in the direct social environment.

In short, a social-cognitive configuration is essentially a 'solid association of people',Aggregates are characterized by one of the following qualities:

low degree of organizing; absence of or little communication,absence of nodal points of information and consultation,no formal consultations; it is unclear if and when they meet each other formally;communication with others is ad hoc and informal,separate themes, few connections,varying constitution,aggregates are not recognizable and open to others,aggregates are weakly or not embedded in the direct social environment.

In short, an aggregate is essentially a collection of people. Nevertheless, the aggregateconcept can be refined. De Laat (1983) defines an aggregate, in contrast to thecollective structure or configuration, as the absence of differentiation and integrationof the elements. An aggregate is characterized by an absence of interaction. In DeLaat's opinion, the defining of aggregates does not exclude aggregates from functio­ning as a basis for the unfolding of configurations. Beyond the configuration anaggregate exists in an abstract sense, it's true, but is meaningless. De Laat argues thatthe state of aggregating can change in an interactive process or configuring, and theother way around. Following De Laat, it seems plausible to elaborate on the notion of'elements without ongoing interaction' a little more. His 'elements' should also beunderstood as 'shared social structures without density in the interaction patterns'('social aggregate') or 'shared cognitive structures without density in interaction­patterns' ('cognitive aggregate'). In addition, these elements refer to a situation inwhich cognitive-social structures could be described without density in the interactionpatterns ('cognitive-social aggregate'). Finally, situations could be described as 'social­cognitive aggregates' by nature. The last two aggregate situations differ in that in thefirst one, the cognitive structure comes to the fore, while in the latter, the socialstructure is prominent. In all four situations the 'elements' are without ongoinginteraction.

As argued, both concepts refer to extremes in social processing. Between bothextremes we found situations which were neither aggregates nor configurations. Wecall these transitional situations 'configurations-in-development'. This type of configu­ration distinguishes itself by one of the following qualities:

low degree of organizing; channels of communication are locatable or still underconstruction, but do not function properly.some formal consultations, especially between nodal persons; communicationwith others still does not have a distinguishable pattern.some parts of this type of configuration are recognizable and open to others; asa whole however, this type of configuration is typified as 'developing'.this type of configuration still has a weak embedding in the direct socialenvironment.

61

Another crucial concept used is the notion of partial inclusion. This can be under­stood to be "the extent to which the individual's definition of reality, contentwise,converges with a specific social configuration. The word 'partial' refers to thephenomenon that the individual's reality definitions do not totally overlap with theones used by others, because of the fact that he belongs to other configurations aswell. And this is why we may be similar and different at the same time" (VanDongen, 1988).

Based on these theoretical points of view, the posture of decentering formulatedoffers the possibility of intervening in situations of blocking. Being aware of thepossibility of decentering means the ability to analyze and integrate the distinctionsproduced in various contexts, the boundaries of (especially) dominant definitions, thevariety of definitions of reality in related contexts for sameness, and possibilities ofredefining. In so doing, each situation of decentering can be understood as anongoing invitation to explore 'more', 'different', 'varied' possible situations and testthem on sameness. When the rules of construction are expected to be real in terms oftheir consequences, the result is that a process of centering based on the principle oforganizing gaming starts, in which definitions of reality are consensually validated. Inthis process actors make distinctions which are related, because of referring, topossible others who hold a different perspective on reality. Gaming is the name ofthat 'game'. The moment this relating to possible others (and other meanings) isblocked, we have to 'play'.

6. Case: context and discussion of elements in the process of metaphorization

To understand the case as an illustration of the posture of decentering, I willelaborate on some elements from the context of the case, and the process of metap­horization.

6.1 Coruext of the case

For 25 years the task of Informing and Advising (I&A) SMEs in the Netherlands hasbeen partly a governmental task, partly commercialized. Several organizations play animportant role in the I&A process; some started in earlier times (e.g., the Chambersof Commerce),' most are of the recent past. Since 1965 several governmental commit­tees have written reports suggesting drastic and less drastic interventions. The resultof this historically based process was that in 1989, the I&A field for SMEs became anattractive business. An inventory of this field shows a variety of organizations, groupsand actors, private and public organizations, all conceivable kinds of legitimate issues,with complex mutual relations among themselves and to many other fields of interest(such as education, labour market, environment, health service, international affairs,etc.). Obviously, boundaries of systems are vague. In this situation, it is not easy todetermine the extent of organizing in the business. Everybody and everything can bedescribed as related and not, depending on the position of the observer. In thissituation the Dutch Ministry of Economic Affairs produced a policy document inwhich the leading question was how to increase the transparency of the processes ofinforming and advising. As a point of departure, the Ministry proposed several modelsof 'ondernemingshuizen' ('business houses').

62

6.2 Towards a posture of decentering

At that time a research team (of which I was a member) was asked to advise. Abroad evaluation of this document in the I&A field endorsed the position of theMinistry, not to become a party in the field, but to create suitable conditions (VanDongen et al., 1990). For a policymaker this social and cognitive variety obviouslymeans that he or she seldom has one point of action which is homogeneous for allparties. Moreover, it is virtually impossible to clearly describe the visions, attitudes,opinions, interests, relations, linkages, variables, etc. As a result, the dysfunctionscannot be solved and effective strategies cannot be developed. In other words,especially for officials of the Ministry, it became necessary to distance themselves, todecenter from the complexity of the field.

At the beginning of the research we decided to decenter as well, but in anotherway: we distanced ourselves from the participating organizations. The reason was thatin some cases internal differences of opinion proved to be an enormous fixation. Thechange of position is another characteristic of the posture of decentering: shiftattention from the internal fixation to the external I&A field. As the interviewsprogressed, the findings puzzled the research team more and more. As a team (and insmaller groups) we discussed the matter frequently, and related our newest insightsfrom written sources and interviews to the bits and pieces we had, which led to newdiscussions, to new puzzling, and so on. As a team we understood that we had to copewith the widest variety of social and cognitive structures we had ever studied.Simultaneously, most participants in the I&A field made it clear to us that we had tobe careful with their opinions or social structure(s). Linking, connecting and relatingcontents and social structures became our method of reasoning. After three months oftough work exploring all aspects and social contexts in the I&A field, we smelIed aweak perfume of sameness: it was, to use a poetic expression 'like the volatile whiff ofroses perfume in a garden'.

6.3 The posture of decentering unfolds

AlI the participants appreciated the facilitating role of the Ministry's officials. In orderto strengthen this facilitator's role, we started (based on our interviews and othersources) a process of metaphorization and negotiation with officials of the Ministrywhich ended in the abovementioned report. In this interaction process betweenconsultancy team and officials several metaphors were developed and used to get intouch with other possible realities. I describe here briefly the process of learning,negotiating and metaphorization we, as a research team, went through. An importantmetaphor touched on the differences in organizing we found between hierarchy andpartnership in the I&A field. Basic to this notion was that, with some exceptions, themost obvious way of organizing is the hierarchical one, which is opposed to the notionof association or partnership of lawyers, chartered accountants or physicians, forexample. In time these partnerships eventually have problems: the partners want touse the partnership model to manage, while maintenance of communality calls for ahierarchical structure. Time is needed to formulate acceptable formulas. Actors haveto rely on negotiating; solutions cannot be forced. In this situation the extent oforganization is an important criterion of evaluation to determine in which stage asection is. Using this notion, we developed an integrating model that integrates

63

several points of view, offers good prospects for changing the perspective and relatesspecific management and policy options to the methods and limits of organization. Asa relevant category of analysis and policy, we introduced the concepts of configurati­on, aggregate, and configuration-in-development, which stands here for the extent towhich an organization is or is not part of a solid field. The participants in the I&Afield (advisors, industrial federations, regional organizations and representatives of theMinistry, confederations, unions, and other Ministries) can be understood in terms ofconfigurations, developing configurations and (a number of different types ot)aggregates. In actuality these kinds of associations appear to differ in their views onthe role of the Ministry, the complexity of the field and possible methods of interve­ning in the I&A field. A variety of organizations was active in the I&A field. Ouranalysis was oriented to the public organizations and the public-related ones, whichcover the entire country, regional parts or confederations of industries. Some of themwere defined as aggregates, others as configurations-in-development, and still othersas configurations. The differences between these kinds of organizations can belikened to the differences between the strong and broken thread for a weaver sittingat his loom (one of the participants let slip a remark in this direction). Strong threadsare conditional for good cloth and help the weaver produce 'rainbows of cloth withmagical ease'. This metaphor was directed at strengthening the existing configurationsand existing links between configuration-like associations.

In this complex field a major difficulty was to open the thinking (and interac­ting) process of the participants to the possibility of relating or connecting activitiesbetween the regionally active organizations and the confederations of industries, forexample. To encourage them to think beyond the traditionally focussed activities, arelated strategy was directed at an emerging of approachable intermediate-actors,links or crystallization points between associations in order to facilitate networking,initially at the regional and industrial federation level. Finally, we elaborated on theweaving metaphor by pointing at the simple technique used. All it is is a techniquefor laying a cross-thread, called the weft or woof, between parallel long threads,called the warp, at right angles. To do this, it is enough to part the warp threads sothat the weft can be passed through, and then close them again so that they lock theweft in place. To get a pattern in the I&A field, we compared the organizations at theregional level with the weft and the organizations at the confederational level withthe warp. By combining the regional and confederational levels, intersectionalrelations originate and the transparency becomes optimal. As I once heard a weaversay: it is like a dance, one way, then another, hands and feet together. We continuedour figure of speech. Weaving this way, you will also find holes. Preventing them isbetter than mending them. Especially in concrete situations where there is a little orno regional organization, and there are several competing administrative bodies oraggregate-like organizations, it is crucial to set up conditions which allow these linksto emerge and to strengthen the linking relations.

In conclusion, in this posture of decentering it appears to be crucial that theI&A field organizes itself; various kinds of configuring are allowed, the forms ofnetworking are not standardized. The process of metaphorization helped us and theofficials of the Ministry to make linking texts at the social-structural (organizational)level as well as at the cognitive level (themes). It helped us to instrumentalize andimplement the situation of 'playing', which we defined earlier as an ongoing processof metaphorization of possible other 'games of reality'.

64

6.4 Designjor a posture oj centering

For the officials of the Ministry this stage of playing was inherently related to thestage of 'gaming' in which organizations could internally as well as externally consen­sually validate the definitions of reality. To engineer the 'gaming' situation in acomplex field like this, the strategic design of the interventions is crucial (Cummings,1989; Nadler and Tushman, 1988). It means awareness of crucial initial factors, suchas:

initial conditions, especially at the social-structural level. The facilitating role ofa Ministry is crucial. We emphasized the use of the principle of guiding negati­vely (see Van Dongen, this volume), the invitation strategy to the participants,the conditions to specify (among others the creation of a third party for aspecific configuration), the initiating role in an aggregate-like situation, and therole of facilitating links;a strategy of inviting three specific participants (in a Dutch context: confederati­ons, federations and regional organizations) (meant to inform and to discuss);planning of regional links or crystallization points (the choice of pilots, concretequestions and issues; here again the Ministry's facilitating role is important);the role of a process consultant (to organize, to initiate, to evaluate and toinform).

7. Conclusion

Finally, I want to make an intermediate assessment of the state of the art (better:balance) of thinking on this posture of decentering and centering. Methodologically,the posture of decentering assumes distance from particular models, definitions orways of organizing (a field). The core of the strategy is to set up conditions for peopleto consider other possible ways of modelling, defining or organizing. In that sense, thestrategy consequently broadens the domain of knowledge. Moreover, the strategyoffers the possibility of analyzing these models in their mutual relations. Essentially,that is why this strategy can be defined as a posture of invitation, which will have thefunction of opening, allowing new contributions, new models, new definitions, newpossibilities of organizing, and so on. Through the process of metaphorization, thestrategy invites one to analyze the variety of knowledge through other eyes. Using thismethod, we have the potential to learn to study an incident, problem or complex fieldsimultaneously and synchronically from various points of view. By relating thesecompeting points of view, we can acquire a coherent body of knowledge (on organi­zing and defining), which is the starting point for our posture of centering, in whichorganizations are invited to consensually validate the definitions of reality, internallyas well as (in a later stage) externally. This posture is based on a social integratingway of thinking (or clustering-reclustering approach) which opposes the conventionalcategorizing one. As usual, further research is required!

65

Notes

1. I am indebted to H. Bolk, H. van Dongen, C. Faucheux and C. Luscuere whose ideas and criticalcomments contributed to creating this posture of decentering and centering.

2. Derrida (1989, p.488) uses the word metaphorization in the original Greek sense as 'to carrybeyond' or 'infmitely receding to the horizon', implying, as Cooper deserves, "the act of beingtransported here and there in a vehicle that has no substance, by a driver whom one cannot seeand to a destination that one can never know".

66

Chapter 6

MANAGING OF SOCIAL COGNITIVECONFlGURATIONS

IN A MULTIPLE CONI'EXT

A.A. Voogt

1. Basic starting points

In this chapter a method for managing organlzmg processes is presented (section 2)and subsequently illustrated in the framework of a case history (section 3). Thispresentation is preceded by a description of the main starting points, the paradigms ofthe method (section 1). For a more extensive description of the subject the reader isreferred to the author's dissertation 'Managing in a multiple context' (Voogt, 1990).

The basis of our view to organizing and the organizing process is the processapproach in organization theory, in particular the theory of social integration. Therelevant starting points of this approach are:1. Knowledge or 'the' reality are created and changed in social processes or interaction.Each individual provides his own definitions and viewpoints for studying situations.Certain aspects of an observed phenomenon are placed in the foreground ofperception, while others are relegated to the background. A choice of 'seeing' encom­passes a choice of 'not seeing'. In social interaction people exchange outlooks, theirindividual realities, and attempt to arrive at a common reality which then serves as abasis for further actions. This construction and augmentation of knowledge is referredto by us as 'organizing' in this paper. It is a social dynamic and constitutes the essenceof what people 'do'in organizations.2. Variety in opinions and interpretations and variety on the ideological level constitutesthe 'engine'for change. If people cause and interpret occurrences in different ways,then it would seem logical to assume that people in organizations have differentmotivations and objectives. This variety is the sole source of progress. That varietymust not be suppressed. On the other hand, agreement on 'what one sees and has todo' is a necessity as otherwise there will be no 'results'.3. Social structure and knowledge are interdependent. In interaction, a distinction can bemade between a social and a knowledge component. A social relationship, the 'Who'side, is in first instance the constituting factor which determines whether people areprepared to agree on common knowledge. The latter 'What' is a product of theprocess. The social and the knowledge structure are interdependent, intertwined, as itwere, in a double helix.

67

4. Ongoing interaction is a condition for constructing and reconstructing reality. Ongoinginteraction is required for the construction and reconstruction of knowledge. When inan interaction the social couplings become stronger, the possibility for ongoing develop­ment of knowledge becomes greater. On the other hand, as the overlap of knowledgebecomes greater, the social relation strengthens and thereby the ongoing interacting.Individuals do not know in advance what knowledge will be produced. Thereforeinteraction cannot be managed on the basis of a knowledge criterion.5. People are multiply included. People participate simultaneously in various socialcontexts. The knowledge structure in each of these contexts is often different. Eachindividual is multiply included and "the" reality of one social context is not necessarily'the' reality of another. In any given interaction individuals try to agree on commonknowledge, whereby negotiation about realities or contents plays an important role('negotiated order'). The multiple inclusions, also called 'thirds', constitute theenvironment of a given interaction and openness to the environment is necessary forchange. These thirds, which cannot be defined in advance, are to be considered asencompassing both knowledge or realities and persons or parties.

Below a number of statements on management will be given that have beenformulated on the basis of the outlook mentioned above.1. The managing of processes whereby the product knowledge is made and changed

should be the focus of attention.2. Managing involves safeguarding/maintaining variety in order to both guarantee

change and achieve agreed upon knowledge.3. Steering of an organizing process must focus on both the social and the knowledge

component of interaction.4. As a content yardstick is lacking (for example, an objective), management aimed at

the support and stimulation of ongoing interaction would appear to be the correctpolicy.

5. Managing interaction should encompass a protection of the openness for thirds inorder to safeguard change.

The aforementioned starting points are the basis of a new management method whichhas been developed. Before dealing with this method, however, some attention will begiven to the organizing process from a systemic perspective.

2. System in organizing

Organizing is the process of ongoing interaction whereby both a construction andreconstruction of knowledge takes place. Interaction is the essential dynamic by whichdevelopment and change of knowledge take place.Below are some definitions which are relevant to 'our' system of organizing:

Interacting system: an ongoing interaction whereby participants develop a social­cognitive structure which shall be denoted as a social cognitive configuration;Organizing system: the Interacting system that is being considered and within whichorganizing takes place;Context systems: all other Interacting systems within which the participants of theOrganizing system are included or will be included.

68

The organizing process from a systemic perspective can thus be described as follows.Each participant brings knowledge derived from Context systems into the Organizingsystem. In the Context systems different realities are made with regard to the samesituation (or occurrence, person, etc.) and this multiplicity becomes apparent in theOrganizing system. The participants will strive to couple the realities constructed in theirrespective Context systems. In the coupling of (multiple) realities in the Organizinrsystem the participants will have to make sure that the common or negotiated definitionof reality does not conflict with their participation in the Context systems. This maintainsthe possibility of continued testing of definitions of reality with other people.

Agreement on a reality in an Organizing system does not always mean that theparticipants adhere to the same interpretation of a situation. Differences will continueto exist in view of the participation in various Context systems with an often conflictingvariety of realities. It is, therefore, best not to speak of negotiating realities, butpreferably of negotiating couplings of (inclusion) realities. By phrasing it in this manner,it is stressed that commonalities in different social cognitive configurations (Contextsystems) form the basis for generally accepted knowledge in a considered configurationor Organizing system. The latter knowledge, 'the' reality, being the output of theOrganizing system, changes the social relationships in the Context systems: some par­ticipants of the Context systems recognize themselves to a greater and others to a lesserdegree in the constructed reality. The consequence of this is often that the knowledgeconstructed in the Context systems is adapted and this knowledge constitutes the basisof a changed input for the Organizing system in the second phase. This input may leadto a reconstruction in the sense of uncoupling of (inclusion) realities and this in turn mayeffectuate new couplings in the Organizing system. This is a cyclical process whereby theparticipants concerned are continuously engaged in constructing and reconstructingreality with and by means of their multiple inclusions or Context systems. 'Cyclical' in thiscase indicates that the participants should not be in a circular or closed process with thesame inclusions, but in one with continuously changing inclusions.

3. Method of management

Steering the organizing process on the basis of prescribed rules is neither desirable norpossible. The method described in this article concentrates on the dynamics of the socialprocess and is in itself 'empty'. In other words, a viewpoint or ideological starting pointis not prescribed and the focus of the managing lies in facilitating and assemblingongoing interaction. This is achieved by guarding the conditions of the couplings betweenindividuals within an Organizing system and with those in the Context systems.

In the following section an attempt will be made to describe the method in layman'sterms. After each segment a summary will be given, but this time in a systems' language.

Unhealthiness of ongoing interaction must be guarded againstIn general an interaction is considered to be healthy when the criterion of social qualityis met. Social quality means that there is variety in social relationships and knowledge,that there is freedom to make choices concerning realities. This is necessary to insurethat the 'space' for negotiating coupling of knowledge is as large as possible. Under thesecircumstances ongoing interaction can take place in an uninterrupted way facilitatingconfrontation, interpretation and reinterpretation of definitions of reality.

69

However, healthiness of an interaction is difficult to define in a positive unequivocal way.It is easier to get people to agree on interaction disturbance than on interactionhealthiness. Therefore it is preferable to speak of unhealthiness or lack of social quality,for example, in the case of realities which have been imposed through dominantbehavior. Or to give another example, when standard procedures must be followedand/or participants have locked themselves into one viewpoint.

In systems' language: The Organizing system is, in the case of dominance ofbehavior, not a 'system' as no interacting takes place. With irreflective procedures orfixation a closed system threatens to develop as the coupling with Context systems isreduced or even absent.

Ensuring the entry of thirdsThe organizing process is not only directed at constructing 'the' reality but also atchanging or reconstructing that reality. Thus, on the one hand achieving agreement onreality is aimed at while on the other hand such agreements are broken. Essential to theprocess of constructing and reconstructing is that the connection with the multipleinclusions is kept open. In that case the participants can form 'new' inclusions: thirdpersons, meanings and realities are allowed to become participants of the interaction ina noninterfering manner.

These thirds are conditional for change or the reconstruction of reality. However,thirds cannot be defined in advance. Participants in an interaction do not know withwhich inclusions they will form a 'new'social cognitive configuration. These inclusionscan contribute to a change of social structure and knowledge in a considered interaction.Only in the course of the organizing process can participants agree which thirds in termsof 'Who' (person and/or social relationship) or 'What' (opinion of a person) areinvolved. In view of the fact that thirds are not known in advance, a regular testing ina negative sense is advisable: from 'Who' or 'What' are people cut off? Which thirdshave no entrance possibilities?

Fixed role patterns and standard procedures often work as a constraint on theentrance of other persons and viewpoints. 'New' knowledge is not discovered becauseeach situation is being interpreted in the framework of 'old' perspectives and realities.The approach of classical organization theory whereby the focus was on what happenedwithin an organization ('In here') and not on the environment ('Out there') is an exampleof such a fixed way of looking.

In systems' language: Change requires the Organizing system to be completely opento Context systems.

Language clarificationIn an initial interaction the manager is confronted with an assortment of definitions oroutlooks which are the result of the multiple inclusions of participants. There is only athin overlap of reality definitions. In order to make the space for the search process andthe coupling of definitions as large as possible, an analysis of the multiple inclusions isrecommended.

'What' or 'Who' are people connected to outside of a given interaction? A managerwill try to find the contrasts and the commonalities in the multiple (knowledge)inclusions. He can then assist in finding a language or grammar in relation to theinclusions. In other words, he may generate a content structure that will enable the par­ticipants to continue communicating with their inclusions. The variety of often conflicting

70

inclusions causes conflicts and it is the task of a manager to see that these are functional,that is, of a heuristic nature. The manager must also avoid dysfunctional situations, thatis, a breakdown in interaction. Essential to the whole process is the presence of socialquality, particularly of a sound reflective atmosphere.

In systems' language: The manager is instrumental in coupling the knowledgestructures of the Context systems; he facilitates and assists in the search for a commonlanguage within the Organizing system.

Assembling couplings for Play and GamePlay is defined as the heuristic process whereby the organizing members define the rulesor the framework for finding 'the' reality. This activity is most often required in abeginning interaction when knowledge or realities are not or only loosely coupled. Gameis seen as a social activity based on the content rules agreed upon in Play.

In Play the social component of the interaction will be in the foreground. Themanager will initiate a process on the basis of social relationships (for example, mutualdependency, sympathy, common norms, availability, etc.). This activity is denoted associal assembling and the characteristic of managing is defined as negative steering:intervention only when the ongoing interaction is disturbed. With this minimal mainte­nance procedure, the (ideological) space for 'filling in' reality is as large as possible.As soon as the rules of the search process, the perspective, have been agreed upon, theGame can start. Interaction can now be organized on the basis of content criteria (forexample, an objective) as set in Play. The rules of the Game serve as a basis for cognitiveassembling and positive steering.

Change means that new definitions of reality are found which require reflection onand changing of the rules of the Game. It must be possible that each Game can bereplaced by another Game with new rules. This reconstructing requires Play each time,but this is only possible when thirds (persons, meanings) are introduced in a consideredinteraction. Reflection on the rules has to be accompanied by reflection on the couplings.Which couplings have to change? Where should a cognitive assembling be replaced bya social assembling? In which couplings are thirds not allowed to enter?

In systems' language: Organizing involves alternating attention between the socialand the cognitive subsystems in the Organizing system in order to facilitate Play andGame respectively. An assembling of the interaction for Play requires changing couplingswith the Context systems.

Intervention in the case of cognitive and social fixationFixation is present when participants are fixed on one view or perspective or on onebehavioral and social pattern. The multiple inclusions of the participants in aninteraction constitute a starting point for intervention and a starting point for 'new'Games.

In the case of a cognitive fixation one knowledge structure is so dominant that theparticipants of an interaction apply 'unchangeable' rules to find reality. Only marginalchanges occur but there is no real change in the definition of reality. Play with reflectionon reality does not take place. To facilitate Play, the entrance of thirds is required. Asthe cognitive side does not allow variety, the choice of thirds has to be found in thesocial component. A fixation can be broken by involving a third person in the interaction.This intervention changes the social structure, and this change of context may alter thefixed knowledge structure. This is called social assembling in terms of thirds.

71

Participants with a social fixation or ritualistic behavior are not prepared to change theirrelationship. A behavioral Game with fixed social rules exists. Change through the socialside is blocked because the participants allow no variety in terms of third persons. Theadvised method of remedying this situation is the introducing of a third reality. Such achange in the knowledge structure, which is the context of the social structure, may wipeout the social fixation. This is called cognitive assembling in terms of thirds.

In systems' language: In both fixation situations one can speak of a closed (Organi­zing) system. To 'open up' the system, a coupling with Context systems is required in asocial or cognitive sense. Managing is, in this case, the steering of the couplings of theOrganizing system with the Context systems.

In general, the method encompasses a facilitating of couplings between individuals,combined with a guarding of the social quality and an openness to the environment. Inaddition, a clarifying of language and intervention may be necessary. The method isdirected at 'learning both for learning and unlearning'; the acquisition of skills to steerconditions in such a manner that they result in production (learning) and reproduction(unlearning) of knowledge.

4. Case history

The case history describes and analyses the development of a project called 'Ositrans',a project involving the transportation company Incotrans BY in Rotterdam. This wasdone from the perspective of organizing and of managing the organizing process. Theproject encompassed the design and implementation of a telematics system incooperation with the Dutch Postal Service (PIT). The purpose of the system was tocompile, produce and transmit the information necessary for the drawing up ofdocuments for the exportation of flowerbulbs. The case history is described in detailelsewhere (Yoogt, 1990). It comprises a description of the organizing process in four'blocs' of consecutive periods of six months during the years 1985-1986. In the descriptionand elaboration a number of social cognitive configurations involving a limited numberof central persons in the strategy formulation and development of Ositrans, wereconcentrated on. The objective of the research was to come to a methodical account ofa number of developed theoretical notions and the method of management described insection two. In particular, effort was directed at illustrating that variety in the sense ofmultiplicity of reality definitions can be mapped out and recognized.

In the space of this chapter it is impossible to reiterate the complete case historyand therefore only a few salient segments will be described. Firstly a field description oftwo 'snapshots' of processes of interaction, denoted as Episodes. Hereby 'facts'are shownwhich developed on desaggregated level. In the next section a reflection on the fielddescription is given. With the assistance of the organizing language, an effort is made tomake visible the knowledge construction and reconstruction process. The concept ofmultiple inclusions and the managing of these form the nucleus. The conclusions arecompared to the developed notions and method of managing. The actors mentioned byname in the relevant Episodes are: P.I. Pieters, general manager Electronic DataProcessing (EDP) ofIncotrans BY, H.I. van Dongen, professor of the Faculty of BusinessAdministration of the Erasmus University Rotterdam and member of the supervisoryboard of Incotrans BY and A.A. Yoogt, member of the executive board of Incotrans BYand manager of the Ositrans project.

72

4.1 Field description

Episode 1 (first half 1985) Pieters/Voogt - Sales staff: Incotrans and TelematicsVoogt and Pieters considered 'the'telematics reality to be, to a large degree, a technicalissue. They viewed it as a matter of computers, transmission standards, software, etc.Moreover Voogt considered telematics, more so than Pieters, as a tool for processinnovation.

In 1984 Pieters designed an information services package to be offered to clientsusing transportation services performed by Incotrans. Using this model Voogt organized,by means of project groups, discussions with the Sales staff in order to hear their opinionof the model. The model was accepted in a passive sense by the Sales staff: theimportance of telematics was acknowledged, but this recognition was not heartfelt.Telematics was also viewed by this group as being a technical matter and therefore theresponsibility of the EDP department. A reflection of the low acceptance level amongthe Sales staff was their hesitation to start communicating with customers on the subject.Finally the model and services offered were discussed with a number of clients by boththe Sales staff and Pieters. These discussions did, however, concentrate on technicalissues, for example, the kind of computers, software and standards required fortransmission. No follow-up of the project was initiated by either Incotrans or the clientsinvolved.

Pieters saw technical limitations as the principle reason for lack of progress: "thecapacity of the present computer is too small, our computer is not compatible with theclients' computers, there is insufficient EDP manpower available", etc. In Pieters'opinion, a new computer would alleviate the technical problems to a considerabledegree. As a second reason for the failure of the project Pieters mentioned theinsufficient involvement of the Sales staff, "it does not interest them at all and they don'tunderstand the issues involved".

Voogt was of the opinion that, because the new computer would only becomeavailable in the course of 1986, the danger of a backlog for Incotrans in the area oftelematics in relation to clients was clearly present. However, he was not forthcomingwith arguments and proposals for ways of dealing with the issue and therefore can besaid to have implicitly accepted Pieters' interpretation of the problems at hand.

Episode 2 (first half 1985) Van Dongen - Voogt: Cooperation with the PITVan Dongen pointed out that Incotrans, aided by telematics, could redefine theiractivities. He saw the developments as a means of realizing couplings within transpor­tation chains. By doing so, added value could be created, to the benefit of all partici­pants. Companies like Incotrans with their extensive relational network would, in thisplan, be able to obtain a leading position in the market place.

Voogt pointed out that there were still too many technical limitations facingtelematic projects involving clients: "the present computer is not suitable/adequate, thereis a shortage of EDP manpower, with the new computer our ability to communicateelectronically with clients will be increased", etc.

Voogt qualified Van Dongen and his opinions as follows: "very nice, but all theory;I do not understand his views; Van Dongen is very unclear". Moreover, Van Dongen, inVoogt's opinion, overestimated the ability of Incotrans to handle large and complexprojects in the area of telematics. The difference in perspective between both actors canbe summarized as follows: Van Dongen looked 'from the outside' at Incotrans as part

73

of a (social) network, while Voogt held to a more technical and internally directedIncotrans perspective.

Van Oongen advised Voogt to start negotiations with the PTT to come to a coop­erative arrangement with them in the area of telematics. Voogt did not say no to thisidea, but did not understand why Van Oongen had suggested it. What, after all, did thePTT have to do with computers and information processing in transportation? Moreover,Incotrans had a small computer staff and in a cooperative effort with a large organizationlike the PTT, the project would no doubt be on such a large scale that Incotrans wouldin the long run not be able to continue participating. 1

4.2 Reflection on the Episodes

Episode 1 - Pieters/Voogt - Sales staff: Incotrans and Telematics1. The inclusion of Pieters in the EOP department was dominant and characterizedby a technical definition: telematics = hardware, software, lack of EOP manpower, etc.2. Voogt was tightly coupled to Incotrans as a whole and in this configurationtelematics was defined as process innovation, whereby the (internal) organization wasplaced in the foreground together with a broad interaction between all disciplines withinthe company.3. In the configuration Pieters!Voogt, the main inclusions of each of them (EOPdepartment and Incotrans) and the definitions belonging to these inclusions (technicaldefinition and process innovation) conflicted.4. Voogt was heavily dependent on Pieters to replace the computer. This socialrelationship was the main reason that Voogt accepted Pieters' technical definition.5. In the interaction Pieters-Voogt, one could not speak of multiple inclusions orcoupling of knowledge: there was only one inclusion, namely with the EOP department,which blocked change and led to a fixation of the knowledge structure.6. Pieters and Voogt organized an interaction with the Sales staff based on technical,rational rules: stringent planning, project groups, analysis of present procedures, etc.These strict rules were not accepted by the Sales staff and no discussion took placeconcerning the rules (of the Game).7. This cognitive assembling was not successful: the interaCtion EOP-Sales staff froze.A coupling of realities did not materialize owing to the fact that the social relationshipwas too weak. There was no mutual dependency relationship between the groups: thedevelopment of telematics was not a requirement for the improvement of the short termresults (inclusion Sales staft).8. Pieters/Voogt were not confronted with new inclusions. Their Game based on oneinclusion continued and the entrance of third parties (like the Sales staft) was impossible.Pieters' remark ("they do not understand anything about it") was, from his perspective,quite "logical".9. The actors did not reach the Play stage as this required stronger multiple inclusions.

Episode 2 - Van Dongen - Voogt: cooperation with the PIT1. The entrance of Van Dongen created a new inclusion for Voogt. This third person,with whom a strong social relationship existed, introduced a new reality: telematics =social networks = value adding.

74

2. The new definition existed in Voogt's opinion parallel to the existing technical one.But Voogt's simultaneous realities conflicted: Van Dongen and Pieters each had adifferent grammar. The subject was the same, but the meanings which were coupled toit were totally different.3. Van Dongen and Voogt did not come to a coupling of knowledge. The telematicdefinitions (networks and technique) remained separate and this separation made Voogtuncertain. The ongoing interaction was founded on the social component: the inclusionof both actors in Incotrans and Voogt's opinion that Van Dongen was an expert in thearea of telematics.4. In view of the existing cognitive fixation in the relationship VoogtiPieters and theloose knowledge coupling Voogt-Van Dongen, it is understandable that Voogt did notat first accept a new third meaning (namely the PIT). This party was 'caught' in theexisting definition of reality; I: telematics = computers and 2: the PIT does not knowmuch about that subject.5. Voogt's qualification of Van Dongen's opinions as "very nice, but all theory" can beseen as a result of the clear difference between the two actors' outlooks and the varietyof their inclusions.

Reflection in social cognitive meaningSocial cognitive configurations can be described in terms of the knowledge side of theinteraction and can be expressed in the language of the social component. In thepreviously described approach to the organizing process certain moments were observedfrom the cognitive perspective, while at other times the social dimension was placed inthe foreground. In order to show the double helix character of the development and inparticular the change of the social cognitive configurations, the reflections from thisperspective are summarized in the following section.

Episode I (PieterslVoogt - Sales staff)

Social Cognitive

a. Confrontation inclusions of Pieters b. Confrontation realities:(EDP Dept.) and Voogt (Incotrans) technique and process innovation

c. PieterslVoogt tightly coupled d. Cognitive fixation on one definition:with the EDP Department telematics = technique

e. Third persons (Sales staff) f. Assembling of the interaction as perintroduced in the interaction precoded rules = cognitive assembling

g. Weak coupling between Sales staff h. Cognitive assembling not successful;and PieterslVoogt the fixation remained owing to lack of

multiple inclusions of PieterslVoogt

Conclusion: the effort to break through the cogmtlve fixation by introducing thirds(persons) is not successful as the social relationship with those thirds is too weak.

75

Episode 2 - (Voogt-Van Dongen)

Social Cognitive

a. Third person (Van Dongen) enters b. New definition: telematics =the interaction (external) networks

c. Voogt experiences confrontation of d. Confrontation realities: techniqueinclusions: Pieters and Van Dongen and networks

e. Strong relationship f. Weak knowledge couplingVoogt - Van Dongen Voogt - Van Dongen

g. Voogt has doubts concerning h. Cognitive fixation unchanged:the operative definition entrance closed to third meaning (PIT)

Conclusion: In the case of cogmtlve fixation a third person with a strong socialrelationship leads to doubts and uncertainty.

4.3 Conclusions

The validation of theoretical notionsIn the reflections described in section 4.2a number of theoretical notions that have beendeveloped in this article come to the foreground:

Cognitive elements and processes ('What') are reflected in the social dimension('Who') and vice versa;Social and cognitive processes and assembling alternate in the foreground of theattention process, which is a condition for, and consequence of, change.Dynamics arise through the introduction of thirds, either through the social or thecognitive dimension. This ensures that the 'collision' of inclusions and knowledgecan (continue to) take place.The interaction process continues, while variety in realities remains.

Specifically, in the Episodes:Cognitive assembling is not successful if the social relationship is too weak(Episode 1).A new third person with whom a strong social relationship exists causes doubt anddynamics, even when there is no knowledge coupling (Episode 2).In a cognitive fixation third realities (PIT) cannot enter (Episode 2).

Play and GameThe period started with a fixation or Game created by Voogt and Pieters: telematics =computers, software, standards, etc. Third parties (Van Dongen and PIT) were hardlyable to enter the technically oriented coupling Pieters-Voogt, but were able to causedoubt and dynamics in the sense that they gave rise to the beginning of Play. Variousrealities and rules concerning realities were discussed: telematics = process innovation(Voogt) = computers/technique (VoogtiPieters) = networks (van Dongen) = projectgroups (pieters), etc. However, this variety did not lead to a new common reality in thecoupling Pieters-Voogt. There was little willingness to deviate from the operative

76

technical definition. Despite the thin overlap in realities, the interaction continued owingto the relatively strong social relationship between the actors concerned.

ManagementThe main aspects of managing, or lack of it, in the two Episodes were:Episode 1: Voogt assembled a coupling with the Sales staff, bringing another inclusionto the foreground, to break through the existing cognitive fixation. Cognitive assemblingwas not successful. In the case at hand, social assembling without strict rules on definingreality would probably have been more successful.Episode 2: The main aspect of management that should be seen in this Episode is theintroduction of a third person (Van Dongen) who supported a definition of telematicswhich conflicted with the existing one. Moreover, Van Dongen tried to assemble in termsof a third meaning but the effort was unsuccessful. These 'actions' resulted in a newsocial dynamic, which is shown in later Episodes (not included in this article).

S. Summary and conclusions

The objective of this article was to show the possibilities of the configuration approachin an operational sense. This was done by developing a number of theoretical notionsand a method of managing, and subsequently testing these in a real-life organizationalsituation. It is the author's opinion that with a configurational point of view the focus isdirected to wholes, processes and structures. Thereby one is 'forced' to take into accountan open relationship with the environment and thirds, without resorting to a prescribedcontent perspective. The 'temptation' is clearly less than it would be in the frameworkof the more classical organization theories to make concessions to the complexity oforganizational life by artificially reducing realities or knowledge in advance and toconsider organizations as 'facts' which are disconnected from a social context.

Notes

1. It should be noted that in 1985 the Rotterdam community did not wholeheartedly accept Van Dongen'sperspective on telematics either.

77

Chapter 7

KNOWLEDGE,RED1CATION AND ORGANIZATION;

SOME CRITICALCOMMENfS

B. Kee

1. Introduction

Typical of the configuration approach is its strong opposition to reification. Attemptsat identifying or defining 'the' organization are invariably remonstrated. Process andchange are posited over against the immutable and the thing-like. Why do configurati­onists do this? Are their arguments sound? The motive is that people produce theirreality in knowing and acting. Nor should they ever forget this. If they do, stagnationand fossilization will occur and development will become impossible. Theconfiguration approach works out this motive in two directions: in epistemology andin the field of action, i.e. the human capacity to direct social processes. I mean toanalyze the arguments predicated on this motive in both the theory of knowledge andthat of action. It turns out that the case for giving preference to 'process' cannot reston the 'reification' diagnosis. Let us begin with the epistemological motive.

2. Epistemology

My treatment of the epistemological argument will be in four stages. In each of theseI present my interpretation of an element essential to the configuration approach andpose my questions concerning that element.

2.1 The Method of De-reijication

Although the configuration approach is a theory about organization, publicationsoften open with a discussion on knowing. There are good grounds for doing so. Manyorganization theories treat the organization as a thing, investigate its properties andso on. If objections are to be made to this, one will need to argue on the level ofepistemology. To marshal these arguments configurationists close ranks with thegrowing criticism of the Cartesian subject-object basis, in which the knowing subjectconfronts a subject-independent object. But how can the subject bridge the gap, howcan accurate knowledge of the object be gained? The critique leveled against this

79

basic dualism turns on the notion of 'accurate knowledge'. Accurate knowledge couldbe had if, in Rorty's words, the subject were 'a mirror of nature', reflecting the object.But this is not what the subject is. A fact increasingly emphasized nowadays is thatknowledge has a social context. That is to say, there is a social process in whichknowledge is gained, and this context cannot be eliminated via some kind of reflectivethought which would leave us with 'pure' knowledge. This is why the idea of accurateknowledge is abandoned. Knowledge is 'socially constructed', or, if you like, reality is(re)produced in social processes.

It is here that reification is brought in. "We can speak of reification if knowledge,situations, institutions, etc. are no longer viewed as human constructs, but are lookedupon as given, as independent of human beings and hence unaffected by them"(Voogt, 1990, p.24). Reification, then, means that the knowing subject takes hisdefinition to be the accurate representation of an independently given reality, areality whose valid order does not originate with the subject. Remarkably, configura­tionists call reification a matter of methodology: they advocate elaboration of amethod of de-reification (Maas, 1988, p.13). Why is this remarkable? Well,methodology builds on epistemology and presumably it is on the level of epistemologythat reification occurs. It seems, then, that we have to do with methodology in aspecial sense, i.e. not as the complex of rules for the accumulation of knowledge butas a tool to combat or prevent epistemological misconceptions--a methodology meantto counteract the human propensity to think of knowledge as corresponding with theworld as objectively given. What sort of content would this methodology have? Itsmaterial point of entry is the social context in which knowing occurs. According to theconfiguration approach reification can be countered through continual change of thiscontext; the social context is, first of all, a changeable context.

Is this a tenable view? I think not. Van Dijk (1989, p.?l) remarks that one way"to escape reification of reality lies in the capacity, never wholly lost, to reflect. Thisability to reflect, to doubt, and hence to discredit 'old' interpretations remains presentin reifications, be it only latently". But this statement requires no-one to abandonCartesianism. After all, the thesis of an independently existing world still allows me totake my knowledge as an interpretation and hence as an approximate representationof it. Nor does it exclude the possibility that my social context affects my interpretati­on. This is why I think that method alone will not wean a person from his epistemolo­gical stance. To speak of a human propensity or inclination to reckon with an objectiveworld -inclinations can be countered through method, as in the case of combatingsmoking- is to rule out a priori that there are those who would very consciously denythe existence of an inclination affecting their thought and action. One might reply thatI am making things too complicated. Reification can mean that people take a specificinterpretation for 'the' reality, and de-reification is to make them give up thisinterpretation. Surely, however, this will not do in the case of organization theory. Toquestion the view that the organization is a thing is to raise the issue of the choicebetween objectivism or subjectivism concerning the subject-object dichotomy.

2.2 Know/edge as socially constructed

Let us tum to the notion of knowledge as 'socially constructed'. What is the meaningof the adjective here? What is this social process? On this point we meet with adiversity of philosophical elaborations. Jorgen Habermas, for instance, posits rational

80

communication and a coherence theory of truth and finds speech acts to figureprominently. More generally, for many it is the use of language which confers weighton the social dimension of the knowing process. 1 The next step is embark upon post­modernism, but this is not where we find the configuration approach. This approachaccords no central role to language. Schematically one can say that the configurationapproach is part of the broad anti-positivistic stream which 'deconstructs' the object infavor of the subject.

At stake is not the organization as datum but organizing as action on the part ofsubjects. In this broad stream the configuration approach typically takes the socialcharacter of the subjects seriously-radically so, to the point that eventually theydisappear in it. Knowledge, configurationists say, implies that individuals share acommon reality socially agreed upon. This means that we are referred to the processby which reality is constructed and reconstructed (Van Twist and Termeer, in thisvolume). Insight into this process alone allows us to ask about an organization. Whenthis is applied to organization theories it becomes possible to break the deadlockamong the different paradigms. It will not be possible to decide which paradigm isaccurate (nor to determine that none of them are correct and that a new one isneeded). What can be done is to study "the course and nature of the discussion whichcan be delineated between representatives of differing viewpoints in the debates onconcepts like 'organization' and 'science"·. This course and nature of the discussion isthat "definitions of reality come about through negotiations" (Bolk, 1989, p.17).

The configuration approach loads the adjective 'socially' with negotiations amongindividuals regarding definitions of reality. Its concern is not with the content of thedefinitions but with the procedure, the on-going contact between individuals. Conti­nuity of the process of knowing requires a dynamic and changeable context. How,then, should we understand this process among individuals in order to see its dynamicmoment? Before turning to this dynamic interaction I want to take critical stock ofthe argument so far. Bracketing the content of knowledge and concentrating onprocedure evokes a dialectics. Viewed as negotiation the process must achieve aresult, at which point the talks are concluded, otherwise the terminology makes nosense. At the same time, though, knowing knows no end, that is, "the possibilities toorganize, order and interpret the world around us are unlimited" (Maas, 1988, p.ll),and all of those possibilities must be realized without giving preference to a singleone of them (in this respect knowledge does not accumulate). A result achieved ir;negotiation must immediately make way for further negotiation. This dialectics is notof the Hegelian sort. Nothing is 'aufgehoben'. It is mere progression. As I see it, thisdialectics arises out of the procedural approach. By leaving aside the question of theterms under which a choice can be made among paradigms and by bypassingrationality of whatever type, the paradigms are no longer 'natural pauses' in theprocess. Nevertheless, the paradigms are needed, otherwise there would be no processof development of knowledge at all.

2.3 Dynamics of the interaction

What is the dynamic moment in the interaction between individuals? It comes outwhen the concept of interaction is specified as 'development of double-interacts'. Notthe person but the process is taken as unit of analysis (De Laat, 1983, p.29). Thedouble-interact is the minimal basic unit in the analysis of the interaction process.

81

A second look teaches that this does not guarantee dynamism, since in the double­interact A and B may arrive at a shared interpretation of the reality (is not this theobjective of negotiation?). Configurationists are aware of this possibility and straight­way call this reification, since for A and B "reality is frozen as 'facticity'" (Van Twistand Termeer, in this volume). I think that this is not necessarily so and perhaps it isimpossible. When A and B know that they are negotiating concerning their realitydefinitions they will not interpret the outcome as independent of themselves. If theydid, they would not think they were negotiating, and the researcher, the outsider,would gain little if he nevertheless acted as if they did negotiate. Once again we mustask if there can be a method of de-reification.

How can interaction be prevented from halting? The configuration approachintroduces the notion of 'a third'. But the appearance of a concrete third (person,meaning, action or interaction) can always yield a shared view of the world. In otherwords, if the dynamism in knowledge presumably arises from the dynamism in theorganization of knowledge, and if this organizational dynamism can fall still, then thedynamics of knowledge cannot be guaranteed through organizing. Or in other wordsstill: the openness which is essential to interaction is not organizational in kind andundermines the presupposition that knowledge (and hence reality) cannot be outsideof the social context. The introduction of a 'third' raises another difficulty as well.Assuming that re-definition of the double-interact into a triadic structure willguarantee dynamism, then what can reification mean? Reification can no longeroccur. It seems that the method of de-reification is replaced by a reconsideration ofthe structure of the minimal basic unit of analysis. This would mean that the problemI referred to earlier in connection with this methodology, namely, the extent to whichthe subjective epistemological choice can be influenced through it, returns here in aspecific form. In the triadic structure subjects are no longer free to make epistemolo­gical choices. The implication of the re-definition is that the subjects themselves areswallowed up by the process.

2.4 Social configurations

Although there is much talk of dynamism, configurationists also speak of configurati­ons, of socio-cognitive configurations. These are characterized by incessant andintensive interaction among people whose reality definitions are congruent. However,the meaning given to these configurations harbors a tension. On the one hand, 'socio­cognitive configuration' is called a basic concept in the understanding of processes ofsocial change. The concept allows us to deduce which forms of social change canoccur (Bolk, 1989, pp.36-37). On the other hand, a socio-cognitive configuration iscalled a reification, since the continuous dynamism has fallen still. The paradigms ofscience are good examples of this.

It seems to me that this tension is a dialectics within the configuration approachitself. The socia-cognitive configurations are needed so that the process can be seenas a process of change, but simultaneously the process as such precludes suchconfigurations. One can object to this and say that these configurations should beconsidered as snapshots lifted out of the process of reality construction (Bolk, 1989,p.35). This metaphor would prevent breaking up the process. To this can be retortedthat it makes talk of reification meaningless, and that, secondly, it runs afoul of thecharacterization as 'relatively stable patterns of interaction'.

82

Via the deconstruction of the object, which connotes givenness and constancy, theconfiguration approach runs up against the thorny problem of the mutable and theimmutable. It remains somewhat unclear how the fixed and the abiding functions inrelation to the configurations. One line of argument says that configurations exist inreality (the organizations we encounter). But the changeable, too, exists, and is givenprimacy. The mutable lies at the root of all fixedness and can alter it again. As notedearlier, however, the problem is that we humans can gain understanding in terms oforder alone, and order has to do with the fixed and the reliable. In other words, whilethe fixed may not be dissolved in the changeable, the fixed may not endure.

Another line seems suggested by the idea of the snapshot, the configuration seenas a virtual point. Reality is process; in it, nothing is fixed. But our understandingneeds assistance by way of constructs in which the process is halted or carried throughto its limit. In economics some use the concept of equilibrium in this way. But thisraises the question of the relation between process and construct. If it is to be of realhelp it has to have some sort of affinity with the nature of the process, and this meansthat 'process' implies more than mere progression - which is the very thing configura­tionists want to exclude: "the theory does not predict and explain why certain con­tents/cognitions exist, originate, or change. The theory... is essentially empty" (Bolk,1989, p.36).

3. Organization Theory

Up to this point I have treated the configuration approach as theory of knowledge.Organization was discussed in this context, and it turned out that most of theproblems can be raised without going beyond it. But the approach pretends to beorganization theory, a theory about the activity of organizing. The step from theory ofknowledge to theory of organizing is a transition which as such calls for criticalcomment. Prima facie the transition seems to present no problem; after all, state­ments about the organization of knowledge can be generalized, can they not? Thisquestion is answered in the affirmative, or, rather, neither question nor answer areexplicitly mentioned. From the start, organization has the general sense of interactionbetween individuals, since organizing means "social construction and reconstruction ofknowledge" (Voogt, 1990, p.?), "how reality is constructed" (Maas, 1988, p.39). Thissays that the organization of the interaction of knowing is not essentially differentfrom knowing the organization as interaction. What this thesis implies is that theorganizational conditions to be met if knowledge is to progress are now declared toconstitute the 'essence' of the organization. But it remains unclear why this isnecessarily so. Knowing occurs in a social context; so does the building of automobi­les. Surely the conditions for the progression of knowledge are not those in terms ofwhich cars come off the line? These conditions would be identical if and only ifDetroit meant to produce knowledge rather than cars.This view on the 'essence' of the organization raises questions. One of them is howthe unqualified identification of knowing and organizing can explain why organiza­tions can vary so greatly, as do enterprise, church and state. Another question is howan enterprise can exist if it is (or should be) a continuous process of change. Theconfiguration approach calls an enterprise a relatively stable interaction pattern. Ithad better be if the company is to produce automobiles. I mean: the enterprise is

83

possible only if it achieves that which the interaction as negotiatIOn was after, namely,to establish a common reality view. An apt illustration of this is the approach tostrategic management as advocated by Smircich and Stubbart (1985). Following KarlWeick, they do not see strategic management as the analyzing of an objectively givensituation harboring opportunities and hazards; rather, topmanagement enacts aninterpretation. The essence of strategic management is to mobilize the greatestnumber of people to share the interpretation and so to arrive at a common definitionof reality.

Set off against this view it becomes clear that with respect to social relationshipsthe configuration approach espouses a norm: interaction ought to be on-goinginteraction. At the same time we see that the transition from theory of knowledge totheory of organization is not without its hurdles. Of knowledge it can justifiably besaid that per definition change (including gains in scope and in depth) is proper to it.Knowing is indeed a process, and speaking of it in this way can make sense. But this'is' ,transposed to organization, transmutes into an 'ought'.

4. Alienation

Some exponents of the configuration approach start from the notion of alienationrather than from epistemology (De Laat, 1983). Alienation means that people nolonger recognize social processes like organizations as their own creations. Thiserodes their capacity to control these social processes. Why is it that they no longersee them as their own products? Because they are inclined to forget that the conceptsthey use are abstractions, products of the human mind. This leads them to ascribe to'organization' the character of concretely existing reality, a given. This configurationistangle of approach is open to comments similar to the ones made above, which allowsme to be brief about them.

Once again we are reminded of the inclination to reification. Even moreemphatically than before we must ask about this inclination's provenance, and ifperhaps its stubbornness does not hint at something more than exclusively negativesignificance. Configurationists claim that the ability to direct processes is in irreconci­lable contrast with the inclination to forget this.

The hint at something more might be picked up from Berger and Pullberg(1966) and Berger and Luckmann (1971), whose view is cited with approbation inconfiguration-approach publications. Berger c.s. distinguish between 'objectification'and 'reification'. Objectification is "the moment in the process of objectivation inwhich man establishes distance from his producing and its product, such that he cantake cognizance of it and make of it an object of his consciousness" (Berger andPullberg, 1966, p.60). Berger c.s. go on to say that this is anthropologically necessary;it is intrinsic to human existence. In this line of thought human inclination takes on adifferent significance, at least in part. But the configuration approach is disinclined tomake room for it because it wants to conceive of 'process' as radically as possible.The configuration approach tends to get in its own way when it is called upon to

justify its position. It operates with two positivism-flavored assertions. The first positsthe human propensity or inclination to reification and the second stipulates that whatis real is not 'the' organization but the process of organizing. Taken strictly, thesepropositions contain concepts to which is ascribed the character of concretely existing

84

reality. It would be easier (and perhaps more consistent?) to say that they are·objectifications'. But there is a price attached to this: the problem of the relationbetween the fixed and the changeable remains unsolved.Perhaps, though, this price should be paid, at least if the core of the configuration

approach is to be retained. This core, as I see it, is the humanism in this approach,expressed in the norm or the ideal of openness among interaction participants, c.q.the human capacity to direct social processes. How much of this would be left if, inorder to guarantee openness and dynamism, •process, is seen as "self-constitutive ...whose progression cannot be explained in terms of the attitudes and intentions ofindividuals" (De Laat, 1983, p.27)?

Notes

1. Kenneth Gergen, for example, asks: If knowledge as representation is a viable option no longer,then what shall we say knowledge is? He then turns to ·what passes as knowledge in human affairs.At least one major candidate is that of linguistic rendering. We generally count as knowledge thatwhich is represented in linguistic propositions· (Gergen, 1985). This leads him to be sympathetic tothe post-modernism of Derrida c.s. (Gergen, 1990).

85

Chapter 8

THE FEASIBILITYOF DEVEWPING A POLmCAL

STEERING THEORYBASED ON PROCESS-ORIENfED

CONFIGURATIONTHEORIES

N.L. Rade

1. Introduction

1.1 Political-philosophical problems

The essence of process-oriented configuration theories is Van Dongen's social integrationtheory. Placed in the context of state and society (against the background of thesteering-issue), we may discuss the use of this theory as a possible political steeringtheory. Even though it was not primarily developed for that purpose, it neverthelessimplies a conception of (minimal) steering.

A number of (political) philosophical subjects are relevant in this connection, suchas the meta-theory of the process-oriented configuration theories and the ontology (bothas an alternative for the positivist approach to science and reality), and above all theethics of social integration. Are these not the issues that should be investigated if onewishes to develop a political steering-theory based on these theories? The politicalphilosopher will surely say 'yes' (Lehning and Van Schendelen, 1981); the politician ormanager, however, will say: 'so what?'. But what do the political scientist and the publicadministration scientist say?

Though it may be important for further theoretical development to reconstruct themeta-theoretical substructure which is closely linked to the alternative (radical process­oriented) conception of reality (or ontology), this article will focus on a limited numberof issues. In this article two lines of reasoning, the ethical and the ontological, will beanalyzed as these are the essence of social integration theory. The ethical and theontological line of reasoning interfere with each other and it is often unclear whether thenormative ethical position relies on ontological reasoning or whether an ontologicalpostulate is at the same time also an ethical imperative or if perhaps it follows naturallyfrom it. Since the concept of 'the third' plays a crucial part, this concept will playacentral role in this analysis.

87

1.2 Definition of the problem: the analysis of a dilemma

The development of thinking about social (dis-)integration is characterized by two(central) phases. In each phase ethical and ontological lines of reasoning are recognized.In the second phase, however, a new crucial concept is brought to the fore which seemsto fundamentally change both the ontology and ethics of social integration theory. At thesame time a dilemma occurs which is related to the attempt to validate the ethicalprocedure of social integration by using Levinas' philosophy.

This analysis will concentrate primarily on the crucial concept of 'the third' whichplays a central part in both the ethical and ontological lines of reasoning, particularly inthe second (main) phase of the development of a social integration theory. Beforeanalyzing that concept further by comparing it to the related idea of 'the Other' in thephilosophy of Levinas, we shall first discuss the two lines of reasoning in the two phasesof the development of thinking about social (dis-)integration.

2. The ethical line in social integration theory:the first stage in the ethical line of thinking

The ethical line in social integration theory has clear implications for the developmentof a political steering theory (see Van Twist and Termeer in this volume). In the contextof the definition of the problem and in view of the analysis of the dilemma an attemptwill be made to provide an outline of the core of the interpretation of social integrationand of the two central concepts of 'unvalue' and 'ethical space', to which the pair ofconcepts 'ongoing interaction' and 'the third' are related in the second phase.

2.1 'Unvalue' and 'ethical space'

This discussion will be restricted to what in the configuration theories is called a'minimal ethical procedure' for steering (management, intervention, etc.) (Maas, 1988,pp.68-70). The concept of 'social integration' is extremely important. Social integrationis considered possible even if there is a conflict of values. A consensus of values orsystem of shared values is not considered vital to social integration. The moment ofsocial integration of organizing is more readily characterized by the 'shared unvalues',even if it is a case of 'unshared values'. Naturally, the central concept of 'unvalue'requires an explanation. The 'unvalue' is an ethical threshold value which can be (moreprecisely) formulated in negative terms as a distinction and a contrast to the ideologicalvalue-concept which in positive terms indicates what future social reality should look like.Whereas the unvalue has a process protection function, the ideological value gives aninterpretation to a desired outcome which is more content related. The unvalue is, as itwere, the threshold value of an ethical space within which discussions on conflictingideological constructions of the future are possible.

2.2 The meaning of social integration

Van Dongen (1983, pp.1l3-114) defines social integration in the first phase of thedevelopment of a social integration theory as follows: "... as a process of makingfundamental human values more explicit. Both interpretations and conduct are separate

88

dimensions. Furthermore, these can be defined both on an individual and a social level."He clarified this by stating that: "Social integration is considered as a process ofdeveloping a consensually valid grammar of fundamental human values (= unvalues) forthe various domains of human existence" (Van Dongen, 1983, p.119). Van Dongenchooses the phrase 'fundamental human values' to introduce the issue of the universalmoral code and the existence of such a code. This led Van Dongen to postulate againstthe denial of universal values by, among other things, utilizing Kohlberg' s empiricalconclusions with regard to moral development in a number of cross-cultural studies. VanDongen, however, amends Kohlberg's conclusions. He points out that the statements onthe highest (sixth) level of moral development have mainly been formulated in negativeterms and affirm his interpretation of unvalues.

2.3 'Social dis-integration' (as an ethical problem ofpolitical steering theory)

Van Dongen's postulation against the (complete) denial of absolute universal values asunvalues also implies prevention of the absoluteness of ideological values. Theseunvalues are boundaries of the ethical space (De Laat, 1983, pp.126-l27). Van Dongen:"The ethical space can be operationally tested by the unvalue, which is the negation offreedom, equality and justice" (Van Dongen, 1983, p.123; author's translation). In astronger formulation he states: "The postulation is that the violation of ethical spaceleads to dysfunctioning, dysproportionalities or, more simply stated, negative growth. Thepostulation is that ·violation ofethical space leads to social disintegration" (Van Dongen,1983, p.123; translated from Dutch; my italics). The ethical procedure for (minimal)steering is to prevent the violation of 'ethical space', to prevent transgressing the'unvalue' .

Kohlberg's cross-cultural studies of moral development offer the first explanationof the concepts 'unvalue' and 'ethical space'. Van Dongen also seems to find startingpoints for his point of view in the discussion between Nozick and Rawls (Van Dongen,1983). But elaboration and specification of the employed central concepts are not soughtin this field of political philosophy and normative ethics. Another line, 'the ontologicalline' in the development of thinking about social integration, exists, which better clarifiesthe exact meaning of the concepts 'social integration', 'ethical space' and 'unvalue' inrelation to the concepts 'on-going interaction' and 'the third'.

3. Introduction of 'the third': two stages in the ontological line of thinking

The ontological line in the discussion of social integration is characterized by aradicalization of a process-oriented and multiple conception of reality. In the first phasethe concepts 'double interact' and 'interlocking of behaviors' are central. In the secondphase the concepts 'on-going interaction' and 'the third' come to the fore. The meaningof the last concept is clarified by the philosophy of Levinas. Here serious inconsistenciesappear in the formulation of the theory on social integration. In Van Twist andTermeer's contribution, a review is given of the central concepts of the conception ofreality which lies at the root of the theory of social integration. Within the frameworkof the definition of the problem as given in this chapter, an attempt will be made todiscuss the crucial concept of 'the third', and primarily the relation maintainedconcerning the idea of 'the Other' in the philosophy of Levinas.

89

3.1 Main concepts; 'double interact', 'interlocking of behaviors', 'on-going interaction ';the first stage

According to Maas (1988, p.67), the starting point must be to see the process ofinteraction as the basis of social phenomena. It is impossible to live in a social vacuum.It is an ontological postulate (or in his terms a 'primitive') of the radical processorientation of social reality. "Besides, from this point of view it is unthinkable that valuesand cognitions can develop without social interactions." "In other words, a centralprimitive in the reasoning is that existence is not possible without interlocking ofbehaviors. "(Maas, p.67, referring to Van Dongen, 1983, translated from Dutch, author's italics).

The social process is specified in the first stage of the development of the socialintegration theory by means of the 'symbolic interaction' concept, and through theinterpretation of 'double interact' (and 'interlocking of behaviors'), one reaches thesecond stage, that of 'on-going interaction'. The other conceptions are enclosed in thatlast interpretation, trying to strip them from their hidden reifications (see Van Twist andTermeer in this volume).

3.2 The crucial concepts 'the third' and 'the Other'; the second stage

"With the introduction of 'the Third'{'the other person(s)', the double interact isredefined as a continuous process always offering the opportunity to involve another,'third' person: (...)it is the third of infinity, undefined, undifferentiated, never to encircle(... ).So the third, being another other, being another, a not pre-defined, not to encirclemeaning is essential in the encounter, in the interaction." (Maas, 1989, p.69). Accordingto Maas, the continuous defining and redefining are protected by the 'undefined', by theun-precoded 'third' meaning, person or (inter)action, offering the opportunity to generatean undefined number of constructions (of the reality).

Maas (1988, p.69) stresses: "This position evokes an epistemological and ontologicaldiscussion linking up with Levinas' 'infini', 'endless' (Levinas, 1987, pp.48 ff.) which inthe social integration theory is specified as 'undefined' and 'undifferentiated' .... ".Referring to Van Dongen (in print), Maas says in this connection, "Itmight be clear thatthe other(s) in symbolic interactionism, in Sullivan, Sartre, and the object-subject-relationtheory do differfrom the Other ofLevinas (.. ).The not-there-part of the other in the viewof Levinas, however, is absolutely not absent, on the contrary, it is the third of infinity,undefined, undifferentiated, never to encircle, but essentially present in the face. It isexactly there where the limits of the cognitive defining possibilities are. The third cannotbe defined nor differentiated, nor precoded. It is the third in front of a we (... )" .(Maas1988, p.107, author's italics).

3.3 The triadic foundation of the process of organizing

In the second phase the concept 'double interact' is extended and revised and obtains atriadic foundation. Voogt (1990, p.60): "The double interact is further revised andextended: an on-going interaction between two or more persons always being open to a'third'who beforehand cannot be defined. This may be a 'third' person, but also a 'third'meaning or a 'third' interaction". According to Voogt, the concept 'third' is a prerequisitefor a radical process-oriented approach of that social reality. "If it is a question of

change, a third is required in the sense of reality, and/or person and/or interaction. Theadmittance of an a priori unknown third in an interaction will always be possible andmust not be closed off in advance". (... ) "The possibility of redefining 'the' reality andchanging the rules to observe the reality is continuously available in this triadicconception (Voogt, 1990, p.61).

Therefore, Bolk used this 'triadic conception' as the starting point of his theory andit is explicitly linked to the interpretation of 'the other', 'the Other'. (Bolk, 1989, pp. 34­35, p.42). This 'on-going interaction' with openness for 'the third' forms the (socialintegrational) base of the process of organizing. This radicalization of a process-orientedconception of reality also leads to a specification of the (ethical) interpretation of socialdisintegration.

4. Social integration and openness for 'the third':the second stage in the ethical line of thinking

The redefinition of 'social integration' also has implications for the political, philosop­hical and ethical interpretation of the (minimal) steering of (social) organizing processes.Although, in the first phase the concept 'on-going- interaction' in the meaning of 'doubleinteract' (De Laat, 1983) already played a role as a substructure for the concepts 'ethicalspace' and 'unvalue', the meaning of the concept changes fundamentally with theintroduction of the concept 'third' in their conception of social reality (ontology). In thesecond phase the reality conception which has been explicitly developed becomesnormative through the introduction of the concept 'third' - initially an ethical conceptrelated to the (ethical) philosophy of Levinas. The (normative ethical) codes whichprevent social disintegration such as the non-(permittance of) violation of 'ethical space'and the non-(permittance of) transgression of the 'unvalues' will,above all, start to mean:the non-(pennittaru:e oj) obstruction of 'ongoing interactions', although more in the senseof having to remain open to 'the third' (or 'the Other') or perhaps even better - fromtheir point of view - in the meaning of the non-(pennittance oj) prevention ofadmittanceof the 'third'.This is the new meaning of the concept 'unvalue'!

4.1 The same or another conception of steering?

The conception of steering may also be formulated differently. For instance, Voogt(1990, p.147) speaks of a method of steering or - in his terms - managing as: "Safeguar­ding against unhealthiness of the on-going interaction". As an explanation: "In generalwe regard an interaction as healthy if it complies with the criteria of social quality,meaning that on-going interaction may take place peacefully". Voogt (1990, p.147) sees"protecting the admittance of thirds" as a method of steering or managing. This point ofview is expressed in various ways (sometimes partially metaphorically) in the variousversions of process-oriented configuration theories, but it also occurs in other expressionssuch as 'social quality' and 'dysfunctionality', 'plurality', which can all be converted to therelated pair of concepts 'on-going interaction' and 'the third'.

In Maas' interpretation (1988, p.42,p.60), a minimal ethical procedure or a steeringmethod (intervention, management) will also have to be primarily directed towardsdetecting and removing disturbance in 'on-going interaction'. Maas rightly observes adiagnostic problem: how does one determine such a fundamental disturbance? It appears

91

that research, diagnosis proceeds to experimenting with the introduction of 'the third'(Maas, 1988, p.llI, p.214). In this manner Bolk (1989, p.46) sees an opportunity forsteering or intervening (with a view to the second order change) in which "theconfrontation with 'theother'isorganized". He does not make a clear distinction between'the Other' (capital letter!), 'third person', 'third parties', and 'the other'. Variousindications of the third or the other as interchangeable concepts (Bolk, 1989, pp.34-35)are involved. Nor is a sufficiently sharp distinction made between the various indicationsof the third in the other process oriented theories.

4.2 Inconsistencies and contradictions

At least two indications of 'the third' can be seen in configuration theories; one refersto Levinas' idea of 'the Other', especially Van Dongen's interpretation. At the same timethe concept of 'the third' is often used in a second entirely different meaning in thevarious configuration theories. Changing the first (ethical) meaning of the concept 'third'to a second meaning which fundamentally differs from it implies that Levinas' philosophyis actually abandoned, and it is in fact that second meaning which is used when onespeaks of an ethical procedure or a method of management, in short, when a conceptionof (minimal) steering is developed.

Besides inconsistencies there are also fundamental contradictions within theprocess-oriented configuration theories. In various configuration theories theseinterpretations of (minimal) steering possibilities appear in terms of introduction of 'thethird'or confrontation with 'the other' or even the organization in a more friendly manner,of a 'meeting with the third'.The fact that this 'encounter', 'introduction' or 'confrontation'can be organized, however, seems to be contrary to Levinas' philosophy. This will beelaborated on below.

5. The concept of 'the third' in process oriented configuration theoriesand the idea of 'the Other'

The first question is whether the effect of the intended radicalization of a process­oriented reality conception is really in line with the philosophy of Levinas. The secondquestion is whether the result of a successful radicalization by means of the philosophyof Levinas really does lead to the foundation of a (normative ethical) procedure toprevent social disintegration. In dealing with this second question an important dilemmaappears, which will be discussed in the last section.

5.1 Some fUndamental differences

To illustrate the difference in meaning between the concept of 'the third' and the ideaof 'the Other' a comparison will be made between the concept of 'the third' used in theprocess oriented configuration theories and the ethical concept of 'the Other' as it isused in the philosophy of Levinas. I must appeal to Levinas' experts such as De Boer andStrasser who have done several translations and have commented at length on them.1. The concept of 'third' in the process-oriented configuration theories varies from a

meaning which attempts to link the ethical idea of 'the Other' to an (ontological)concept of the unknown other in the neutral sense 'any third person'. "With

92

Levinas, the Other is not any fellow man, but the other-who suffers injustice." (DeBoer, 1980, p.163, author's italics).

2. While it may seem as though there is a certain equality between the interactionpartners in the double interaction, and 'the Third' (as 'the Other'), it is in fact un­equality that is involved. The other as a disenfranchised stranger begs and orders.There is a double asymmetry between the 'I' and 'the Other', and this precedes the(double) interaction.

3. The openness towards the Third is seen as an activity. In truth, however, it is apassive/passioned experience that 'I' have of the Other? And, 'I' really cannotorganize that openness actively.

4. Becoming aware of the third (as the Other) is not primarily indicated in theprocess-oriented configuration theories by the awakening of the conscience incontradiction to Levinas' "The brightening of the self-consciousness coincides withthe awakening of the concerned conscience" (De Boer, 1976, p.1S).

5. The appearance of the Third is seen as a guarantee for on-going interaction. Butto Levinas, the appearance of the Other is a disturbance of my peaceful existenceand an ethical appeal to me (De Boer, 1976, p.S2).

6. In social integration theory a categorical imperative is mentioned (thoughnegatively formulated) which universally applies to everybody, including myself.Strasser says on an imperative of the personal responsibility: "I vouch for the Otherbut in no respect has the Other committed himself to me." (Strasser, 1981, p.141).According to Strasser, Levinas calls this (ethical) 'substitution'. Categorically, thatimperative applies more to 'myself than to others.

7. Process oriented configuration theories mention organizing encounters, confrontati­ons with the Third. Apart from the fact that it involves a passivity and not anactivity, and 'I' cannot organize a similar experience for another, here in fact itinvolves an 'epiphany'rather than an encounter, making 'me' hear the shocking cryof the Other in need, who brings about a tum in the enjoyable existence of the 'I'.(Strasser, p.147).

The above mentioned selection of references is somewhat arbitrary and certainly limited;many other quotations are also applicable.

5.2 'I' and 'the Other' The ethical disturbance of 'my' peace

The fact that an ethical experience of the 'I' is described by Levinas is insufficientlyexpressed in the process oriented configuration theories. In my view, Levinas regards thisethical relation between the 'I' and 'the Other' fundamental for human existence. In theexperience which 'the I' (whether or not in an 'on-going interaction') has of the injusticedone to the Other, the Other is making an appeal to 'me', an appeal which implies acategorical imperative and which 'I'read on 'the human face' of the Other. It is not thecase that one should be open to the other, but rather the 'J'is seized by the fate of theOther. It is a question of being overcome. It is not a question of an encounter, a dialogue(in an 'on-going interaction') but rather a question of an ethical disturbance of 'my peace'.The ethical sense question of 'my responsibility' towards the Other breaks through everymeaning. (Levinas, 1971, pp.192-207; see also Levinas, 1987, 1989; Duyndam, 1984;Burggraeve, 1977; Van de Velde, 1977; Bakker, 1981; Bakker and De Graaf, 1979).

93

5.3 Problems with the introduction of 'the third '; St. Francis ofAssisi as an example ofachange agent?

The example of St. Francis of Assisi illustrates the essence of my criticism on the conceptof 'the third' as it is used in the process-oriented configuration theories of Voogt, Maas,Bolk, and others. The introduction of the third reminds me of St. Francis of Assisi who,shocked by the injustice done to the lepers (after a long period during which heovercame his fear of infection), helped them and consequently reintroduced them intotheir own village community. In the ethical shock experienced by St. Francis, the leperis the Other who is humiliated and expelled, whose plea is at the same time experiencedas a command by St. Francis. ('Love thy neighbor', or: 'love thy enemies'.) The leper wasat that moment an ethical point of Archimedes for him, because he read the voice ofGod on 'the face' of the leper.

St. Francis never wanted nor was be able to organize that 'ethical shock' with thevillage people, but he did organize support. The re-introduction of the leper may be seenas the introduction of 'a third', but then surely not any third person. St. Francis wasdriven by support to this 'third' as the Other rather than by organizing an 'on-goinginteraction' in the village community (to reach dereification and social innovation). Andwhen that did occur, it still was not certain that they saw the lepers as St. Francis sawthem. It is more likely that an 'on-going interaction' with the lepers in the villagecommunity did not happen at all, but being under the influence of the authority of St.Francis, they just obeyed and offered some help.

An on-going interaction did not come about. The lepers were still avoided as muchas possible, but one felt obliged to help them a little because of St. Francis' authority.The life of the lepers became a little more bearable through 'obligation' and 'obedience'.A start on social change had been made. One could see St. Francis as the interveningchange agent. This example is more spiritual than philosophical. Naturally, we need apolitical philosophy to be able to say more about such problems of steering socialchanges, knowing the fundamental diversity of human beings. Instead of an elaboratetheoretical explanation, the differences can be illustrated with a second example that atfirst sight will probably cause surprise.

6. The problem of 'on-going interaction' in the Milgram experiments.Where is 'the third'? Who is 'the Other'?

Unquestioning blind obedience can be seen as a disturbance of the 'double interaction'or the 'on-going interaction'. Milgram examined blind obedience to authority by havingtestees who thought that they were merely research-assistants punish a 'pupil' (who wasa fellow worker of Milgram's without the testees knowing this) in a laboratory settingwith electric shocks varying from 15 Volts to 450 Volts. The question why so manypeople continued to give electric shocks to a 'pupil' as a punishment for mistakes is justas interesting as why people did not continue.

6.1 'My' experience of 'the Other'

Considering the following example of the experience of a testee, as described byMilgram (1974). In the first place obedience to the authority of the test-leader (Le.,

94

Milgram) dominated; gradually, however, the 'victim'obtained a 'face' for some and theyrefused to execute Milgram's orders. One testee then said, pointing to the 'victim': "Ifhedoesn't want to continue, I am taking orders from him" (Milgram, 1974, p.48). Milgram,however, demanded that the testee continue for the sake of the research on learningprocesses, making this testee react by saying: "Here he doesn't want to go on, and youthink the experiment is more important? Have you examined him? Do you know whathis physical state is?" (Milgram, 1974, p.48). Isn't this a case of an experience that 'theI' has of the Other? That experience, however, cannot be organized. It is not a matterof organizing an 'encounter' with 'the third'.

How does Milgram respond to this reaction? "Finally, it is interesting that heinitially justified his breaking off the experiment not by asserting disobedience but byasserting that he would then take orders from the victim." (Milgram, 1974, p.49). Appa­rently, Milgram regards it as strange that the testee based his response on the orders ofthe victim/'my' victim. This appears to be a perfect illustration of Levinas' idea of theasymmetry between 'the I' and 'the Other-who-suffers' whether or not through my doing,but for which 'I' am responsible. Suddenly 'the human face' breaks through with thetestee who reports from his ('my') experience.

6.2 'My' responsibilityfor 'the Other'

Milgram, however, says: "Thus, he speaks of an equivalence between the experimentersand the learner's orders and does not disobey so much as shifts the person from whichhe will take orders" (Milgram, 1974, p.49). Apparently, Milgram is so fixed on thephenomenon of obedience that he no longer sees the difference between unquestioningblind obedience to the demand to continue giving electrical shocks (which the testee doesnot do) and conscientiously reacting to the victim's plea to stop. Furthermore, it was ­in this case - the first time that the victim asked him to stop (which was at 150 Volts)and immediately the testee reacted.

We can see that the testee no longer obeyed Milgram, but responded to the victim.However, it was not just a choice for one or the other (third) person. It was not a freeact, but even before one is aware of the freedom to choose, there is an immediateresponsibility which suddenly emerges and disturbs the peace of the testee. The victim("the pupil") is then the ethical point of Archimedes. To the testee (a religious man) 'thevoice of God' became audible in the victim's call. It strengthened him in his resistanceto Milgram. "If one had as one's ultimate authority God, then it trivializes humanauthority", says the testee (Milgram, 1974, p.49). Milgram stated: "Again, the answer forthis man lies not in the repudiation of authority but in substitution of good - that is,divine - authority for bad" (Milgram, 1974, p.49). Isn't it much more in line with theethics of Levinas that we can find a trace of God in the 'face' of the Other-who-suffers,and that the humiliated Other orders with 'divine authority', at least as it is experiencedby 'the I' whose disturbed conscience lights up? It is not possible to give a definiteanswer to these questions. It does, however, seem sufficiently clear, which differences areinvolved when we speak about 'the third' or 'the Other'.

We may conclude that it is not the organizing of an encounter with 'the third', butthe experience of 'the Other' that breaks through (unquestioning) blind obedience toauthority. In the enlightenment of the conscience the victim (that is, the Other for thetestee) obtained a 'human face', and 'on-going interaction' (the double interact withMilgram) became principally possible. We see that Milgram then obstructs that 'on-going

95

interaction' (in terms of the process-oriented configuration theories) by continuing todefine the testee in terms of (unquestioning) blind obedience. It would be interesting tofurther analyze the triadic substructure in the Milgram experiment both with Levinas'philosophy and with the social integration theory of Van Dongen and the process­oriented configuration theories mentioned above. But that is a different project.

7. A dilemma in the development of a (political philosophical) interpretationof (minimal) steering

How to organize for 'myself that basic experience of the 'Other (the 'Third' in terms ofthe process-oriented configuration theories) is a big problem and it is also not in linewith the philosophy of Levinas. But how it can be organized for others in interaction isa puzzle and seems to be at right angles with the ethical thinking of Levinas. The onlyexplanation for why this introduction of 'the Third' is regarded possible in configurationtheories (which are based on a radical process-oriented social integration theory), seemsto be that the concept 'third' in configuration theory has at least two meanings and thatthe meaning of 'the third' used with steering is no longer related to Levinas' philosophyand is inconsistent with the other meaning which is mainly used when Van Dongen isquoted. Why do the process-oriented configuration theories still need the philosophy ofLevinas? Would it not be better to let this philosophy rest in view of the internalinconsistencies and contradictions? In my opinion it is a matter of a dilemma. The factthat the ethical philosophy of Levinas has been referred to, probably has two reasons.

Firstly (following the ontological line) one wishes to radically dereificate theapproach to social reality. It is in that connection that Maas (1988, p.62) refers toLevinas. The philosophy of Levinas is surely able to contribute to drastically dereifica­tioning, but then primarily from ethical experience.A second reason (following the ethical line) could be that one is searching for a betterfoundation for the ethical conception as regards the social (dis-)integration, and the(minimal) ethical procedure for steering. One does not look for the defence of thatethical position in normative ethics or political philosophy, in which one cannot avoidmaking the reality conception or, rather, the concept of man more explicit (De Graaf,1972). The philosophy of Levinas seems to have been chosen because Levinas starts withethically problematizing a 'concept of man'. At least, his ethical philosophy seems toprovide the opportunity to thoroughly examine ethics and ontology in their limitations,for instance, whether the theory also allows people not to comply with the (implicit)concept of man. In that sense, Van Dongen's reference to Levinas is logical (see Maas,1989, p.IO?).

To base the crucial concept of 'the third' on Levinas' philosophy seems necessarybecause it plays such a central role both in the ontological line and the ethical line ofthe development of thinking about social integration. But if the concept 'the third' nolonger has anything to do with the philosophy of Levinas through inconsistency in itselaboration, an appeal to that philosophy seems inadequate to me. However, if onechooses for consistency in the use of the concept 'the third' in the sense of 'the Other'Levinas' philosophy is contradicted, as I have just pointed out, because encountering theOther cannot be organized. On the other hand, if one does not want to contradict it, oneshould then be inconsistent, or what seems to be more logical, one should let go of thephilosophy of Levinas. If his philosophy is abandoned, how drastic will that dereification

96

of the process-oriented reality conception still be and how would one account for thenormativity of the concept 'the third' which again implies a concept of man? In shon, adilemma! If one chooses to ethicize the ontology of 'the on-going interaction' and 'thethird' by means of Levinas, uncertainty will remain about what (normative) procedurewill prevent social disintegration. In any case such a procedure will be different from thepresent one, because it is not possible to experiment with 'the experience of the Other'.Then the (political philosophical) conception of steering with that starting point alsocannot be maintained for it contradicts the philosophy of Levinas.

One can of course maintain the present conception of steering and the presenttheory of social integration, and one can choose for the foundation of another line thanthrough the philosophy of Levinas. But then the present (normative) ethical procedureto prevent social disintegration is (for the meantime) unfounded because the ontologyfrom the first phase of the development of a social integration conception was notsufficient as a foundation which was the reason that the concept of 'the third' wasintroduced. But if one wishes - in an ethical sense - to support that concept 'the third',wouldn't one then need the philosophy of Levinas..... IThis is a dilemma, along the linesof the process-oriented configuration theories towards a (political) steering theory.

8. Conclusion and final remark

We may, for the meantime, assume that there is a dilemma in the development of thetheory of social integration as the basis for a possible political philosophy of steering.The theory is characterized by a number of ethical and ontological concepts such as'ethical space', 'unvalue', 'ongoing interaction' and 'the third'. For the foundation oneseems to want to rely on Levinas' philosophy. At the same time, I have attempted toshow that the translation of a central concept from that philosophy to a (minimal) ethicalprocedure for steering (intervention, management, etc.) leads to a point of view thatseems to be in contradiction with Levinas' philosophy. Social integration theory defendsa conception of minimal steering, but if one tries to avoid that contradiction, to stay inline with Levinas' philosophy, then the possibility of steering is even less than minimal.

As a final remark I wish to say that it has been inspiring to let Levinas' philosophyfertilize my thinking about ethical issues, but at the same time, it seems to me an almostimpossible task to translate this through the (process-oriented) configuration theories toa political steering theory. Yet - despite the criticisms - I admire the attempt undertakenby Van Dongen and others to seek alliance with the philosophy of Levinas. For incombination with this philosophy, social integration theory and the (process-oriented)configuration theory seem to offer interesting possibilities, for instance, to analyze the'on-going interaction' in the Milgram experiments. But then too, it remains necessary tosafeguard the concepts used against internal inconsistencies and contradictions.

97

Chapter 9

EFFECTIVE ROLICYMAKING IN A MULTI-ACTOR

SE1TING: NETWORKS AND STEERING

E.H. KlijnG.R. Teisman

1. Introduction

There has been a growing tendency in policy science to incorporate parts of the policyenvironment in theories on policy making (Snellen, 1987). Studies on implementationhave shown that the government when implementing its policy must take into accountthe reactions of several organizations in the policy field, organizations which eachhave their own aims and goals. An analyst starting from this multi-actor point of view,must analyze decision making as 'joint action' (Pressman and Wildavsky, 1983). In thismethod of analysis the complexity of joint action is accepted. A priori mutualadjustment and consensus building thus become new areas for research. Governmentis not a monolithic entity. Joint policy making calls for different conceptual approa­ches to analyze the policy process other than the 'one-actor-approach'. The configura­tion theory can be seen as one of the theories that fit in this multi-actor approach topublic policy. It is, however, not the only one. In policy science the network approachto the analysis of policy decisions has gained many followers in the last ten years. Inthis article we will introduce the network perspective on public policy and elaborateon the consequences of this perspective for analysis, evaluation and improvement ofpublic policy. We will also contrast these views with some of the views from theconfiguration perspective. Three questions will be dealt with:

What are the main features of the nerwork perspective?How can this perspective be used to analyze policy processes and their outcomes?What are some similarities and dissimilarities berween the nerwork perspective andthe corifiguration theory?

Before we answer these questions, we will, briefly state the main differences betweenunicentric views and pluricentric views on analyzing and evaluating policy and policymaking.

99

2. Uni- and pluricentrism: two perspectives to analyze and evaluatesteering performance

Government is not a private organization. In many ways it is the complement of theprivate sector. Services required to fulfill public and private needs which cannot beproduced within the free market system must be carried out by governmentalagencies. This may be the result of insufficient coordination on the part of the marketor of normative views on the desired structure of society. As not only self-interest butalso collective aims are at stake, the task of government is more complex than that ofa private company. Government agencies have to deal with the questions (1) what iscollective profit and (2) how can it be achieved. These questions can be approachedusing two different points of view on government and its production process. Thesedifferent views not only lead to different analytical perspectives on policy making, butalso to different perspectives on the adequate evaluation criteria and perspectives onthe improvement of public policy.

2.1 Analyzing public policy

The first perspective, the unicentric approach, views the government as a more or lesshierarchical autonomous organization. In contrast to the private market which consistsof a large number of autonomous organizations (a multicentered decision-makingsituation), the public sector is seen as a single organization. This organization has afairly clear organization principle, on the basis of which tasks are divided. Thepoliticians at the top, chosen by the citizens, decide upon the desired collective goalsand the means to reach these goals. They are in the best position to know what thegeneral interest is. The administration, including lower level governmental agencies,help to formulate and to implement the adopted policy. The policy-making freedomof the local government is, however, restricted by the national government. This viewon government and the policy-making process is adequately described by the metap­hor 'the system', with each part possessing a specific purpose in service of the wholeof the system.

The second perspective, the pluricentric approach, rejects the monolithicstructure of the government. The government is not seen as a single unit, but as aninterwoven network of organizations. The governmental sector has an organizationalprinciple, tasks are divided and there is a certain formal relationship between theactors in the public network. This, however, does not mean that it is a hierarchy inevery respect. A formal juridical hierarchy may exist, but means such as money,knowledge and commitment/legitimacy can be and often are divided in a non­hierarchical way. Actors within such a policy network are interdependent. This is alsothe important difference between the public sector and the private sector. In theprivate sector all enterprises and consumers can decide autonomously to produce andto consume. 1n the interwoven public sector it is almost always necessary to work togetherwhen a given actor wants to realize a cenain policy project. In Diagram 1 some of themain features of the unicentric and pluricentric view on government and policyprocesses are summarized. 1

100

Diagram 1. The main axioms about the policy field from two perspectives

unicentric approach pluricentric approach

principle to defme the politicians on the central level decisions on goals and means arecommon interest decide upon goals and means taken in a network

structure of the the public sector is hierarchically government is a complex con-public sector divided into functional parts stellation of actors

process in the starts with problem on national several actors in the network canpublic sector agenda, an optimal solution is adop- take initiatives and others can

ted there and implemented support or oppose them

most adequate the system, in which every element a network, in which actors strug-metaphor has a functional task in the whole gle to influence policies

2.2 The evaluation and improvement of existing policy processes

Both perspectives hold that policy-making processes often take an unexpected courseand lead to unexpected results. Both perspectives acknowledge that several actors canbe involved in policy making. The manner in which this phenomenon is judged fromeach perspective does however differ.

The analyst working from a unicentric perspective on policy making will empha­size the disturbing effect on the process. References will be made to the central goalsand the autonomy of the central unit and these will be used as the basis for judgingand improving public policy. From this point of view, the question will be asked howit was possible that the goals set were not reached. It is then important to have aclear formulation of goals. This can be supported by a scientific effort to find theoptimal solution for a certain goal. When the central decision-making unit has formu­lated these goals and is not able to reach them, the quality of the government can beimproved by increasing the steering capacity of the central unit. Instruments popularin the unicentric point of view are reorganization, improved coordination and bettercontrol mechanisms. Coordination can be reached through laws and planning.

From a pluricentric perspective, other criteria for the evaluation and improve­ment of the policy process are used. This perspective emphasizes the role and policyfreedom of other actors. Policy, defined as the result of the policy making process, isno longer the outcome of one central decision, but the sum of several decisions takenby more than one organization. The policy-making process becomes a process inwhich the different goals of several organizations have been intertwined to become anacceptable compromise. Different organizations may actively participate in the policy­making process for their own reasons; goals they themselves have set, their formalposition, or the fact that they possess indispensable information or means. When anorganization provides such means, it can be expected that the organization wantssomething in return. When several organizations make this demand, a situationdevelops in which policy-making freedom is no longer connected to one, but to allactors. Policy freedom then becomes one of the main criteria for successful policy ma­king. Only in a policy process where all actors have the possibility of obtaining acertain amount of policy freedom will all the goals which can be reached be articula­ted and all the means be mobilized.

101

The Starting point for improvements is therefore the acceptance, optimalization andutilization of the policy-making freedom of all required actors in order to optimalizethe influence of actors. The pluricentric perspective accepts the existing organizationalstructure. In all organizations structures there will always be a high degree of interde­pendence. This interdependency has to be used, not denied. Therefore the advicegiven will be to develop tailor-made structures for every new policy project whichrequires more than one actor. This arrangement has to be facilitated by clear bilateralagreements on the formal arrangement and the division of tasks. It must also beaided by accessible lines of communication and by interorganizational managementsupport. Increasing the consensus building efforts to find a satisfactory solution is thebest tactic to insure the adequate performance of government. The differencesbetween the unicentric and pluricentric way of looking at the evaluation and improve­ment of policy processes are summarized in Diagram 2.

Diagram 2: Variables for evaluating and improving the policy-making process

unicentric approach pluricentric approach

perfonnance indi- the a priori goal of the national the a posteriori satisfaction of allcator decision-making unit actors that provided means

cause of low effec- lack of information about society inadequate process of developing ativity and inadequate planning satisfactory compromise

cause of low effi- lack of central steering capacity; lack of arrangements for inter-ciency obstruction by others action between actors involved

organizational Im- better division of tasks, improved acceptance of existing organizalprovements coordination- and control mecha- structure and development of

nisms tailor-made structures

process improve- improves implementation plans better lines of communication,ments and sophisticated planning sustained by management support

policy improve- increase scientific efforts to find increase the consensus-buildingments the optimal solution based on a efforts to find a satisfactory solu-

national perspective tion

3. Networks, configurations and steering

The network perspective and the configuration theory can be seen as examples of thepluricentric view on public policy. The network perspective in policy science wasdeveloped in the 1970s, and was a reaction to the unicentric view of policy makingthat then dominated the field of public policy analysis. The network perspectiveemphasizes the variety of actors involved in policy making, and also stresses therelatively autonomous position of those actors. On the other hand, however, thenetwork perspective focusses on the interdependencies of these relatively autonomousactors. Generally this interdependency is conceptualized as a resource dependency(Aldrich, 1979; Scharpf, 1978; Benson, 1978; Crozier and Friedberg, 1980). The strate-

102

gies of actors in the network and the patterns of interactions between actors thatevolve out of these strategies are a consequence of the actors coping with theseinterdependencies.

The concept of networks is basically a metaphor, a way to analyze social process­es. It also refers in some way to realities perceived by actors in the field, (Aldrich andWhetten, 1981; Godfroy, 1981) and to organizational settings and arrangements thathave developed during the interactions of organizations. A policy network can beroughly defined as "a social system which is involved with policy problems and/orpolicy programs, and which consists of (semi) autonomous actors who find themselvesin a dependency relation which is based on the possession of resources that arevalued by (one or more) other actors".

3.1 Steering, networks and configurations

A perspective on steering presupposes a vision and conceptualization of social proces­ses: how they are organized, changed and influenced. According to the networkperspective, networks exist because of the mutual dependence of actors. The mainreason for this interdependence is that actors need resources that they do not havethemselves. To reach their goals or to survive as a unit, they have to engage in aresource exchange process with other actors (White, Levine and Vlasak, 1975; Aikenand Hage, 1968; Benson, 1978; Scharpf, 1978). Experts on network analysis also agreeon the partial autonomy of organizations within the network. Therefore the conceptof organizational domains is frequently used to describe this feature of networks (Vander Ven, Emmett and Koening, 1975; Benson, 1978). Organizations can develop theirown goals in separate domains. In the network they are part of, they can try toinfluence decision making according to their own goals. Although networks areusually characterized as 'loosely joined systems' (Aldrich and Whetten, 1981), theyshow a certain stability. Actors create through repeated patterns of interaction,structural and cultural conditions which influence further interaction in the networkand form a starting point for new interactions. These conditions can function asstrategic opportunities for the actor as well as have a constraining effect. The networkperspective thus focusses on the interrelation between the strategic actions of separateactors and the consequences these action have for the relation pattern and thestructures that belong to these patterns between actors. Knowledge of steering can begenerated by studying the process of policy making and consensus building betweenactors, the process of institutionalizing of policy processes, and analyzing the strategicconduct of separate actors in the network. In the next two sections we will elaborateon some of these themes.

Central in the configuration theory is the idea that knowledge and reality arechanged and created in social processes. There is an interdependence betweencognitive and social processes. Ongoing interaction is needed to allow for change; thevariety of cognitive views is the engine for those changes (Voogt and Termeer, VanTwist). This view has certain advantages. One of them, which it shares with thenetwork approach, is that it rejects classical unicentric views on steering in which onecentral actor tries to realize goals by steering other actors. Like the network ap­proach, the configuration approach opens the possibility of looking at policy processesfrom a pluricentric point of view, analyzing interaction processes between different 1actors or sets of actors.

103

This means, according to Termeer and Van Twist (this volume), that steering is seenas a process activity, as a way of influencing interaction processes. Two perspectivesare of great importance within the configuration theory: management andintervention. Management means the provision of conditions for continuing interac­tion. Intervention by a third is needed when interaction is blocked.

3.2 Networks and configurations: similarities and dissimilarities

The network and the configuration perspective can be seen as examples of thepluralistic view on public policy. Both approaches focus on the interaction processesbetween actors and the patterns these processes generate. Both perspectives stress thedynamics of these processes and pay a good deal of attention to cognitive processes.

There are, however, some dissimilarities. One of the most important dissimilari­ties is found in the theoretical roots of both perspectives. These theoretical roots aredifferent. The configuration theory evolved out of a more hermeneutic and actioncentered theoretical paradigm on social science. It has its roots in symbolic interactio­nalism. The network perspective has its roots in a quite different tradition. Originallyit evolved out of the interorganization sociology and political science of the sixtiesand seventies. These were dominated by functionalism and a preoccupation withempirical (statistical) science. The network perspective was a result of the idea thatsystems are open to their environment and that other organizations are an importantpart of that organization (Negandhi, 1975; Rogers and Whetten (eds.), 1982). Thenetwork approach thus has its roots in a paradigm which is strong on structure butweak on action, and the configuration theory has its roots in a paradigm which isstrong on action but has tittle to offer for the analysis of structures (Giddens, 1984;1985; Burrell and Morgan, 1979). The configuration approach has a preoccupationwith action and processes. Although some attention is given to impediments on actorsto participate in those interaction processes, there is not much systematic analysisaddressed to the rules and resources which are implied in the reproduction ofpractices between actors and the patterns they create.2 Structural constraints that arethe result of patterns and rules in organizations, for instance, are almost neglected. Inthis regard the network perspective has, in our view, more to offer. In line with thetradition it originates from, more attention is paid to organization structures and theinfluence these have on the interaction processes (constraints as wen as opportuni­ties). This greater attention to structures in the network approach is also reflected inthe focus on interaction processes. The network perspective focusses on interactionswhich are generated because of resource dependencies. The configuration theoryseems to focus on the interactions between actors in general.

These different perspectives on interaction processes can also be seen by lookingat the way both perspectives conceptualize cognitive processes between actors.Generalizing, in network theory there is a tendency to view these processes as a (partof a) bargaining process in which information is being exchanged and in which actorstry to reach consensus about goals or an adequate package of goals (search for goalcongruency). The role of brokers can be important in this process. Although theconfiguration theory also pays attention to this facilitator's role (Maas, in thisvolume), there is a tendency to conceptualize cognitive processes as a rather freeexchange of information in which the discovery and eliminating of cognitive blocksare important.3 Although the latter view is very interesting, and probably useful in

104

managements settings, it is rather descriptive in nature and as a theoretical frame­work of organizing and policy processes it seems to be lacking something. Thequestion why actors organize in the way they do is less extensively addressed. 4 It paysless attention than the network perspective to the function of power differencesbetween actors in processes of decision making, of reaching consensus and of theframing and reframing of policy problems (SchOn, 1983).

In the following sections we will look at what the network perspective has to sayabout policy processes and the steering of those processes.

4. Policy networks: analyzing complex decision making

The concept of policy networks can be used for analyzing complex decision making inpublic policy. We will elaborate on this theme in this paragraph.

4.1 Policy shapes the inter-organizational arena

A network can crystalize around a certain policy program or certain problems in apolicy field (Scharpf, 1976). Policy problems then set the stage for certain actors.They shape certain inter-dependencies between those actors. Problem solving in thiscase often means 'constructing' a network of relevant actors that have problem solvingcapacities around a certain policy topic (Scharpf, 1976). In this way the policy pro­blem determines which actors are likely to be included in the policy network throughwhich attempts to solve the problem are channelled. From this angle, the analystemphasizes the origin of networks.

The network concept is used, among other things, for policy arrangements whichconcern incidentally implemented projects, for example, in the field of infrastructures.In this situation the network is in general not very tightly structured. The only interde­pendence between the actors consists of their desire to attack a certain problem ortheir involvement in attacking a certain problem. The arrangements are a result ofnegotiation between actors who are involved because they have 'problem solvingcapacities' which cannot be excluded. It is of course possible that out of these kinds ofarrangements more stable and well-structured networks evolve. This is especially thecase as governments set up policy programs which create long lasting interdependen­cies between actors.

4.2 Networks constrain the result of inter-organizational policy making

It is unlikely that the setting for policy problems is completely unstructured. In mostcases there is already some kind of network in existence in the policy field. Thismeans that policy networks are the action systems through which policy problems andsolutions are processed. They are the setting in which the formulation of the problemtakes shape and through which policy is implemented and modified. In this method ofanalysis, networks are a more stable setting within which purposeful actions of actorsconcerning the definition, redefinition and processing of policy problems take place.From this perspective, policy implementation and formation are strongly influencedby the characteristics of the policy network. The analyst emphasizes the conditioningof the policy by the network. Students of the contingency theories concentrate on

105

structural characteristics of the network, with variables like professionalism, fragmen­tation, and degree of stability, (Koppenjan, 1987; Mintzberg, 1983) in contrast toothers who concentrate on the strategic behavior of actors in a network as being themost important explanatory variable for the policy result (Crozier and Friedberg,1980). Analyses of policy sectors such as health and housing which use networks oftenstart with an existing network.5 The regular interactions in those policy fieldsbetween government agencies and different private organizations lead to complexinteraction networks. The policy program has an important function in maintainingand structuring those policy networks. The tasks and available resources connectedwith those programs create interdependencies between actors.

The frequent interactions between actors who have their own goals lead tocertain repeated patterns in games and strategies (Crozier and Friedberg, 1980).These repeated patterns facilitate the existence of structures that enable and constrainfurther interactions. Organizational and institutional arrangements come into being.Cultural patterns constrain and may even guide the behavior of actors. The availableresources and their distribution create an 'internal political economy' which determi­nes the allocation of values to the various participants involved (Milward andWamsley, 1985). Policy instruments are part of that 'political economy'. But they canalso be the object through which actors try to change the political economy of thenetwork. This explains why these instruments are often the objects of intense bargai­ning (Klijn, 1990).

In reality both angles clarify certain aspects of inter-organizational policymaking.Networks often do exist, but every concrete policy problem has its own policy arenawithin the network and can therefore partially reshape a network. 6 Policy problemsdo create their own arenas, but within an existing and changing network. Networkscan be dissected into partial networks or action-sets (Aldrich and Whetten, 1981).Implementation of the policy of one of the actors can be seen as 'a seamless' web ofconnected elements and actors (pressman and Wildavsky, 1983). Many decisions, byseveral actors, have to be made in order to achieve program implementation. Thequestion that must be dealt with is how an analyst can deal with this efficiency pro­blem. In the classical unicentric view on policy making, this complex situation of manydecisions and actors must result in inefficiency. From the network perspective, a"analyst not only illustrates the problems in a network (slowness, goal postponementand extensive administrative costs), but also the advantages of a network. From anetwork perspective the analyst can see the diversity and the character of 'looselycoupling' as a way of making the network more open to the goals of other actors inthe network (Weick, 1976). The actors in a network have greater flexibility and moreopportunity to adapt than in a system with more hierarchical control (Aldrich andWhetten, 1981). Networks can facilitate the intertwining of the goals of several actorsand are therefore capable ofproducing satisfactory solutions for dijjicult problems.

5. Government in an interwoven network and adequate task perfonnance

Policy making takes place in an interwoven network of actors, which leads tocomplex interaction. This cannot easily be controlled. In order to produce satisfactorysolutions, the interwoven network is a necessary structural characteristic. The questionis how government can perform adequately in such a network? First the network has

106

to be activated by the actor who wants to realize his own goals. The strategy of'selective activation' is not a single actor activity (Scharpf, 1978). All actors involvedhave a certain degree of policy freedom. This means that policy making is more thanchoosing the best solution. It is a process of communication and consensus-building.Policy freedom of all actors makes a satisfactory solution possible, a fact which cannotbe denied when analyses of steering are made. This freedom can however easily leadto blocked decision making. Therefore attention has to be paid to self-blockage. Thesolution to self-blockage will be sought in the asymmetric structure of the policynetwork.

5.1 Selective activation

A network consists of a wide variety of actors, arrangements and interdependencies.Initiatives can come from any actor in the network. The success of any given initiativedepends on the structural position of the actor in the network (resources, centrality inthe network, etc.) and the strategic decision whether or not to participate in aconcrete decision-making process. Attention is limited. Actors choose certain issues orcertain arenas, mainly because they want to gain something. The attention structuresof actors can take various forms (March and Olsen, 1978). March and Olsen distin­guish three criteria for participation. Attention can be the result of obligation (duty,tradition, routines), of cultural beliefs about who and what is important, or the resultof rational action (March and Olsen, 1978). A rational choice is based on theassumption that time for involvement in decision- making is a scarce resource.Therefore an actor calculates the marginal gain he gets from participating in decisionmaking situations.

Public policy is therefore shaped by activating certain parts of the network.Successful activation depends on "the correct identification of necessary participantsin policy-congruent networks, upon their willingness to invest time and resources in aparticular inter-organizational policy process (in competition with other interests anddemands), and upon the relative disinterest of other parties with potential influenceover one or more necessary participants" (Scharpf, 1978). The number of actors afocal actor has to activate depends on his goals. The more ambitious his goals, themore actors he has to activate. This is one of the reasons why government, because ofits goals (see section 2), often has to activate more actors than other (private) actorsin the network who are collectively trying to achieve less ambitious goals.

As Scharpf (1978) correctly observes, the selective activation of networks can befacilitated by the existence of a tightly integrated permanent structure with strongmutual dependence links among the participants. The character of the loosely couplednetworks we discussed earlier still makes the games of bargaining, consensus-buildingand coalition very important (Scharpf, 1976). In these processes policy freedom andthe goals of each of the actors involved have to be taken into account.

5.2 Policy freedom and the permeability of organizational boundaries

Goals in a policy network are only achieved when they are articulated by an actor.The policy-making process in a network is seen not only as the implementation of apriori goals, but also as the development of the goals and the intertwining of a varietyof goals. This means that the policy freedom of the actors is not only a reality which

107

has to be taken into account, but is also an important starting-point for improving thesteering capacity. The role of information and the strategic action possibilities of thevarious actors are important.

It is not possible to acquire complete information at the central level. Informa­tion is available but possessed by several actors. The main trigger for good policyresults is therefore the generation of this information. This can only be realized byinteraction between all actors that have crucial information.

Inter-organizational interaction requires strategic behavior. Strategic behavior is,however, limited to persons or groups. The organization does not maintain relationswith surrounding actors. This can only be done by representatives. In order toimprove communication with other actors, communication which is necessary to gainenough information for an adequate performance, the organization has to personalizethese communication tasks. To do so, gatekeeping managers are needed.

Gatekeepers (Crozier and Friedburg, 1980) or brokers playa double role in thedecision-making process. On the one hand the gatekeeper-broker is a representativeof the external actor in his own organization (gatekeeper), on the other hand he iscommitted to his own organization (broker). This makes the role of gatekeeper­broker attractive for the person/group and for the organization. The more successfulthe broker is in generating information from the external source, the more thestrategic impact of the organization increases. This is, in fact, the main reason tocreate brokers. The influence of the gatekeeper on internal strategy formulationincreases as his capability to generate external information grows. The policy freedomof the gatekeeper is therefore a function of his value to the organization as a whole.The policy freedom of the broker is not unlimited. To be successful he must be dedi­cated to his organization. This means that broker and organization must have at leastsome common interests. Secondly, the gatekeeper does not have a monopoly position.He masters one part of the external relations, but is not the only gatekeeper. Evenwhen we are talking about a bilateral relationship, there is often more than one inter­mediating line. The organization can therefore obtain information through differentlines. The governmental sector then becomes a network of communication lines,facilitated and maintained by gatekeepers/brokers. The organizational boundarieslose their dividing function. They become permeable formal lines, overruled byappointed brokers. The optimal policy freedom of these brokers is based on theirsupport of the strategy of the organization as a whole. This networking or reticulistfunction is an indispensable part of an adequate functioning government.

5.3 Asymmetry of steering positions as an expedient against self-blockage

Networks are important for generating information about goals, means, problems andselection criteria. They do, however, lack a clear command structure. Thereforenetworks can be expected to have difficulties making decisions. In contrast to a hier­archical system, decisions are made on the basis of consensus. This means that anydecision that has important negative consequences for one of the main participantswill not be taken. Empirical research shows that horizontal networks are not capableof making fast and painful decisions (Godfroy, 1981).

From a pluricentric point of view, the solution to this problem is sought in theasymmetric structure of the interwoven network. Society as a whole is neither aperpetuum mobile nor a centrally steered company. Change does not take place as a

108

result of invisible governors nor by central steering orders. The conservative horizon­tal policy networks too easily lead to rigidity, the top-down orientation to overestima­tion of the steering capacity and of central knowledge about common interests.Networks must be seen as interdependent constellations of several organizations, butalso as a constellation of different types of organizations. Not every organization playsthe same role in a network. In other words a network, however interdependent, has acertain degree of asymmetry. The asymmetry can be seen as an important motor fordecision making which does not lead to a hierarchical relationship. "Some processesinvolve substantial elements of dominance in superior - subordinate relationships, butthe way these processes are linked together establishes such fundamental patterns ofinterdependence that it is impossible to identify a single center of authority that isabsolute, indivisible and unlimited" (Ostrom, 1985). The asymmetry of parties oftenhas to do with their access to means. Some actors have the formal right to adoptcertain proposals. They can select that which they prefer most. This does not meanthat they possess absolute power. They cannot develop proposals. Their choicedepends upon the variety that already exists within the group of initiators. Initiativesoften come into being because of the efforts of street level professionals. These actorshave knowledge about the products and the market and can therefore see theproblems arising from market developments or from the effect of the product on themarket. The initiatives will not be presented directly to the policy adopters. Theissues will first be debated by the different agencies involved. A proposed initiativewill be compared to the goals set and will be scrutinised for weaknesses. Eventually, ifa proposed initiative passes through enough levels, it will reach a point where formaladoption may take place. After adoption the further policy-making process is oftenthe task of the initiators. This means that the center of gravity of decision making isat the concrete level of daily policy development. At the same time it can be statedthat the ability of central actors to select certain policy projects has a tremendous im­pact on the process.

Asymmetry also alleviates the danger of bureaucratic politics and other forms ofnegotiations being brought to a standstill. The asymmetry makes it possible for certainactors in the network to implement certain formal instruments such as laws andreorganization. As a threat, these formal policies can put pressure on interactingorganizations to solve certain problems and to take initiatives. Interwoven networksnot only consist of interdependent actors, but also of dissimilar actors, some of whomhave the ability to interact and play the game, while others have the ability to selectand to intervene. The role of referee is important for speeding up the process, butdoes not give the referee the ability to take over playing. For this he does not possessthe prerequisite skills.

6. Conclusions

In this chapter we have introduced two perspectives for the analysis and evaluation ofpublic policy: the unicentric and the pluricentric perspective. In the pluricentricperspective in contrast to the unicentric perspective, the government is seen as acomplex constellation of actors. Decisions on means and goals are made in a networkin which actors struggle to influence decision-making processes.

A policy network is a group of actors who form a social system which is involved

109

with policy problems and/or policy programs, and where interdependency existsbetween the actors which is based on the possession of resources. On the one hand,actors take initiatives and activate certain parts of a network and, consequently,reshape the network. On the other hand, the network has certain cultural andstructural characteristics, which constrain or facilitate policy initiatives and influencepolicy results. Even though networks may be stable, all actors within a network have acertain amount of policy-making freedom. This freedom is an important means todevelop certain specific goals and to fuse these goals. One of the dangers of anetwork structure in a policy field is that of stalemate. Policy advisors using apluricentric perspective must pay attention to this weakness in a network. Asymmetryin an interwoven structure is one means of preventing stalemate. Asymmetry introdu­ces actors who are capable of intervening and arbitrating without becoming directlyinvolved. Accepting the policy freedom of all actors, a pluricentrically orientatedadviser must also accept the existing organizational network and develop tailor-madestructures to improve the process of concrete decision making. In order to facilitatethe decision-making process, borderlines between actors must be permeable. Thispermeability can be heightened by adequate inter-organizational management supportat the most practical level as well as through political interaction. The main goal ofpolicy making is no longer the exact implementation of the policy of one actor, butthe intertwining of several goals and the aim to reach a satisfactory solution for all ofthem. The main difference between this network perspective and the configurationtheory is that the network perspective focusses on interactions generated because ofresource dependencies. Network theory also pays more attention to structures andtheir influence on the interaction process and on the role of power differences in theoutcome of interaction processes. Nevertheless, we think that both perspectives havea lot in common despite their different theoretical roots and each can contribute tothe analysis of public policy and the interaction processes concerned.

Notes

1. Of course these descriptions of the two perspectives on government are ideal types. In themonocentric approach, analists have found several ways of conceptualizing the influence of otheractors into the policy-making process. One of the more interesting ways of achieving the notion ofinter-dependency of government on other actors within the monocentric view on government isElmore's backward mapping (Elmore, 1985). Backward mapping tries to incorporate the strategicchoices and possibilities of the implementing actors into the policy implementation program of thefocal organization. The central government does, however remain the starting point for theanalysis, evaluation, and steering of public policy.

2. Structure is conceptualized bere, in accordance witb Giddens (1984), as "Rules and resources,recursively implicated in the reproduction of social systems". Structure, according to Giddens, existsonly as memor:; traces and manifests itself in practices and systems. Giddens mostly makes thecomparison to language. We speak and write (social practices) but use grammar rules wbicbenables us to do tbat. We learn tbese rules in our dealings with others. In this way structures are aprcconditioo for action. But we also modify tbese rules through our exchanges witb actors and

110

without these exchanges these structures wouldn't exist. Giddens speaks of this phenomenon as 'theduality of structure' ("The essential recursiveness of social life as constituted in social practices;structure is both medium and outcome of the reproduction of practices" Giddens, 1979; pp.5).There is no need for us to really can make these structures discursive. Most people do not knowthe grammar language they use so often although they do possess knowledge of it. Practices andstructures are connected by social systems. Giddens means by social systems "the patterning ofsocial relations across time-space, understood as reproduced practices" (Giddens, 1984; 377).Contrary to some network theorists (Lauman and Pappi, 1976, Aldrich and Whetten, 1981) we donot consider the structure of a network to be a stable pattern of relations between social positions.In our (and Giddens) view this is the definition of a social system. Systems do have structuralproperties but are not structures. In infonnal and formal patterns, however, (what normally iscalled the organisational structure) you can find structures that influence the interaction patterns ofactors.

3. See the chapters by Maas and Voogt. In both chapters one finds strong assumptions of freeinteraction among participants (for example the concept of 'game' in Voogt which assumes thatpeople agree on concrete rules. What is missing here is a theoretical assumption about theinfluence of already existing rules on the rules for game; •As soon as the rules of the searchprocess, the perspective, have been agreed upon, the Game can start. Interaction can now beorganized on the basis of content criteria (for example, an objective) as set in Play.· This way ofdescribing interaction processes strongly resembles a market-like free interaction situation). Bothchapters reveal an interest in information flows and the improvement of those information flows.

4. To be sure; in the network perspective there is an important group of analists who are alsopreoccupied with the mapping of interactions and connections between social roles (f.i. Laumanand Pappi, 1976). Our use of the word network is less descriptive and we favor an actor perspecti­ve while some network theorists favor something which comes close to a system perspective (f.i.Van der Yen, Emmett and Koening, 1975).

5. A variety of terms have been used to describe these policy networks such as subsystems (Milwardand Wamsley, 1985), policy-sectors (Benson, 1982), meso-corporatism (Cawson, 1986). Anotherterm relevant to this discussion is Heclo's 'issue network'.

6. In this sense the network structure is both a necessity for the actor as the outcome of the action ofseparate actors. See for this 'duality of structure' Giddens, 1984.

111

Chapter 10

CONSISfENCY,CONFIGURATION,CWSURE AND CHANGE

G. Broekstra

1. Social autopoiesis: a controversy

Autopoiesis is a typical biological systems concept. As developed in the seventies bythe Chilean scientists, Humberto Maturana and Francisco Varela (1980), it stands forthe autonomous, self-referring, identity-maintaining and self-producing organization ofliving systems in relation to their unitary/holistic character. An autopoietic system isfundamentally a homeostatic machine maintaining invariantly, as the critical systemicvariable, its own organization, defined as a network of relations (Varela, 1979). Theyare "self-contained unities whose only reference is to themselves" (Maturana andVarela, 1980). The image of these self-referring situations is the mythical one of thesnake eating its own tail.

Maturana and Varela's approach to the phenomenon of biological autonomy interms of self-referential, recursive processes is very much in the tradition of the 'logicof pure mechanism' as proposed by the cybernetician Ross Ashby (1952), who studiedthe ability of the nervous system to produce adaptive behavior. To understand thephenomenology of biological systems, the approach does not resort to the domain ofdescription by introducing non-material organizing or functionalist principles, such asgoal, purpose or function. The focus of interest is not the particular embodiment ofmechanism, and thus not the properties of components, but the processes andrelations between processes realized through components.

Autopoiesis attracted a great deal of attention in the second half of theseventies, and led to some heated debates at various Systems Conferences held inthat period. Although the first publication in English by Maturana, Varela and Uribeappeared in 1974, my first encounter with the concept was at the InternationalSystems Conference held at Binghamton, New York, in August 1977, where Varelapresented a challenging paper on the notion of the autonomy of living systems. Heunfolded the generalized notion of organizational closure in terms of his "ClosureThesis: every autonomous system is organizationally closed." He suggested that thisthesis could provide heuristic guidance that would give more specific content to theintuitive notion of self-referring autonomy. This thesis holds that "ifyou are interestedin the autonomous aspects of a system, go and look for the way in which itsorganization closes onto itself' (Varela, 1978). As the then President of the DutchSystems Society, I invited Varela to speak at the 1980 International Anniversary

113

Systems Society in Volendam. This was right after the publication of Varela's book(1979), and his ideas then became the subject of some intense discussions among systemsscientists in HoIland. It may also be of interest to know that the German sociologistNiklas Luhmann, who later became an enthusiastic supporter of the autopoiesis conceptfor social systems (Luhmann, 1984), was another invited speaker -though on a differenttopic- at that meeting.

Because of my background in control engineering and involvement in theapplication of systems and cybernetics to problems of organizational change, I was deeplyinterested in Maturana and Varela's approach to the nature of the living organizationofbiologicaI systems. The less understood phenomenon of autonomy, internal or self-law,and the weIl-charted concept of control, external or other-law, are in fact mirror images.Initially I felt that, taken together as complementary concepts, they could also providedeeper insights into the nature and behavior of social systems. Be that as it may, thenotion of social autopoiesis, in particular, has not gone unchallenged.

When Beer (1980) and Zeleny (1980) proposed that certain social systems couldbe understood as autopoietic, Varela himself called this a "category mistake: they confuseautopoiesis with autonomy." Varela elaborated:

The relations that characterize autopoiesis are relations of productions ofcomponents. Further, this idea of component production has, as its fundamentalreferent, chemical production. Given this notion of production of components, itfoIlows that the cases of autopoiesis we can actually exhibit, such as livingsystems.... , have as a criterion of distinction a topological boundary, and theprocesses that define them occur in a physical-like space, actual or simulated in acomputer (my emphases).Thus the idea of autopoiesis is, by definition, restricted to relations of productionsof some kind, and refers to topological boundaries. These two conditions are clearlyunsatisfactory for other systems exhibiting autonomy. Consider for example ananimal society: certainly the unity's boundaries are not topological, and it seemsvery farfetched to describe social interactions in terms of 'production' ofcomponents (Varela, 1979).

Varela then continued to say that, nevertheless, the lessons offered by the autonomy ofliving systems, such as the ceIl, may be converted "into an operational characterizationof autonomy in general, living and otherwise." This led him to his definition oforganizational closure. I agreed with Varela then, and I still think that the applicationof the weIl-defined and restricted notion of autopoiesis to social systems leads to ratherwoolly generalizations.

More recently, in 1989, the debate flared up again in systems circles because of thepublication of, and ensuing discussion on, a lucidly written article on autopoiesis byMingers (1989). I fully concur with Minger's emphatic assertion: "This term (autopoiesis)is now widely used, although often without a proper understanding of the detail of thetheory. Moreover, the concept has been applied to other systems, such as societies andinstitutions, in a rather naive manner" (Mingers, 1989). The interested reader is referredto this debate to draw his or her own conclusions. I will not repeat the arguments here.It occurs to me that people sometimes succumb to a mystical sounding word like

114

autopOlesls to impress their audiences with the magical wands of the priesthood ofscientism, rather than taking the trouble to critically examine the content and domain ofapplication of such concepts. I have been using the admittedly less exotic sounding notionof organizational closure, which I believe has more interesting and less controversialconsequences for social systems theory. I would like to demonstrate this in what follows.

2. Organizational closure: an alternative

To fully comprehend the concept of organizational closure it is important to understandthe terms organization and structure,as used in a somewhat uncommon way by Maturanaand Varela. These terms occur in the context of the recursive distinction by an observerof a composite unity (system or whole), as an entity distinct from a background. Therelations between the components that define a unity as one of a particular kindconstitute its organization. Only those properties of components are viewed thatparticipate in the network of relations that constitutes an integrated whole. Otherproperties of the components that are not required by the realization of the organizationof the unity are irrelevant for the definition of the organization.

The actual components with all their properties and the actual relations between-them that concretely realize a system as a particular member of a given class constituteits structure. Mingers uses the example of a car. The organization of "a car may bedefined by describing the necessary relations between components such as steering,brakes, seating, power, etc. If a unity has such an organization, then it may be identifiedas a car since this particular organization would produce the properties we expect in acar as a whole" (Mingers, 1989). The actual properties and actual relations constitute thestructure, 'such as the rusty blue Mini in my drive". Mingers concludes that thedistinction between structure and organization, as the distinction between the reality ofan actual example and the abstract generality lying behind all such examples, is stronglyreminiscent of classical structuralism "in which an empirical surface structure of eventsis related to an unobservable deep structure of basic relationships" (Mingers, 1989). Tofurther clarify the important distinction of structure and organization, we will pursue theexample of Mingers' car a little further.

An existing unity therefore has both a structure and an organization. There will bemany different structures which can realize the same organization and the structurewill have many properties and relations not specified by the organization, andessentially irrelevant to it, e.g., the shape, color, size, and material of a particularcar. Moreover, the structure can change or be changed without necessarily alteringthe organization, for example, as the car ages, has new parts, and gets resprayed,it still retains its identity as a car. Some changes, however, will not be compatiblewith the maintenance of the organization, e.g. ,a crash which changes the car to awreck (Mingers, 1989).

Since the application of the terms organization and structure, as defined by Maturanaand Varela in the context of biological systems, gives rise to an even greater Babylonianconfusion of tongues in the realm of organization and management theory, I will

115

henceforth maintain the term structure as proposed by these authors and replace theterm organization by that of deep structure. It is clear now that the invariance of theorganization/deep-structure of a system and its identity are closely related. While thestructure may change within limits determined by its deep structure, the system's identitymay remain invariant. If, however, the deep structure of a system changes, its identitychanges and it becomes one of another kind. Varela's definition of organizational closurecaptures the essence of autonomy.

We shall say that autonomous systems are organizationally closed. That is. theirorganization is characterized by processes such that (1) the processes are related as anetwork, so that they recursivelydepend on each other in the generation and realizationof the processes themselves. and (2) they constitute the system as a unity recognizablein the space (domain) in which the processes exist (Varela, 1979).

Clearly, autopoiesis is a particular instance of organizational closure, that is, when theprocesses are processes of production of (physical) components. It is also important tonote that when the "circular concatenation of processes that constitute an interdependentnetwork" arises, the network forms a "self-computing organization, which attainscoherence through its own operation, and not through the intervention of contingenciesfrom the environment" (Varela, 1979). So, in describing autonomy and related conceptssuch as identity and configuration (see below), closure definitely refers to the absenceof descriptions of systems in terms of inputs and outputs which embody their externalcontrol and constraints. This, as Varela remarks, makes the Wienerian concept offeedback that requires a reference external to the system quite distinct from the idea oforganizational closure.

Furthermore, of particular importance for autonomous social systems is therealization that the recursive interdependence of processes that constitute organizationalclosure may contain the observer himself as part of the process specifying the system,"not only by describing it, but by being one link in the network of processes that definesthe system" (Varela, 1979). Varela explicitly refers to the phenomenon of the verydescription of the system inducing changes in the system through the 'hermeneutic circleof interpretation-action'. A system establishes through its recursive, self-referringinteractions an identity which generates, in Varela's words, a cognitive domain. At thispoint we can connect the idea of organizational closure with the conception of thesocial-eognitive configuration as developed by van Dongen (1991), and with my ownideas related to the so-called Consistency Model (Broekstra, 1984). In the latter, thedominant coalition represents the connection with the social dimension, at the heart ofan otherwise cognitive deep structure of key strategic processes (see below.)

There is one more issue that should be emphasized in the case of business or, forthat matter, other types of organizations. The standard view of organizations is so muchdominated by the open-systems control model of input-transformation-output that it willappear quite strange to speak of organizational closure. It is therefore important torealize that a system which exhibits organizational closure in its deep structure is notnecessarily interactively closed, since structurally it may have rich interactions with itsenvironment, for example, through the exchange of matter, energy, and information. Thisleads to Maturana's ideas about structural coupling (Maturana and Varela, 1980), where

116

perturbations in the environment are viewed as triggers of changes in the state of asystem that tries to accommodate structural changes in such a way that it maintains itsidentity through organizational closure.

In an organizationally closed system, all changes are subordinate to themaintenance of the system's autonomy and identity. Obviously, this view can havedramatic consequences for our description of organizational functioning and ourconception of the management of change. As Varela notes: "If the corporation exhibitsclosure, no matter what our description of the system's purpose is, its behavior will besuch that all perturbations and changes will be subordinated to the maintenance of thesystem's identity. This is so even when we may treat perturbations from the environmentas controlling inputs. Such controlling inputs belong to an alternative description of thesystem, revealing a phenomenology that is complementary, but not reducible, toautonomous behavior" (Varela, 1979). In addition, this perspective leads to areinterpretation of the concept of information as constructive rather thanrepresentational or instructive; Varela continues: "For such systems, all apparentinformational exchanges with its environment will be, and can only be, treated asperturbations within the processes that define its closure, and thus no 'instructions' or'programming' can possibly exist" (Varela, 1979). Clearly, we are becoming quite awarenow of the duality and complementarity of the concepts of control/steering andautopoiesis/closure. The first pair refers to the influence that may be exerted on a systemat the structural level through structural coupling. The second refers to the self-referringand closed character of the deep structure of the system involved in the maintenance ofits autonomy and identity, resisting any attempts at change. This, in turn, is reflected inthe structural phenomenology of the system, much, of course, to the dismay of the 'socialengineer'. For, more often than not, after an initial phase of euphoria, change efforts,especially of the 'grand-design' or programmatic type, shipwreck resulting in a situationin which positions are more entrenched and hardened than ever before. This may partlyexplain what we now see happening in the Soviet Union.

Likewise, Mingers notes the far-reaching consequences of the self-referringautonomy-view in saying: "it is wrong therefore to suggest that the environmentdetermines or specifies what will be the changes of state of the system" (Mingers, 1989).Where this may lead has been vividly described by Morgan (1986), who suggests thatfirms may understand their environments only as a product of their own identity, and thatin order to change significantly they have to construct totally different conceptions ofthemselves. And, this is the hardest part of all, generally overlooked by 'social tinkerers'.We have to limit ourselves to these introductory remarks on Maturana's and Varela'sfascinating theories. We will now proceed to show some of their implications in termsof the Consistency Model developed by this author.

3. The Consistency Model

More than anything else, the Consistency Model, or C-Model, constitutes a researchprogram that I started in the early eighties to develop a systemic or holistic approach toorganizational change and effectiveness. The first publication appeared in the systemsliterature in 1984. Since then I have elaborated on its applications and implications in

117

a series of articles (Broekstra, 1984-1991), the more recent of which proposed a newterm for the holistic perspective on management entailed by this approach:h%managemenr (Broekstra, 1990). The underlying rationale of the C-Model and theassociated management logic is that the quality of top management, overallorganizational performance and 'health', and other indicators of effective, strategicmanagement are closely related to the degree of holistic conceptualization oforganizational reality. This, of course, becomes a particularly compelling perspective inthe light of the logic of organizational closure. For the purpose of further reference, ashort synopsis of the key features of the C-Model will be presented here. For a moreextensive treatment and further references, the reader may consult previous publications.

Taking as its starting point the idea of a holistic vision, the model logically unfoldsfrom considering two fundamental concepts embedded in a vision: system and complexity(Broekstra, 199Ib). The first essentially results from a distinction between environmentand system, separated by a boundary (i.e., a triad). It therefore entails the dimensionexternal/internal. This we translate into two fundamental choices, one referring to thefundamental business idea, relating the organization to its construed environment, theso-called business choice, and one referring the organization to itself, the organizationalchoice. These represent the upper and lower halves of Figure I, respectively.

External

VarietyAmplification

Internal

VarietyAttenuation

Figure 1. The organizationally-closed deep structure of the Consistency Model in atwo-dimensional plane.

The second dimension relates to the fundamental measure of complexity, that is, variety.The second axis refers to variety amplification and variety attenuation. Incidentally, theaforementioned dimensions show a remarkable similarity to those used in Quinn's studies

118

on effectiveness (Broekstra, 1991b). This translates into a further subdivision of theabove two choices into two subchoices leading to the four Cardinal Generic Systems(CGSs) shown in Figure 1, that is, CGS1: Entrepreneurial System (ES), CGS2:Technological System (TS), CGS3: Administrative System (AS), and CGS4: HumanResource System (HRS).

The four interrelated systems may be regarded as the immaterial assets of anorganization that, in combination, constitute the basis for its effectiveness. Each of themrepresents a number of related variables or organizational competences. The C-Mode1can be applied at various levels (including the level of the individual (Broekstra, 1984»,and has also been applied to organizations in the public sector. For reasons ofconvenience, we will continue to refer to a business unit in the private sector. Thefol1owing questions may help to clarify the nature of the CGSs.

ES: what products/services to offer to what customers (product/market matrix), whatmarketing mixes, service image, market position, innovation rate, etc.?

TS: what value-chain, productive axis, service delivery system, know-how, businesslogistics, productivity, etc.?

AS: what formal structure, planning systems, control systems, reward systems,decision making processes, degree of differentiation/integration, degree offlexibility, etc.?

HRS: what skil1s, educational levels, management styles, morale, informal structure,commitment, etc.?

These four cardinal systems form an interrelated network, in which each CGS should,ideally, be related to each other through a consistency relationship, which has beencharacterized in the literature by a variety of synonyms such as fit, match, congruence,alignment, etc. The general idea is that to form an effectively functioning autonomouswhole the CGSs constitute a number of mutually supportive subsystems, containingprocesses that "recursively depend on each other in the generation and realization of theprocesses themselves" (cf. the definition of organizational closure, above.)

To illustrate, it is intuitively clear that a successful penetration into a rapidlychanging market with innovative products of ever shortening product-life cycles, such asthe consumer electronics market, needs to be supported, or realized, by an equallyinnovative technology, and consistent with that, a flexible learning organization, andhighly skil1ed technicians. In effect, what we are briefly, and incompletely, describing hereis the essence of a configuration formed by the consistency relations that configure theCGSs to form an integrated whole thus ensuring organizational effectiveness. In otherwords, the four interrelated CGSs, representing the systemic immaterial assets andassociated processes, constitute (part of) the deep structure of any organization(Broekstra, 1991a).

In order to emphasize the central role of the dominant coalition of an organizationin formulating a holistic vision, which is cognitively elaborated into the four interrelatedCGSs, and in emphasizing the link of the cognitive structure formed by the CGS systemwith the social actor perspective, the Dominant Coalition (DC) is symbolically placed atthe center of the model. The social system dimension is expanded by adding two moresubsystems of interrelated variables: the Political System (PS) and the Cultural System

119

(CS). The first represents the power and influence distributions that may productivelyenhance or oppressively inhibit organizational functioning. Finally, the cultural systemrepresents the value and belief system or SOUL (soft organizational unconscious level)capping the C-model. Both systems are pictorially represented as circles concentric withthe DC, to indicate their connection with the social dimension of this configuration, andsymbolically accentuate their importance in the implementation part of the totalC-Model. This also supports Weick's idea that Strategy, represented by the four coherentCGSs unfolding from a holistic vision, and Culture are two sides of the same coin(Weick, 1985).

It may now be clear from what has been said above on the subject of organizationalclosure that I propose that the network of relations connecting the processes representedby the seven subsystems of the C-Model answers the definition of organizational closure.The ensuing configuration is then entirely characterized by its autonomous aspects, andindeed functions as a self-referring whole primarily involved in maintaining its identityagainst a backdrop of environmental disturbances. To fully understand the import ofthese conclusions, some additional comments on the concept of configuration arenecessary.

4. Configuration

Among the first to make a clear case for configuration were Miles and Snow (1978).Their seminal work constitutes one of the pillars of the C-Model. In studying the processof organizational adaptation, they demonstrated that managerial choices, thoughnumerous and complex, can be usefully categorized into an adaptive cycle, consisting ofthree subsystems which they called entrepreneurial, engineering, and administrative.Organizational effectiveness can then be judged by examining the consistency relationsbetween them and the way management has enacted the organization's environment.Starting from the premises of Child's concept of strategic choice and Weick's concept ofenvironmental enactment, it is quite relevant to the present discussion to summarize thefeatures of the adaptive cycle as described by Miles and Snow (1978):

1. The adaptive cycle is a general physiology of organizational behavior.2. The three adaptive problems -entrepreneurial, engineering, administrative- are

intricately interwoven.3. Adaptation frequently occurs by moving sequentially through the three phases, but

the cycle can be triggered at anyone of these points.4. Adaptive decisions made today tend to harden and become aspects of tomorrow's

structure.

In my view, the first three items clearly point in the direction of an underlying intuitiveunderstanding of the concept of organizational closure, and of the wayan organizationresponds through its deep structure to environmental disturbances as discussed above.In addition, viewed in the light of the concept of organizational closure, Miles andSnow'sideas gain strong reinforcement, and their insights are meaningfully enriched. Thefourth point enhances these conclusions when read in terms of how the deep structure

120

of the organization as represented by the adaptive cycle, or for that matter, the C-Model,unfolds into a manifest structure. However, there is not an unlimited range of adaptivepatterns. On the contrary, Miles and Snow claim that the 'hardening' results in only a fewarchetypes. They identified a conceptual typology of four organization types, each ofwhich "has its own strategy for responding to the environment, and each has a particularconfiguration of technology, structure and process that is consistent with its strategy"(Miles and Snow, 1978). Three well-known stable and consistently pure strategic typeswere distinguished: defenders, prospectors, and analyzers, and one unstable andinconsistent type : reactors. Viewing organizations as integrated wholes or as 'totalsystems' ,Miles and Snow were thus able to show why managers embarking on programsof organizational change frequently make mistakes and jeopardize organizationaleffectiveness by a fragmented or piecemeal approach which does not take into accountthe interrelated aspects of the change process and the consistent integration ofmanagerial choices. In other words, the still widespread lack of 'configuration-thinking'among managers, particularly those who are preoccupied with the short-term and quickfixes, seriously endangers organizational effectiveness.

Our second example concerns Henry Mintzberg's (1983, 1989) work onconfiguration and that of his colleagues at McGill University, Miller and Friesen (1984),in the first half of the eighties. Mintzberg was concerned with the design of effectiveorganizational structures and argued that the elements of structure -coordinatingmechanisms and design parameters- should be selected in such a way that internalconsistency or harmony is achieved, as well as a basic consistency with the organization'ssituation. In his chapter on design as configuration, Mintzberg observed that the elementsof structure "all seem to fall into natural clusters, or configurations" (Mintzberg, 1983).Indeed, Mintzberg's configuration hypothesis "postulates that effective organizationsachieve an internal consistency among their design parameters as well as compatibilitywith their situational factors,- in effect, configuration." The effective organization "got itall together: consistency in its internal characteristics, harmony in its processes, fit withits context" (Mintzberg, 1989). Mintzberg developed a typology of five, later extended toseven, pure structural configurations: the entrepreneurial, machine, diversified,professional, innovative, missionary and political organizational forms. Later he alsoadded the idea of a system of forces (re)configuring the forms (Mintzberg, 1989).

However, it should be realized that Mintzberg's configurations are of the structuraltype, rather than related to variables that make up the underlying deep structure, in thesense discussed above. It appears that the failure to distinguish between structure anddeep structure contaminates Mintzberg's findings on configurations. The concept oforganizational closure, therefore, does not seem to apply to Mintzberg's configurations.In fact, this might be the result of a confusion of paradigms, which could be describedas the difference between the contingency approach, based on the open-system model,and the configuration approach, based on a closed-system model. At times, this confusioncan also be found in Miles and Snow's work. Implicitly, it appears that the authorsoccasionally submit to the standard open-system model directly under the influence ofits environment. On the other hand, it is the thesis of this paper that configurationrequires organizational closure of the system's deep structure. And, therefore,paraphrasing Varela's Closure Thesis: "ifyou are interested in the configuration aspectof a system, go and look for the way in which its deep structure closes onto itself'. This

121

thesis may offer heuristic guidance (Varela, 1978) in discovering the Essence of aperceived configuration.

Miller and Friesen's (1984) work is of a different kind, and their results qualify asgenuine configurational deep structures held together by organizational closure(Broekstra, 199Ia). They took the empirical road towards discovering configurations. Byusing Q-type factor analysis applied to 52 cases, they came up with a taxonomy of 10archetypes based on clustering 31 variables pertaining to environment, organization,strategy-making, and success. In their vanguard book, they presented their quantum viewof organizations treated as "complex entities whose elements of structure, strategy, andenvironment have a natural tendency to coalesce into quantum states or configurations.These configurations are composed of tightly interdependent and mutually supportiveelements such that the importance of each element can best be understood by makingreference to the whole configuration. Organizational structures, production systems,information-processing procedures, strategies, and environments all tend to influenceeach other. Our thesis is that they do so in a manner that gives rise to a small numberof extremely common and sometimes discretely different configurations" (Miller andFriesen, 1984).

Their arguments against the widespread analytic approach in organizationalresearch, leading more often than not to conflicting trivializations, and in favor of theapproach of synthesis, leading to a fulfillment of the need to obtain a holistic, integratedimage of reality, are quite compelling. Seeing the world in 'systems terms': "the approachof synthesis is really the search for networks of causation. "This again is very similar tothe organizational closure perspective. And, again, I contend that Miller and Friesen'sfindings could be much enhanced and clarified by applying the 'Configuration Thesis'.This is the subject of future research" (Broekstra, 1991a).

s. Change

From the notion of the consistency relation between components of the deep structureof a system, as exemplified in the C-Model, and through the concept of organizationalclosure as the self-referring network of such relations, and, subsequently, the 'hardening'of the system's closure into some configuration, we finally arrive at the problem ofchange, and the management of change. In their book, Miller and Friesen dealtextensively with the dramatic consequences of the configuration perspective, and hence,as we have argued, of the organizational-closure perspective, for the approach toorganizational change. They advanced the notion of 'quantum change'- long periods ofthe maintenance of a given configuration, punctuated by brief periods of multifacetedand concerted transition to a new one. They dealt extensively with the reasons for'sluggish adaptation' and 'resistance to change', or why organizations often opt for themaintenance of internal consistency and retaining the internal configuration as long aspossible, rather than continually adapting to the environment. That sudden and longoverdue renewal of a 'tried and true' configuration, often initiated by a new tornado-typechief executive, can have the character of a revolution is exemplified by the presentpredicament of the Dutch electronics giant Philips.

122

As Miller and Friesen argued, the dilemma is between the cost of the structuralconfiguration being out of kilter with the environment or strategy (CI), and the cost ofcausing imbalances among the mutually supportive elements of the internal configuration(C2). C2 costs will often be high because many elements must change together to formanother configuration. As Miller and Friesen concluded their case for revolutionarychange: "such quantum structural change must therefore be delayed until the anticipatedlong-run CI costs are larger than the C2 costs. Often these delays can be considerable"(Miller and Friesen, 1984). This is particularly true for some public organizations inwhich, for lack of a fiercely competitive environment, bureaucracies are the dominantconfigurations that are heavily organizationally closed. Though these 'sluggishthermostats' are dramatically and increasingly incompetent to cope with the complexitiesof a greatly incongruent environment, a critical state where society is ready to accept thatthe Cl costs have exceeded the C2 costs has apparently not yet been reached. It can beanticipated on the basis of the argument of organizational closure that change, when itfinally comes, may have to be of a disruptive revolutionary nature.

Organizational closure occurs in many guises. One of them is the idea of resistanceto change. Rather than viewing it as some innate social evil that needs to be overcome,preferably in a forceful way, Goldstein (1986) has argued that a new perspective ofresistance to change can be arrived at when it is viewed through the lenses ofautopoiesis, or, rather, organizational closure. Ifresistance to change is seen as a 'survivalmechanism' ensured by the self-referential character of organizational closure, one canbegin to think of alternative strategies to change. The essence of this is that rather thanworking against the resistance, one starts to work with the resistance and attempts toreframe it. Here other systems concepts may play an important metaphoric role, such asthat of Prigogine's far-from-equilibrium systems and those from chaotic systems theory(Broekstra, 1991c). As an example, Goldstein suggests the technique of differencequestioning, known from family therapy, where the system is not inundated withinformation from the outside, as the steering or control view would propose, but where,instead, "the system itself must be induced to generate more information about itself"(Goldstein, 1986). This concurs with the constructive rather than instructive nature ofinformation that emerges under the banner of the organizational-closure perspective.Information is generated to in-form, to form within, and may be used to createfar-from-equilibrium conditions where the system becomes ultrasensitive to even smallfluctuations (Lorenz's Butterfly Effect) and reveals a new orderly dissipative structure.This, incidentally, raises the intriguing question of the relationship between the notionsof organizational closure and configuration, as developed here, and the appearance ofcoherent structures found in turbulent environments, as studied by chaos theory.

6. Conclusions

The objective of this paper is to indicate how the important systems concept oforganizational closure can provide a new perspective to reinterpret the organizationalphenomena of configuration and (resistance to) change. One thing is clear. If change isviewed solely from the popular control or steering perspective, trouble is imminent. Or,as Varela put it: "But unless we take into account that there is an autonomous side to

123

many natural and social systems, we run into troubles, not only in the specifics ofresearch and formalizations, but in the wider scale of our dealings with sentient beings,with life, with the environment, and in human communication" (Varela, 1979).

It is also quite clear that research on the implications of the concept oforganizational closure for the configuration perspective and the problems of quantumchange is far from complete. In particular, I expect that some new insights may be gainedfrom the exciting new developments in chaotic systems theory. This is the subject offurther research.

Furthermore, although we may better understand how self-reference lies at the rootof the autonomy of wholes, the fundamental question of what drives systems towardsorganizational closure in the first place remains unanswered; a clear case of G6del'sTheorem of Incompleteness. In this respect, it may suffice to refer to the distinctionbetween the concepts of accidental and substantial wholeness (Broekstra, 1990). Thisleads to the conclusion that even the modern systems concept may be largely based ona mechanistic-organicistic position. And, this entails that the concept of wholeness cannever be comprehended without the metaphysical anchorage of the phenomenon ofwholeness in the order of being. It has long been known, for example, from the Vedicscriptures of ancient India, that the (deep) structure of Being is also of a self-referring,holistic nature. Would it then come as a surprise that manifestations of Being -or,Buddha, Tao, Pure Consciousness, Universal Intelligence- are also intrinsically strivingtowards wholeness and organizational closure? I am sure that the Buddhist Varela wouldhave no problem in answering that question.

124

Partm

AUTOPOIESIS THEORY

Chapter 11

AlITOPOIESIS AND STEERING: HOW POLmCS PROFIT

FROM TIlE NORMATIVE SURPLUS OF CAPITAL

G.C.M. Teubner

1. Introduction

I shall begin with a question that is trivial as it is fundamental: why, in order to steersociety, does politics take the detour through law? Why does it not rule directly intosociety at its own boundary points, say by authoritatively ordering money payments, orby using economic interest groups as 'juridifying' instruments of political control?Instead, politics takes the laborious path of norming its policies through legislation,adjudication and administration, only in the end to find how great expectations inWashington are dashed in Oakland.

Consensus? Efficiency? Rationality? The legitimacy of legality has somewhat lostits savour, particularly if one thinks of the real acceptance problems of the law (seeTeubner, 1991a). The effectiveness of law as a steering device is in tum ratherskeptically assessed by the wisest among its legal sociological advocates (Rottleuthner,1991). And not much is left of the rationality of bureaucracy in the implementation ofpolicy-oriented legal norms (since Pressman and Wildavsky, 1973). It failed becauseof self-created perverse effects, implementation deficits and self-fulfilling as well asself-defeating prophecies. I Why then, despite all the well-known failure of law, isthere this unfailing clinging to law as a means of political control?

Perhaps we can take help from a hint from history - specifically the history ofBritish colonial law in Africa. In the 19th century the British colonial masters met thebitter experience that their attempts to force British law on the African colonies wereending in chaos. Their expedient, in the early 20th century, was a 'soft law' methodthat proved to be highly successful. They instructed their legal staffs to apply the'indigenous law' of the Colonies themselves as far as possible (Ranger, 1983,pp.250ft). The colonial masters were therefore no longer handing down their own law,but instead incorporating indigenous law into their official law. They retained onlyone minor power. They deprived customary law of force where it proved incompatiblewith fundamental British legal principles. The limit to validity was the repugnancyprinciple: indigenous law did not apply where it was 'repugnant to natural justice,equity, and good conscience' or was manifestly 'inconsistent with any written law'(Merry, 1988, p.870; Okoth-Ogendo, 1979, p.l60; Adewoye, 1986, p.60; Bentsi-Enchill,

127

1969). Wise self-restraint by enlightened colonialism? Quite the contrary. Criticalanthropologists have recently succeeded in unmasking the secret dominance mecha­nism of this soft law. Through laborious detailed research they were able to show thatindigenous law as such did not exist anywhere. The whole thing was a bluff, a purefiction by the British colonial administration and its compliant anthropologists(Snyder, 1981a, 1981b; Ranger, 1983; Moore, 1986; Starr and Collier, 1989). It was inthis very lie that the trick lay: indigenous law or customary law were not at all, asassumed by traditional anthropologists, "rules that trace back to the habits, customs,and practices of the people" (Al1ott, 1960, p.62), but "were constructs of the Europeanexpansion and capitalist transformations", and therefore nothing but a "myth of thecolonial era" (Merry, 1988, pp.875ff). Out of a multiplicity of entirely heterogeneouscultural sediments, the British jurists cobbled together only the elements that suitedthem, into a collage they then presented as existing 'indigenous law' in order to give itthe stamp of colonial legal authority. The British thereby opened up for themselves, anew, richly yielding source of law with its origin ostensibly lying in the actual socialpractice of oppressed peoples - but usable by them for political manipulation. Theeffect was a stronger commitment of the 'indigenes' to their own pseudo-indigenouslaw, greater legitimacy and acceptance for the British colonial administration, andhigher effectiveness for 'modernizing' colonial policy from the customary law thusdoubly manipulated (Snyder, 1981b, pp.74ff). What does this cynical self-embedmentof British colonial administration in the indigenous culture of Africa, in order betterto control, manipulate and exploit it, teach us? It gives us a partial answer to theinitial question, why despite all failure of law, politics continues to hold to law as aninstrument of steering. Summarizing in the form of theses:

I. Politics embeds itself in structural couplings of law and society in order to cream offthe normative surplus-value of social self-reproduction for its control purposes.

2. I suggest to distinguish between structural coupling and linkage mechanisms inoperationally closed systems: structural coupling is dependent on specific linkagemechanisms that decide its duration. quality. intensity and institutionalization. Ofpanicular interest are the cyclical linkages that bring aboul a kind of ecologicalrecursiveness.

3. In a suitable situation these linkage mechanisms may set off systematic regulatoryeffects which in tum. can be exploited by regulatory politics. or even artificiallyinstitutionalized.

2. Modes of interpretation

Critical observers perceive the last position of autopoiesis theory as "radical regulato­ry pessimism" (Scharpf, 1989, p.IO): all political and legal regulation is regarded asrunning up against the 'intrinsic logic' of social sub-systems. Legal norms 'break' onthe code of the regulated system and can only 'perturb', 'modulate', and 'stimulate' itin some obscure way. Theoretical statements on regulation get stuck in vagueness andgenerality without being able to point to specific criteria for regulatory practice oreven only for the detailed observation of regulation processes (pointedly expressed byRottleuthner, 1989, p.281). Is it really impossible for autopoiesis theory to specifyideas on regulation, particularly those on structural coupling, sufficiently to "make

128

specific differences of typical problem situations visible" (Kaufmann, 1988, pp.71ft)?Let us do a sample test. Let us take Max Weber's classical, simple example of

legal regulation of the economy, 'price taxes'. Recently in particular, with theimmoderate petrol price increases by the oil multinationals taking advantage of theGulf crisis, 'price taxes' by European Governments, particularly in France but alsoregulatory res~nses from the Japanese MIT! and the World Bank, have againbecome topical. 2 And immediately the effects of legal price controls are broughtbefore our eyes again: in Southern France petrol stations closed because of the pricecontrols. In the U.S. the President by legal regulation ordered the partial release ofoil reserves to bring down market prices, with the result that because of this measurethey rose. 3 If here we replace simple causal control models (price regulation - >height of sanction -> motivation of actor -> compliance with or departure fromnorm), as still used in current mainstream sociology of law (Opp, 1973: pp.190ff;Diekmann, 1980, pp.32ff; Rottleuthner, 1987, pp.54ft) by multiple interwovenness ofautonomous closed discourses, this forces upon us the world view of the sort ofoccasionalism once advocated by Geulincx (1624-1669). The world of law and theeconomy would then seem to us like a mechanism of clocks running in parallel,except that we could today no longer see them as mechanical machines but ashermetically closed self-reproducing discourses and that the clockwork would nolonger be synchronized by God-ordered harmony but by the more or less chaoticdissonances of blind evolution. Evolution brings informational events in closedsystems into such a relationship as to make some of them into 'causae occasionales'of the others. The running of the clocks of law is not an effective cause of the runningof the clocks of the economy, but only its 'causa occasionalis' (Vor1<i"nder, 1966,pp.25ff; Jonas, 1969: IV, p.168).

We would then have to transfer the linear causal relations of Gulf blockade,price forcing by the multinationals, protests by the public, 'price taxes' by govern­ments, and the avoidance behavior of economic actors into an acausal parallelprocessing of legal and economic information. What appears in law with beautifulregularity as positive or negative value of the binary code legal/illegal shows upsimultaneously - in completely uncontrolled fashion - on different screens in thecomplex network of differences of the economy. These signals appear at quitedifferent places, depending on how the economy's sensors happen to have gone off:now in the economic constructs of reality, now in the economic code itself, now inprice signals, now in the most diverse economic control programs in market andorganization. In other words, there are various 'modes of interpretation' in which theeconomic discourse observes the price taxes of legal discourse, i.e. reconstructs themin the economy's own language as economic information.

Interpretation as 'accounting items' in the economic calculus

This is the current interpretation today. Rational economic actors interpret price taxesnot as legal norms bidding obedience and regarded as unbreakable, but instead makestrategic use of them. Legal price controls are the object of cost-benefit calculations,the outcome of which as net benefit decides whether they are to be complied with ornot. 'Amount of sanction multiplied by likelihood of sanction' - that is, ultimately, thecalculation. And the compliance or otherwise of rational actors with law can veryappositely be characterized by the formula 'efficient breach of law'. Indeed, progres-

129

sive legal economists are today already talking about a legal duty of efficient breach oflegal rules (Easterbrook and Fischel, 1982, p.Il77). But, and this is often overlooked,even this economic interpretation of legal price controls which it already alienatedfrom the interpretation of norms inherent in law is only one among many economicinterpretations. In technical terms, the accounting-item interpretation sets up structur­al coupling between law and the economy in such a way that the legal event of aprice tax acts as a perturbation of economic processes and leads to the alteration oftransaction cost calculations, that is, of program structures in the economic system.This coupling with cost calculations is, however, as we have said, not necessary. Thesame legal norm may also, without this being predictable or controllable externally,be used to indicate quite different distinctions proper to the economy.

Interpretation as 'property rights'

Not every legal norm is automatically interpreted by economic actors in terms ofcost/benefit calculations. The responses of Japanese oil multinationals to the silent'price taxes' of MITI show that there are other ways too. 4 Many legal norms arecoupled not with program structures but with the property code of 'having or nothaving'. The economy then no longer interprets them as an object of avoidancestrategies, but on the contrary as fixed limits to actual leeway for action. In thecalculations of rational actors, legal norms start to act as constraints and no longer aschoices depending on how strong the 'guile' of 'opportunism' is (Williamson, 1985).They are now regarded as a modification of 'property' understood in economic (notlegal!) terms, of property rights, of assets, of regimes - all interpreted as opportunitiesfor action of a purely factual nature. Correspondingly, norms are recorded in theeconomic text neither as manipulable objects, nor as normative precepts; but insteadas genuinely economic expectations of a cognitive and not normative nature. Theirbases are physical, biological and cultural factual situations, which may include legalnorms too, and even our price taxes.

Interpretation as 'bargaining chip'

Quite different room for manoeuvre opens up where legal norms are exploited for allsorts of manoeuvres and extortions in other contexts than those aimed at by the legalnorm. 'Trade in justice' (Schumann, 1977) faces us with a further interpretation oflegal events that goes far beyond penal law practice. In "bargaining in the shadow oflaw" (Mnookin and Kornhauser, 1978), actors do not apply norms of conduct butemploy them. They use them as a means of pressure in order to attain other goals."The 'letter' of the law becomes the keystone of the negotiating position of parties tothe interaction" (Winter, 1990, p.329). Rights are not exercised, procedural positionsnot taken up, possibilities of legal action not exploited; instead their employment ismerely threatened in order to build up a negotiating position in other economiccontexts. Practice in antitrust law shows many cases where given price abuses or othersituations of relevance in antitrust law, the competitors or even the antitrust authori­ties have allowed themselves to be well paid for not exercising their right of legalaction or their powers of intervention.5 And more recent control strategies in legalpolicy even deliberately employ negotiating and mediating mechanisms based on thissort of economic interpretation of the law (Hoffmann-Riem, 1989).

130

Interpretation in the context of economic 'selfregulation programs'of economic organizations

The need for detailed empirical studies emerges particularly with regard to thequestion of which specific 'program' is being pursued by the individual firms in theeconomic regulatory area and correspondingly leads to a particular economicinterpretation of law. Here it is by now means enough to start a priori from programsof profit maximization to which the cost calculations of legal compliance are subordi­nated. Instead, it needs close observation to find out which programs of self-regula­tion are actually being pursued in the regulatory field: organizational growth strate­gies, mere survival strategies by firms, priority to guaranteeing jobs, managementinterests or interests of institutional investors, programs of risk minimization insteadof profit increase, or avoidance of losses of reputation.6 The specific shape of suchstrategies decides the central question whether legal control and economic control can'meet' or not. It is just as dependent on the degree of concentration of the marketconcerned, the specific power positions and the prevailing interaction patterns as onthe rigid organizational hierarchies or else soft organizational cultures of the firmsinvolved. The oil multinationals thus give different readings to petrol 'price taxes' bythe French governmental authorities, MITIor the American president. If themultinationals are concerned primarily with minimizing losses of reputation, then fora happy moment the self-regulation programs of law and the economy can 'meet'.And Mr. Nahamowitz (1990, pp.7ft) can again record a success for regulation byinterventionist law. "Once the work was done the dwarf went about boasting every­where he had done it himself' (Kaupen, 1975, p.34).

Interpretation as 'change of preferences'

This case is often neglected by economists: de gustibus non est disputandum (Stiglerand Becker, 1977). They prefer to deal with preferences, their formation and changesas a sort of black box, calling on other academic disciplines to shed light on it. Forsystems theory by contrast, this case is particularly interesting (Luhmann, 1988: 275ff,280ff; Baecker, 1988: pp.126ff, pp.318ft). It will treat preferences of economic actorsnot only as external mental motivations in the heads of those involved, but asgenuinely social expectations, as structures internal to the economy, attributedcommunicatively to the individuals or to the collective actors as semantic artefacts.Preference changing is after all the (alas, all too rare) ideal case of regulation of theeconomy by law. And if Max Weber (1978, pp.319ft) is to be believed in saying thatin the economy too 'education to obedience' has considerably increased in moderntimes, then it may be expected today that it is most likely in the work sphere and theenvironment sphere that real preference changes will come about on the basis of'rational' didactic legal provisions - even in the case of 'corporate actors' withconcrete heads!7 However one must also take legally induced preference changes ofa quite different nature into account. Nippon-style managers with great readiness foradaptation and compromise can in the face of bureaucratic legal regulations suddenlytum into super-tough US litigators, or even Bavarian legal hotheads.

131

Interpretation as 'price signal'

Structural coupling of the simplest type is present where the enactment of a legalnorm is not interpreted at all as an economic structure operating in the longer termbut quite simply as economically relevant event that depresses the price or sends itup. This instrument is, as successful practice with 'talking down' the dollar has shown,already being deliberately used by political regulatory agencies. Unfortunately, theGermans had to go and talk too loud again! As already mentioned earlier, in themost recent oil crisis too the US president employed this 'psychological effect of legalregulation deliberately, but with contra-intuitive effect, as a 'price tax'. The tranquil­lizing signal of releasing the reserves was misunderstood as signalling worry, and theoil market reacted with panic price increases.

The Case of 'Non-Interpretability' of Legal Norms should not be forgotten.What we then have is an indifference of economic operations to legal norms. Wherethe legal event can be converted into a structural part only by setting the economiccode itself at naught (rather unlikely for the whole economy, but quite realistic insubsets of it), then structural coupling is not possible. The economy then practicescivil disobedience, appeals to the highest values of the economic institution and takesrefuge in black markets. The outcome of price taxes is then price increases to absorbthe greater legal risk of the black market. And of course then there will be the publicprocurator and the police! If, say, particular prohibitions on economic action areimposed by main force, then in that area the economic code has been replaced by thepower code. Economic satisfaction of needs has been replaced by political satisfactionof needs, and the advantages and drawbacks of an economy based on force can bediscovered. But is the scare resource of the power of the bayonet really what interestsus in the topic of social regulation through law?

If the humble price control gives rise to such a range of economically specificinterpretations, then Max Weber's verdict that from a historical point of view "pricetaxes" have always been precarious but today on the whole have "still less chances ofsuccess than formerly" is hardly surprising (Weber, 1925, p.197). The diverse econom­ic interpretations of law listed here could certainly be refined still further. What isimportant in our connection is that they are the theoretical statements on regulationarising out of the conceptual framework of autopoiesis theory. Developed further,they can supply specific criteria for regulatory practices and for the detailedobservation of regulatory processes. We can see from the specific grammar of thediscourses where and how they reconstruct distinctions of other discourse in their ownlanguage: in codes, structures, programs, reality constructs etc. It is at this point thatactual hypotheses must set in, presenting assumptions as to what types of legal normcan be interpreted in what type of economic interpretation under what conditions.8

But can it then still seriously be said that an autopoiesis viewpoint does not gobeyond the vague pictorial language of stimulation, perturbation or modulation ofclosed systems?

But what is so different about this language game by comparison with causalchains reaching from the political goals via juridification and implementation up tothe social effects? I shall attempt to give two brief answers and a third rather longerone.9 Firstly: In this sort of analysis of discourse we do not, like Vilhelm Aubert(1967), get lost in the tangle of causal chains preceding from legislative acts asprimary, secondary or tertiary consequences. Instead we look for typical rules in the

132

discourse involved, say their degree of openness or closure towards the legal dis­course, which may range from hermetic closure up to a fairly low threshold ofreconstruction. By contrast with individual causal analyses, this allows us generaliza­tions. Secondly: acausal parallel processing of differing discourses places recursiveness- in the systems and between the systems - at the center of analysis. Recursivenessmeans a constant micro-variation of meanings in applying operations to the outcomesof similar operations, which is simply forbidden in a causal analysis of the type ofOpp/Diekmann/Rottleuthner (Diekmann, 1980, pp.67ff; Rottleuthner, 1987, pp.56ff),which needs a clear definition of the dependent and independent variables, stablethrough time.

Thirdly: Even if sociology can show only a limited amount of stable regularitiesin parallel processing, then it should pay more attention to society itself, particularlyto stability-creating processes of social self-organization. Mere structural coupling oflegal norms to other social fields does not yet lead to any systematic coordination.But occasionally social practice itself sets up firm links between the chaoticallycoupled discourses which confer duration, acceptance and directional indications onthe transitory 'tangential responses' of structural coupling. And the fact that some(not all!) legal norms have such firm social links available is what provides anopportunity for political regulation through plural law.

3. Structural coupling

The concept of structural coupling in autopoiesis theory denotes a void between thesystems. Humberto Maturana seeks to fill it with tautologies. As a necessaryappendage to operational closure, according to Maturana, structural coupling of thesystem with its niche is present where both co-exist; and they co-exist where they arestructurally coupled. The outcome is evolution as structural drift (Maturana, 1980,pp.102ff; Maturana and Varela, 1987, pp.l13ff).

Niklas Luhmann has analyzed the phenomenon of structural coupling in moredetail. A system is structurally coupled to its environment when it uses events in theenvironment as perturbations in order to build up its own structure. Where a systemhas internally available the distinction between self-reference and hetero-reference, itmay via structural coupling make itself dependent upon its environment by usingexternal events as conditions for its own operations, as irritations or even asopportunities (Luhmann, 1989, p.8; 1990, pp.29ff, pp.163ff).

In my view two types of structural coupling will have to be distinguished accord­ing to whether the coupled systems belong to different areas of phenomena (e.g.body-soul or consciousness-communication) or whether they belong to the same areaof phenomena, as second-order-autopoietic Systems, (e.g. law economy). The secondtype of structural coupling which I call interference, is distinguished by the fact thatevery event in the functional sub-system is at the same time always communication tothe whole of society and is therefore 'linked' to the events in the coupled system in aquite specific way (Teubner, 1991b, Chapter 5). This 'all-society linkage' of inprinciple separate, autonomous system events puts us on the track of a more generaldistinction, that between coupling and linkage.

While structural coupling denotes the mechanism of actual intersystem contact,namely use of perturbation of one system to build the structure of the other, linkage

133

denotes the set of conditions necessary to make structural couplings possible.Without linkage, structural coupling would be confined to the extreme case of singlechance contacts in which a single event acts as a perturbation and affects a singlestructural formation. Linkages are responsible for the fact that structural couplingscan take on different values of duration, intensity, quality and institutionalization.The very material, organic and mental infrastructure of society ensures that function­ally specialized communications are to some extent linked with each other, as does, aswe said, the fact of society itself. But the really interesting 'close' links are createdonly with the social institutions specialized for that. We need not think immediatelyof the institution of marriage; instead, the most important glue in functionallydifferentiated society is probably the formal organizations, to the extent that they are'multilingual' (thus Mayntz, 1987, pp.l00ff; Scharpf, 1989, p.15). The really effectiveaspect of this mediation is not even the fact that the multilingual organizations'understand' various sub-systems, but the mere fact that they link them to each other.They thereby force the functional sub-systems each to process their own informationon the same events in parallel over the long term. Interest associations, for instance,are communicatively involved in two or more functional sub-systems. They ensure thatfunctionally specified communications of politics, the economy and law no longermeet each other merely randomly, transitorily and tangentially, but are systematicallysynchronized (Teubner, 1978, pp.14lff). Formal-organizational links systematicallychange the duration, quality and intensity of structural couplings. 1O With an ade­quate range of such intersystem organizations, their micro-synchronizations ultimatelylead to a position where the functional subsystems tending to drift apart are broughton to a common path of development, even though one cannot speak of 'intentionalregulation' in the sense of Mayntz and Scharpf.

This is interesting enough by itself. But it becomes really exciting where it ispossible to identify intersystem links that squeeze structural coupling into a directionsuch that systems act on each other catalytically in cyclical fashion. We then have theinteresting case that processes of self-reproduction would, without the systemsinvolved losing their autopoietic closure, operate outside the boundaries of autopo­ietic systems. This would be the case of ecological (not: systemic!) recursiveness. Theautopoiesis of the systems involved is not impaired; instead, it is being exploited tobuild up ecological cycles respecting system boundaries, even though crossing theseboundaries.

4. Ultracyclical linkage

But this is not just the sick spawn of a brain suffering from epidemic systemosis(diagnosis by Ballweg, 1972, p.580), as we learn once again from the British colonialmasters. Their manipulation of 'indigenous law' presents us paradigmatically with thepicture of how today so-called 'new legal pluralism' - the pluralist law of transactions,organizations and networks - is day-in and day-out linking the structural couplings ofeconomy and the law.

The normal case of non-recursive coupling of economy and law is for economicactors to bring their conflicts before the court. The legal discourse allows itself to beperturbed by the economic discourse, reconstructs the conflict as a legal case inspecial legal language, checks it against existing sets of norms produced elsewhere,

134

where necessary invents a new norm and processes it up to a court decision which ­see above - is read by economic autopoiesis in the most divergent interpretations, oreven not read at all. This is the usual, the merely transitory, structural coupling ofeconomy and law.

Something different happens in the modem contract law. Here jurists work, asonce British colonial rulers did, with a bluff. The law, by providing the economy withthe favorable opportunity of 'nuda pacta', namely of concluding absolutely formlesscontracts of any content whatever, provides itself with the favorable opportunity oftreating more or less any economic transaction as a source of law, as the conclusionof a contract creating legal norms, even where the actions were meant quitedifferently. The law 'observes', 'construes', the economic process that has actuallytaken place once again, in its own language. It thereby creates the 'fiction' that theeconomic process itself produces its legal norms. The law paints the quality of being alegal norm onto economic transactions by asserting that there is such a thing as 'self­created law of the economy' (GroBmann-Doerth). And then the law uses this actualfiction as the third source of law alongside statute and case law.

Really this self-created law of the economy exists just as little as did the indige­nous law of the British colonies. It is a mere construct of its inventors. Since Macau­lay's pathbreaking studies, empirical legal sociology has repeatedly demonstrated thateconomic transactions take place free of law, that deals are struck, deliveries carriedout, payments made, and long-term business relationships carried on, without legalnorms being evoked or even merely insinuated (Macaulay, 1963; Daintith, 1986,pp.185ff). Transactions are made on the basis of actual chances for action in respectof changing them and projected into the future on a basis of mutual trust. But out ofthe actual chances for action and the economic trust the law cobbles the factors thatsuit it into a collage that it then presents as legal property and contractual dutiescreated by the economy itself. This reading by law is based on a single big misunder­standing - a creative misunderstanding. Its creativity lies not in the pure invention ofsomething new out of the blue but in the productive utilization of social material.Making a variation on Polanyi's famous distinction, the legal misunderstanding iscreative to the extent that it builds up its explicit legal knowledge on the basis ofimplicit social knowledge (Polanyi, 1958, pp.69 ff). For all the misgivings, misunder­standing still remains a kind of understanding! What does Portuguese legal sociologistBoaventura de Sousa Santos, who wishes to set up a post-modern theory of legalpluralism, have to say? "Law. A Map of Misreading" (Santos, 1987). The law systemat­ically misunderstands economic transactions as sources of law, as legal acts that createlegal norms. Be it noted that this is not just as legal acts the law judges using normsproduced elsewhere, but as legal acts that themselves produce norms. Napoleon knewwhat he was saying: "Les conventions legalement formees tiennent lieu de loi a cequi les ont faites" (Art. 1134 Code Civil). Through this real fiction, the law is todaycreating in the institution of the modem contract, as the British once did throughindigenous law, a new, richly yielding source of law that in creativity and dynamicsgoes far beyond the other classical sources of law, statute and case law.

Taking the link between economy and law the other way round, where economyreads legal constructions (new legal forms of contract or newly created property rightsin intellectual products, computer programs or genetic techniques) as favorableopportunities to open up new markets and exponentially increase economic transac­tions, we can see the mutual exploitation. The circle is complete: the economy

135

nonchalantly makes profit from legal norms created for quite different purposes, andthe law, undisturbed, creams off the normative surplus-value of money circulation.

In somewhat different language, this is not just mutual reconstruction of systemoperations but in the strict sense a mutual catalysis of growth processes (Eigen andSchuster, 1979; Eigen, 1987, pp.225ff; Zeleny, 1981, pp.lOOff). Transactions operateauto-catalytically on the reproduction of transactions and at the same time hetero­catalytically on the reproduction of legal acts that create norms. Conversely, legal actsoperate auto-catalytically on the reproduction of legal acts and hetero-catalytically onthe reproduction of economic transactions. We may here speak of a linkage through'ultra-cyclical' processes, more or less in the sense described, continuing ManfredEigen's ideas, by Ballmer and Weizsacker (1974). The socio-Iegal generalization andrespecification of ultra-cyclical relations can in my view be formulated as follows:1. Self-reproductive processes in law and the economy join together cyclically

through the institutions of contract and property, into a self-reproductiveprocess.

2. The ultra-cycle operates to accelerate growth as the economy produces transac­tions which at the same time act auto-catalytically for their own growth andhetero-catalytically for the growth of the production of legal norms, andconversely.

3. However, this ultra-cyclical linkage of law and the economy does not lead tothe emergence of a new operatively closed system, but is based on the separa­tion and autonomy of the systems involved. Indeed it exploits just that funda­mental difference of the system operations (transactions and legal acts) forheterocatalysis. Accordingly there is no economico-legal hypercycle, but anultracycle that crosses the limits of law and economy, a circular relationship ofreinforcement between system and niche: as we already said, a kind of ecologi­cal recursiveness.

5. Politics

The evolutionary advantages of a cyclical linkage of law with the economy throughthe institution of the contract are obvious. Natural scientists, were they to look overthe fence, would perhaps recommend a look at the acceleration in rate of growth ofthe communication populations in the economy and the law (see Eigen, 1987,pp.227ff). Social scientists at the same time see other evolutionary advantages of theultracycle. It makes the law gain in responsiveness to the economy. As it wereautomatically, without need for a training in economic analysis of law, law adapts tothe intrinsic logic of economic processes through incorporating contractual norms(Luhmann, 1989). With contract law, the law has further available highly effective'self-executing norms', since it develops directly out of economic motivations. But thesecret trick in the contractual mechanism is a kind of self-binding of the economy.While the economic discourse initially disposes freely as to the content of contractualnorms, it then loses control and gets tied up in the self-set linkages of law. For thepower of definition is now transferred to the legal discourse which, in the interest ofinternal legal consistency, disposes autocratically over the contractual norms, definingwhat the actual will of the parties has to be, manipulating contractual contentarbitrarily through 'implied conditions' in contractual interpretation, declaring

136

contractual norms invalid in accordance with 'public policy' and setting completelynew contractual norms under the general clause of 'good faith'. And as long as thelegal discourse does not overstrain structural coupling, the economic discoursewillingly follows the legal corrections to the self-created law of the economy.

Now strikes the hour of politics! If it has hitherto had to admit that directpolitical regulation of the economy tended to be like the 'journey into the unknown'romanticized about by Josef Esser, never knowing where one might ultimately land, itnow suddenly finds in the self-created law of the economy a handy controlmechanism, easily manipulable and, to the delight of the regulations, also actuallyable to bite. For structural coupling institutionalized and made durable through thecontractual links continue to function even once the contractual norms are effectivelychanged with regulatory intentions. And the difficult conversion of political decisionsinto economic transactions now becomes mere self-regulation of law. The policy­oriented law deliberately regulates some norms of contract law. It is just as simple asthat. One just has to watch out that the tie between the structural coupling of theeconomy and the law does not break as happened - as private lawyers know to theircost - in the case of, for instance, the legal policy control of financed purchasingcontracts (more generally on the efficacy of legal policy in contractual law, see Raiser,1990).

But things ought not to stop at this parasitical exploitation of the symbiosis ofeconomy and the law for purposes of political regulation. For the mechanism ofpolicy-oriented legal self-regulation plus twofold structural coupling can begeneralized. In the new legal pluralism, politics do not exploit only the lex mercatoria,the self-created law of capital, but cream off, for regulatory purposes, the normativesurplus-value of quite different processes of social reproduction.

At the center are the decisions of formal organizations. In the internal law ofassociations and organizations, the law is bound to the self-reproductive processes notonly of the economy, but of quite different social sub-systems like health, the massmedia, religion or culture. Here it is the entry/exit mechanism that is fundamental forthe cyclical linkage (Teubner, 1988b, pp.66ft), which the law in tum creativelymisunderstands as a law-producing mechanism. Organizational structures in economicenterprises, interest associations, trade unions, press organizations, hospitals, andcultural organizations are through this legal fiction made into a rich source of 'sociallaw' that can be taken up, disciplined and controlled by official law. The legaldiscourse here ventures even on extremely far-reaching norm formation - just think ofthe law of the Moon sects - that no legislator or judge would ever venture toformulate. And politics finds a new fertile field for regulation that allows it togovern right into closed societies. For once again it is just changing legal norms bylegal norms which as it were 'by themselves' become converted into social praxis.This also confirms and at the same time extends the well-established thesis of the'legal affinity' of formal organizations (Selznick, 1969: pp.32ff; Mayntz, 1987: p.103;Kaufmann, 1988, pp.82ff; Scharpf, 1987, pp.117ff; 1989, p.16; Edelman, 1990:pp.1406ff-1435ft). This affinity is not based only on the structural similarity of legalnorms and formal organizational structures, but additionally on the close cyclicallinkage between law and organization. The law takes responsibility for this because itstylizes the similarity into identity by creatively misunderstanding processes within theorganization as production of genuine law and thereby coupling them closely to eachother.

137

The dramatic extension of 'due process' in US private organizations since the sixtiessupplies illustrative material of the new introduction of this sort of 'plural' productionof law (Selznick, 1969, pp.183ff; Foulkes, 1980; Westin and Feliu 1988; Edelman,1990). Without statutory provisions here having prescribed the introduction ofprocedures under the rule of law, a self-accelerating process got under way in whichcourts misunderstood previously 'law-free' internal organizational decisions as legaldecisions and the organizations in tum formalized the grievance procedures andcontinually extended them. The end of the development is a private machinery fornorm production supported by the official law, representing a new 'source of law',comparable with the contractual mechanism. In the ultracycle of the mutualmisreading of law and organization, the law gained a new source of law and theorganization a new source of legitimation.

6. Legal pluralism

Modem legal pluralism has long outgrown its onglOs, contract and organization.General terms of business, agreements by the umbrella organizations of enterprisesand banks and other private arrangements in organized markets, the system ofcollective bargaining and other forms of collective law in industrial relations all makethe classical individual contract look pretty old. The classical social mechanisms ofnorm formation, contract and organization, are far from being exhaustive today; theyare superseded today by an elaborate private inter-organization law and intercontrac­tual law. Indeed, one has to speak of a new sort of 'network' of these legal discoursesin a societal, non-statal discourse, a 'legal dogmatics of private Kautelasjuris­prudenz,.1l By contrast with classical private law, we are here facing a new type oflinkage of social and legal episodes. 12 A whole network of collective actors operatesas 'social legislator': firms, associations, trade unions, chambers of trade, antitrustboards, quangos, municipalities, lawyers' offices, institutes of commercial law. Theformation of law takes place not in a vast number of individual contracts but in'private ordering' through collective negotiations and strategic communications, inbrief: through power relationships in organized markets (see Bercusson, 1987,pp.50ff). And it is not only the parliamentary legislatures that are involved a posterioriin the political regulation of this semi-autonomous law, but, very early on in theprocess of its formation, the regulatory agencies, quasi-governmental associations,anti-trust board and other authorities, and in particular courts. State law is embed­ding itself ever deeper into these social non-juridical forms of conflict settlement(Auerbach, 1983; Arthurs, 1985; also Nader, 1984; Harrington, 1985). With suchhappy coinages as 'practice as procedure for disclosure' (Joerges, 1981: pp.132ff) or'private justice' as legal counterpart to 'private government' (Henry, 1983; 1987;Macaulay, 1986), but also 'alongside the State' (Ronge, 1980), jurists and sociologistsare engaged in illuminating various aspects of the modem social proto-law. Howeverwe are still a long way away from a systematic, empirical, theoretical or doctrinalscrutiny of this self-created law of society.

Yet despite all the research deficits, the outlines of the new legal pluralism canbe discerned. The differences from Eugen Ehrlich's 'living law' (1913) can beprovisionally formulated as follows:

138

1. The new legal pluralism does not focus primarily on the local law of ethniccommunities as the old legal pluralism did. The focus is rather on the proto-lawof specialized organizational and functional systems. The new living lawtherefore lives not from stores of tradition but from the ongoing self-repro­duction of highly specialized, often formally organized systems of an economic,academic or technical nature.

2. The new social law is not formed autonomously by social processes butconstituted first through the legal discourse itself, which reconstructs socialprocesses as legal norm production. In this creative misunderstanding lies thelegal system's own achievement in producing 'social law'.

3. It is not the 'acceptance' of social law by those concerned that is the basis for itssocial implementation but its specific interwovenness with its social environment(cyclical linkage of structural coupling). It guarantees that social operations canbe continued under the dominance of the legally reconstructed social law.

4. It is not the contrast with state law that characterizes the new social law, but itsinstrumentalization for purposes of political regulation, which even goes so farthat politics in tum initiates artificial procedures of social norm production.

7. Conclusions

Today we see a whole wave of institutional experiments with pluralist law formationcoming upon us. Under the suggestive power of unburdening the state, decentraliza­tion, closeness to society and self-organization, everywhere ethical commissions, roundtables, micro-corporatist bodies and negotiating rounds supported by mediators arebeing set up. In Community Europe the 'new approach' to legal harmonization iscreating a furor: whole policy areas are being transferred to a complex procedure ofprivate standard setting where, under the umbrella control of European Communitylaw, substantive norm production is negotiated in pluralistically composed bodies ofprivate associations (Bruha, 1986).

Can the British colonial masters teach us a lesson here too? After all, theparallel with their creative misunderstanding of indigenous law is all too clear. Butthere is something new about the 'artificial' pluralist bodies. As against thesubsequent political and legal exploitation of spontaneous quasi-norming, social normproduction itself is being taken under political control. But this is just what the Britishskeptics would have regarded with suspicion. For the success of their soft law did notlie in the aspects of pluralist law production that the German debate is so fond ofbringing out: democratic legitimacy through representative social relevant groups, thecozy warmth of decentralized round tables, the advantage of inter-system discoursesof mutually closed conceptual worlds. With their sense of reality and aversion to allfar-fetched theory, the British give us one message above all: Pay attention to theultra-cyclical links in the structural couplings!

What this lasting legacy of European colonialism means for us today is some­thing we can merely vaguely speculate about. Its clarification will have to be left tolater empirical research. I think that in the institutional design of pluralist normingbodies we should take to heart the following warnings from British colonial teaching:

Create de facto linkages! There is little sense in setting up pluralist bodies wheretheir decision-making processes are not at the same time closely 'linked' in the sense

139

described above to the real elementary operations of functional sub-systems. In orderfor their misunderstandings to be truly constructive, care must be taken to ensure thatthey do not invent their norms freely but in fact 'read' their standards off fromeconomic transactions, organizational acts, technical processes and research resultsthat really arise. The test question is: Which real social processes are normatively'read' in the artificially institutionalized pluralist law production - constructivelymisunderstood - the results of which can in tum be 'read' by the real social processes?The famous ethical commissions of our day do not look too good from this viewpoint.Nor do court psychiatric procedures, since the 'woe of psychiatry' lies just in the factthat the trans-scientific questions of jurists to psychiatrists and their answers can nolonger be linked up with cognitive acts of psychiatry as an academic discipline (seePrins, 1980, chapter 2). The standard-setting bodies of DIN and other safety standardscommittees can be given a rather better prognosis from this viewpoint. Howeverproblematic their social representativity may be, at any rate through their attempts atstandardization they stick closely to technical and economic processes that areactually taking place.

Pay attention to the limits of structural coupling! The logic of constructive misun­derstandings runs as if 'automatically', but only within definite limits. Pluralist normproduction is entirely controllable politically, but only to the extent that its results aretaken up in real social processes without a State implementing apparatus having tocontrol their application in the individual case. The test question is: Are the politicalmanipulations of pluralist law still moving within the motivational leeway of the socialprocesses concerned? Here lies the deeper reason why substantive regulation ofpluralist law by politics is practicable only within very narrow limits, while proceduralregulation, the political redefinition of control rights, property positions, participationrights, decision-making procedures and rules of evidence have much better chances ofbeing taken up. Collective labor law offers a paradigmatic example of very far­reaching political control of social processes of norm production through proceduralmeans. Where the limits to the motivation for take-up lie can ultimately be deter­mined only in practice. But legal sociology and legal economics should be useful forthis question, insofar as they are able to make generalizations on the readiness foracceptance in a specific social context.

Pay attention to institutional separation! The historically successful examples ofpolitically manipulated social law (contract, organization, bargaining system) at anyrate show that the pluralist norm production regularly takes place in two stages. Thelaw's constructive misunderstanding of social processes is the first stage, followed bythe political manipulation of its product as the second stage. By contrast the morerecent experiments with pluralist norming bodies are frequently distinguished by thefact that the first stage is already 'politicized'. The political interest representationand the regulatory intentions are already nested inside the social process of normformation. For comprehensive 'clearing of interests' this may be advantageous. It mayhowever also be the case that this sort of compact procedure brings together the twostructural couplings (law with social sub-system and law with politics) involved tooquickly, thereby overloading both. The test question is whether in such a situation thecyclical linkages, which set up the coupling lastingly, mutually interfere, or evencannot be formed at all. The escape would be, in the case of artificially createdpluralist procedures too, to pay heed to institutional separation, by separating eitherin time or organization the social norm formation and its political control. The 'new

140

approach' in European safety law seems to have intuitively grasped this range ofissues by providing for procedural separation of safety standardization in 'privatestandardization bodies' and its political and administrative control by national andEuropean authorities (Bruha, 1986).

In any case we should pay attention to historically tested conditions of success inadvocating extension of pluralist institutions for creating law, and not simply letourselves be carried away by the romanticism of decentralized social dialogue or thehermeneutics of intersystem discourses. This is meant self-critically too (Teubner,1987: 33ft) and is at the same time addressed to more recent pleas for 'reflexive'forms of control (Kirsch, 1988; Rosanvallon, 1988; Gffe, 1989, pp.16ft). Not that Ihave anything against the new 'constitutional patriotism', the patriotism of democrati­cally constituted social micro-systems! And I am all in favor of local discourse with alearning capacity! But if we do not take sober account of their rootedness in realsocial processes - of the ultra-cyclical linkages of structural coupling - then we are alltoo likely to let our pluralist micro-bodies decay into the 'talking shops' thatdemocratically conceived institutions have already been denounced as in our century.

Notes

I. Luhmann, 1988, pp.329ff; see also the cautiously sceptical assessments by Offe, 1989, pp.9ff andthe brutally negative judgments of Posner, 1987,p.761,pp.769ff.

2. Financial Times 22.8,23.8,25.8,28.9,29.9.1990;FAZ 24.8,26.9,28.9,29.9.1990.

3. FAZ 28 September: "Resource to America's strategic reserves dampens oil price"; FAZ 29September: "Bush'ssignal fails to bring down oil price".

4. "The ministry's control over the industry is notorious, even down to the minutest details", FinancialTimes, 23.8.1990,p.3.

5. On bargaining in antitrust law see Gotthold and Vieth, 1982.

6. On this see the empirical fmdings in Budde, Child, Francis and Kieser, 1982.

7. For one impressive empirical study on legally induced learning by corporate actors in the field ofcorporate governance, see Edelman, 1990.

8. A similar attempt to specify the problems of regulation is made by Kaufmann, 1988, pp.85ff. Thechances of control through law are held to be specifiable only with the assistance of the 'context ofdiffering forms of social State intervention'.

9. For more on the two brief ones see Teubner, 1991c.

10. For a theoretical interpretation of relevant empirical research see Hutter, 1989, pp.9Off, pp.127ff.

II. For a first systematic presentation see Rehbinder, 1982, pp.llff.

12. On linkage of episodes see Teubner, 1988a, pp.432ff.

141

Chapter 12

ENVIRONMENfALMEDIATION: AN EXAMPLE OF

APPLIED AUfOPOIESIS?

I.J. Koppen

1. Introduction

The paradigm of environmental regulation has changed. From its inception in the latesixties until the beginning of the eighties, environmental policies were anchored incommand-and-control regulation. The legalistic prerogative of unilaterally imposedrules dominated governmental steering in this area. A vast number of legal rules wereadopted without questioning of their effectiveness in terms of pollution abatement.When government agencies, however, started to monitor the implementation ofenvironmental programs and when the first policy effectiveness studies came out, thedifference between law-in-books and law-in-action became apparent, and theshortcomings of the approach were recognized. It is not surprising therefore thatduring the eighties, environmental policy was characterized by a search for alternativemodes of intervention. Two considerations have played a major role in reflections onenvironmental re-regulation: finding ways to make environmental policy instrumentscompatible with market forces and assuring social consensus about conflict situationsprior to the legal implementation of policy decisions. In practice this has meant anincreased use of economic incentives and the institutionalization of consultation andconsensus building mechanisms in decision making processes.

The theory of autopoiesis adds an interesting perspective to the analysis of thesedevelopments. First of all, the theory helps explain why state intervention isproblematic and often unsuccessful: the cognitive-analytic function of autopoiesis.Secondly, the theory indicates ways to improve the steering process: the normative­prescriptive function. The argumentation below draws upon both functions. Recenttrends in environmental regulation are, in other words, put to 'an autopoietic test' inorder to find out if they get 'the autopoietic seal of approval. '

The focus of the analysis is on one of the new instruments of environmentalpolicy that can be regarded as indicative of the trend towards consensual steering. Itis an instrument that was developed in the United States to resolve long-lastingenvironmental disputes and to assist government agencies in reaching agreementabout controversial environmental policy goals. Environmental mediation is thegeneral term for different intervention mechanisms that rely on the role of a neutral

143

intermediary to resolve conflict situations through a process of assisted negotiations.Representatives of all the groups that have an interest in the outcome of a particularconflict, generally referred to as stakeholder groups, are asked to participate in thenegotiations. The principal features of mediation that make an autopoietic analysisinteresting are the specific function performed by the mediator and the process ofinJeractive bargaining in which the parties engage.

This chapter is organized in four parts. In section 2, a number of preliminarypoints are raised that clarify some of the underlying assumptions of the argumentpresented here. The first deals with the question what kind of theory autopoiesis is orwants to be and briefly reviews the two functions of the theory: is it legitimate to useautopoiesis for the evaluation of public policy or is its scope restricted to thecognitive-analytic function? The second is on the state-of-the-art of the policy scienceswith respect to environmental regulation; what is the exact problem that we aresubmitting to autopoietic analysis? The third is about the linguistic complexity of thetheory, which might be an obstacle to its practical application. Having brieflycontemplated these issues, section 3 then gives an autopoietic account ofenvironmental regulation. The social subsystems that play a role in the regulation ofenvironmental problems are described and the consequences of autopoietic closurefor regulatory intervention are assessed.

The normative-prescriptive analysis of environmental regulation developed insection 4 shows that consultation and consensus building mechanisms can be seen asways to interfere with autopoietic closure. Finally, in section 5, environmentalmediation is analyzed and evaluated against the backdrop of autopoietic principles.At the surface we find an incompatibility between the analytic structures ofautopoiesis and mediation; the social subsystems of autopoiesis do not match thestakeholder groups of mediation. Two lines of reasoning are proposed in order tomake the units of analysis compatible. The question, however, whether environmentalmediation can be seen as a 'multi-dimensional super-discourse' remains unanswered.

2. Applying autopoiesis

The theory of autopoiesis has only recently been applied to issues of law and policy.The elaboration of concrete examples was impeded by the complexity of the theory.Since this chapter attempts to develop such an example, a few reservations must bemade at the outset.

First of all, we must note that the theoretical nature of autopoiesis is somewhatambiguous. Autopoiesis can be perceived as a theory of cognition, seekingfundamental answers to problems of knowledge. In this view the autopoietic closurebetween social subsystems is an element of our perception of social interaction thatcan neither be challenged nor altered (see Luhmann 1986a, pp.185ff). Some authorsrefer to it as the unyielding severity of autopoiesis, "die unbiegsame Harte" (Teubner1989a, p.95). Similarly, the operative closure of the legal system as a social subsystemis often analyzed as absolute: the system has no output (Luhmann 1986a, p.174).When God laughs in the Hebrew parable cited by Teubner (Teubner 1989b, ppA01,408), it is because man has not understood the ineluctable self-reference of the law,which is a real problem of social systems, not an error in the mental reconstruction ofsocial reality (Teubner 1989b, pA07): "reality has a circular structure, independently

144

of its cognition" (Luhmann 1986b, as cited in Teubner 1989b, pA07). Law has noaccess to reality, and outside facts, it produces legal constructions of reality, validatedonly by internal legal norms: an epistemic trap (Teubner 1989c, pp.742-746). Thisjustifies the interpretation of autopoiesis as a cognitive-analytic theory. The closedfunctioning and the self-referentiality of social subsystems are the parameters thatdefine the theory rather than its variable components. Efforts to interfere with theseparameters either become themselves an autopoietically closed discourse or falsify thetheory.

Recently, however, some of the same authors have used autopoietic analysis tounderstand common failures of State intervention and to design strategies to resolvethem (Teubner, 1990). Autopoietic closure is indeed regarded as a variable, aproblem for which the theory can formulate solutions. In the context of steering,policy and regulation, autopoiesis assumes a double role, a "normativ-analytischeDoppelstatus" (Stout and Stoop, 1991, p.5): it analyzes and prescribes. Thisapplication of the theory is still experimental and not without criticism. Indeed sometake no note of the theory because of its complicated name. For the moment we willtake autopoiesis as a theory in transition, a theory that is changing and growing, witha prescriptive side that needs to materialize; a theory with the potential to contributeto the treatment of the social ills of our time, environmental problems being one ofthem.

When we apply autopoiesis as a normative-prescriptive theory to issues ofenvironmental regulation, we must formulate the kinds of problems that need to besolved as clearly as possible. It is not enough to assume 'a regulatory crisis.' After all,the question whether or not a regulatory crisis exist is answered differently bydifferent authors (Scharpf, 1988; Rottleuthner, 1989; Teubner, 1984, p.381). In 't Veldspeaks of the decried State, "de verguisde Staat", emphasizing the adaptability ofdifferent steering-systems (In 't Veld, 1989). Some environmental programs haveindeed been quite successful (Bressers, 1983, for water quality policy in theNetherlands). Without embarking upon that discussion, I propose to view steering asa learning process: regulation fails only where we do not learn from the problems thatarise.

The problems that have come up in environmental regulation relate to the highdegree of complexity and the new parameters introduced by this policy field. Mostenvironmental policy decisions involve difficult trade-affs between different sociallydesirable goals; the conflict of affected interests is one aspect of the complexity. Thesecond aspect is the uncertainty about the underlying scientific data. Environmentaldecisions are based on probability assessments of scientific cause-effect relationships.Science no longer offers certainty, not even about the management of the uncertainty(Ravetz, 1990). Adjustments in the steering process must especially be directedtowards these two aspects of environmental decision making and make account forthe special regulatory requirements they pose.

The final point I want to raise here about applying autopoiesis, concerns the attimes inaccessible language of the autopoietic discourse. A practical application ofautopoiesis challenges this very aspect of the theory. Translating its conceptualcomplexity into practical terms might imply negating part of the theory. For, asTeubner himself states, autopoiesis constructs a socio-legal reality that cannot beadequately expressed by ordinary language without compromising the content of itsmessage (Teubner 1989c, p.728). He even questions "whether the language is complex

145

enough to match the complexity of the subject matter" (ibid). A practical example ofautopoiesis might be a contradiction in terms, located in the middle of two extremepositions. On the one hand there are the critics, who maintain that autopoiesis is'Jabberwocky' (Teubner 1989c, p.727), while on the other hand its proponents holdthat the linguistic complexity is the theory's strength. The real problem is maybe thatautopoiesis itself functions as an operatively closed system and is not interested in itsapplicability. Both refutations and confirmations reproduce the theory within its ownacademic boundaries and the 'world out there' remains unchanged. We must beaware of the risk that autopoietic explanations of obstacles to social interaction maythemselves turn into obstacles to interaction.

3. An autopoietic account of environmental regulation

The discussion about environmental re-regulation in recent years has centred aroundthe difference between the command-and-control approach, based on coercivemeasures, and the incentive approach, based on consultative measures. Differentarguments have been presented to show that the command-and-control approach isnot suited to environmental problems. The alternatives that have been developedinclude economic incentives and a consensual rather than an adversarial approach topolicy formulation and conflict resolution. The legal rules that are formulated areflexible and focus on procedure rather than on substantive norms.

If we look at this discussion in the light of the two principal features ofenvironmental regulation indicated above, we get the following picture. Difficultiessurrounding interest representation and scientific uncertainty put a strain on steeringprocesses that use coercive unilaterally imposed rules. Traditional decision makingprocedures cannot accommodate the regulatory requirements that follow from thesenew parameters. Environmental interests end up being under-represented while thecomplex scientific questions involved cause long delays and dangerous mistakes indecision making. Shortcomings in legal rules that do not accommodate the specificchallenges posed by environmental issues, result in: increasing numbers ofadministrative court cases; non-legal adjustments of the rules in the implementationand enforcement phase of the regulatory process; illegal behavior and lack ofcompliance (Hucke, 1978; Bardach and Kagan, 1982; Richardson, 1983; Susskind andCruikshank, 1987).

From an autopoietic point of view it is the operative closure between thedifferent functional social subsystems involved in environmental regulation thatrenders the steering of environmental parameters problematic. Four subsystems areinvolved: politics, economy, science and law. Each system is autopoietically closed andself-referential. It can receive information from the outside, but it can only react tothat information by reproducing it in its own internal code. The information thatscientists pass on to lawyers, economists and politicians about the state of theenvironment, enters into their respective subsystems. It causes 'noise': initiatives aredeveloped, activities of various kinds take place. Typically, lawyers develop legaldoctrine, start a court case or propose new rules; economists want to find out howmuch protective measures cost; politicians appoint expert committees and put an issueon the political agenda. But the information then stays there, trapped as it were inthe subsystem. It reproduces new forms of information that are compatible with the

146

codes of the subsystem, but there is no output, no feed back to the original source ofinformation, or very little at most. And, after many years, neither law, nor politics, noreconomics has achieved considerable ecological improvements. Because of theautopoietic closure between the subsystems the effectiveness of environmentalregulation remains necessarily limited. This is one way of explaining the lack of logicin the responses of different functional subsystems to steering activities intended toimprove the environmental condition.

4. Breaking the autopoietic impasse

Greater involvement and participation of all affected parties - the stakeholders - in adecision making process that allows for more flexible and easily adaptable decisions isone way to deal with the specific requirements of environmental regulation.Consultation and negotiation can be seen as mechanisms to enhance communicationbetween different social subsystems, the involvement of representatives of eachsubsystem in a continuous interactive communication being a first step to treat someaspects of autopoietic closure. The question remains what kind of interactivecommunication would be able to play this crucial role.

In the theory of negotiation two different styles of negotiation are usuallyidentified. The first approach looks at negotiation as a zero-sum game. Parties enterthe negotiation assuming they have opposing and mutually exclusive interests. Theattitude towards the other party is typically adversarial and hostile: whatever I gain,you loose and vice versa. Parties are enemies. This attitude was common amongenvironmental groups and industry in the seventies and eighties.

The second approach assumes that negotiation will include some possibilities formutual gain. Parties might be able to trade issues that are of different value to them:I gain more than you loose and vice versa. In this way the negotiation becomes apositive-sum game. The total value of satisfied interests increases in the process ofnegotiating. A win-loose perception is changed into a win-win situation (See Lewickiand Utterer, 1985, Chapters 4 and 5; Fisher and Ury, 1981, Part II).

Over the past few years there has been an increasing tendency to regardenvironmental protection as a common interest. But, in order to achieve the secondapproach, parties have to be educated, both about negotiation techniques and aboutthe substance of the negotiation. As long as parties are not aware of all the differentissues that play a role in the negotiation it remains difficult to identify all thepossibilities for mutual gain. Parties in negotiations are inhibited from exploringmutually advantageous solutions because of a false perception of their own interests.Rigid political positions are often responsible for this attitude.

Environmental mediation takes the win-win approach to negotiation, alsoreferred to as mutual gains approach or integrative bargaining, as a starting point.Mediation was developed in the US as a steering tool used in situations of enduringenvironmental conflict (Bingham, 1985). By offering an alternative to traditionalmethods of conflict resolution, mediation was able to achieve better outcomes faster.Some environmental policy conflicts were resolved through mediation after they hadbeen stuck in fruitless litigation for many years (Susskind and Cruikshank, 1987).Mediation has also been applied to processes of policy formulation and theestablishment of policy goals, the so-called negotiated rulemaking procedures (Harter,

147

1982; Susskind and McMahon, 1985).The essence of mediation lies in the role of the mediator. The mediator assists

the parties in a legal dispute or the stakeholder groups in a policy debate to reachconsensus about the optimal joint solution to the problem that keeps them divided.The mediator guides the negotiations. Unlike a judge or an arbitrator, the mediatordoes not impose a solution. The mediator helps parties decide, both about proceduraland about substantive issues (Susskind, 1987, Chapter 5; Raiffa, 1982, Chapters 15and 21; Bacow and Wheeler, Chapters 8 and 9). It is important that all partiesperceive the mediator as an outsider and neutral to the issues at stake.

The role of the mediator includes educating the parties involved in anenvironmental conflict or policy debate about the win-win approach to negotiation.Parties are assisted in formulating their interests beyond rigid positions that aretypically taken to provoke the opponent. "I am against what you are proposing" is notan interest one has, but a position one takes. The interest might be "I do not want theair quality to deteriorate as a result of the factory that is being proposed" or "I wantto protect myself against the risk of a serious accident in the chemical plant you wantto build". When interests are formulated in such a manner the likelihood thatdifferent interests might be reconciled increases dramatically.

Guided negotiations are also used in situations where scientific evidence isuncertain and parties decide together which evidence they will use as the basis fortheir decision making (Ozawa and Susskind, 1985). Also in situations where sciencehas no answer, the so-called 'trans-scientific' questions (Majone 1989, p.3), the processof establishing preliminary facts in a joint discourse benefits from the presence of amediator.

Another important aspect of the mediators role is to make sure that all groupsand individuals that have an interest in the outcome of the conflict are included inthe negotiation process. If one interest is not represented, there is a chance that theoutcome of the negotiation will be unstable since the group that did not participatemight try to block its implementation. Thus the conclusion is justified that mediationaddresses some of the most pressing problems of environmental steering.

5. Environmental mediation as a 'multidimensional super-discourse'

The question we ultimately want to answer is whether environmental mediation solvesthe problem of autopoietic closure with respect to the steering of environmentalproblems. In order to make such an assessment we have to deal, above all, with theobvious differences between the autopoietic analysis and the mediation approach.

Environmental mediation is based on an interest analysis of the parties involvedin a conflict over the allocation of environmental costs and benefits. Its contributionto environmental conflict resolution lies in the application of integrative bargainingtechniques and in assisting the contesting parties to move their attention frompositions to interests. The mediator opens up communication channels, and makespeople talk to one another. The mediator helps parties to better formulate their owninterests and fine-tunes the participants to one another's interests; a mediator can beviewed as a 'policy therapist'. Mediation identifies the following stakeholder-groups,typically involved in environmental conflicts: government; industry; environmentalgroups.

148

From an autopoietic point of view the interest analysis becomes secondary to asystems analysis of the conflict situation. Instead of looking at the parties in thenegotiation as representatives of different interests in the conflict, the negotiation isseen as so-called interference between the different social subsystems involved.Teubner calls it a reformulation of the "balancing of interests", which must no longerbe interpreted with reference to individual or collective actors, but instead "cast interms of a conflict between communicative networks" (feubner, 1989b, pA18). It isnot the interests of the stakeholders that determines the interaction in thenegotiation, but the interest of the social subsystem that they belong to. The interestof a social subsystem is expressed in its internal coding: if I talk law and you talkpolitics we have a different interest even if we are both green.

Mediation submits that a broader definition of the interests of the stakeholdergroups in an environmental conflict enables parties to achieve a greater joint degreeof goal achievement. The assistance of a neutral intermediary, trained in integrativebargaining, is viewed as crucial to this consensus building process. Autopoiesis on theother hand, submits that instead of focusing on the reconciliation of interests,attention must be focussed on differences in the system-specific codes of the socialsubsystems involved. The codes must be deciphered and translated as it were, into acode that is comprehensible for the other systems.

In both analyses communication plays a central role. In mediation it is thecommunication between parties that have a different interest in environmentalconflicts, divided into representatives of government, industry and environmentgroups. Autopoiesis specifies law, politics, economics and science as the socialsubsystems that need to interact in order to improve environmental regulation. Thisdifference in analytic structure precludes a simple correlation between autopoiesisand mediation. The conclusion that mediation solves autopoietic closure is, at thisstage, not justified. Neither has autopoiesis so far provided clear theoretical supportfor mediation.

I see two ways to merge the two structures and to define a relationship betweenenvironmental mediation and autopoiesis. The first is based on the cognitive-analyticfunction of autopoiesis. With this interpretation, mediation is secondary to thetheoretical predominance of autopoiesis and the analysis remains entirely within theclosed boundaries of the theory of autopoiesis. We look at environmental interests asthe internal coding of a (not yet mentioned) functional social subsystem, 'theenvironmental policy discourse.' In this subsystem the opposition of the interest topollute versus the interest to protect the environment is the binary code that directsthe operations of the system (compare Luhmann, 1990). The negotiation language ofmediation is defined by this binary code and all mediation activities are internal noiseof the social subsystem. The analysis confirms the theory in that it recognizesautopoietic closure and the self-referentiality of social subsystems. It does not,however, give any practical advice for solving regulatory and environmental problems.It adds a theoretical perspective to mediation but does not make a closer linkbetween reality and autopoiesis.

The second interpretation takes the normative-prescriptive side of autopoiesis asa starting point. It aims to diversify the stakeholder groups indicated by mediationinto analytic units which autopoiesis may recognize. The different groups identified bymediation according to their interests are to be viewed as agglomerates of bits andpieces of autopoietically closed subsystems. Law, politics, economics and science are

149

all represented in each stakeholder group. Thus autopOletic closure does not onlyexist between the groups, but also within each group. The interest analysis and thesystems analysis run parallel, one is not a solution for the other. Mediation does notsolve the problem of autopoietic closure. Rather, autopoiesis adds further insight intothe dynamics of mediation.

The practical contribution of mediation to existing problems of steering hasbeen the great sensitivity it has added to regulatory communication, a process that istypically characterized by fierce antagonism. In the literature about mediation, themediator has jokingly been pictured as a supranatural being capable of doing acombination of almost mutually exclusive things. Phrases like "the mediator must beflexible but firm" and "must be an outsider, but well informed about the issue" areillustrative. It seems that to this 'wish-list' for the perfect mediator one more item cannow be added: must be trained in autopoietic analysis and master the internal codesof all social subsystems involved in the issue at stake. The practical application ofautopoiesis to problems of steering can be to take mediation one step further. Toadd, in other words, to the further sensibilization of the mediation process toproblems of communication, not only between groups with different environmentalinterests, but also between systems with different internal codes. If the mediationprocess is capable of responding to this autopoietic advice, mediation might become amultidimensional superdiscourse.

150

Chapter 13

AUfOPOIESIS AND STEERING: TIlE ROLE OF IDENTITY

I.Th.M. Snellen

1. Introduction

Metaphors can lead you by the hand, but they can also pull your leg. Fruitfulmetaphors normally do both at the same time. It will be some time before a finalverdict is reached on whether the autopoiesis metaphor is doing one or the other.Originally developed for the field of biology, autopoiesis is difficult to convert to thesocial sciences. In the opinion of the author, terms like 'operational closure' and'biological drift' are not easy to adapt for use in the social science sphere and mightappear more of a hindrance than a help. In this chapter the correct translation ofautopoiesis into the domain of public administration will not be at issue. A concisedescription of the manner in which the author views and interprets autopoiesis andself-referentiality in social systems of meaning, such as the legal, the economic, thepolitical and the differentiated sectoral systems and in social organizations, maysuffice.

Before a connection is made between autopoiesis on the one hand and steeringon the other, some related subjects have to be clarified. A distinction will be madebetween autopoiesis in systems of meaning and in social organizations. In the firstplace, this distinction is essential to clearly understand the different rates of changewithin and between systems of meaning and organizations. Braudel's model of 'layers'in historical developments, each with its own rate of change, may be fruitful in thisrespect. In the second place the author's view on steering as a combination ofmovement and direction, which will be briefly summarized, fits in well with such adistinction. However, 'steering' is also a metaphor possessing the same dangerousqualities common to all metaphors.

The autopoietic and self-referential nature of organizations and social systems ofmeaning makes the effectiveness of attempts to steer them problematic. It isquestionable, however, whether all possible steering deficits are equally relevant inconnection with the autopoiesis approach, or only specific ones. Are chronicunemployment among minorities, massive violation of velocity regulations onhighways, evasion of value added tax by bars and restaurants, the failure to get thelegally prescribed quota of invalids employed in enterprises and governmentadministrations, etc. without exception steering problems that have to do withautopoietic mechanisms? Is autopoiesis a concept 'having to do too much work'? Is

151

more selection necessary? Are there, in fact, any steering problems that can beexplained by autopoiesis? Or are Renate Mayntz and the many others who share heropinion correct in their assessment that steering deficits can be explained by otherfactors without resorting to the solutions offered by Teubner and Willke, solutionswhich are based on their belief in the working of autopoiesis? In this chapter amiddle position will be taken in which special importance is given to the identity ofthe systems and organizations that are to be steered.

2. Autopoiesis in social systems of meaning and in organizations

The basic tenets of autopoietic thinking as applied to social systems are that socialsystems maintain their identity by (re-)producing themselves, that they perceive andreact to their environment as a reflection of this identity and not as a reflection of theway in which they themselves are perceived by the environment, and that they as self­referential systems are, to a certain extent, closed to influences from the outsideworld. I do not have to go further into this, because the concepts concerned areexplained in the chapter by Van Twist and Schaap in this volume.

However, in this context it is, in my opinion, essential to make a distinctionbetween social systems as systems of meaning and as organizations. Self-referentiality,operational closure and structural coupling have a different significance for systems ofmeaning such as the legal system or the economic system, which pervade all sectors oflife transversally, than for organizational bodies such as courts, enterprises, welfareinstitutions or schools, which are active in only one sector of society. The sectors towhich these organizations belong are also characterized by their own self-referentialsystem of meaning. Thus, not only the legal, political and economic system havedifferentiated themselves in a self-referential sense with respect to each other, but thesame is true for the sectoral systems of meaning. This is aptly demonstrated byRosewitz and Schimansk in a recent chapter on the way in which sectors of society inthe past gained their independence and how this affects the political possibilities tosteer them.

The self-referentiality of a system of meaning entails that all social phenomenaare interpreted according to the structure of relevance of the system concerned. Thelegal system, for example, distinguishes between legally relevant and legally irrelevant(or merely factual) acts and facts, reads the meaning of the legally relevant acts andfacts in a juridical manner, relating them to juridical sources, norms, systems of normsand consequences. A juridical consequence cannot exist without a juridical cause,consisting of a juridically read act or fact, interpreted according to juridical normswhich are consistent within a juridical system. Thus, systems of meaning such as thelegal system and sectoral systems are separate worlds of self-referentially interpretedexistences. This also holds true for the economic system which distinguishes betweeneconomically relevant and irrelevant acts and facts and reads and interprets the thuscreated symbolic world according to the axiomatic points of departure of economicprinciple and economic motive. As I shall try to clarify later on in this chapter, theautopoiesis of social organizations is more volatile and volitional than that of systemsof meaning. The rate of change of intra-organizational self-referential patterns thatresults from this makes it worthwhile to give special attention to the elements ofmeaning of which the identity of organizations at any moment consists.

152

3. Rates of change in social systems; Brandeis 'layers in history'

The French historian and founder of the 'Annales' school, Ferdinand Braudel, wrotea thesis on the history of the Mediterranean under the reign of Philip II of Spain. Inthis Magnum Opus he distinguished three layers in historical developments: a'structural' level, a 'conjunctural' level and an 'evenemential' level respectively.Braudel characterized these layers or levels as to their time-scale, to be able to takeinto account the differences in rate of change between, among others, the physicalenvironment as background of the historical developments concerned, thegenerational sequence of eb and flood, for example, in economic tides and longcycles, and the hectic pace of everyday occurrences.

This conceptual framework can be fruitfully applied to the systems of meaningwhich I distinguished above: the legal, economic and political systems on the onehand and the sectoral systems on the other. On the level of occurrences('evenements'), a system of meaning reproduces itself from day to day. For example:legal practitioners (and the citizens acting in a juridical sense) confirm and recreateday after day the legal system, with its consistent structure, through its forms andnorms (contracts, deeds, mortgages, among others).

At the same time the legal system, through its structure recreates its elements incontracts, mortgages and deeds, i.e. ,juridical devices. Jurists function as guardians ofthe consistency of the legal system and its reliability. Important changes take place,but only after protracted process and after the eventual breaches in the normativesystem are gauged and conceptually remedied. Such changes take place at theconjunctural level of social systems. At this level new legal devices and concepts aredeveloped, legal principles systematized and refined, and new categories of juridicalfacts recognized. The rate of change at this level is slow and can be measured indecades.

Through the conjunctural level of changes, with their cyclical or spiral-shapeappearance, a structural age long trend becomes apparent. At this level very slowshifts take place, for example, the shift from legal protection of employers' intereststo legal protection of employee's interests. Some of these shifts may take hundreds ofyears (for example, the shifts in the laws concerning real estate users and owners).

In the economic system as a system of economic concepts, meanings andprevailing theorems and theories, the same three levels at which changes take placecan be discerned, but these will not be elaborated on at this time. This short overviewmakes it clear that, from the autopoietic viewpoint, systems of meaning are self­referential and operationally closed at the level of events (evenements) and arestructurally linked to the environment at the conjunctural and structural levels.Adaptations to societal requirements take place at the last two levels, maybe as aconsequence of re-interpretations of societal requirements originating in the systemsof meaning themselves.

This analysis in terms of layers in history or of different rates of change is, in myopinion, necessary as an addition to Luhmann's treatment of the self-referentiality ofsocial systems, understood as systems of communication, as well as to Rosewitz's andSchimansk's idea of the gaining of independence of sectors in society.

153

4. Changes in systems of meaning and social organization

As I indicated before, it is necessary to distinguish between the autopoietic characterof systems of meaning as described above and that of individual organizations. Thebiological metaphor might pull our leg. First, the 'operational closure' oforganizations does not relate to the exchanges between organization and environmentin terms of goods, energy and data, but to the symbolic interactions that are an aspectof each of these exchanges. In terms of business science, it might therefore be moreappropriate to speak of 'strategic' closure in the daily operations (at the level ofevents). Secondly, changes in the identity, the self-interpretation or self-referentialityof an organization may take place through different channels. At the conjuncturallevel the replacement of a functionary by an individual with a different disciplinary orsocial background or of a different generation, or the creation of a new departmentmay change the view of the organization concerning its environment. The creation ofa public affairs department, staffed by political scientists, will affect the self-imageand the self-referential activities of a multinational enterprise. The symbolicinteractions, and the meanings with which they are pervaded, are derived from thesystems of meaning described above.

Explicit strategy formulation may be another channel through which the basis ofself-referentiality of an organization and its interpretation of the environment maychange. The same is true of so-called 'emergent' strategies, which come into beingduring the course of operational activities. An intuitive reaction to a request at theoperational level of an organization may appear to be (in hindsight) the start of anew strategy and of the gradual development of a new self-image. Thus autopoieticinterpretations of change of social systems have to be differentiated when applied toorganizations. The rate of change is different, and the identity of organizations is of adifferent nature than the rationality of systems of meaning within the legal, economic,political or sectoral spheres.

The identities of organizations within a sectoral system, such as universities inthe educational system, or local theaters in the cultural system, may reflect the basicprinciples that dominate the sector to which they belong. The rate of change of theseidentities may be the same as the pace of their sector, but may also be quicker. Localtheaters have different funds from which they can derive the components of theiridentity: education, recreation, social participation, culture as beauty and art as socialexperimentation. Each mix of these components will lead to a specific identity anddifferent kinds of reactions to developments in the environment. Identity and reactionare thus more volatile and volitional than a strict application of autopoietic thinkingwould suggest.

5. Steering as movement and direction

If we want to determine what kinds (if any) of governmental steering are appropriateto autopoietic social systems, as described above, an attempt must be made todiscover the essence of social steering. The metaphor of 'steering of a ship' is oftenused in this context and is still very helpful in explaining the basic aspects andproblems of governmental steering. The basic aspects that will be stressed here andwhich are neglected in traditional debates on steering are:

154

1. the distinction between driving and directing;2. the different media through which driving and directing may take place.

It has to be recognized that effective driving (at least any movement) is a necessaryprerequisite for any form of steering. A ship that is not moving, that is simply floating,cannot be directed or kept on course. Thus, giving direction assumes motion. Or touse another nautical metaphor: the rudder will only be effective if the screw propelleris turning. In discussions of societal steering, the elements of driving and givingdirection are mixed up. One does not realize that possibilities of giving direction aredependent on existing motion, which is a consequence of driving by other actors orfactors. Actors or factors that are responsible for driving are not necessarily the sameas those that are giving direction. In most considerations of problems of societalsteering by governments the focus is on the aspects of direction giving. Driving forcesare seen as given. Normally it remains unclear in what way the forces of givingdirection and of driving are thought to be deployed and to relate to each other.

A consequence of the neglect of the difference between driving and givingdirection is that it is not taken into consideration that both may take place throughdifferent media. The ship metaphor may clarify this. A sailboat is driven by themedium air and kept on course by the medium water. For the steering of a ship, therelevant media are those environmental conditions which are 'structurally linked' tothose parts of the ship that have a driving or directing function.

Translated to an organization such as an enterprise, a school or a hospital, themedia through which driving and giving direction take place are those systems ofmeaning which are connected to the structures of relevancies in which the identity ofthe organization concerned manifests itself. The identity of economic organizations,such as enterprises, is mainly determined by economic, technical and social-scientificconsiderations. Technical-scientific considerations will predominate in their research,development and production activities; economic ones in their buying and selling andtheir financial business. Social-scientific considerations are an ingredient in theirpersonnel, marketing and industrial policies. Organizations which have a semi­governmental character, such as welfare, health care and educational institutions,generally have an identity in which legal, economic and (technical- and social-)scientific traits co-exist. The legal trait in their identity has to do with the fact thatthese semi-governmental institutions generally execute governmental tasks which arerelated to basic rights such as schooling, minimum income, etc. The identity of publicauthorities is, however, determined by legal, economic, (social- and technical-)scientific as well as political considerations simultaneously. Their structure ofrelevancies is therefore much more complex than that of economic or semi­governmental organizations.

Governmental steering of organizations in a society can take place through diffe­rent systems of meaning such as the economic or legal system or through thedominant system of meaning (medical, educational, welfare) in a sector. It is alsopossible that governments only play a role with respect to driving or giving direction.The government may supply the money for research but leave the choice of researchthemes or research direction to the discretion of a peer group. In that case the drivingmedium is economic while the directing medium is intellectual and immaterial. Byway of the first medium, research instruments can be acquired and researchersappointed; through the second medium the direction in which the researchers will

155

point their intellectual efforts is determined. Scientific research is a good example ofthe relative ineffectiveness of governments to steer through some (especiallyintellectual) systems of meaning. Researchers can hardly be prevented from pointingtheir research in directions which they themselves find the most relevant. Mutatismutandis the same is true for the other sectors of society.

6. Governmental steering of systems of meaning

It has been shown that, in systems of meaning, three layers can be discerned. Each ofthese layers possesses its own rate of change and its own degree of closure. At thelevel of occurrences (evenements), systems of meaning recreate themselvescontinuously without changing any of their (basic) premises. The steering possibilitiesof governments are limited to confirmation and reinforcement of the self-referentialoutcomes of the systems of meaning concerned. After all, the behavior of thegovernment authorities themselves is also regulated by the systems of meaning whichwere indicated in the previous section. The daily routines of government, as well asthose of non-governmental institutions therefore belong to the evenemential level ofthe systems of meaning.

At the conjunctural level of systems of meaning, the steering capacities ofgovernments may be substantial. In the legal system, governments can change lawsand regulations but have to remain within the confines of the existing system ofnorms. They can try to change existing jurisprudence by starting a juridical test case.They can try to influence 'legal opinions. In some countries, the nomination of courtjudges by the executive may have the same effect. In the economic system govern­ments can shift the basis of their policies, for example, from Keynesianism to supply­side economics, and thereby instigate changes in middle range theorizing. But heretoo, the basic premises of economics and public finance have to be accepted as given.

In the sectoral systems of meaning, mainly developed by the 'gamma' disciplines(applied social sciences), governments may stimulate some subjects or points of focusby funding research in specific directions. But as has been seen in the previoussection, the government will be more effective (less ineffective) in creating researchfacilities, than in procuring required research outcomes.

At the structural level, systems of meaning may change extremely slowly.Governments will have almost no influence on these changes. As far as new sectoralsystems of meaning are concerned, for example, concern for the environment, socialand technical efforts directed towards the natural environment might be a point ofattachment for government steering. The government will not be able to determinebasic premises. It can, however, determine the research efforts concerning fundamen­tal issues thereby influencing the system of meaning, if only indirectly.

7. Governmental steering of organizations

The importance of determining in what respect governments can steer systems ofmeaning derives from the fact that the self-referentiality of social organizations,reflected in their identity, is linked to those systems of meaning. Systems of meaningare thus the medium through which social organizations steer themselves and can in

156

principle be susceptible to steering from the outside. However, apart from the lock ofthe self-referentiality they share with all social systems, their entrenchment in self­referential systems of meaning is a second lock, which bars them from governmentsteering. However, the changes in the identity of organizations are more volatile, andvolitional, than strict autopoietic thinking according to biological insights wouldwarrant.

The 'ports d'entn!e' of government steering in self-referential social organizati­ons are diverse. Prescriptions and incentives may affect the operational level. However,in view of their 'strategic' closure at the operational level, prescriptions and incentivesmay be relatively ineffective. They are better instruments for setting things in motionthan for moving them in a distinct direction. Recognition of the relativeimpenetrability of social organizations may lead governments to limit themselves tothe indication of some key parameters that have to be reached, leaving it to theorganizations themselves to determine in which fashion they intend to reach them.

It would be more time consuming, but perhaps more effective, for the govern­ment to try and develop a commonly shared 'definition of the situation '. The 'portd'entn!e' of governmental steering would in this case be the interpretation whichleads to a structural linking of the social organization and the (symbolic) environ­ment. The strategy formulation of an organization is an obvious instrument forcreating not only shared definitions of the situation, but also the necessary commit­ment to draw the consequences thereof. Governments can further strategic planningefforts in organizations and supply information that hopefully may affect their strategyformulations in the desired direction.

The structural linking of organizations with the symbolic environment might alsobe realized by a prescribed proceduralization of decision making. Procedures can beprescribed which ensure that the basic premises of a dominant system of meaning aretaken into consideration. Proceduralization of decisions on euthanasia, for example,may ensure that the principles of medical practice are strictly adhered to.

A government undertakes an attempt to approach the identity of a socialorganization directly when it forces the organization to accept one or more 'invading'systems in its structure. Structuralization may be an effective way of steering socialorganizations. When the invading system is accepted and becomes part of theorganization, it will playa role in the organization's autopoietic and self-referentialcircularity and co-determine its identity. It may, however, take years before a newstructural element is accepted. A good example is the enforced participation ofworkers' representatives on the boards of supervisors of German enterprises. Normal­ly, at least one generation of functionaries will have to pass from the conjuncturallevel before the invading system can become an integral part of the identity of theorganization and take part in the circular processes of 'defining situations for theorganization'. A comparable method of structuralization may be followed for anetwork of organizations. In this case the participation of one or more invadingsystems in the network will be prescribed and thereby a channel created. Outcomes ofpolicy networks may be effectively influenced by such changes in the composition ofparticipation. Autopoietic approaches to governmental steering may therefore takeplace in the following ways:a. traditional regulation accompanied by sanctionsb. common definition of the situationc. indication of key parameters

157

d. proceduralization of decision makinge. structuralization of an individual organization or of an arena of organizations.As a matter of course, combinations of these approaches are possible.

8. The configuration approach and steering

The configuration approach, developed by Van Dongen and his coworkers, focusseson the role of interaction in social cognition. Definitions of a situation such as theone mentioned above are formed and formulated through social interactions. Thecontinuation of social interactions depends on a structural aspect, a certain fit inoutward physical behavior and on a symbolic aspect, a congruence in concepts andmeaning which can be ascribed to outward behavior. Interactions take place insocial-cognitive configurations. Actors in an interaction are more or less included inthose configurations. This inclusion may be dominantly cognitive or dominantlysocial, I and it results in a common 'definition of the situation'. Changes in such adefinition may thus originate from new social inclusions as well as from new cognitiveinclusions.

From the viewpoint of steering, the configurative theory is interesting. It mayunderline the notion of steering as being on the one hand driving (the social inclusionaspect) and on the other hand direction giving (the cognitive inclusion aspect). TheDutch Ministry of Education and Science succeeded in forcing researchers in differentuniversity departments, who were not used to working together, to cooperate withinresearch programs. In many cases this social inclusion in 'conditionally financed'programs resulted in fruitful research programs.

However, social inclusion as such may not be enough to become a driving force.The configuration approach fills in some parts of steering according to the autopoieticapproach mentioned in the foregoing section. Apart from explaining situationdefinitions as an effect of continuous streams of interaction on the basis of social­cognitive inclusion, it clarifies how structuralization lays a foundation for socialinclusion, and through this, for cognitive inclusion. The configuration approach maythus help in recognizing the identity of organizations as object and outcome ofsocially negotiated constructions.

9. Steering and identity of social organizations

A question that still has to be answered for social organizations is which mode ofautopoietic steering might be preferable under what circumstances. In the line ofreasoning of autopoiesis, the identity of the organization in relation to the self­referentiality of its (part of the) sector has to be decisive for this answer. It issurprising that the issue of identity of organizations has so far received so littleattention. Chapters on autopoietic steering are therefore often unduly abstract.

The identity of organizations is, as we have seen above, not as uni-dimensionalas Teubner suggests in his contribution to this book. Economic organizations, such asprivate enterprises, are not only following economic considerations and norms in theirsocial behavior. They may, and often will, translate legal prescriptions into economicvariables, such as cost factors, competitive edge, etc. However, many years of

158

experience in a legal department of a multinational enterprise has taught me thatlegal departments in enterprises regularly act as legal conscience and have an intrinsicinfluence on the behavior of the firm.

Medical specialists in hospitals do not 'translate' all external developments andsignals, such as political statements, economic incentives and legal regulations, intomedical considerations. Neither do teachers 'translate' them exclusively in educationalterms. While individuals and organizations may have a dominant (medical,educational, welfare,legal , economic) orientation, this does not mean that they adheresolely to one (sectoral) system of meaning. Modem citizens and organizations areused to living simultaneously as inhabitants of different worlds in which diverselanguages are spoken. The more they are versed in these languages, the more theythink (and even dream) in them. In 'Public Sector Management, Systems and Ethics'Louis Gawthrop gives a very useful analysis of the main components of identity.Although Gawthrop does not adhere to the autopoietic approach, his analysis enablesus to link autopoiesis and steering through his concept of identity. It might help toestablish what mix of elements from diverse systems of meaning will likely determinethe behavior of the organizations concerned for steering.

Identity, according to Gawthrop, is ontologically characterized by a sense ofpurpose, a sense of consequence, a sense of history and a sense of order. "The mostattractive feature of these four ontological components is, that if viewed in terms oforganizational responses to the forces of change, such can be used for different modesof analysis". (Gawthrop, 1984, p.51). In his discussion of reactions to change,Gawthrop is mainly interested in the way these ontological components result in anincremental versus a non-incremental approach. For our purpose the relationshipbetween autopoietic steering and identity of organizations, sense of purpose, ofconsequence, of time and of order are interesting as ontological components becausethey refer to dimensions of relevance which explain possible outcomes of self­referential behavior. Sense of purpose manifests itself in the lens through whichorganizations in a sector conceptualize their objectives and goals (e.g., profit, income,size, quality of work, professional criteria). This sense also expresses itself in anorientation to the future or the maintenance of acquired positions. The sense ofpurpose of an organization may be derived from a transversal or from a sectoralsystem of meaning. Sense of consequence appears in the degree of interest in theoutcomes of one's behavior in time and space. It is shown in a more atomistic or amore holistic responsibility, (main focus on direct consequences and on the localenvironment versus on future generations and global ramifications). The sense ofconsequence may be limited to relevant values in a specific system of meaning (asTeubner in his contribution seems to suppose), but may also take into considerationvalues and norms from other systems of meaning, as stated above. Sense of historyreflects an attitude to the timeframe of the organizational behavior. It may stress thecontinuum aspects of this behavior or be characterized by a more opportunisticlocalized hic et nunc orientation. Sense of order denotes the perspective in which theactivities of the organization are visualized, a horizontal sector-overarchingperspective or a more narrow vertical sector-bound perspective. In the firstperspective organizational power is conceived more as a public duty, in the secondone, more as a private prerogative. Careful analysis of social organizations alongthese dimensions of identity, and acceptance of the probability of 'mixed motives' isnecessary to achieve (a bit less in-) effective steering.

159

10. Conclusion

The autopoietic approach is useful for the development of concepts of steering inpublic administration. It may contribute to the self-interpretation of the professionalactivities of public officials in their relation to organizations and sectors of society.

If one is interested only in workable solutions to problems of governmentalsteering, a general idea of the impenetrability of social systems of meaning, socialsectors and social organizations may suffice. But for a further development ofprinciples of steering and a further fine-tuning of steering instruments, the autopoieticapproach seems promising. However, a distinction between rates of change indifferent layers of history, between driving and directing as components of steeringand between elements of identity is necessary for a substantial improvement of theautopoietic approach to governmental steering.

Notes

1. Although social and cognitive inclusion always go together, I understand from Bolk (1989, p.46)that cognitive inclusion or social inclusion may be a point of departure. As to intervention, heremarks: "The effect of reframing, of a directed way of organizing contacts, of guaranteeing thepossibility of being subjected to alternative definitions, is dynamizing the substantial and socialstructural developmp.nt" of "social-cognitive configurations. "

160

Chapter 14

POLICY INSTRUMENfS FOR SI'EERING

AUfOPOIETIC ACTORS

I.A. de BruijnE.F. ten Heuvelhof

1. Introduction

To consider the problem of governmental steering intelligently, we must have someinsight into the tools or instruments that might be used by government. The instru­ments of government can be considered as the means by which it attempts to steerthe behavior of non-governmental actors. It seems plausible that these instrumentswill be an influential variable in the process of governmental steering. In daily life,using a carrot will lead to other results than using a stick or a sermon. This is also thecase with policy instruments and the policy analyst is therefore required to havetheories on policy instruments at his disposal. One element of these theories shouldbe a hypothesis about the autopoietic character of the target group which is to besteered. The autopoietic character of the target group which is to be influencedmakes it difficult for the steering actor to evaluate the effects of his steering efforts.

This chapter contains a review of the main themes in the literature on policyinstruments. It is interesting to note that theories on policy instruments and strategiesfor using them are surprisingly underdeveloped areas in policy analysis (Hogwood andPeters, 1985, p.178). This seems to be particularly marked in the Anglo-Saxontradition. The German and the Dutch traditions contain more relevant insights. Themain explanation for this phenomenon may be that vital elements of the Dutch andGerman public administration system stem from a judicial tradition, suggesting afamiliarity with the concept of the policy instrument. This situation implies that theconcepts presented here will be derived from German-Dutch public administration.Special attention will be given to the problems which arise because of the autopoieticcharacter of the target groups which are to be steered.

2. Main themes in the literature on policy instruments

The main themes in the literature on policy instruments can be categorized into threetopics, which will be presented here. As it is not possible to elaborate the themes indetail, only the central concepts will be dealt with.

161

2.1 Some definitory problems

The first topic has to do with the problem of defining the concept. The qualification'policy instrument' is used for divergent phenomena such as regulation, subsidization,the state, the organization and political rhetoric. An investigation of definitionsindicates that there are broad and narrow definitions of instruments (De Bruijn,1990). In the narrow definition policy instruments are considered to be things (e.g.,rules, money, contracts). In the broad definition, instruments are activities (e.g.,applying rules, transferring money, finalizing contracts). Both definitions have theirdrawbacks. Using a narrow definition can lead to formalization and reification. Usinga broad definition may in the end lead to superfluity of the instrumental approach. Itis for this reason that many scholars prefer a more or less narrow definition: a policyinstrument is a thing, whether formal or informal.

2.2 Characteristics of instruments

How do policy instruments influence the outcomes of the policy process? In theliterature on policy instruments two approaches can be distinguished in order toanswer this question. The first approach will be presented in this section, in section2.3 attention will be given to the second approach.

In many studies there is a strong emphasis on the intrinsic characteristics ofinstruments, which is assumed to be of decisive influence on policy outcomes. Take,for example, a capital levy, which constrains the alternatives of a firm and willtherefore in one way or another guide the firm's behavior. If policy instruments havecertain characteristics which influence the outcomes of the policy process, it isnecessary to determine how different instruments effect policy outcomes. The nextstep is to categorize different instruments and to develop typologies of instruments.Each category of instruments has certain intrinsic characteristics that have a certaineffect on policy outcomes. Indeed many typologies of instruments have beendeveloped (Hood, 1983; Mitnick, 1980). While a broad overview of the manytypologies of instruments will not be presented is this chapter, the typology developedby the Dutch scholar Van der Doelen, who based his typology on many others,including Hood's and Mitnick's, will be discussed.

Van der Doelen distinguishes three main types of instruments. Each category isqualified by the authors as a family of instruments. The three main families are thefamily of judicial instruments, the family of financial instruments and the family ofcommunicative instruments. How do these families influence the policy process?There are three answers to this question. Each family has instruments which broadenthe alternatives of the target group and instruments which constrain the alternativesof the target group. A subsidy broadens alternatives; a capital levy constrainsalternatives. Each family has general and particular instruments. General instrumentsare focused on the target group as a whole, particular instruments are focused on theindividual actor. Each family has constitutive and directive instruments. Constitutiveinstruments are used for institutionalization; directive instruments are used fordirecting actors.

Although one can criticize this typology, it can be useful in the discussion ofpolicy instruments. It makes clear that instruments have certain characteristics thatwill have at least some effect on policy outcomes.

162

2.3 Context: from determinists to voluntarists

At this point it can again be asked how policy instruments influence the outcomes ofthe policy process. The first answer was presented in section 2.2: policy instrumentshave intrinsic characteristics which influence the process of steering. However, theway in which these characteristics influence the policy process also depends on thecontext in which they are used. Each instrument has an organizational, political,judicial and economic context, which will influence the use of instruments and whichwill therefore influence the outcomes of the policy process (De Bruijn, 1990; Hufen,1990). A second implication of the relevance of contextual variables is the dynamiccharacter of instruments. Contextual variables will force the government to change itsinstruments if they do not fit into a certain context. There are numerous reasons foradapting instruments to their contexts (e.g., a certain behavior of the target group,certain characteristics of the implementing organization etcetera). Berman hasqualified these processes as technological learning: government will learn how to useits instruments ('technology') adequately in a certain context (Berman, 1978).

Some scholars stress the intrinsic characteristics of instruments while othersstress the importance of the context. The intrinsic characteristic approach can bequalified as deterministic: the characteristics of a policy instrument determines thepolicy process. Stressing the importance of the context can lead to voluntarism.Voluntarism denies the importance of the instrumental variable. Contextual variablesare far more important; it is the context and nothing but the context.

3. The autopoietic target group as context for using instruments

A central assumption in the discussion of policy instruments is that a target group issensitive to the signals that are generated by an instrument. The theory of autopoiesis,however, teaches us that autopoietic social systems are self-producing and that theyare relatively unresponsive to signals from their environment. Used as a metaphor(Morgan, 1986, p.240), the theory of autopoiesis implies that a target group is alwaysattempting to achieve a form of self-referential closure in relation to its environment.The non-penetrability of the target group will frustrate the effectiveness of usingpolicy instruments. Government will have to develop steering strategies that fit withina context of an autopoietic target group. In general, such new strategies should meetone or more of the following requirements:1. They have to be able to penetrate the self-referential closure of a target group;2. They have to fit within the 'identity', 'image' or 'perceptual filter' of the auto­

poietic target group.3. They have to be able to change the 'identity', 'image' or 'perceptual filter' of the

autopoietic target group.

These principles are the rationale of a catalogue of instruments which is presented insection 4. Before presenting these instruments, a few more general remarks about theuse of instruments in a context of autopoiesis are made.

163

3.1 The use of regulatory instruments

The autopoietic character of target groups makes it very difficult to forecast theconsequences of using an instrument. The target group which has to be steered willgive meaning to steering signals in a way which cannot be anticipated. The minimalimplication of this is that instruments will need a certain flexibility and are to be usedexpediently. Some instruments are characterized by rigidity and can be considered asinadequate for steering autopoietic target groups. Well-known examples are manykinds of formal regulation now in existence.

It is important to note that owing to the nature of an autopoietic target group,the use of unilateral regulatory instruments proves difficult, at best. The use ofregulatory instruments will provoke an autopoietic target group as soon as the normsbehind regulation do not fit within its identity. A well known consequence of usingsuch instruments is that the target group strengthens its self-referential closure. Inthat case the use of regulatory instruments is counter-productive.

Of course this does not mean that traditional regulatory instruments are inade­quate in a context of autopoietic target groups. The values underlying regulatoryinstruments may fit within the identity of an autopoietic target group. Some regulatoryinstruments will have a deterrent effect on the target group and may break up itsclosure. However, a government using only regulatory instruments will have a verylimited capacity for steering.

It seems that in administrative practice, as well in administrative theory, a searchfor new instruments that fit within the identity autopoietic target groups is takingplace. Such instruments are financial incentives, contracts and various kinds offacultative instruments. Although such instruments are not entirely new, they willoften be used in an unorthodox fashion and can therefore be qualified as 'new'instruments. As said before, these instruments will be presented in section 4.

3.2 The use of instrumental mixtures

The confrontation with an autopoietic target group will often require government touse a differentiated set of instruments or an instrumental mixture. Some autopoieticactors can be guided by using a carrot, others require a sermon or a stick. An instru­mental mixture is a configuration of instruments that are used simultaneously. Amixture may be an adequate instrument for steering autopoietic target groups becausetheir identity is often multi-dimensional and complex. Identities are hardly everconsistent and monolithic, but often consist of competing values. An instrumentalmixture can be used for dealing with this complicated identity. Government uses acombination of steering signals that all have a different influence on the identity ofthe target group. Although the isolated instruments may not have much influence onthe target group, it is the combination of instruments that influences the target group.

An example of this is the implementation by government of incentives, legalrules and communicative instruments. The communicative instruments can be used tocreate a shared image between government and non-governmental actors. Further­more, the incentives can be used to gain entrance into a target group. The use ofthese instruments may lead to a relationship between government and the non­governmental actor such that government can use formal rules to guide the non­governmental actor.

164

3.3 Managing the environment of the target group

The identity of an autopoietic target group is responsible for a selective interpretationof its environment. The autopoietic target group is sensitive to only a few variables inits environment. In general, this notions limits the steering capacity of government.However under certain conditions there are also opportunities for effective steering.It is up to government to influence the environmental variables in such a way that thetarget group becomes sensitive to them. Suppose a target group is very sensitive tovariable X. By managing an environment in such a way that variable X changes,government can influence the behavior of an autopoietic target group.

In other words, using instruments for steering an autopoietic target group mayrequire some form of 'context management' by government. Context management canbe viewed as a technique that creates an environment in which an autopoietic targetgroup is more sensitive to steering signals. Context management is a technique ofmaking use of the self-referential closure of the autopoietic target group.

Take, for example, the use of communicative instruments. It is a well knownphenomenon that target groups are most receptive to information when their environ­ment is changing or when they need to know something specific about an issue thatconcerns them (Meltsner and Bellevita, 1983, pA2). If government wants its communi­cative instruments to be effective, it is a matter of context management to create asituation that will be perceived as a changing environment by a target group or tocreate a situation that will lead to a need for information. Furthermore, governmentcan use its communicative instruments, which will then be linked by the target group'sneed for information.

A species of the genus 'context management' can be qualified as 'networkmanagement' (De Bruijn and Ten Heuvelhof, 1991; Mandell, 1989). Networkmanagement means that government uses interdependencies between different targetgroups. For example, government can create interdependencies between two targetgroups by changing their financial, juridical or political relations. The increase ofinterdependencies between target groups is a very important technique for steering atarget group, for interdependence means that one has to give up its closed nature.

3.4 Information, instrumental interference and fine tuning

Finally, it is important to point out some general implications of the use of theinstruments, to be presented in section 4. These instruments fit within a context of anautopoietic target group because they meet the three principles mentioned above.

The first important implication is that the instruments in section 4 often requiredetailed information about the target group to be steered. Government will needinformation about the 'identity' of the autopoietic target group. Without suchinformation government will not know what kind of signals the target group issensitive to. Getting enough information about the target group's identity may proveto be very difficult. Information about the identity is internal information of the targetgroup and will therefore be protected by its closed nature.The second implication can be qualified as 'instrumental interference'. As said before,government may use an instrumental mixture in order to meet the multi-value identityof an autopoietic target group. Using a mixture may lead to interaction or interfer­ence between the instruments. Instrument A may frustrate or neutralize the effects of

165

instrument B. Especially in very complicated situations, where several governmentsare using instrumental mixtures in order to steer several target groups, such processesof instrumental interference can arise. In such a situation the ultimate result of theinstruments depends on the way government manages the process of usinginstrumental mixtures. It will also depend on good or bad luck. Instrumentalinterference can lead to a high degree of unpredictability concerning theconsequences of applying instruments.

Thirdly, the instruments in section 4 have steering capacity, but it is a limitedcapacity. On the one hand these instruments recognize the autopoietic character ofthe target group. On the other hand, because of this recognition, they can influenceonly very specific 'reinforcers' of the behavior of the target group. A target group witha closed nature will allow only very specific instruments to have any steering capacity.In other words, using the instruments in section 4 will be a matter of fine tuning.Instruments have to be used very precisely and therefore run the risk of missing theirgoal.

4. Autopoiesis-proof instruments

Which instruments can be qualified for use in situations of autopoiesis? To answerthis question, a catalogue of instruments will be presented. It is important to note thatthe adjectives of the instruments being presented do not say anything about thecharacter of the instruments, but about the way they are used. This difference isgenerally recognized in instrumental theory. The qualification 'judicial' says somethingabout the character of an instrument, the qualification 'indirect' (see below) sayssomething about the way instruments are used.

4.1 Indirect instruments

One implication of instruments being used for steering an autopOletlc target group isthe importance of indirect instruments. Indirect instruments do not try to steer thebehavior of an actor directly, but by means of other actors. Indirect instruments maybe useful for steering autopoietic and powerful actors that are not sensitive togovernmental signals, but may be sensitive to the steering signals of other actors. Ifgovernment uses these other actors to send steering signals, the chance of success maybe improved.Familiar examples are the steering of private firms through massive information

campaigns (communicative instruments) aimed at consumers. This technique has beenused successfully to influence the behavior of public enterprises (Ramanadham, 1984,pp.219-225; Ferner, 1988). These enterprises were barely receptive to governmentaldirectives, but were very receptive to the wishes of consumers. An alternative way ofusing indirect instruments is influencing the individual members of an organization inthe hope that they will further be able to guide the behavior of the firm as a whole.Using indirect instruments means that government is not controlling the output of atarget group, but that it is trying to influence the input. Changing a target group'sinput may subsequently lead to another output.

166

4.2 Facultative instruments

Facultative instruments are instruments that present a high degree of voluntarism tothe target group which is to be steered. The decision to use the instrument is taken bythis actor. He may even play an important role in the process of conceiving the policyinstrument. The rationale for the use of facultative instruments is that the targetgroup will be willing to accept the consequences of the use of an instrument becauseof its voluntary agreement to use it.

A generally recognized example of a facultative instrument is the economicsubsidy. Private firms are free to accept or ignore subsidization facilities. As soon asthey accept a subsidy and the conditions that apply to it, it is very probable that theywill be willing to implement these conditions. Facultative instruments are especiallyuseful when government is confronted with powerful and autopoietic target groups.Facultative instruments take into account that target groups are autopoietic and holdpowerful positions. It puts the initiative to break through its closed character on thetarget group itself.

4.3 Communicative instruments

The autopoietic character of target groups that are to be steered implies that they willonly be receptive to signals that fit within their 'perceptual filter' or 'image'. There­fore intelligent steering will be focused on influencing the composition of theperceptual filter. Communicative instruments can play an important role in thisprocess. The use of communicative instruments consists on the one hand of thetransfer of information. This is a rather soft instrument that will be accepted rathereasily by the target group. The transfer of information can influence the perceptualfilter of the target group and can make this actor more receptive to other steeringsignals. In the words of Meltsner and Bellevita: "Ideas first take shape in the mind.Then they are communicated to others and altered, strengthened, or abandoned.Eventually, under the right mix of planning, circumstances, and luck, successful ideasinfluence the way that organizations involved in public or private sector policybehave" (Meltsner and Bellevita, 1983, p.25).

On the other hand, the use of communicative instruments can consist ofacquiring information about actors to be steered. This process can influence govern­ment's perception of the target group. Furthermore, it can be helpful in creatingshared images between government and target group. Obviously, shared images are ofvital importance for steering activities in networks. Therefore communicativeinstruments can be considered not only as the government's 'eyes', but also as its'ears' .

4.4 Multilateral instruments

A central characteristic of many regulatory instruments is their unilaterality.Government directs the non-governmental actors on how to behave and they areexpected to do so. As said before, such instruments do not fit within an autopoietictarget group. Autopoiesis and unilateral instruments will seldom be harmonious.

Multilateral instruments recognize the autopoietic character of the actors to besteered. Well-known examples of multilateral instruments are contracts, gentlemen's

167

agreements and various types of convenants. The use of these instruments presuppos­es that government and target groups will negotiate the extensions to be delivered. Byallowing non-governmental actors to negotiate, governments respects their position ofpower. Bargaining processes between government and the autopoietic target groupwill establish relations that can lead to value congruence, mutual trust and mutualrespect. By bargaining with government, the autopoietic target group is breaking upits closed nature. This implies that it is far easier for government to gain entrance tothe autopoietic target group.

4.5 Incentives

Financial incentives can also be considered as instruments that are well suited to anautopoietic context. Incentives have a voluntary character: it is up to the target groupto react or not to react. Take for, example, a capital levy in environmental policy. Aprivate firm which produces a large amount of pollution will be obliged to pay acapital levy, but not be stopped from polluting the environment in any other fashion.Besides such negative financial incentives, there is also a positive one. Governmentcan subsidize a target group. Again this financial incentive has a voluntary character.It is up to the target group to accept or reject the subsidy.

The voluntary character of incentives makes them appropriate instruments forsteering autopoietic target groups. Because of their voluntary character they willrather easily be accepted by a target group. Incentives have the additional advantagethat costs and benefits are hardly ever a constitutive element in the perceptual filterof a target group, whether private or public. Instruments like incentives influence therelation between costs and benefits can be expected to fit within the identity of theactor to be steered.

s. Instruments as an organization's technology

Ideally, a flexible and responsive government should have the capacity to use theinstruments that are most adequate in a certain situation. Therefore a continuingprocess of choosing the right policy instrument should be characteristic ofgovernment. In practice this is often not the case. Governmental organization areaccustomed to certain instruments, have experience in dealing with them and areoften unwilling or unable the use new ones. It is therefore doubtful that governmentwill always be able to choose the most adequate instruments. It does not seem veryrealistic to expect that government will be able to switch instruments as soon as this isrequired.

Why is this the case? Is this just a matter of conservatism and bureaucraticpathology, or is it an inevitable aspect of using instruments? To answer this questionit should be taken into account that instruments can be considered as an organiza­tion's technology (Salamon, 1989, p.7). If one compares the main definitions of policyinstruments with the main definitions of technology, the resemblance is surprising.Both concepts refer to the means of an organization. Both concepts can be definednarrowly and broadly. The main difference seems to be the disciplines in which theconcepts originate. Policy instruments stem from policy analysis; technology stemsfrom organization theory.

168

The literature on technology leads to three main insights on the relation betweenpolicy instruments and governmental organizations:

I. Policy instruments always put certain requirements on the governmental organi­zation. Different instruments have to be handled in different ways and thereforeimpose different knowledge requirements on organizations. This implies thatchoosing new policy instruments can have important consequences for intra­organizational relations. Certain knowledge will suddenly be outdated. A newbody of knowledge will have to be built up. More often than not, the existingimplementation structure for instruments will have to be altered. All theserequirements for using a certain instrument imply that introducing new instru­ments will take considerable time and require considerable organizational effort.

2. Policy instruments structure the governmental organization. A governmentalorganization that is confronted with instrumental requirements will have to reactto such requirements. As soon as such a reaction is standardized, it can beconcluded that policy instruments in one way or another structure the govern­mental organization. Of course there is no linear relation between instrumentsand organizational structure. This relation is always multifaceted. However,within a certain range organizational structure can be qualified as adequate orinadequate for using certain instruments. It is clear that choosing new policyinstruments has implications for the organizational structure and is thereforeinvariably a slow process.

3. The relationship between instruments as technology on the one hand and thegovernmental organization on the other hand is twofold. Instruments putrequirements on the governmental organization, but the governmental organi­zation can also perceive instruments in a way that is convenient to it. It is a well­known phenomenon that administrators adapt new instruments to existingstandard operating procedures. Amending President Johnson's familiar words,one can state that the best way to kill a new instrument is to put it in an old lineagency. A switch to a new instrument could bring new departmental and agencyactors, with their differing perspectives, into the policy-making arena, a possibili­ty that the governmental organization may choose to resist (Woodside, 1986,p.787). In addition, it can be pointed out that a 'new' policy is often imple­mented with the help of already existing instruments. In other words: (therequest for) new instruments can be perceived by the governmental organizationin such a way that intra-organizational modification will be unnecessary and thenewness of such instruments of a limited character.

To consider instruments as an organization's technology makes it clear that there maybe a tension between the policy instruments that are required by autopoietic targetgroups and the possibilities a government has to put these requirements into practice.The requirements imposed by using new instruments can be difficult for any organiza­tion to meet. Therefore it is probable that these of requirements will surpass organi­zational limits.

The implication of this statement is twofold. First of all, the governmentalorganization will often be able to meet just a few of the above mentionedrequirements. Secondly, using instruments in the confrontation with the autopoietictarget group places requirements on the governmental organization that will challenge

169

traditional bureaucratic organizations and will require new kinds of organizations,such as the adhocracy, project organization, etc.

6. What can we learn from the instrumental approach?

The role of instruments in the process of steering has been elaborated on in theprevious sections. In this section the contribution of the instrumental approach topolicy analysis will be summarized. The instrumental approach plays three roles in theanalysis of the steering process.

As stated earlier, policy instruments play a role in the process of steering.Theory on policy instruments focusses on a variable in the process of steering that issometimes a core variable and sometimes a variable in the periphery of the steeringprocess. However, it always plays a role that must be reckoned with. The way this roleis played depends on the rate of closure of the autopoietic target group. In manycases a large number of instruments will be ineffective in a confrontation with anautopoietic target group. There are, however, some instruments which have arelatively high likelihood of success in steering autopoietic actors. These instrumentshave been presented in section 4. Policy analysts who deny the instrumental variablein their analysis are reducing the complexity of the policy process in an unacceptablemanner.

In addition, theory on policy instruments also pays attention to the technology ofa governmental organization. When instruments are considered as being the technolo­gy of an organization, it becomes clear that there are important organizational limitsto a flexible use of instruments in networks.

A second important aspect of the instrumental approach is its role of 'enlighten­ment'. To ask the instrumental question is to ask how policy aspirations will betranslated into action. The instrumental approach focusses on the repertoire ofgovernment. By definition the instrumental repertory of government is limited.Numerous policy proposals have to be made to fit into this limited repertoire and canbe unmasked by asking the instrumental questions. Pretentious policy proposals may,on the instrumental level, prove to be nothing more than existing policy.

This role of enlightenment is not limited to the practice of public policy. Many'steering styles', 'steering concepts', 'steering philosophies', etcetera, may prove to beof little worth once they reach the instrumental level. Different philosophies may leadto the use of the same instruments and thus have any meaning for the way policyinstruments are used.

Thirdly, it is important to note that an instrumental approach links up with thepractice of public policy making. In this practice administrators often deal with theinstrumental question: how to translate intentions into practice. Theories that link upwith such questions have the advantage of being able to bridge the gap betweenadministrative theory and administrative practice. They can, for example, mitigate thetop-down perspective of the administrator, without being so critical that theadministrator loses sight of the relationship between theoretical insights and practice.

170

Chapter 15

AUfOPOIESIS, LEARNING AND GOVERNMENTAL m'EERlNG

J.F.M. KoppenjanJ.A.M. Hufen

1. Introduction

Political theories have for some time portrayed public organizations as omnipotentphilosopher-kings. However, public administrators now find themselves in changingtimes. They are confronted with complex, wicked problems and face dynamic,unwilling environments. The disappointing results of governmental policies aimed atinfluencing societal development create a need for reflection and opportunities fornew theoretical approaches to public administration. Two of these approaches arelearning theories and theories about autopoietic social systems.

It is easy to understand why theories about learning by governmental organiza­tions are quite popular. This approach legitimizes failure and contributes to theoveroptimistic assumption 'practice makes perfect'. As Pressman and Wildavsky (1984,p.245) remark: "Learning is a golden concept, everybody is for it". Less reason foroptimism is offered by the theory of autopoietic social systems. Because socialsubsystems are relatively closed, prospects for governmental steering are not verypromising. Moreover, the political subsystem is an autopoietic system by itself:possibilities for learning from mistakes are not self-evident.

In this contribution the contradiction between the aforementioned approaches, isdiscussed. Can we learn from both theories about the causes of policy failures andabout the possibilities to improve the quality of governmental steering as a reactionto these failures?

The investigation starts with a short description of both approaches. Firstly, wewill direct attention to learning approaches. Hereafter, we will focus on theories ofautopoietic social systems. These descriptions will refer to the explanations for policyfailure these theories offer, to the learning behavior these theories distinguish and tosome critical remarks with respect to the basic concepts of both approaches. Thirdly,the question is raised to what extent a synthesis of both approaches can contribute toa better understanding of restrictions and possibilities of learning by governmentalorganizations with respect to steering processes.

171

2. Learning theories: a hopeful approach to governmental steering

2.1 Learning theories and steering problems

According to learning approaches, policy failures are the result of a lack of informati­on about societal developments and the possibilities of directed change. Knowledgeabout the relation between policy goals and the application and effects of instrumentsis often inadequate. Instruments are used in unstable, unpredictable environments.These environments, which are the focus of steering attempts, pursue there ownobjectives. They can resist government policies or try to take counter measures. Evenif policy-makers use more or less elaborated policy theories about goals and theeffects of policy instruments, these theories may become outdated, owing to learningby parts of their environment (In 't Veld, 1985).

Learning theories point out that detailed policy designs and blueprints fordesired societal developments do not work. Learning theories support the criticism ofincrementalist and bottom-up approaches of the policy process. Instead of using a top­down approach by imposing ambitious and far-reaching policy programs upon society,an incremental approach is advocated. Learning theories suggest a focus on interacti­on between policy-agent and environment. An action of the policy agent results inenvironmental change. Environmental change generates information about the extentto which the policy actor's actions are effective. If necessary, the agent will takefurther action in order to improve his performance. This process of more or lesscontrolled 'trial and error' can be seen as the carrying out of policy experiments (VanVught, 1984). Central in this cyclical conception of the steering process is the gradualimprovement of administrative action.

The feedback mechanism is essential to learning approaches. It is the lack ofadequate feedback that causes major steering disasters and which prohibits learningby public administrators. By pointing to this mechanism, learning theories alsoprovide a hopeful perspective. Firstly, this approach helps to decrease exaggeratedexpectations about governmental steering. Furthermore, it explains how to overcomefailure by feedback and adapted behavior. So, even in the case of disappointing policyoutcomes, public administrators can point to improving performance.

2.2 What learning is about

Learning has been described as discovering and correcting errors (Argyris and SchOn,1978). However, there is no reason why learning only occurs in cases of policy failure.It is sensible to interpret organizations that are doing better than others as learning(or faster learning) organizations. Learning refers not only to correcting failure, butalso to systematic attempts to improve performance in general. Anticipating activitiesthat prevent future mistakes or decline in performance can be considered a moreadvanced way of learning.

Learning is always accompanied by changes in belief systems that underliebehavior. A belief system can be described as a set of basic values, causal assumpti­ons and problem-perceptions (Sabatier, 1988). Learning is not confined to cognitions,but also includes adaptations of goals, norms and values. Especially in more complexforms of learning, the most 'painful' and 'difficult' adaptations to new situations dealwith the acceptance of new goals, norms and values. Problems with learning seem to

172

have the same causes as the so-caned 'future shock', since this concept introduced byToffler refers in the first place to the problem of unlearning traditional norms andaccepting new goals, norms and values (Toffler, 1968; Hedberg, 1981).

2.3 The units of learning: systems, organisations, or individuals?

Learning theories in public administration are frequently 'borrowed' from psychologi­cal theories about learning. Psychological metaphors interpret administrative behaviorin terms of memories, cognitive faculties and learning. These applications of psycholo­gical theories tend to consider organizations as the units that learn. Organizations areconsidered to have an organizational memory and information systems that help toselect stimuli from the environment, and to chose new courses of action. Theprevalent sources of knowledge, through-put capacity and selection mechanismsdetermine the learning capacity of organizations.

If it is true that modern society is an institutional society characterized byorganizations which deal with every aspect of social life, this approach to organizatio­nal learning would imply a pessimistic image of the future: men outlearned by'intelligent' organizations. Doubts about the validity of the applications of thesetheories to explain the administrative behavior of individuals, groups, organizations,networks or social systems as a whole, seem to be legitimate. In the strict sense,organizations do not learn. An organization learns to the extent that its members orgroups of members learn. The structure and culture of organizations are the crystalli­zations of learning experiences and communication patterns of its members. In thisway, organizations are more than their component parts.

Organizations, in their turn, influence, not determine, the behavior of theirmembers. The learning behavior of persons is influenced by the dominant beliefsystem of the organization. The belief system of an organization is more or less stablebecause people experience the existing structure and culture as an investment in theirfuture. A change in the belief system would jeopardize the expected dividend.Pressure from an environment to change the belief system of an organization wouldbe interpreted by members of that organization as an attack against them and wouldtherefore meet opposition by stakeholders within and outside the organization.

The image of an organization as an advocacy coalition of individuals and groupsmeans that learning processes are heterogenous and asynchronous. A learning street­level bureaucrat does not guarantee changes within his agency. Learning at the topdoes not produce enlightenment in all rank and files. Established interest will try toresist change. Learning is accompanied by mobilization and conflict. Only if supportfor change exceeds certain thresholds, will these changes be realized. As a result,learning is not a smooth and continuous process, but disruptive and quarrelsome(Cangelosi and Dill, 1965).

In recent literature the focus of attention has begun moving from organizationallearning to policy-oriented learning, learning within advocacy coalitions that cut acrossorganizational boundaries, and learning within interorganizational networks ('t Hart,Hufen and Van Duin, 1988; Sabatier, 1988) If individuals and groups within anorganization can learn regardless of formal organizational goals, norms and values, itis also possible to learn across organizational boundaries. This type of learning is theresult of interactions between organizations at different levels, overlapping members-

173

hips and exchange of personnel between institutions. Often, members of interactingorganizations have similar professional backgrounds and share a common beliefsystem. Sometimes advocacy coalitions are stronger than organizational loyalty. Forinstance, in the seventies the agency within the Dutch Ministry of Home Affairs,which was responsible for the financial facilities of lower governments, was involvedin legislation, but lacked the necessary skills. Instead of asking legislative specialistswithin the Ministry of Home Affairs, the agency turned for assistance to specialists inthe Ministry of Finance. The ties between groups are not necessarily contingent uponorganizational boundaries.

The shift of interest to inter-organizational boundaries is interesting because thisangle seems to explain a lot of administrative problems. Sabatier can be seen as anproponent of this theoretical approach, since he tried to combine a policy-orientedlearning approach and theories on organizational relations (Sabatier, 1988). Hesuggests focusing on the learning behavior of advocacy coalitions as a part of a policycommunity or subsystem that addresses a policy problem. The replacement of theorganization by the advocacy coalition as a unit of analysis has, at least in Sabatier'sapproach, no repercussions for the way learning processes can be dealt with. Asorganizations, coalitions have a belief system that is affected by learning. Coalitionsconsist of individuals and subgroups with an independent identity that does not alwaysshare the belief system of the dominant coalition in the network.

In this policy-oriented approach additional attention is paid to learning acrossadvocacy coalitions. The focus of attention shifts from learning within organizations orwithin advocacy coalitions, to learning across organizations and across coalitions. Suchan approach seems to be especially promising for analysis of public policy processesand governmental steering in interorganizational networks. It combines 'policylearning' with the popular idea of interorganizational cooperation and struggle. Itoffers a grip on developing a dynamic policy approach that combines goal-directedbehavior of administrators with the limits and possibilities of interorganizationalreality of actors operating in networks.

Thus, the question 'who learns?' can be answered by pointing at three levels ofanalysis with respect to learning behavior: firstly, 'audience learning' by individualsand subgroups within organizations and advocacy coalitions; secondly, learning byorganizations and coalitions; and thirdly, learning across organizations and advocacycoalitions.

2.4 Single-loop, double loop and deutero learning

Several forms of learning can be discerned. Single loop learning refers to instrumentalor goal-oriented learning. Individuals, coalitions, organizations of subsystems learnhow to reach their goals in a more effective and efficient way. It includes the learningof groups or organizations about how to improve their internal, administrativeprocesses. Single loop learning is restricted to the adaptation of peripheral parts ofthe belief system. For example: single loop learning occurs when an organizationsucceeds in improving water quality by adapting its instruments.

Double loop learning concerns learning about goals, values and concepts. Thiskind of learning puts into question the use of standard procedures, the priorities ofcertain problems, the central goals and values at stake, and sometimes even theidentity of learning actors. Double loop learning aims at adapting the very core of the

174

belief system involved. Double loop learning organizations try to rethink and react tosome fundamental threats in the environment. For example: double loop learningoccurs when old problem definitions and instruments are evaluated and exchanged fornew problem definitions and instruments. Learning across borders of organizationsand coalitions will almost by definition be double loop learning.

Deutero learning is an advanced and complex form of learning. Deuterolearning occurs when organizations learn to learn. It deals with attempts to restructu­re organization in such a way that it becomes continually self-adapting, self-transfor­ming. During deutero learning processes, organizations try to define procedures thatmake smooth and rigorous adaptation possible. Deutero learning is not directed atone or some policy problems, but deals with the adaptation potential as such.

If ambitious colleagues, missionary agencies, alarmed politicians, competingagencies or aroused constituencies are learning, it becomes inevitable not to learn, inorder to maintain at least the previous level of performance. Sabatier states: "In aworld of scarce resources, those who do not learn are at a competitive disadvantagein realizing their goals" (Sabatier, 1988, p.151). Learning is evoked by threats, entice­ments and competition. It results in adapting to perceived (future) changes inperformance and environmental change. Survival in dynamic environments demandsactive learning processes within organizations. Not only will organizations have toadapt their technical or instrumental logic, but also their fundamental values asmanifest in belief systems.

2.5 Limits of learning

At first sight learning theories seem to be optlmlstlc about the possibilities to guidesocietal change. Closer inspection points out that theorists on organizational andpolicy-oriented learning are quite pessimistic about possibilities to learn. Organizati­ons or advocacy coalitions tend to resist change. The perception of learning entities isbiased and therefore any need for change is denied. The selection of information isguided by existing goals, values and opinions. The predominant belief system operatesas a filter. This perceptual filter not only remodels unwanted information, but alsoignores information that does not fit within existing categories. Cyert and Marchpointed at the existence of these filters in their study 'A behavioral theory of the firm'which was published in 1963. Political science theories about agenda building empha­size that unwanted issues are organized out of the arena (Schattschneider, 1960;Bachrach and Baratz, 1960; Allison, 1970; Cobb and Elder, 1972; Kingdon, 1984).March and Olson describe weak links in the feedback loops that prevent learning(March and Olsen, 1976), for example the problem that relevant information does notreach responsible decision-makers. In fact these disturbed learning processes arevariations on a familiar theme: bottom-up officials produce distorted information forthe top. Unpopular and disturbing information is filtered out and distorted as a resultof anticipation by lower rank officials. Middle management officials tend to distortinformation in order to avoid the risk of punishment or big policy changes. Accordingto March and Olsen, restrictions of role-definitions, organizational rules and procedu­res limit the learning capacities of individuals. 'Audience learning' occurs whenindividuals and groups within an organization are adapting their belief system, but theorganization does not. 'Superstitious learning' occurs if links between organizationalactions and effects are not properly perceived. The link between environmental

175

signals and the perceptions of individuals is sometimes weak. Individuals and groupswithin the organization perceive different realities. In these cases 'learning underambiguity' takes place.

In research on economic policy it has been suggested that interrupted feedbackloops are artificially restored by pseudo-signals which replace information about 'real'effects. A budget which is spent can be wrongly interpreted as a policy success.'Pseudo signals' are a form of 'superstitious learning'. The implementation practice ineconomic policy shows the existence of relevant information about effects andoutcomes is the exception rather than the rule. The need for relevant information bymiddle management and the top is satisfied by pseudo-signals (Hufen, 1990).

Rosenthal warns us of another pitfall of learning by feedback. The content offeedback may change over time. It is possible to learn too fast. Instead of reactingalmost mechanically to compelling incentives derived by specific recent situations, oneshould try to learn from more general, and less recent experiences. In the case of amajor policy disaster, public discussion focusses almost exclusively upon questionsconcerning political accountability. The inclination is to concentrate on the ritual ofpunishment or even replacement of political leaders. However, the less obviousstructural causes of policy failure are not discussed (Rosenthal, 1989; Ringeling,1989). A summary of the insights of learning theories leads to the conclusion thatthese theories are not at all optimistic about possibilities of organizational and policy­oriented learning. The inclination of organizations, coalitions or networks to adapt themore crucial parts of their belief systems is estimated to be almost zero. Furthermore,there are serious doubts about the effectiveness of one of the most prominent ways oflearning: learning by feedback. In the next section we will investigate what autopoiesistheory may add to this discussion of learning theories.

3. Autopoiesis: An anti-learning approach?

3.1 Autopoiesis and steering problems: vanishing hopes?

Problems concerning directing societal change are, according to the autopoieticperspective, not the result of a lack of knowledge or inadequate applications ofinstruments as learning theories suggest. Failures in governmental steering are theresult of intrinsic features of social subsystems that build up society. Social subsystemsare considered to be closed: there is no input or output. Changes in internal operati­ons of subsystems are not determined by external stimuli. Mechanism such as self­observation, self-description, self-reflection and self-steering replace the idea of thesubsequence: input, throughput and output. This line of thought seems to lead to theconclusion that directed societal change is difficult, if not impossible. The objects ofadministrative action are condemned to interpret governmental initiatives in theirown language. It is not possible for these organizations to understand the meaning ofthe messages of government, even if this would be profitable for them. Furthermore,the political system can be interpreted as an autopoietic system that does notunderstand the reactions of the objects that are to be influenced. Governments arehardly capable of learning from previous experiences and failures. Knowledge of theinternal processes and reaction in non-governmental systems is no more than aninterpretation and reflection of the internal operation of the interpreting system,

176

according to autopoiesis theorists. The basis of autopoletlc thought seems to beexpressed in two key concepts: closure and self-referentiality.

3.2 Autopoiesis and social systems: closure and self-re/erentiality

The theory of autopoieses was developed by the biologists Maturana and Varela inorder to find a way of describing 'living systems'. In contrast with non-living systems,living systems are able to reproduce. In the reproduction process the parts of thesystem do not matter. The specific organization of the relation between the compo­nents as translated in genetic material is duplicated. Therefore, living systems are notdependent upon their environment; they are organizationally independent or closed.In the process of self-reproduction, autopoietic systems refer only to themselves.Environmental disturbances can be the cause of changes in the structure of a system,but these disturbances cannot determine nor direct these changes (Maturana andVarela, 1980; Kratky and Wallner, 1990).

Some autopoiesis theorists have tried to adapt this theory to social systems. Inthese attempts it is presupposed that not only living organisms but also social systemsare relatively closed or independent of their environment. Furthermore, changes insocial organization rather than the environment. Systems use communications in orderto reproduce their own meaning. In this approach communications, not individuals,actions or behavior, are the central units of analysis. In communication, these systemsalways refer to themselves (Luhman, 1984; Weick 1976). Recently, attempts havebeen made to apply autopoiesis theory to the fields of public policy making andgovernmental steering (Schaap, Van Twist and In 't Veld, 1990).

The autopoiesis perspective is used to describe processes of differentiation andspecialization in society. Subsystems in society, such as law, economy, politics andeducation, have developed in the course of time their own languages and rationalities.These subsystems have become increasingly autonomous as a result of ongoingspecialization and differentiation. As a result of this closure and self-referentiality,these subsystems create their own reality. This systematic independency makes itincreasingly difficult to fulfil public functions which depend on the cooperation ofseveral subsystems. Public policy making is like building the tower of Babylon. Inorder to do the job, cooperation is necessary, but a common language which wouldmake communication possible is lacking. At the same time there seems to be somedependency between subsystems to fulfil their tasks.

In this approach feedback loops are non-existent, because subsystems aresupposed to have no input or output. This means that learning by feedback will beimpossible. Objective knowledge about developments 'outside' the subsystem isimpossible. Adaptations of internal processes will be guided by internal, narcissticinterpretations. Subsystems may learn, but these processes are not determined ordirected by external events. Communications between subsystems are impossible as islearning across subsystems.

3.3 Two questions: closure and internal operations

Applications of theories to new fields of investigation must do justice to the crucialelements of these theories. This principle should not be used to safeguard oldertheories against new ideas. The application of autopoiesis theory to public administra-

177

tion give rise to at least two forms of reconsideration or criticism of autopoiesistheories.

Firstly, the concept of closure is inconsistent with the established fact of theinterdependency of actors in subsystems. The functional dependence on externaldevelopments influences elements in every subsystem. The existence of functionalinterdependence implies that communication and interaction between different actorsis possible. In this sense there is input, just as there is output. However, the evaluati­on of these disturbances is determined by internal factors. As a result of this self­referentiality, for the outside world there may be no obvious connection betweeninput and output. As Kickert rightly points out, the concept of 'closure' needsreconsideration (Kickert, in this volume). Self-referentiality seems to be the crucialconcept, not closure. Furthermore, by eliminating the idea of closure, the applicabilityof autopoiesis theory to the field of governmental steering increases substantially.Nowadays closed subsystems are hard to find. However, self-referential subsystemsseem to be omnipresent. Dialogues among the deaf can be found in almost anysection of public administration: between policymaker and policy executor, betweenstreetlevel bureaucrat and client, between generalist and specialist, between politicianand civil servant, between environmentalist and economist, between one section of aministry and another.

Secondly, autopoiesis theory is not very explicit about the way changes ininternal operations take place. Perhaps these internal changes can in some circum­stances be interpreted as single loop learning processes in disguise. This seems tohold when subsystems are trying to adapt to external disturbances. Systems will,according to autopoiesis theory, try to improve their internal operations in order toinfluence the environment. The perception and the action manifest themselves in theframework of the prevailing rationality. However, according to the autopoiesisapproach, components of the subsystems do not matter. This contradicts the idea andwell-documented fact that individuals, groups and organizations within a subsystemare crucial in learning processes. Autopoiesis theorists are confronted with a riddle: ifit is true that subsystems are closed and components within the subsystem do notmatter, where does change come from?

4. Learning approaches and autopoiesis theory: a confrontation

In this section we will compare the concepts in learning theories and autopoleSIStheory. We will discuss some mutual insights as well as some differences in opinion.

4.1 The limits of learning processes

The autopoiesis perspective suggests the concept of learning by feedback is not valid.The idea of feedback overestimates the rationality of the subsystem which intends tosteer and its capability to learn. Policy learning is only possible if there is an inter­subjective language for the participants in the policy process. However, this inter­subjective language does not exist, according to the autopoiesis theory. Learningacross subsystems would require an absent meta-discourse between autopoieticsystems. Such a discourse is inconsistent with the assumptions which are at the basisof autopoietic theory (compare Koppen, in this volume). Attempts to create a

178

communication system which would provide conditions for such a discourse would infact create a further subsystem. Instead of creating an inter-subjective language, thiswould result in more misunderstanding and an even stronger need for an intersubjec­tive language. This means that learning processes are not able to affect the core ofthe communication system itself.

Internal factors will determine the way information about steering effects iscommunicated. The same is true for the way subsystems, which were the object ofsteering attempts, will react. The possibility of learning by feedback is thereforerestricted. These limitations of learning processes are well-known in the literature onlearning. Pessimism about the possibilities of feedback learning is not exclusivelyinspired by the notion of self-referentiality. Ideas about perceptive filters of organiza­tions and advocacy coalitions lead to the same conclusion. The selection of informati­on by a subsystem is guided by internal goals, values and preferences. Unwantedinformation becomes remodelled or is ignored. That is why Dery puts forward thecrucial question: "How can one perceive a premise while observing the world throughthat premise's lenses?" (Dery, 1984, p.98) His answer is plain: one cannot. Argyris andSchOn suggest organizations will not be able to adapt their belief systems by themsel­ves. Change should be triggered from outside (Argyris and Schon, 1978). Autopoiesistheory emphasizes this restriction of feedback learning and tries to explain theimportance of internal factors in the interaction between subsystems.

The limits of learning processes are well-documented in the literature. Aninteresting question is whether autopoiesis theory and learning theories can benefitfrom mutual insights about factors that limit learning processes. Is autopoiesis theorycapable of specifying the rather abstract metaphor of autopoietic systems? Or dolearning theories provide the substance to reconsider or reformulate some ideasconcerning the limits of communication? The limitations of learning processes seemto be common ground in both approaches.

4.2 Selfreferentiality and alternatives to feedback learning

Autopoiesis theories emphasize the limits of learning by feedback. As we argued inthe last section, learning processes may occur even in self-referential subsystems. Onefactor that contributes to learning processes was the availability of alternatives toexisting courses of action, perceptions and values. Alternatives in systems can begenerated by comparing. Self-referentiality does not exclude subsystems comparingthemselves with other systems who are part of their frame of reference. For example,a municipality may compare its performance with an other municipality.

Important changes in public policy are often a result of imitation of successfulperformance in other areas, regardless the effectiveness of existing policies. Thislearning by imitation or analogy may result in uncritically parroting slogans likeprivatization, management by contract, public private partnership. For instance,privatization was one of the trendy priorities of the Dutch center-right cabinetLubbers I that came to power in 1983. However, applying the concept to the producti­on and distribution of passports was all but self-evident. Constitutionally, the state isobliged to guarantee a person's right to travel to other countries. By privatization,fulfilling this right becomes dependent on the strategic behavior of private companiesthat may jeopardize it. This was one of the reasons why the introduction of the newpassportsystem failed (Koppenjan, 1991).

179

Imitation does not always result in invalid analogies and anachronisms. Copyingsuccessful arrangements from other organizations or systems is one of the majorsources of innovation or learning. Feedback may deliver information about ineffecti­veness; it does not generate information about the correctness or desirability ofprevailing perceptions, goals and values. This kind of self-evaluation is stimulated bycomparing the performance of persons, organizations, networks or subsystems withthe way others perform. Alternatives are often arrived at by imitation and spill-overfrom other areas. Both phenomena contribute to internal variation and to possibilitiesto learn and to change (Elias, 1982; Kingdon, 1984).

4.3 Additions to autopoiesis theory

The autopoiesis perspective seems especially fruitful for indicating some mechanismsof learning and criticizing the idea that learning is a self evident and linear process.In so far the perspective can perhaps add some remarks to the long list of inhibitinglearning factors in the learning approach itself. If we know what factors inhibitlearning, we also have a clue about what must done in order to stimulate learning.This list of stimulating learning factors seems to be of great interest to autopoiesistheory. It shows under what conditions operational closure or self-reference is by nomeans the dominant transformation mechanism. Case studies on binding or couplingbetween subsystems seem to be good material to enrich autopoiesis theory. Further­more, the learning literature provides some insight into the circumstances in whichpolicy learning in multi-actor networks does occur. The intersubjective languageseems to depend on some shared mutual interests of actors involved. Discomfortstress or crisis can also be seen an contingent stimulating factors.

The dominance of stubborn or self-reflective organizations does not alwayscount. The relevance of these factors seems to be contingent upon several conditionsthat can be deduced out of empirical research into learning.

5. Conclusion: conditions of learning

At first sight learning theories seem to stress possibilities of learning about govern­mental steering. The autopoiesis perspective appears to point at limitations ofgovernmental steering and learning behavior. A closer examination of both approa­ches proved these prejudices to be wrong. Learning theories are not at all optimisticabout the inclination of organizations and subsystems to learn. On the other hand,autopoiesis theory does not all together exclude possibilities to learn about govern­mental steering.

This confrontation between ideas in learning theories and autopoiesis theoryresults in three conclusions. Firstly, the confrontation between both approaches leadsto the joint assertion that the learning process is not a rational, linear process ofknowledge building. Every organization has the drive to continuously refer to olderwisdoms of the organization rather than to new developments in the environment.This self-referentiality of organizations or systems prevent them from learning byfeedback and learning across subsystems. Empirical research investigates the contin­gent factors that inhibit learning processes. The value of this research for autopoiesistheory is that factors that inhibit learning processes can be identified. It offers some

180

handhold to understand internal processes in systems and organizations.Secondly, autopoiesis theory points to a field of interest that might be relevant

for learning theories as well. According to autopoiesis theory, learning is directed onlyby internal factors. If this is correct, this idea has important repercussions with respectto efforts to improve learning by governmental organizations and policy sectors. Thismeans attempts to create a common language between subsystems are doomed to fail.Instead of this, efforts should be directed at improving learning conditions bystimulating internal variety within systems and external variety in their surroundings.

Thirdly, learning theories raise some interesting questions for adherents ofautopoiesis theory. Learning theory specifies the factors that stimulate policy learning.In other words, learning theories indicate the conditions in which the coupling orlinking of subsystems is possible. This seems to be a fundamental addition to theolder versions of autopoietic theory .

181

Chapter 16

COMMUNICATION AND STEERING:

WHATTIIE SCIENCE OF PUBLIC ADMINISfRATION

COULD LEARN FROM IlABERMASAND LUHMANN

A.R. Edwards

1. Introduction

Social problems are, in general, the result of an interplay of the actions of manyactors who formulate their goals and develop their strategies within different systemiccontexts and lifeworlds. Attempts by the state to steer the conduct of actors often runup against the 'logic' or resistance of these social systems and lifeworlds.Such observations have led to an increasing interest in general social theory

within the study of public administration. Among the most outstanding social theoriesformulated in the last two decades are those of Anthony Giddens, Jurgen Habermasand Niklas Luhmann. These authors combine a formal, meta-theoretical project onthe constitution of social life with a substantive theoretical project on modemWestern society. In their meta-theoretical projects they tried to bridge traditionaldividing lines in social theory, notably between 'objectivist' and 'interpretative'approaches to social action and social order.Central in Giddens' structuration theory is the concept of the 'duality of struc­

ture'. This means "that the structural properties of social systems are both themedium and the outcome of the practices that constitute those systems" (Giddens,1979, p.69). Social practices are the ongoing accomplishment of capable and knowl­edgeable agents who draw upon social rules and resources (structure) as partlyunacknowledged conditions of action. Social practices are ordered across time andspace. Modernity is characterized by the separation of time and space, and by thedisembedding of social systems. Social relations are 'lifted out' from local contexts ofinteraction and restructured across indefinite spans of time-space (Giddens, 1990,p.21).Habermas (1981) developed a communication-theoretical approach to society, in

which systems theory is placed on a foundation of action theory. The core of hisaction theory is a communicative conception of rationality. A formal analysis ofcommunicative action exposes the basic structure of the way in which interdependentactors can arrive at a common definition of the situation. In communicative actioncertain validity claims on truth, rightness and sincerity are implied. Actors draw upon

183

the resources of the lifeworld which are in tum being reproduced in communicativeaction. This is called the 'symbolic reproduction' of society, explained within a action­theoretical framework. Systems theory is used in the account of the 'material reprodu­ction' of society. In modem societies material reproduction is secured by the function­ing of autonomous economic and political-administrative subsystems, coordinated bysteering media (and not by rational discursive processes). These systemic mechanismsare, however, tied to normative institutions, codified in modem law. Central inHabermas' theory of modem society is the thesis of colonization. The functioning ofthe subsystems tends to subjugate the symbolic reproduction, disturbing the communi­cative processes taking place in the lifeworld.Luhmann (1984) offers a systems theory of society. His starting point is the

complexity of possibilities in the world. The most flexible strategy human beings haveto cope with complexity is giving meaning, a strategy of reduction of complexity inwhich the selections not chosen are not destroyed but 'stored', and the selectionschosen are continuously revisable. Giving meaning always manifests itself as coher­ence within (psychic or social) systems. In his account of the constitution of meaning­steered systems, particularly social systems, Luhmann uses the concept of 'autopoie­sis'. The constitution of social systems cannot be understood solely as adaptation ofinternal complexity to the complexity of the environment, but also as self-adaptationto internal complexity. Social systems have to organize and control their internalfunctioning selectively with regard to the relationing of their elements. (The elementsof social systems are communications). The gathering and processing of informationfrom the environment takes place within the specific frame of relevance of the systemin question. This autopoietic view of social systems is used for exposing the conditionof modern societies. Modern societies can be characterized by the primacy offunctional differentiation. AIl 'consequential' communication in society is structuredand channelled along functional lines. Sectors like the economy, politics, educationand science function as 'self-referential' social systems.In this chapter some of Habermas' and Luhmann's notions on societal consensus

building and steering will be discussed. Their views on these matters have manyinterfaces and are highly competing in the way they use systems theory and deal withquestions of societal communication and differentiation. 1 The views of Giddens, whostrongly opposes functionalist approaches and puts Habermas in this respect almoston a par with Luhmann (Giddens 1984, pp.xxxvi-xxxvii)are relatively more influencedby structuralism and lend themselves less to direct comparison for my purposes.2

The aim of our discussion is to investigate what the implications are of theadoption of Habermas' or Luhmann's social theories with regard to the identificationand approach of research questions pertaining to the science of public administration.Our conclusion will be that the theory of self-referential social systems has probablymore to offer for the study of questions of steering. However, Habermas offers aperspective for the conceptualization of certain disfunctions of steering, which isexcluded in Luhmann's line of theorizing. As, on the other hand, Habermas' socialtheory does not exclude a way of theorizing on steering in terms of 'self-referentiality',the adoption of Luhmann's theory (with the implication of dismissing Habermas' one)seems to be the more restrictive choice.Societal steering will be dealt with against the backdrop of societal consensus

building. As Habermas (1985, p.4l5) points out, to be able to speak of the 'impact ofa society on itself, one has to assume, firstly, a 'reflexive center' for (intersubjective)

184

self-observation and, secondly, an administrative system that can act on behalf ofsociety and can act upon it. The conceptualization of a 'reflexive center' centersaround the concept of 'public sphere' or 'Oeffentlichkeit'.All steering conceptions within the study of public administration are still based

on the assumption that steering programs have a certain legitimacy, in the sense ofbeing related in some way to societal consensus building. This assumption has anormative bearing as well, as it is founded in the political thinking of Westerndemocracies. Obviously, the basic conception of steering would have to bereconsidered if the assumption of a 'reflexive center' no (longer) held, as Luhmannsuggests. Equally important are the implications for the study of public administrationof an analysis showing that the administrative system, instead of acting on behalf ofsociety, has developed its own bureaucratic impetus disturbing the processes ofsocietal consensus building, as Habermas suggests.

2. Public sphere and consensus building

In his book 'Strukturwandel der Oeffentlichkeit' (1962) Habermas analyzed the liberalmodel of the civil public sphere as it developed in Western societies at the end of the18th century. This public sphere constituted a specific social domain between theprivate sphere and the state, in which the public of citizens discussed questions ofgeneral interest. The public sphere differentiated itself into a political and culturalpart. In the political public sphere ('Politische Oeffentlichkeit'), a 'public opinion'could originate, as a source of rational criticism and control of the citizenry vis-a.-visthe state. Habermas describes the decline of the liberal public sphere in the lastdecades of the 19th century, but he upholds the normative conception of public life asa sphere of unrestrained exchange of opinions about questions of general interest vis­a-vis the state.This conception of the public sphere was criticized by Luhmann. He stated that

the functional differentiation of society has advanced so far that an integration ofsociety as a totality by public opinions, which are not related to a specific societalsubsystem, has become highly improbable (Luhmann, 1970, p.18). Therefore, 'publicsphere' as well as 'public opinion' have to be reconceptualized at a higher level ofabstraction. The point of reference in his reconceptualization is the problem ofcontingency, the complexity of possibilities in politics. The function of public opinionis the thematic structuring of political communication. In the liberal society of theearly 19th century this function could indeed be performed by citizens, forming smallcircles, meeting each other as 'subjects' (Luhmann, 1970, pp.4-5). In modern societypublic opinion has its place within the political subsystem. The selective function ofpublic opinion is steered by specific 'rules of attention' of the political system. 'Publicsphere', then, refers to a diversity of 'situations' created by the political system for thestructuring of political communication by public opinion (Luhmann, 1970, p.18).Luhmann considered this account as a tentative one. At the end of his article he

warns against an 'exclusive assignment' of public opinion to the political system. Thenotion that public opinion has lost every function for society as a whole is rejected aspremature, awaiting a further elaboration of a general theory of social systems.Questions of intersystem communication and of interdependencies between systemsalso have to be treated as still open (Luhmann, 1970, pp.24-25).

185

In Habermas' 'Theorie des kommunikativen Handelns' (1981) the concept of (politi­cal) public sphere is placed within the framework of the lifeworld/system duality. Thepublic sphere is one of the lifeworld institutions in which the symbolic reproduction ofsociety takes place. In critical discussions the normative limiting conditions areformulated for the functioning of the economy and the state. The conceptualization ofthe public sphere as a domain of rational discourse in societies is grounded on twotheoretical lines of argument. The first line of argument entails a 'horizontal'reconstruction of the rationality embedded in ordinary language use; the second one,a 'vertical' reconstruction of the process of rationalization of societies. On the basis ofspeech act theory, Habermas showed that ordinary communicative action is foundedon the assumption that certain validity claims on truth, rightness and sincerity can beredeemed. If required, each validity claim can be subjected to critical debate. Asystem of social rules can be reconstructed to be followed by discussants in order tofacilitate the resolution of their conflict of opinion (Van Eemeren and Grootendorst,1984). The truth of propositions (or, the effectiveness of policy proposals) is examinedin theoretical discourse; in practical discourse the normative assumptions ofpropositions or proposals are subjected to criticism. The (political) public sphere isthe institutionalization of practical discourse.Against the backdrop of this 'horizontal' reconstruction, Habermas reconstructed

the 'logic' of societal development, which can be characterized as rationalization. Onefeature of rationalization is that the validity of world views is increasingly beingestablished by argumentation and critical debate (and not by tradition). This logic ofdevelopment provides a framework for the analysis of the real dynamics of thedevelopment of Western societies. Both lines of argument together could provide arational basis for normative criticism for those observers who take the normativeideals of the 'Enlightenment' seriously.In modem society the public sphere is a highly differentiated network of different

public spheres: local or supra-regional, cultural, scientific or political, innerparty orassociational, media-dependent or subcultural (Habermas, 1985, pA17). However,these different spheres are not closed to each other: "In Oeffentlichkeiten werdenProzesse der Meinungs- und Willensbildung institutionalisiert, die so spezialisiert siesein magen, auf Diffusion und wechselseitige Durchdringung angelegt sind. DieGrenzen sind durchIcfssig, jede Oeffentlichket ist zu anderen Oeffentlichkeiten hinauch ge6 ffnet" (Habermas, 1985, pp0417-418).Habermas upholds the view that modem societies still have a 'virtual center' for

self-reflexivity: "Dieses Zentrum ist freilich auch eine Projektion, aber eine wirksame(oo.). Auch in modernen Gesellschaften formt sich aus den polyphonen und trubenTotaliti"tsentwu'rfen ein diffuses Gemeinbewusstsein. Dieses lasst sich anhand spezifi­scher Themen und geordneter Beit"fge konzentrieren und Idarer artikulieren; in derhaherstufigen und verdichteten Kommunikationsprocessen einer Oeffentlichtkeitgelangt es zu gr6sserer Klarheit" (Habermas, 1985, p.417).In his recently developed theory of self-referential social systems, Luhmann

rejects any idea referring to identity-formation on the level of a societal totality(Luhmann, 1984, p.629). All 'consequential' communication takes place withinfunctionally differentiated social systems. A 'lifeworld' refers only to communication'in the streets' and the like (Luhmann, 1989, p.36). We can surmise that the reserva­tions expressed by Luhmann at the end of his article on public opinion (1970) nolonger provide any starting-point for bridging his views with those of Habermas.

186

I want to make some further remarks on both views, bearing on the relation betweensystems theory and action theory (underlying the conceptualization of 'public sphere')and on the role of language. In addition to meta-theoretical considerations, I want tomention some crucial empirical implications.

2.1 The relation between systems theory and action theory

To a certain extent the theories of Habermas and Luhmann could be seen asintegrations of systems theory and action theory. What Habermas does is placesystems theory on a foundation of action theory. Berger (1987, p.132) characterizesLuhmann's theory as systems-theoretically reconstructed action theory. What is themeaning of these different characterizations?Kunneman (1990) has emphasized that Habermas gives primacy to action theory

above systems theory, but only on the methodological and meta-theoretical level, noton the empirical level. The methodological primacy of lifeworld is grounded on theview that a system can only be described by an observer in his theoretical language onthe basis of a hermeneutical grasp of the descriptions of relevant actors (for example,in order to establish the relevant system borders). The meta-theoretical primacy oflifeworld entails that what has to be reproduced materially has to be formulated incommunicative processes, drawing on cultural resources of a society. Habermas,however, leaves open the possibility that social development leads to a dominance of'system integration' above 'social integration'. In this perspective, Luhmann's theorycould be incorporated within Habermas' theory, as a theoretical anticipation of ahistorical possibility. But it is exactly on this point, of course, that Habermas criticizesLuhmann for 'deciding on' questions of substantive theory on the level of meta-theory(Habermas, 1981, II, pp.461-462). Berger's characterization of the theory of self­referential social systems as systems theoretically reconstructed action theory can bemisleading. It is important to note that Luhmann's conceptualization of 'social system'differs fundamentally from Giddens' use of the term, notwithstanding Giddens'occasional reference to autopoiesis (1979, pp.75-76). No use is made of the action­theoretical notion of mutual background knowledge of social rules and resourcesdrawn upon by knowledgeable and capable agents. I would prefer, therefore, to speakof hermeneutically adapted systems theory. This version of systems theory is indeedricher than the version used by Habermas. It enables the researcher to go deeper intothe differentiation of societal subsystems in view of their frames of relevance. Thisperspective is potentially very fruitful for questions of steering. Habermas does notoffer many clues for the conceptualization of problems of the effectiveness of steeringmeasures by the state, for example vis-a.-vis the economy, and of the role played bylaw and science in this context. We discuss these questions related to steering furtherin the following section.Here we conclude that the superiority (or perhaps one should say: the relevance)

of Habermas' theoretical perspective on societal consensus building should be deter­mined not only on methodological and meta-theoretical but also on empiricalgrounds.Is it only an 'anticipation of a historical possibility' that Luhmann offers? To what

extent does Habermas' conceptualization of 'lifeworld' obtain in empirical research?An important research task is the identification of public spheres and the theoreticaldescription of discursive processes. In my research into communicative strategies of

187

farmers and environmentalists in a land development project, it was possible toidentify a public sphere on the regional level, activated by one group of youngfarmers pressing for a more democratic planning procedure. The communicativeactivities of this group, which contributed to (using Luhmann's terminology) the'thematic structuring' of political communication within the region, were not interpret­ed along functionalist lines but as symbolic reproduction of lifeworld resources(Edwards, 1990). Broader research into the functioning of public spheres must bedone to show whether the image of a 'differentiated network' as a 'virtual center' ofintersubjective self-observation in modem society is adequate.

2.2 Communication and language

Self-referential social systems are reproduced by communications. The starting pointin Luhmann's view on communication and language is the fundamental unpredictabili­ty of human behavior. As for communication, he distinguishes the improbability thatEgo comprehends what Alter means and the improbability that Ego adopts theinformation communicated by Alter as premiss for his own behavior. The act ofacceptance (or rejection) does not form part of communication itself but is a conjunc­tive act. Luhmann limits the function of language to the level of comprehensibility.The improbability of the adoption of the information communicated by Alter isobviated by communication-media generalized in system-specific binary codes. Theyprovide for linkages between actions ('Handlungsanschlu" sse'), irrespective of consen­sus or dissensus (Luhmann, 1982).The reduction of the role of language to a medium of communication on the

level of comprehensibility is contested by Habermas. But, here again, this is a meta­theoretical discussion linked up with the ontological lifeworld/system duality and theprimacy of the lifeworld concept in that context of theorizing. On the basis of speechact theory, Habermas showed that in the case of communicative action, the act ofacceptance is made through agreement with validity claims, embedded in the speechact itself. These validity claims are backed by cultural background knowledge,assumed by the speaker to be shared with the listener. However, within systemdomains the role of language is reduced to comprehensibility and the coordination ofaction is attained by way of 'steering-media' (money and power). This is explainedagainst the backdrop of the high costs of communication and risks of dissensus incontexts of continuous pressure to act. The steering media code the goal-directeddealing with calculable values and make possible the generalized strategic influenceon the decisions of other participants in interaction, without consensus building bylanguage use (Habermas, 1981, II, p.273).With regard to relations or communication between systems, Habermas concurs

with Parsons' concept of 'double exchange'. According to Habermas, power can beassimilated to money. Therefore, the relations between the political-administrativeand economic subsystems can be conceptualized unproblematically as the doubleexchange of power and money (Habermas, 1981, II: p.256, pp.399-4(0). In the theoryof self-referential social systems, the concept of 'structural coupling' between systemsis an important and interesting one. Eichmann (1989) uses this concept in the contextof intersystemic conflicts. In that case the problem is the effectuation of linkagesbetween actions belonging to systems, between which a fusion of frames of relevanceis impossible.

188

Discursive processes taking place in lifeworld situations are in principle notdistinguished by continuous pressure to act. The public sphere is a forum of ongoingsocietal debate. Indeed, reaching consensus is in no way implied in Habermas'account of discursive processes, but only the following of social rules which makeconsensus possible. A rationalized lifeworld is rather more characterized by criticismand (rational) dissensus, the refusal to accept validity claims, than by consensus.As Habermas points out, differentiation is also taking place within the lifeworld.

Fields such as science, law and the arts develop as separate domains, specializing inone 'rationality' or validity claim. This development leads to fragmentation, by whichordinary practical consciousness is cut off from the results of rational leaming­processes taking place within these domains (Habermas, 1981, II: pp.521-522).Kunneman (1986) has emphasized that one could speak as well of the subjection ofordinary practical consciousness to diverse communicative forms of influence andcontrol by experts functioning in Iifeworld domains. Using Piaget's concept of'reflexive abstractions', Kunneman outlines a 'performative discourse' in which thelegitimacy of certain interactions between the different validity claims are examined,for example, the permissibility of normative points of view within science. The publicsphere can be a forum for horizontal interaction of cognitive, normative and expres­sive points of view. Social movements, representing a program of the 'good life' in theform of a specific constellation of these points of view, have an important role.In my research 'Planning in dispute' (Edwards, 1990), the situation under research

was characterized by the refusal of groups of farmers and environmentalists to acceptthe propositions and proposals formulated by the planning authorities. It was possibleto identify communicative strategies in which agents appeal to the argumentativeacceptability of their validity claims and other communicative strategies backed bypower, money or the instrumental use of language. One group of farmers with landdesignated for nature conservation confronted the planning authorities in purelyutilitarian terms and threatened to use their power as users of the land. Their casegot a hearing at the time that attention rules of the political system were activated. Inthis case the conflict could be depicted as intersystemic (economy-state) communica­tion. A coalition of environmentalists formulated an alternative plan in which a newconception of 'land development' and 'agriculture' was elaborated. This formulationcould be interpreted as a contribution to a broader conception of the relationbetween man and nature in which cognitive, normative and expressive points of vieware interrelated.

3. Systems and steering

Habermas (1981) sees the process of modernization as a twofold process of differenti­ation. There is a process of differentiation within both the lifeworld and the systemand there is a process of differentiation between the lifeworld and the system. Theprocess of differentiation within the lifeworld can be understood as rationalization inwhich discursive processes gradually take over the role formerly played by traditionand religion. Different societal domains develop, each specializing in one 'rationality'or validity claim. The rationalization of the lifeworld is a necessary precondition forthe development of autonomous economic and administrative subsystems. Thesesubsystems are not integrated by discursive processes, but through the systemic

189

mechanisms of exchange and the exercise of power.While Habermas and Luhmann agree that differentiation is a primary characteris­

tic of modem societies, they disagree about the mechanisms underlying it. In Haber­mas' view, the relation between, for example, the state and the sciences cannot beapproached on a par with the relation between the state and the economy. Therelation between the state and the economy is an intersystemic relationship, while therelation between the state and the sciences should be seen both as an (inter)systemicrelationship and as a system-lifeworld relationship. In so far as the modem sciencescontribute to the material reproduction of society, the relationship between the stateand the sciences can be approached in terms of steering. However, science (particu­larly social science) can also contribute to the symbolic reproduction of society, i.e., tothe ongoing societal debate about the definition of social problems and about policystrategies. In this respect, the adoption of a steering perspective could weaken thecritical function of science and support the colonization of the lifeworld.The relationship between the state and the economy can be conceptualized as a

problem of steering. On the one hand, state and economy are self-referential systems,each with their own frame of relevance. On the other hand, both institutions arecoordinated by steering media. Power is assimilated to money. This implies that theeffects of state steering measures directed at the economy can be calculated, providedthat an adequate theory is available of what Teubner (in this volume) calls the'modes of interpretation' of such measures by the economic system. I alreadymentioned Eichmann's use of concepts like 'structural coupling'. He draws attentionto what he calls 'systemic discourses' and contextual intervention (Eichmann, 1990).The theory of self-referential systems provides new and promising conceptualizationsfor theory-building about steering.System/lifeworld relationships (as understood by Habermas), however, cannot be

approached primarily as steering problems. Another approach is needed here, anapproach in which the disturbing effects of bureaucratization and commercializationon public life (and ordinary, private life as well), are taken into account.So, with regard of problems of steering, the science of public administration

operates on a field of tensions. First of all, there is the classical problem of effectiveand efficient steering of the economy and of other sectors contributing to the materialreproduction of society. Owing to the self-referentiality of functionally differentiatedsocial systems, this classical problem has to be approached in a new way. But,secondly, there is the problem of the protection of the lifeworld against bureaucratiza­tion and other 'pathologies' of the state. The field of public administration couldmake a contribution by suggesting possibilities of cautious reforms, which - assuggested by Habermas - should be situated within the 'domain of transition' betweensystem and lifeworld. He speaks about built-in checks and sensors to monitortransactions between the lifeworld and the system. Such cautious reforms shouldprovide room for 'innovative combinations of power and intelligent self-limitation'(Habermas, 1985, p.423). The lifeworld should be able to exert influence on thesystem. This influence can only be exerted indirectly, because direct influence wouldimply the setting up of a new steering subsystem which would itself develop into aself-referential system. However, the exertion of influence should not disturb theessential mechanism of the steering media. Cohen and Arato (1990, p.501) suggestthe establishment of democratic publics within the firm and the state: "Herein lies asense of political and economic democracy compatible with modemity(.... ). The

190

relatively lower level of democracy within economic and state institutions is(... ) theguarantee against dedifferentiating the steering mechanisms, against reembeddingthem in diffuse societal institutions that would eventually be retraditionalized becauseof the constraints of time and complexity."Another suggestion can be derived from a recent study by Bovens (1990) on

organizational responsibility and bureaucratic disobedience. Civil servants faced withserious abuses within their organization could resort to 'bell-ringing' in the publicsphere as an expression of institutional citizenship. This bell-ringing might result inthe necessary 'resonance' within the political-administrative system.

4. Conclusion

Our aim was to investigate what the implications are of the adoption of Habermas' orLuhmann's social theories with regard to the identification and approach of researchquestions pertaining to the science of public administration. We can ascertain that thetwo theories are highly competing in the way they use systems theory and actiontheory and deal with language, communication and societal differentiation. However,on the level of substantive questions on modern society, Habermas' theory seems tohave more 'incorporating potential'. Whereas the identification of disfunctions ofsteering in terms of 'colonization of the Iifeworld , is excluded in Luhmann's theory,Habermas' theory does not exclude an approach of steering in terms of self-referen­tiality. The meta-theoretical debate has to be accompanied with empirical researchbearing on substantive questions on which the two theories have contrary views. Asyet, the adoption of Luhmann's theory (with the implication of dismissing Habermas'one) seems to be the more restrictive choice.

Notes

1. Concerning the use of systems theory, see: Habermas and Luhmann (197 I).

2. A comparative analysis of the views of Giddens, Habermas and Luhmann on the issues of socialaction, social order and 'Verstehen' is offered by Ruebens (1990).

191

Chapter 17

APPLICABILITY OF AUfOPOIESIS

TO ADMINISfRATION SCIENCE

W.J.M. Kickert

1. Introduction

A remarkable number of applications of the concept of 'autopoiesis' has appeared invarious fields of the social sciences. The popularity of 'autopoiesis' in the socialsciences seems to stem from the work of the sociologist Luhmann in the mid-eighties.He used the model of a living system -autopoiesis- which was formulated by Chileanbiologists (Varela, Maturana and Uribe, 1974) to describe how a living system is ableto generate and regenerate its own organization, in the development of a moregeneral theory of self-referential systems centered around the key concept of'communication'. The law theorist Teubner, who was struggling with paradoxes of self­reference in modem methodological approaches, picked up Luhmann's theory of self­referential systems and communication and developed an autopoietic approach to law(Teubner, 1988). The Marxist political scientist Jessop, interested in the problem whythe capitalist system could survive despite its immanent crisis tendencies andcontinuing class struggle, used the autopoietic theories of Luhmann to develop amodel for radical autonomy in societies (Jessop, 1990). The organization scientistMorgan (1986) used autopoiesis in one of his metaphors about organizations. Morgancreated an interpretation of autopoiesis which lead him to intriguing ideas on self­referential closure, ego-centrism and self-reflective evolution of organizations.

This remarkable popularity of autopoiesis in quite a number of highly differentfields of social science, makes it worthwhile to have a close look at it. The modelseems particularly interesting for the advancement of the science of (public) policyand administration because it seems to offer new insights into the self-governance ofsocial organizations. In a time when top-down direct government steering isincreasingly being replaced by self-responsibility and autonomy of social institutionsand theories on central governance from superordinated positions are being replacedby theories on interorganizational networks and non-compulsive cooperation, theneed for theoretical insights into self-governance is apparent. The autopoiesis modelseems to offer these. This model, which is capable of producing, reproducing andmaintaining its organization, seems interesting and relevant for the science of (public)administration.

193

In this chapter the applicability of the autopoiesis model to administrative science willbe discussed. The model has already stimulated some outstanding social andadministrative scientists to creative thinking about analogies and possible implications.The model apparently triggers lateral thinking and seems a rich source of creativity.In order to assess the applicability of 'autopoiesis', it is, however, not sufficient tojudge whether or not the various interpretations are 'creative'. The question thenarises whether or not these ideas have much to do with the original model. In orderto apply a theory to the social sciences that has been derived from the naturalsciences, it is necessary to fully understand the original and to be cautious inconverting and applying it. I had some doubts about the full understanding andcautious conversion when I considered some social scientific 'applications' of thebiological model of a living organism: autopoiesis. As a matter of fact, the originatorsof the model themselves explicitly opposed the idea of stretching the model beyondthe area of biology (Varela, 1981). It is therefore probably wise to start ourconsideration of the applicability of autopoiesis by first explaining what the naturalscientific model originally meant. The primary fields of origin of autopoiesis arebiology and systems theory. Only after this exploration of the home lands of themodel will we discuss the various applications of autopoiesis to different fields of thesocial sciences and consider its usefulness for (public) administration.

The origin of the model in biology, its development in systems theory, and itsapplication in administrative science reflect to a great extent my own career inscience, being a physicist by training, having arrived via control and systems theory ata dissertation in organization science, and thereafter proceeding to the science andpractice of public administration. This enables me to understand both the 'hardscience' aspects of the model and assess its usefulness for the 'soft' administrativescience. It should be kept in mind that, from the viewpoint of usefulness, it is not soimportant whether a useful idea is an accurate translation of the original naturalscientific model, but rather whether the idea is interesting and relevant foradministrative science.

2. Autopoiesis

2.1 The original model

In 1974 two Chilean biologists Varela, Maturana and the systems theorist Uribe,published an article on living systems. They were in search of the essentialcharacteristic which distinguishes such systems as 'living'systems. In other words, theywere in search of the definition of 'living'. Contrary to the prevalent biologicaldefinition of living, the ability to reproduce, they argued that a living organization canonly be characterized by the network of interactions of components which constitute aliving system as a whole. These relations define a complex system as a unity. Theyconstitute its organization. Instead of looking at what makes a living system reproducethe parts of the system, its components, they looked at the organization of the livingsystem that is reproduced. In the manner of true holistic systems theorists, they didnot emphasize the separate parts but concentrated on the whole.

194

In their view, it is not reproduction as such, but rather the reproduction of theorganization of a living system, the network of interactions between the components,which makes it 'living'. In this manner they arrived at their definition of anautopoietic system as a "network of productions of components which (i) participaterecursively in the same network which produced them and (ii) realize the network ofproductions as a unity" (Varela et al., 1974). Such a system is autonomous: itreproduces itself, its parts and its organization. An autopoietic system isorganizationally closed with no apparent inputs and outputs. This is in contrast to an'allopoietic' system, such as a mechanism in which the product of the operation isdifferent from itself. An allopoietic system is open, receives inputs and producesoutputs.

An autopoietic system is an autonomous and self-sustaining unity with a networkof component-producing processes such that the components in interaction recursivelygenerate the same network of processes which produced them. The process ofcontinuous reproduction of the systemic unity is called autopoiesis. In summary, theoriginal model of autopoiesis has the following essential characteristics:

It is a definition of what makes a living system 'living'.It is an extension of the usual definition of living - a living system is capable ofreproducing itself - in a holistic systems direction. Not the parts, but theorganization of the living system which makes it a unity, a whole, is whatmatters.Living means the reproduction of the organization of the living system, thenetwork of interactions between its components.An autopoietic system is a autonomous unity which consists of a network ofcomponent-producing process such that the interactive components recursivelygenerate the same network of processes.Such a system is organizationally closed with no apparent inputs of outputs.

The original 1974 English paper on autopoiesis was based on a book by Maturanaand Varela written in 1973 which was later translated from Spanish into English(Maturana and Varela, 1980). This remarkable paper, which attracted wide attentionin systems theorists' circles, was followed by a number of publications by the sameauthors (e.g., Maturana, 1975,1981; Varela, 1979,1981; Uribe, 1981) in which variousaspects of the model were further clarified and elaborated. In particular, their articlesin the volume edited by Zeleny (1981) emphasize how autopoiesis should not beinterpreted. According to Maturana (1981), it is not identical to autonomy or to self­organizing. In his treatment of organizational closure - recursively organized systems ­Varela (1981) stresses the limits of autopoiesis: the concept should not be usedbeyond the field of biology. Apparently the originators of the model were quitereluctant about its wider applicability.

2.2 Autopoiesis and biology

The illustrative model of an autopoletlc system that Varela et al. (1974) used clearlyresembles the model of a biological cell. Their simple simulation model of autopoiesisconsists of a substrate in which a catalyst is capable of producing a link which iscapable of bonding with other links or can disintegrate into substrate. The computersimulation shows the spontaneous generation and reproduction of chains of bonded

195

links around a catalyst. Biologically stated, it is a cell in which a catalyst nucleusgenerates and reproduces a membrane around itself.

Genetic theories from biology about cells, DNA molecules and genes haveundoubtedly contributed to the invention of the autopoietic model. A single gene in aDNA molecule contains everything necessary to generate and reproduce a certainproperty of a living being. It is however doubtful whether genetic biologists wouldconsider the autopoietic definition of 'living' new. In genetics, reproduction haseverything to do with structure. A cell has a nucleus. A nucleus containschromosomes. A chromosome consists of strings of DNA and these DNA strings arecomposed of genes, which code certain properties of the living being. A gene is aparticular order of certain nucleic acids. The basic coding element for the propertiesof a living being, the gene, is thus defined by its particular ordering, in other words,its structure or organization as a unique unity. Reproduction has to do withstructurally composed wholes. This is common knowledge to biologists, not a newcontribution of autopoiesis.

2.3 Epigons of autopoiesis in systems theory

More relevant is the model's implication for the advancement of systems theory, itssecondary field of origin. In the late seventies the idea of autopoiesis stimulated anumber of systems theorists to further thought and reflection. The mathematicalelaboration of autopoietic systems, using the automata theory of computer science,received relatively early attention (Zeleny, 1977; Goguen and Varela, 1979). Theoverenthusiastic sweeping remarks about the wide applicability of the concept tohuman relations are mostly 'soft talk by hard scientists'. In fact, these remarks areexamples of the exaggerated pretentions of universal applicability system theoristswere often blamed for, a view which has considerably contributed to the decline inappreciation of the systems theoretical 'paradigm' in the administrative sciences sincethe seventies. In some systems theoretical circles the development of the model ofautopoiesis was welcomed as a possible breakthrough which could restore thedeclining esteem for systems theory in the 'applied' sciences.

Apart from mathematical elaborations of autopoiesis - which for our purposeare neither interesting nor relevant - the model has stimulated other systems theoristsas well. In the volume edited by Zeleny (1981) a number of articles which extend theoriginal model are noticeable.

Jantsch (1981) extends self-reproduction outside the realm of living. Heformulated a general dynamic systems theory with 'dissipative self-organization' as theunifying perspective. Jantsch was afraid that autopoiesis would otherwise becomesterile. He distinguished two forms of self-organization. The first consists ofconservative attractive and repelling forces leading to stable equilibrium. The second ­called 'dissipative structures' - is a self-organizing dynamic order maintained bycontinuous exchange of energy with the environment, spontaneously arising in far­from-equilibrium conditions, establishing a certain autonomy. Continuous autopoiesisdepends on the continuous maintenance of nonequilibria. Jantsch's interpretation ofautopoiesis - called a 'misunderstanding' by Maturana and Varela - related to notionsfrom chaos theory, e.g., about 'order out of chaos' (Prigogine and Stengers, 1985)which we will discuss later on. In the same volume Ben-Eli (1981) took a much moreclassical cybernetic stance towards autopoiesis. Equilibrium, stability, regulation and

196

control were key notions in his elaboration of autopoiesis into a notion of evolution.The concept of self-organization was seen as a basis for the idea of stability in viabledynamic systems. A self-organizing system consists of both the idea of consistentidentity and the idea of dynamic variations essential for its continuous viability.Notice that in the same volume Maturana (1981) states that the invariance oforganization being the essence of an autopoietic system, such a system cannotundergo a change in its organization, that is, the system cannot be 'self-organizing' inan evolutionary sense.

Elsewhere (Geyer and Van der Zouwen, 1986) other renowned systems theoristspresented their ideas about autopoiesis. Laszlo (1986) elaborated on the cyberneticsof social evolution. He distinguished between stability and 'resilience', the latter beingmore important for the long-term persistence of ecosystems. Resilience is the abilityto absorb change and disturbance within a wide range of large and unexpectedfluctuations. A system might be stable for small deviations, but unstable for largeones. The foregoing incomplete sample of systems theoretical publications related toautopoiesis clearly shows that even in its own field of origin, elaborations andapplications of autopoiesis use broad interpretations of the original model, the validityand the accuracy of which is contested by the original authors. How does this relateto the introductory statement that in order to apply a natural scientific model one hasto be cautious in converting and applying it? If one accepts the self-imposedrestriction by Varela that autopoiesis should not be used beyond biology, ourdiscussion about application and transposition is pointless. The same holds if therestriction of mathematical formality of the model is imposed. If one drops biologyand mathematics, the essentials of the model come down to the self-reproduction ofthe organization of the autopoietic system.

But what if even that essence is dropped and replaced by the much broader ideaof self-organizing? A self-organizing system does not necessarily invariably reproduceone and the same organization. Ben-Eli (1981) and Laszlo (1986) use self­organization as an idea for their theories on evolution. Evolution implies a long-termchange in organization. Jantsch' (1981) ideas on order out of fluctuations are evenless compatible with regeneration of the same organization. Autopoiesis then seemsreduced to some very broad idea about self referentiality, and then shows strikingsimilarities with certain modem methodological approaches in the social sciences. Aswe will see from the following discussion on its applications in social science, that isindeed more or less what is left of the cautious transportation of the original modelof autopoiesis.

3. Autopoiesis in social sciences

3.1 Luhmann's aUlopoiesis of social systems

After the popularity of autopoiesis amongst systems theorists in the late seventies andearly eighties, the concept broke through in the mid-eighties in the social scienceswith the publications of the renowned German sociologist Luhmann (1984, 1986). Inhis book on 'social systems' Luhmann (1984) introduced the idea of a paradigmaticshift in systems theory from the holistic notion on parts and whole, via the distinctionof system and environment, towards a theory on self-referential systems. In the

197

constitution of their elements and operations systems refer to themselves. They areself-referentially closed. This new theory of self-referential systems emerged fromtheories on self-organizing systems developed in the early sixties. Autopoiesis is oneof the more recent systems theories he used.

He extended autopoiesis, a biological model of living systems, to the realm ofsocial system by interpreting autopoiesis as a form of system-building using self­referential closure (Luhmann, 1986). In Luhmann's view, communications are thebasic elements of social systems. They are recursively produced and reproduced by anetwork of communications. Communication unities consist of a synthesis ofinformation, utterance and understanding. This synthesis is not produced by someoutside force like language, but by the network of communication itself. It requiresself-reference. Luhmann saw the maintenance of social systems as a self-referentialproduction. Autopoiesis presupposes a continuous need for renewal. Continuingdissolution of the system becomes the cause of its autopoietic reproduction.Maintenance is not a matter of reproducing the same patterns in similarcircumstances, but rather the production of subsequent elements different fromprevious ones. According to Luhmann, an important contribution of autopoiesis wasits epistemological consequence of distinguishing autopoiesis and observation,observing systems being themselves autopoietic systems.

Luhmann's interpretation of autopoiesis can indeed be characterized as verybroad. Self-referentiality is broader than self-organization, which is broader than self­reproduction, which is what seems left of autopoiesis if one drops biology. Moreover,Luhmann uses self-referentiality primarily as a methodological analogue for hisepistemology. So, autopoiesis has in fact become an abstract philosophy. Furthermore,one has to realize that the notion of communication plays a much more central rolein his social systems theory.

3.2 Teubner's autopoietic law

In his development of a constructivist epistemology of law, Teubner (1988, 1989)referred extensively to Luhmann's epistemological interpretations of autopoiesis. Inhis journey towards an epistemology centered around 'social construction of reality'and 'decentering of the subject', Teubner discussed the poststructuralism of Foucaultand the critical theory of Habermas. In his view, both Habermas and Foucaultbecome entrapped in paradoxes of self-reference, be it infinite regression, circularityor tautology. Luhmann's theory of autopoiesis offered him a way out in the sense thatautopoiesis does not treat self-referentiality as a paradox to avoid, but as its usefulessence. Autopoietic systems are based on that very self-referentiality. Teubnerderived a theory of law as an autopoietic system, a network of communications thatrecursively reproduces itself. Following Luhmann, he considered law to be made up oflegal communications. These do not derive their meaning from some outside 'real'world. The outside world itself is a social construction by the law. Persons are mereconstructs, semantic artifacts produced by the legal discourse itself. Autopoiesisprovides Teubner the epistemological basis to support these theses.

Note that Teubner primarily used Luhmann's interpretation of autopoiesis as anepistemological foundation. Hence, my previous comments on Luhmann's interpreta­tion also apply here. The level of abstraction reached by Teubner seems light-yearsremote from the original model of autopoiesis.

198

3.3 Jessop'sautopoiesis in politics and the state

Luhmann's autopoietic ideas on communicative processes also formed the basis ofJessop's (1990) ideas. Jessop viewed autopoiesis as a condition of radical autonomy.An autopoietic system defines its own boundaries, develops its own unifying code,reproduces its own elements, obeys its own law of motion. It is not controlled fromoutside, there is no superordinate governing centre, the interactions between differentautopoietic systems lead to blind co-evolution. Referring to Luhmann, Jessopremarked that the political system in modern societies is no longer a paramountpower, it is differentiated and involved in complex interdependencies with otherautonomous autopoietic systems. It is hard to maintain the myth that the state cansolve societal problems. The state is a collective fiction, a semantic artifact, whichserves as a self-reference of political responsibility in a complex power circuit. Jessopgave the example of welfare politics as a self-referential and closed discourse.

Jessop also discussed the problem of societal guidance. If social systems areradically autonomous autopoietic systems, what is the unity of a society? A societywould then be less than the sum of its parts, a series of self-closed systems engaged ina blind co-evolution. Jessop searched for a solution in the idea of 'structural coupling'between mutually indifferent systems which co-exist and co-evolve in the same eco­system. Such a mechanism could explain how radically autonomous systems can co­evolve and develop into stable, coupled blocs.

Assessing the merits of (Luhmannn's) autopoiesis from a Marxist perspective,Jessop concluded that the notion of operational self-closure of an autopoietic systemand the mechanism of structural coupling between autopoietic systems have shed newlight on the crucial concept of autonomy and its relativity. Autopoiesis serves tosupport, on the one hand, the view that society cannot be organized and controlledfrom a single superordinated center - contrary to usual Marxist notions of class ­because it consists of autonomous part systems and, on the other hand, helps torecognize that there are stable networks in co-evolution in society.

As Jessop also only refers to Luhmann's interpretations, a report of earliercomments is unnecessary. From the viewpoint of usefulness for administrativesciences, Jessop's interpretations are more in line with modern administrative theorieson the limitations of central government steering, the departure from a monocentrictowards a polycentric theory of the state, the recognition that policy-making usuallyconsists of a network of nearly autonomous actor, interorganizational networktheories, etc.

3.4Morgan's logic of self-producing systems

One metaphor in Morgan's (1986) well-known collection of metaphors oforganizations one is that of the organization as 'flux and transformation'. In search ofthe logics of transformation and change for the basic dynamics that generate andsustain organizations, Morgan explored three different images, one of which was theidea of autopoiesis. In contrast to the aforementioned authors, he derived his ideasdirectly from the original publications of Maturana and Varela. Well aware of theirstrong reservations about applying a biological model in the social world, he uses itonly as a metaphor. Morgan introduced the idea of autopoiesis as a basic challenge tothe traditional approach in organization theory in which an organization is typically

199

viewed as an open system in interaction with its changing environment, which has tobe adaptive to survive environmental challenges. An autopoietic system, on thecontrary, is organizationally closed, autonomous and only self-referential. Autonomy,circularity and self-reference are its key features. This opens radically differentperspectives on the relationship between organization and environment. In anautopoietic system relations with the environment are internally determined. This hasimportant implications for the maintenance of the organization's identity. Morgan'screative interpretation of autopoiesis led him to three intriguing conclusions. The firstis self-referential closure, the attempt by organizations to interact with theirenvironment as projections of themselves. The second he called 'egocentrism', theattempt by organizations to try to maintain their own identity against a threateningoutside world. Organizations become preoccupied with and overemphasize theimportance of themselves, underplaying the significance of their wider context. Thethird is self-reflective evolution, the process of organizational change as an evolutionof self-identity in relation to the wider world.

Although Morgan regards the theory of autopoiesis as a major challenge tounderstanding organizations, he also repeatedly underscores for its dubiousimplications. Egocentric corporate cultures may be successful in the short run, butoften do so at the expense of their context, and run the danger of destroying thewhole. Morgan derived as the motto of this approach, "think and act systematically:more self-reflection, less self-centeredness'.

4. Autopoiesis reconsidered

Although the above mentioned interpretations of autopoiesis from the respectivefields of application undeniably led to synergetic surplus values, one should realizethat the various translations and uses of the model of autopoiesis were indeed rathercreative. The transpositions differ so substantially from the original that one mightwonder if they really were correct applications at all. Although it can be argued thatfrom the viewpoint of the field of applications, it is not so important to ascertainwhether a good and useful idea is an accurate transposition of an original idea fromthe natural sciences, the focus of this paper is however also to assess the applicabilityof the autopoietic model. In this chapter we will therefore first return to the questionwhat essential innovations autopoiesis contributed to systems theory. What is thesurplus value of autopoiesis and what are its limitations?

4.1 System maintenance, stability and autopoiesis

Maintenance is at the heart of the concept of 'homeostasis', the cyberneticcounterpart of the control engineering concept of 'servomechanism'. A homeostaticsystem is capable of self-maintenance, restoring from disturbances. Not only intechnical engineering does stability play a crucial role, but in many other fields ofboth natural and social sciences as well. It should not be hard for political andadministrative scientists to imagine situations where one would be interested inknowing whether a particular system will be able to survive, to maintain itself in themidst of disturbing events such as a crisis, or in fact will disintegrate.

200

The concept of 'stability' is defined as the capacity of a system to regain an initialstate after a disturbance. The formal definition of stability is that the differencebetween initial and actual state will in time become zero. This formal definitionconsists of two elements, deviation and time. Using these two elements, it is possibleto distinguish between several types of stability. The time factor provides thedistinction between short and long term stability. Although the statement 'in the longrun we will all be dead' holds true, people might be interested to know what is goingto happen in the short run. The factor deviation provides the distinction betweenstability and equilibrium, and the distinction between a deviation which is 'near' anequilibrium state and one which is 'far from' an equilibrium state. An equilibrium isnot necessarily stable when disturbed. A ball on the top of a hill will roll down thehill when touched. Moreover, it makes a difference to what extent an equilibrium isdisturbed. The system might be stable for small deviations but unstable for large ones.In other words, the system can be stable for a limited range of deviations. It is thiskind of distinction which Laszlo (1986) used in pointing out the difference betweenstability and 'resilience', the ability to absorb change and disturbance within a widerange of large and unexpected fluctuations.

The innovative aspect of autopoiesis, given the tradition in the thinking aboutmaintenance, continuity, stability and so forth, is that it models the self-maintenanceof the organizations of the system and not of the variables, the elements of thesystem. The autopoietic model pictures a system that generates and regenerates itsorganization. The model shows that a certain structural configuration is generated andthat once it exists, notwithstanding perturbations by the rise or decay of elements, theconfiguration is maintained. It is not the usual model of stability of system variables,but a model of organizational stability. It is a stability model at the next higher(meta)level of the organization of the system.

4.2 Organizational closure

A serious limitation of the original formal model of autopoiesis (Varela et al., 1974)is its notion of closedness, albeit closure on higher meta-level of organization. Theexpose on system stability given above made it clear that stability should not beconfused with closeness of a system. It is the very existence of environmentaldisturbances to the system, that is, the openness of the system, which gives rise todiscussions of stability. And yet the model of autopoiesis unfortunately suggested thatorganizational stability had something to do with closure. A closed system has noinputs from and outputs to its environment. It does not interact with the environment.Indeed the original formal autopoietic model lacked inputs and outputs. All elementsneeded to generate and reproduce the system existed within the system.

Later clarifications and interpretations of autopoiesis and its organizationalclosure have loosened this restriction. Varela's (1981) explanation of organizationalclosure no longer suggests that strict closeness. Jantsch (1981) goes much further andconsiders the exchange of matter and energy with the environment, the openness ofthe autopoietic system, as crucial. His interpretation of self-organizing in terms ofdissipative structures is that of the organization of a dynamic order in the midst offluctuations in far-from-equilibrium conditions. Luhmann (1986) stated that the maincontribution of autopoiesis is that it had to be understood as the recursively closedorganization of an open system. It does not return to the old notion of 'closed' versus

201

'open' systems. It links closure and openness. In Luhmann's view, the theory on self­referential systems represents a step forward from the old systems theoreticaldistinction between 'open' and 'closed'; self-referential closure is only possible in anenvironment, it produces openness (Luhmann, 1984).

Indeed in social reality closed systems do not exist. Most social systems are bydefinition open. In the political system a ministry cannot isolate itself from itsenvironment: parliament, interest and pressure groups, social institutions, citizens, etc.This also holds true for the organization of a ministry. The organization of a ministryis a reflection of the needs, interests and institutions in a particular policy area. Strictorganizational closure is an unrealistic notion.

4.3 Structural meta-stability and chaos

An autopoietic system is a model with organizational stability, that is, stability at themeta-level of the structure of the system. Whatever the perturbations are at theobject-level of the elements of the system, at the meta-level the system maintains itsstructure. In general broader terms, at the object-level the system can even beunstable, while at the meta-level the organization of the system remains stable. Chaoscan reign in an organizationally stable system. To a certain extent this can actuallyhappen in biological reality. Some living organisms can be wounded or cut in parts,and will nevertheless recover. Elements of the organism can 'die' but the organism asa whole 'survives'. This idea of the coexistence of chaos and structural meta-stabilityopens interesting perspectives analogous to the intriguing eye-openers which Gleick(1988) vividly describes in his best-seller on 'chaos'. Scientists usually study linear'normal' systems, where small perturbations have predictable 'linear' effects that arecontrollable. Chaos is usually considered 'weird' as a category of non-solvableproblems.

As long as a dynamic system can be modelled by linear differential equations,problems can be solved but non-linear differential equations cannot usually be solved.Using many illustrations from the fields of meteorology, ecology, astronomy, etc.,Gleick pointed out that there can be regular patterns in chaos, that systems which atfirst sight appear utterly chaotic, if studied in a different way, might reveal a certainorder, a structure. Instead of studying relatively undisturbed 'linear' developments,one might obtain intriguing insights by studying the possible structure in utterlychaotic 'nonlinear' phenomena. Search for 'stability' at a meta-level of a system whichis seemingly 'unstable' (chaotic) at the object level.

Autopoiesis is an example of stability at the meta-level of the organization ofthe system. Let us take the parallel with 'order out of chaos' (Prigogine and Stengers,1985) one step further. One of the interesting discoveries in chaos theory is the'bifurcation diagram' which describes a system which has a single steady state up tosome parameter threshold, and possesses two steady states above that threshold.Order out of chaos does not necessarily have to be constrained to one single order.Chaotic systems can reveal different types of 'order', different '(meta)stabilities'.Translating this into terms of organizations and structure, creates an interestingextension of the idea of autopoiesis. In an autopoietic system the system alwaysgenerates and regenerates one and the same 'organization'. A 'bifurcation'-like systemcan generate and regenerate several different 'organizations' in the midst ofperburtations and chaos. The multi-stable self (re)production by the system of various

202

organizations avoids the negative connotation of the mono-stable 'autopoiesis' with'dynamic conservatism' of government which always regenerates the bureaucraticorganization and therefore perpetuates itself (this is discussed later on). The multi­stability conception combines the attractiveness of order out of chaos with thedesirability of a pluralist order.

5. Autopoiesis and administration science

After the previous theoretical reconsideration of the model of autopoiesis and itspossible implications, let us now end with some attempts to perform a more or less'literal' transposition to the science of (public) administration. The first attempt turnsout to have rather controversial implications in public administration. The secondpossible application points in another direction.

5.1 Dynamic conservatism of bureaucracy

One possible translation leads to an ultimate conservatism of bureaucracy, or moreprecisely, to dynamic conservatism. That is a system which possesses the valuedfeatures of flexibility, adaptiveness, changeability, but uses them only to preserve itsown existence. Change of policy objectives, change of functions and instrumentalchange all preserve the existence of the institution. The continuance of the institution,the organization, is the ultimate or only goal. Whatever happens, be it policy changeor budgetary cutbacks, the bureaucracy will never die (Kaufmann, 1976). Theautopoietic model in fact describes precisely that dynamic conservatism. A systemgenerates and reproduces itself, particularly its organization. In translation, theautopoietic model would show government institutions continuously reproducing theirorganization, that is, the bureaucratic form of their organization. The application ofthe concept of 'living' public administration would seem to suggest that thebureaucracy possesses perpetual existence. It can be expected that the model willtherefore hold a strong attraction for those individuals with a dislike for or suspicionof bureaucracies.

The objections given above do not imply any criticism of the model ofautopoiesis. If the choice is made to describe bureaucracy as a self-reproducingsystem, the model cannot be blamed for that characteristic. A model is merely arepresentation of some elements and interactions in reality into a formal system ofvariables and relations. The model cannot be held responsible for the hard facts ofreality.

The negative interpretation of autopoiesis seems related to the strict interpreta­tion of the organizational closure of the system. If interaction with a turbulent outsideworld, the essence of self-maintenance, is added to the model, the interpretationbecomes substantially more positive, putting it in line with Morgan's (1986)interpretations. Note that Morgan also recognized the dangers of self-centeredness oforganizations. If the autopoietic model is extended to a multi-stability model of 'orderout of chaos' (see discussion before), the objection of conservatism also vanishes.

The presumed perpetuity of government organizations has received attention inpractical politics and in administrative science. Kaufmann's analysis showed that in aperiod of some fifty years almost no government organizations had ceased to exist,

203

which prompted him to state that they seemed to be immortal (Kaufmann, 1976). Inpractical politics this presumed perpetuity led to the introduction of time limit clausesin policies and regulations. Examples are 'sunset laws' which expire on apredetermined date. Research on sunset legislation showed, however, that thepresupposed expirations often did not take place and that continuations of supposedtemporary regulations were the rule and not the exception (De Leon, 1978; Behn,1978). In an attempt to refine the thesis on immortality, Hogwood and Peters (1982)made a distinction between policy maintenance, termination, succession andinnovation. Real policy innovation or termination does not take place as often aspolicy succession - a new policy, program or organization for the existing sameproblem. Apparently the 'living' of government organizations cannot be regarded as abinary property (either continuation or termination).

Another way of looking at 'living' organizations worth mentioning here is thestudy of births and deaths of organizations based on the notions of population ecologyin biology (Me Kelvey and Aldrich, 1983). Here too 'living' is considered a binaryproperty: an organization is either alive or dead. In reality gradual change ingovernment organizations occurs much more often than the formation of entirely neworganizations (birth) or the total abolition of organizations (death).

5.2 Autonomy arul self-governance

A second possible application of autopoiesis is as a model for the autonomy and self­governance which many societal systems appear to possess. The idea that governmentis able to exert direct control over the course of affairs in societal sectors, the idea ofgovernment steering society from a powerful control position above and apart fromthe rest of society, has been shown to be unrealistic (Den Hoed et al., 1983).'Government is not the almighty controller of social processes. The control capacity ofgovernment is limited for a number of reasons: lack of legitimacy, complexity ofpolicy processes, complexity and multitude of institutions concerned etc. Governmentis only one of many actors that influence the course of events in a societal system.Government does not have enough power to impose its will on other actors. Othersocial institutions are to a great extent autonomous. They are not controlled by anysingle superordinated actor, not even the government. To a great extent they controlthemselves. Autonomy not only implies freedom, it also implies self-responsibility.Autonomous systems have a much larger degree of freedom of self-governance.

Deregulation, government withdrawal and steering at a distance (Kickert, 1991)are all notions of less direct government regulation and control, and lead to moreautonomy and self-governance for social institutions. The need for theoreticalconcepts on autonomy and self-governance is self-evident. The autopoiesis model canbe considered the next step in the direction of a theory of system autonomy and self­control. Control is crucial in cybernetics. Cybernetics was originally defined by Wiener(1948) as 'the science of control and communications'. Regulation and control are itsbasis themes (Ashby, 1956). All 'hard science' approaches to control assume thehierarchical notion of a controller which controls a controlled system which interactswith an environment. The controller receives information from the system and sendscontrol measures to it.

Hierarchical control often does not exist in social systems. Several entities haveinfluence. They influence each other, and possess a degree of autonomy. The

204

controllability of the system is restricted; there is no unilateral control relation and nosingle controller above others. The subsystems are capable of self-control. The notionof a network of several nearly autonomous actors with different and often conflictinginterests, in which no single actor has enough power to dominate the others and inwhich decision-making is a bargaining process and decisions are compromises, isfundamentally different from the mono-rational model of governance by a singlecontroller which adjusts the subsystems in order to reach a collective objective (Hanfand Scharpf, 1978; Kickert, 1980).

The contribution of autopoiesis is that it not only describes how a network ofautonomous (autopoietic) systems behave, but also how it is able to maintain itsorganization, its identity. Organizational meta-stability is possible. Notice that apartfrom ideological differences, this is also what intrigued Jessop (1990) about the notionof autopoiesis. The ability of a network of autonomous systems to maintain itself andits basic order when confronted with disturbances, and its ability to survive in aturbulent and complex environment, is a vital quality in an administrative situationwhere direct top-down control by government is being steadily replaced by self­governance autonomy of social institutions. It is reassuring to know that theseinstitutions might be able to survive hostile storms they may encounter.

6. Conclusions

The central ideas of the autopoietic model of living systems can be used in a numberof different ways to create concepts and models useful in the social and administrativesciences. Reasoning by analogy is a perfectly legitimate source of creativity. Thevarious creative interpretations of autopoiesis reviewed in this article led toremarkably interesting and relevant ideas. From the social science point of view, onecan be reassured as to its usefulness. The question is whether or not these interpreta­tions have much to do with the original model of autopoiesis.

In this chapter the focus of interest was on the applicability of the autopoiesismodel to the administrative sciences. Applying a natural scientific model to a socialscience is hazardous. Full understanding of the original model is necessary if a correctconversion is to take place. Caution when applying the model is recommended. Thischapter included an extensive description of the original model of autopoiesis and adiscussion of the implications of the model for the systems theoretical concepts ofstability. After an exploration of the home lands of the model, an extensive review ofthe various interpretations of autopoiesis in different fields of the social sciences waspresented. Finally an attempt was made to perform a careful translation and cautiousapplication.

The overall conclusion is that the possibilities of a strict conversion of theautopoiesis model into a valid model that can be used in the administrative sciencesare limited. The usefulness of the model does not seem to lie in strict adherence tothe original and literal translation, but rather in its power as a source of creativelateral thinking. It can inspire highly interesting and relevant ideas. From the socialand administration science point of view, it is not so relevant whether or not aninteresting and useful idea is an accurate translation of some natural scientific model.Finally, note that a 'hard science' model cannot be blamed for not accurately fitting a'soft science' world.

205

Chapter 18

SIMULATION: LEARNING ENVIRONMENTS FOR

(SELF-)STEERING IN SOCIAL SYSTEMS

LH.G. Klabbers

1. Introduction

In the preceding chapters steering, autopoiesis and configuration were dealt with mainlyfrom a theoretical viewpoint. In this chapter several related methodological issues forexperimenting with notions of (self-)steering in social systems will be discussed. 'Socialsystem' will be used as a generic term for various types of human organizations.

In human organizations images playa potentially powerful role in directing themtoward actions. They help in making sense of ongoing activities, and in arranging supportfor dealing with change. They symbolize organizations in many different ways. Logos arespecific images indicating how an organization wishes to express its identity, i.e.,how itrefers to itself both as an organization and as a system of meaning. In an organizationsymbolic interactions between departments and individuals bring forward systems ofmeanings that influence its evolving self-image. This process of self-referencing may leadto operational closure in terms of meaning-processing, as well as in coordinatingorganizational activities.

Self-steering capacities are inherent to any social system as an organization.Whether a social system strives for a shared identity in terms of a system of meaning asa necessary condition for (self-)steering remains to be seen. Autopoiesis, configurationand (self-)steering provide a domain of knowledge that presupposes a continuousswitching of perspective, from the outsider/spectator to insider/participant. As pointedout by Van Dongen (in this volume), social integration and steering are based on actors,boundaries, cognitive and socio-structures. Through interactions and sense making, actorsdevelop possible worlds and multiple realities. Van Dongen makes a distinction betweenanalytical theory and social constructivism. Both approaches represent different scientificcultures. Analytical theory, in stressing a spectator's perspective of making distinctions,leads to a notion of steering of social systems based on a rationalist conception of reality.Social constructivism on the other hand, referring to distinctors, stresses a historicistviewpoint. It is related to the notion of self-steering in social systems.

Elaborating on the abstract notions of autopoiesis and configuration presented inthis volume, I will focus on two related approaches, i.e. ,simulation and gaming. Both arewidely used for the study of (social) systems, utilizing analytical theory and social

207

constructivism respectively as their relevant knowledge domains. Can both simulation andgaming improve our understanding of processes of structural coupling within andbetween autonomous social systems? What kind of learning environment can theyprovide in order to grasp those abstract notions of self-reproduction and configurationthrough learning-by-doing?

Although both are models of social systems, simulation and gaming have quitedifferent connotations. Simulation generally refers to modelling (artificial) systems inrelation to and determined by the complexity of external conditions (Simon, 1969).Games are viewed as social settings shaped by rules and roles. Games are by definitionhuman organizations, as they consist of people (actors), rules, and procedures for takinga variety of actions aimed at reaching a more or less well-defined set of goals. Becauseof their autotelic character, games provide a context for making actors aware that theycan, as Van Dongen sees it, take the opportunity to become distinctors in a process ofsocial construction of reality. Multi-actor simulation, being an integration of simulationand gaming, provides a methodological basis for studying both steering and self-steeringaspects of social systems, as it relates both the autotelic (self-referential) nature ofgaming with the allotelic character of simulation. On the basis of the frame of referenceof multi-actor simulation, three types of learning environments are being distinguished.They differ in the way they facilitate the emergence of cognitive and especially socialstructures and, consequently, in how they support our understanding of and skills inhandling (self-)steering in social systems.

2. Characteristics of Social Systems

Social systems emerge from interactions between individuals and gradually expand tobecome complex collective structures, because of the increasing number of individualsinvolved and/or the connectedness of their interrelationships. The cornerstone of thesecollective structures is, according to Weick, the double-interact (Weick, 1979). Double­interacts constitute cycles that can be maintained only by the continuous efforts of theactors involved. These internal circular processes (cycles) are closed upon themselves(organizational closure) and produce a social system only to the extent that action amongthe actors is interlocked. In this regard social systems are not autopoietic systems per se,as defined by Maturana and Varela (1980). Because of their capacity for organizationalclosure, communication, coordination, and self-steering are built into collective structures.Social systems specify their topological boundaries, distinguishing themselves from theirenvironment, which is important insofar as it hinders or agitates them in their continuingefforts to maintain stable interaction patterns.

Simple collective structures may form stable assemblies, which can constitute evenmore complex stable assemblies. In this way arbitrarily complex stable social systems maybe formed that are nearly decomposable (Simon, 1969). Society is seen as a complexsocial system, which is a set of interacting subsystems, which is a set of interacting sub­sub-systems, etc. From an overall system's viewpoint, interactions between sub-systemsof a system can be ignored. This interconnection of relatively independent social systemsbrings forward a modular structure which is considered a form of self-organization(Heylighen, 1990). The study of social systems deals with the various kinds of studies ofhuman organizations in the public and private domain. In describing social systems threeinterlocking strata are distinguished:

208

1. technology, the whole complex of routine and non-routine procedures to handlematerial processes;

2. structure, vertical and horizontal communication and coordination;3. culture, norms, values, attitudes, ideas, etc. (Klabbers, 1986).

The technology stratum represents all kinds of artificial systems (Simon, 1969) which havea given form and behavior only because they adapt or are adapted to their environment,in reference to a goal or purpose. Artificial systems are designed to fulfill a givenfunction in a given environment. This view can be applied to man-made artifacts likemachines, bureaucracies and enterprises. This general paradigm of classical cyberneticsrefers to steering, control and information processing. The technology stratum is viewedas the object of steering or the environment to be controlled. The structure stratumresults from stable assemblies of double-inter-acts. From a spectator's perspective, itrepresents vertical and horizontal communication and coordination. This stratum isconsidered the domain of procedures, rules, customs that guide interactions between theorganizational (policy) actors. The culture stratum represents values, attitudes, symbolsand icons as products of and conditions for action. It embodies the system of meaningof the respective social system.

Related to Snellen's distinction between driving and directing (in this volume), weargue that the technology stratum represents the driving function, while directing takesplace through the structure stratum, bounded by norms and values embedded in theculture stratum. These three strata are incorporated in multi-actor simulation. Thestructure and culture stratum relate to the actor approach to simulation, while thetechnology stratum stresses the machine approach to simulation.

3. Simulation versus Gaming

Games and simulations are constructs that have the potential to help us symbolize thecomplex world we live in. Since the sixties gaming has provided suitable learningenvironments for actors to mutually understand complex social systems from the pointof view of corporate planning, environmental planning, health care planning, urbanplanning, learning and education, etc. At the same time, a whole variety of classroomgames reflect a fundamental criticism of the way children are traditionally taught atschools. These developments have gradually broadened the scope of gaming. It hasemerged in scientific communities that generally favor a reductionist, narrowly technicalapproach, based on analytical theories. In order to understand its transdisciplinarycharacter, two major scientific cultures are outlined below. Within the scientificcommunity two distinct preoccupations or cultures, Le. ,social (re-)constructions of reality,are distinguished. One culture represents the rationalist, the other the historicistconception of reality.

Western science is based predominantly on the assumptions of universality andeternality, i.e., the context independent character of natural laws, which is the ideal ofrationality. This scientific approach is prevailingly occupied with investigating reversibleprocesses, with the notion of trajectories and of time that can be reversed (Prigogine andStengers, 1985). This conception of science has disclosed a passive nature, a dead world.Once the rules or laws governing nature are established, nature will obey them.

209

Eventually, nature will be trivialized, mastered and controlled. Technology is thedominant vehicle in accomplishing this goal. Classical cybernetics fits into this rationalistconception because of its stress on the mechanistic, causal and externally controlledaspects of general systems (Heylighen, 1989). In this positivist conception of reality it isassumed that knowledge is context independent.

As it turns out, the assumption of reversibility is rather weak for most systems.Basically, chemical, biological, ecological and social processes are irreversible. Naturalas well as social systems are understood to be time-oriented, i.e., they evolve over timethrough processes that are closed upon themselves. They defy analysis simply in termsof their constituent parts. They request that we focus on circular rather than linearschemes, on becoming rather than being. This so-called science of complexity eliminatesthe myth of deterministic predictability. It acknowledges the self-organizing, i.e.,internally controlled character, of complex systems. All features of complexity point atgaming as the basic characteristic of nature (Eigen and Winkler, 1987). Man participatesin this huge evolutionary game with its uncertain results.

The historicist conception of reality implies that: "Historical conditions change, andthat means the facts and their contexts. as well as the scholars with their interest andmethods. Understanding remains always conditioned, because the observerand society.subject and object, are aspects of history and undergo change" (Bendix and Bening.1989).

Simulation models based on classical cybernetics represent linear information processingsystems. Their behavior is adapted in reference to a goal to fulfill a given function in agiven environment. Their aim is to control the system involved, to steer. Accordingly,social systems are reduced to simple basic structures (e.g., causal models) which explaintheir behavior. Simulation models are appropriate tools for studying processes at thetechnology stratum of social systems. Gaming is more receptive to making sense of realityin terms of meaning processing between human beings. The cognitive and socialstructures evolving during games are constructed by the actors themselves. During agame the participants, through ongoing interactions, have the opportunity to create newsocial order and, consequently, they produce situated knowledge. The building blocks ofgames are double interacts between the actors, who engage in producing and re­producing a collective network (more or less stable configurations). Via its autotelicnature, gaming enhances self-steering in the social system it constitutes (see below). Asstated earlier, gaming focusses on the structure and culture stratum of social systems.

In the rationalist traditions of artificial systems, knowledge is composed of abstract,context-independent, formally interconnected domain-specific concepts. In the historicistperspective, related to autopoietic systems, knowledge is context dependent. It is notaccumulative per se, because historical conditions change, and that means the facts andtheir contexts. Understanding always remains conditioned, because the observer andsociety, spectator and nature, subject and object, are part of history and undergo change.As both scientific cultures have different conceptions about knowledge, they pursuedifferent objectives in knowledge transfer, and in sharing learning environments. Againstthe background of the rationalist tradition I will first introduce the machine approach toaccessing knowledge. This will provide the framework for subsequently introducing theactor approach, which is embedded in the historicist perspective.

210

4. The Machine Approach

4.1 The Trivial M(Uhine

A characteristic example of the rationalist conception of reality is the so-called TrivialMachine (TM). Von Foerster (1984) defines a machine as a conceptual device with well­defined rules of operation. The Trivial Machine is defined as a black box f processinginput x to output y. See Figure 1.

Figure 1: Representation 01 the Trivial Machine.

The inference scheme, i.e., the invariable relationship f between x and y provided bytrivial machines, is widely used in science in a whole variety of contexts. The followinglist of names illustrates this point (Von Foerster, 1984):

xinputinputindependent variablecauseminor premisestimulusgoal

foperationsystemfunctionLaw of Naturemajor premiseC.N.S.system

youtputoutputdependent variableeffectconclusionresponseaction

As soon as the x-y-correspondence through f is established, any x will generate a specificy, independent of time. Consequently, TMs are:

predictablehistory independent (universal, respectively context independent knowledge)synthetically deterministic (definition of x-y correspondence)analytically determinable

Most problem solving in education, especially in mathematics, and science educationdeals with learning to use the inference scheme of the Trivial Machine. Related problemsolving tasks are well-articulated, and respective problems are well-defined. Therepresentation of TMs is equivalent to the representation of systems. TMs or systems canbe decomposed or desaggregated into more detailed TMs or subsystems, which can beanalyzed independently subsequently yielding an over-all solution.

Problem solving skills related to TMs can either refer to goal directed strategiesapplied to specific knowledge domains, i.e., task limited specific strategies, or to across­domains general strategies. By utilizing the inference schemata of TM, learners

211

progressively reproduce a body of disciplinary knowledge. Problem solving is thusconsidered mainly a matter of transfer of disciplinary knowledge or acquisition ofprofessional knowledge. TMs are suitable constructs from the viewpoint of a rationalistconception of reality. They enable one to make a clear distinction between the observerand the observed. For the type of learning environment provided by TMs, I haveproposed the term reproductive learning environment (Klabbers, 1989). The TrivialMachine is the realm of (scientific) distinctions mentioned by Van Dongen (in thisvolume).

4.2 The Non-Trivial Machine

We have come to realize that many phenomena cannot fruitfully be investigated via theconstruct of TM. For many phenomena it is impossible to establish a transfer functionf. In many cases, for example, in econometric, sociometric and psychometric research,the error component for variance not explained in the x-ycorrespondence is too high togenerate valid and reliable results. A stronger argument is based on an epistemologicalpoint of view, indicating that as a construct the Trivial Machine is too narrow. The Non­Trivial Machine as presented by Von Foerster (1984) is a more powerful construct forunderstanding complex dynamic systems (see Figure 2). NTMs provide us with interestingterminology, such as recursive processes and autological, self-referential, concepts, Le.,concepts that are embedded in their own domain. For example, to answer the question:'What is language?', we need language to ask the question. 'Learning-to-leam' is anotherwell-known example of a second-order concept.

As can be noted from Figure 2, the driving function y = F(x,s) is similar to the TMin Figure 1. New is the state function s' = S(x,s), which is recursive: s' is defined by s atan earlier stage. Suppose NTM is a discrete system. In that case the state function canbe represented by:

s(t) = S(x(t), s(t-l)s(t-l) = S(x(t-l), s(t-2), etc.

Figure 2: Representation ol1he Non-Trivial Machine.

212

It is obvious that S represents the memory or history of any NTM. The internal stategenerated through S influences the processing of subsequent inputs to outputs as definedby Von Foerster (1984) shows that identification of a large class of NTMs is impossible(transcomputational) because observed sequences of input/output pairs (x,y) cannot beinferred from the machine's driving and state functions. This implies that for these NTMsit is impossible to predict output y from input x.NTMs are characterized as:

synthetically deterministic;history dependent;analytically indeterminable;analytically unpredictable;

NTMs make us conscious of an important distinction between the perspective of theperson knowing the internal structure and the one who is not aware it. For the knower,the identified and its behavior is transparent and rational. For the not-knower(outsider/observer), the dynamic behavior of such a system (NTM) is, except for a verysmall number of simple NTMs, unpredictable and thus chaotic.

Suppose we arranged a problem solving situation based on an NTM, moreover,putting a learner in the position of an observer/experimenter. That would teach him/herthat the simple notion of causality inferred from experiences with the Trivial Machineno longer holds. The behavior of an NTM depends on its internal state, which dependsrecursively on its history. Problem solving in interaction with NTMs will require theapplication of heuristic techniques: continuous planning, monitoring, checking andrevising of options and alternatives. System Dynamics models, which are appropriateexamples of NTMs, provide learners with a well-articulated environment, although theproblem solving situation may be ill-defined, even if learners have access to the detailsof the internal structure of the model.

Problem solving skills related to NTMs may stimulate learners to apply varioussearch strategies and develop reflective awareness or metacognition. They will learn tounderstand the switching of positions between the observer of NTMs behavior and theknower of its internal structure. They will realize that there are no singular answers tothe problem at hand, even if the task is well-defined. As we know from analyzing systemdynamics models, a varying set of measures, representing various scenarios may producesimilar as well as completely different overall results. Learners become aware of theirblind spots. This may improve their reflective awareness of what they know or do notknow or understand of the subject matter. Here we enter the realm of metacognition ofwhat Von Foerster (1984) has defined as second-order problems. Learners will realizethat the problem is not not-knowing, but not knowing that one is not knowing. VonFoerster points out that second-order concepts have an unusual structure, i.e., "doublenegation does not yield affirmation". Not not-knowing does not imply knowing. Learnersinteracting with NTMs will realize that accessing knowledge is context-dependent. Forexample, the behavior of a system dynamics model depends on its initial conditions andon the mix of adjustments in parameters. To search for acceptable policies, alternativescenarios might be suitable. For the type of learning environment generated by NTMs,I have proposed the term heuristic learning environment (Klabbers, 1989).

NTMs are machines, conceptual devices with well-defined rules of operation. Onecould argue that such a mechanistic approach does not seem to be very suitable forlearning about social systems, simply because we do not know their rules of operation.

213

Even if we know them at a certain moment in time, we would have to admit that theyare apt to change in unforeseen ways. One major reason is that the internal states ofsocial systems shift continuously as a result of the efforts of the actors involved. As in therelevant literature problems are defined as situations where one has a good idea aboutwhat to accomplish but no clear idea about how to accomplish it (Davis, 1973; Duncker,1945; Prawat, 1989), it will be obvious that many situations by definition hardly qualifyas problems (prawat, 1989), because, for example, we frequently do not have a good ideaabout what to accomplish. This applies to health care, crime prevention, communityplanning, education, environmental planning, etc. Usually in society we disagree aboutthe desired outcomes in relation to the efforts we are wil1ing to make. Solving problemsof the types discussed in relation to the Trivial- and Non-Trivial Machine is no longersuitable. Rein (1976) has stated that knowledge presupposes a framework to interpretit, but in pluralistic systems there are only competing frameworks. According to Ackoff(1974), we are dealing in this regard with a mess of problems. As will be discussed later,problem setting or problem-framing is a more promising way to cope with these typesof situations (Schon, 1982). The notion of the Non-Trivial Machine obscures our capacityto make distinctions as it confuses our cognitive structuring, because of its indeterminabi­lity and unpredictability. This is especially true of NTMs producing perturbations likedeterministic chaos.

5. The Actor Approach

Progressing from the frame of reference stipulated previously by the machine approach,I win generalize the representation of the NTM. The state function S is replaced byActors and Rules. Accordingly, I win switch from the machine to the actor approach. Inthe actor-system actors develop rules and procedures both for communicating and forintervening in their internal environment F via communication and coordination.Through this recursive processing of information, i.e., conversation and communicationthey continuously transform their col1ective structure. Actors engage in a process of socialconstruction of reality and their col1ective awareness of this process is an indicator of theinternal state of the system. Actors engage in a recursive time-variant process of change,in a flux. They reproduce their social system in an operationally closed setting (Klabbers1986; Luhmann, 1986). They shape order and as a consequence make history.

Figure 3: Illustration of an Actor-System.

214

A next step is rearranging the actor system, as presented in Figure 3, into the referenceframe of a game (see Fig. 4). Through these steps, the linking of the rationalist traditionof the Trivial and Non-Trivial Machine with the historicist tradition of gaming asmentioned in the first section is illustrated. The actor approach allows for taking twoperspectives: the observer/outsider point of view and the participant/insider point ofview.

Viability and resilience of social systems depend in the long run on the capacity ofthe actors involved to switch between both perspectives (Klabbers, 1986). The actorapproach, illustrated by Fig. 3 and 4, is based on the notion of multiple reality andreveals the blind spots of the various actors when dealing with complex social issues,subsequently enhancing their potential for shaping social reality. In this regard multi­actor systems are self-referential social systems. The acts of knowing imply not only thenotion of context-dependent knowledge, they also open the door to those realms, whereactors can cope with multiple contexts. In characterizing multi-actor systems, scientistsmay take the two perspectives into account. From the spectator's viewpoint, multi-actorsystems are:1. synthetically indeterministic;2. history dependent;3. analytically indeterministic;4. analytically unpredictable.

From the participant'S point of view, multi-actor systems are:1. synthetically (structurally) deterministic by the actors, i.e., they mutually

shape order;2. dependent on definitions of reality by the actors (values, beliefs, interests,

rules, etc.);3. comprehensible on the basis of meaning-processing (shared language,

signs, symbols, images, icons, etc.).

Actors are persons or groups of persons with roles, interests, norms, values andcognitions, carrying out one or more activities in a social system. They may use differentsigns and symbols and apply different rules and procedures while intervening in orshaping their natural and/or artificial environment, i.e. ,(re-)constructed reality.This (re)­constructed reality refers to the action domain of the respective actors. By definition,games are social systems, collective structures shaped and maintained through individualsand their interrelationships. With this reference frame it is rather simple to distinguishmany well-known games, for example, Monopoly, Chess, Go, etc. The different boardsrepresent different types of reference systems. Through the respective rules, the actorsintervene in their symbolic environment by moving pieces. They change the situation onthe board in recursive processes of monitoring, planning, checking, revising, moving,monitoring, etc.

215

Figure 4: Basic components of a game.

(Re)constructed reality may be represented with the signs and symbols that are in usein computer simulation, i.e., the formal language of mathematical models. In this case,gaming and simulation are integrated in what we call multi-actor simulation. See Fig. 5.Generally speaking both the actor- and rules-subsystem represent the gamingcomponents, while the reference system, the 'internal' environment, is represented by thesimulation model symbolizing material processes such as chemical, physical, ecologicaland technological flows. Depending on the designer's point of view, they may besymbolized either as Trivial or as Non-Trivial Machines. The actors may take the rolesof management or (public) policy makers. Accordingly, complex multi-level, multi-actorsimulations like Perform (Klabbers, 1986) and Dentist (Klabbers et aI. 1980) may bedeveloped.

Figure 5: Illustration of computer supported multi-actor simulation

216

In the field of gaming two different types of games are distinguished: rigid-rule games,like the classical management games or business simulations, and frame games or free­form games. They require for different contexts of use and problem orientations, relatedto notions of configuration and (self-)steering.

6. Rigid-Rule Games

In rigid rule games it is assumed that rules have some validity, and that they applyequally to the respective actors (player groups). Sometimes those rules are prescriptive,in the sense that the game leader forces the players to take steps accordingly. Sometimesthey are descriptive as they model behavioral equations included in the reference system(Klabbers and van der Waals, 1989-1). Rigid-rule games are closed games. Playersreceive specific role instructions, and their actions are goal oriented. The game operatortells them: "This is the problem: how will you solve it?"

Management games and production scheduling games are examples of this type ofgame. Classical management games developed during the fifties and sixties have had agreat impact on the notion of gaming. Since the sixties a great many students of businessadministration have participated in management gaming. Many training programs oflarge corporations in the USA, in Europe and elsewhere currently use games for trainingand assessing managerial skills. The generic structure of these games is described below.

Three to fifteen competitors share the same market, providing a situation of risksand opportunities. Functional management activities such as, production, financing,marketing and logistics are integrated in a general strategy aimed at outmaneuvering thecompetitors in terms of profit, market share, and dividend. The context of use iseducation and training. Fifteen to two hundred participants, divided into subgroups of2-25 players invest between 8 - 15 hours (or one semester) in running the game. During3 - 8 rounds (cycles) the game evolves on the basis of rigid rules. During each cycle theresults of prior managerial actions are fed back via accounting sheets which are usuallyproduced by a computer.

A review of three decades of business gaming suggests that it is still not clear howeffective they are. Evaluation studies carried out since the early sixties have not beenconclusive (Wolfe, 1985). It is generally recognized that, from a methodological point ofview, games are 'unruly' when it comes to evaluating them (Bredemeier and Greenblat,1981). For several reasons the impact of games is notoriously difficult to assess:

There is a wide difference in the practice of games.Reporting procedures are not standardized.No standardized contexts of use exist for running games.The application of games contains many interrelated elements: cognitive and socialas well as affective and ethical skills.The positivist (reductionist) approach usually most of the time taken in evaluationstudies is too limited to capture the essence of what happens during a game session.Nevertheless some general recommendations can be made on the basis of theevaluation studies.Business games are useful tools for management teaching.The limitations should be recognized.Their value in developing managerial skills appears to be limited (Evans and Sculli,1984).

217

Those general management games of the sixties and seventies are models of amechanistic type of organization functioning in a disturbed reactive environment, whichis characteristic of market conditions of the fifties. Classical management games aretraining tools which emphasize training of managerial skills. In the game it is thealgorithm embedded in the computer that executes the planning, production, sales etc.In evaluation studies of those management games the most important criterion forsuccess is defined in terms of tum-over, profits and market share.

According to their structure, rigid-rule games resemble the setting of a heuristiclearning environment as discussed above. As in rigid-rule games, the rules are given, theyinitially provide the actors with a system of meaning. They help in shaping a specificorganization, for example, for learning to do business.

7. Free-Conn Games

In free-form games, only a scenario and a simple accounting system are provided. Playersare put in an open situation and are asked: 'What will you do?' Playing free-form gameswill serve as a way to better understand the preoccupations of the actors as they emerge.Players engage in a process of creating order through multiple lines of action andcommunication, producing and re-producing their social system. They develop a self­awareness of new opportunities for constructing social reality. Learning is related toshaping order, creating knowledge and improving mutual awareness among theparticipants. These self-organizing learning environments not only show that knowledge iscontext dependent; moreover they highlight the changing meaning of knowledge. Shubikpoints out that one should be able to see immediately that emphasis on the participationand quality of the individuals must be higher when using free-form gaming as comparedwith rigid-rules gaming. "The value of a free-form game may be highly related to theexpertise and sophistication of both the players and the referees" (Shubik, 1983).

Self-organizing learning environments, like free-form games, question the coherenceof the assumptions on which situated knowledge is based. They help in articulating thesubject matter in a continuous process of re-setting the problematic situation. Moore andAndersen (1975) in presenting heuristic principles for designing learning environmentsargue that:1. Learners should be given the opportunity to operate from various perspectives. The

learner should not just be a recipient of information, but should at times be anagent, a referee, and a reciprocator.

2. Activities should contain their own goals and sources of motivation, not justrepresent means to some end (such as grades). That is, in an effective learningenvironment, activities are autotelic.

3. Learners should be freed from a dependence on authority and allowed to reasonfor themselves; they are thus made productive in the learning process.

4. The environment should be responsive to the leamer's activity. Not only shouldthey be given feedback, but they should be helped to be reflexive, evaluating theirown progress.

With regard to free-form games, rules and roles evolve as the game processes over time,shaping a specific collective network and consequently a system of meaning may emergesimultaneously.

218

8. Social construction of reality through gaming

In previous sections I have progressively moved away from analytical theory based on therationalist conception of reality, i.e. ,the machine approach, towards social constructivismbased on the historicist tradition in science, i.e., the actor approach. One of the mainreasons for coupling both epistemological domains is to provide a common methodologi­cal basis for the study of (self-)steering of social systems through simulation and gaming.The framework of multi-actor simulation enables one to the link the rationalist andhistoricist conception of reality, taking into account recent advances in deterministicchaos, producing various bifurcations in non-linear simulation models, as well as in socialconstructivism. From a methodological point of view, I don't consider these theoriesmutually exclusive. Let me elaborate on this fundamental issue.

At the technology stratum of social systems we are dealing with material processeswhich include physical, chemical, ecological, technological and demographic processes,which are basically irreversible (Prigogine and Stengers, 1985; Eigen and Winkler, 1987).In equilibrium conditions we may notice repetitive, universal behavior, represented bylaws of nature and allowing the traditional rationalist approach of classical science.However, moving away from those conditions, we enter the realm of disorder andturbulence. In far-from-equilibrium conditions we may notice transformations of chaosinto orders of higher complexity. In that domain we have moved from the repetitive anduniversal to the specific and unique. What is important in this regard is that matter,through communication and diffusion, is able to generate self-organizing physical systems.Matter becomes self-referential, evolving over time. The point of view of modem scienceon chaos, order and self-organizing systems is that traditional boundaries between socialand physical science seem to vanish as nature is not passive, behaving like an automaton,i.e., a trivial machine. Nature displays self-organizing capacities and seems capable ofmaking distinctions. If this notion is true, the basic question is: "How can we learn tounderstand the cognitive- and socio-structural aspects of those complex systems, referringboth to the social and physical realities that constitute social systems?"

As discussed above, multi-actor simulation provides an appropriate and flexibleapproach to learning to understand the various meanings of autopoiesis with respect tothe class of non-linear systems to which social systems belong. Related learningenvironments for understanding the different worlds of meaning, and as van Dongen seesit, for improving our reflexivity, our ability to change either the distinctor or thedistinction, are the reproductive, heuristic and self-organizing learning environments. Inprofessional education these learning environments have attained different status levels.Schon (1987) presents an interesting example of medical education:

"Physicians are thought to be trained as biotechnical problem solvers by immersion,first in medical science and then in supervised clinical practice where they learn toapply research-based techniques to diagnosis, treatment, and prevention. In thisview of medical education, and its extension in the normative curriculum of otherprofessional schools, there is a hierarchy of knowledge:

Basic science and Applied scienceTechnical skills of day-to-day practice.

The greater one's proximity to basic science, as a rule, the higher one's academicstatus. General, theoretical, propositional knowledge enjoys a privileged position."

219

Paraphrasing this view: The reproductive learning environment is highest, and the self­organizing learning environment is lowest on the ladder of esteem of professionalknowledge. Professional knowledge is specialized, firmly bounded, scientific, i.e., basedon the rationalist conception of reality, and standardized. Accordingly, professionalactivity consists of instrumental problem solving made rigorous by the application ofscientific theory and technique (SchOn, 1982). As stated by Schon (1982, 1987), there isa growing crisis of confidence in this type of professional knowledge. Key problems tobe addressed in social studies will have to take into account:

complexity in the sense of inadequate knowledge of the dynamics of social systems;uncenainty of iriformation in the sense of equivocality, vagueness and randomnessof signals and data;differences in interests, values and ideas, i.e. unique cultural or cognitive differences.

Donald Schon (1982) has argued that traditional professional knowledge is not wellsuited for coping with ill-structured, unique situations, if it is based on a narrowlytechnical rationality. "Professionally designed solutions to public problems have hadunanticipated consequences, sometimes worse than the problems they were designed tosolve. Newly invented technologies, professionally conceived and evaluated, have turnedout to produce unintended side effects unacceptable to large segments of our society"(Schon, 1982).

Technical rationality, being the positivist epistemology of practice (SchOn, 1982),puts the professional in a position in which it is allowed to isolate the object or systemfrom its context, solve the respective problem and subsequently place the object back inits original position. The respective professional remains an expert-outsider/spectator.Nevertheless, complexity and uncertainty regarding the problematic situation are notresolved by transforming an ill-structured problem into a well-articulated task. In thesocial realm, one of the main characteristics of complexity is the existence of competingframeworks or paradigms or reality definitions. Plurality of conceptions cannot beresolved by a narrowly technical approach. In order to be able to cope with complex,uncertain social situations, SchOn proposed the term problem setting. It is a process inwhich, interactively, we name the elements and attributes to which we will attend andframe the contexts in which we will attend to them (Schon, 1982). Problem setting fitswell into social constructivism and is exactly what is happening in free-form gaming. Inproblem setting practitioners deal with complex situations, with uncertain outcomes, andbecome involved in a process of 'knowing-in-action', which as such is a self-referentialprocess. They become participants in a process of framing and re-framing contexts, andin naming the respective elements and attributes to attend to. Free-form games areadequate tools for problem setting, through which context dependent knowledge isshaped and articulated. These self-organizing learning environments offer students aswell as practitioners the opportunity to improve their competence in dealing withcomplex problematic situations with their uncertain outcomes and in coping with valueconflicts based on different interests and competing frameworks. Free-form games aresuitable tools for learning to understand the often chaotic process of bringing forwardvarious systems of meaning.

As many methods of didactic education assume a separation between knowing anddoing, knowledge is treated as an integral, self-sufficient substance, theoreticallyindependent of the situations in which it is learned and used (Brown, Collins andDuguid, 1989). This applies especially to the reproductive learning environment. Brown

220

et al. (1989) point out that "knowledge is situated, being in part a product of the activity,context, and culture in which it is developed and used." Regarding teaching, knowledgeis a product of didactic configurations. Evidently gaming is appropriate for supportingsituated learning. Free-form gaming in particular is a fruitful cognitive apprenticeshipmethod, as it "may enculturate students into authentic practices through activity andsocial interaction in a way similar to that evident - and evidently successful - in craftapprenticeship" (Brown et al. 1989). In free-form gaming students are guided inbecoming reflective-practitioners through cognitive apprenticeship.

Most of the problematic situations we encounter such as environmental degradati­on, drugs and crime prevention, health care, starvation, etc., issues relevant to publicadministration, cannot be resolved by narrowly technical means based on a rationalistconception of reality. The question still remains: "Can they be dealt with through socialconstructivism?"

9. Summary

Suppose we view notions on autopoiesis, steering and social integration as constructs forsocial problem solving. What does steering mean in a context of multiple realities andrationalities, in divergent or convergent configurations? Is steering nonetheless possiblein social settings with regard to reaching weB-defined goals through clear measures ofdriving and directing? What kind of empirical evidence might be available to supportthose notions of steering, as human organizations, i.e., social systems, are constructsthemselves?

Having been involved in a whole variety of multi-actor simulations, games andsimulations, and considering games in particular to be simple social systems, I am veryreluctant about using notions like autopoiesis for providing empirical evidence for socialproblem solving. Autopoiesis is a social construct generated by a configuration ofparticular scientists. Are they sure that they are dealing with real phenomena or with realmyths? Games and simulations are being used widely in education, training and research.Because of their autotelic character, stressing their self-referential nature, games arenotoriously difficult to evaluate (Wolfe, 1985; Bredemeier and Greenblat, 1981, Evansand Sculli, 1984). As stated earlier, they are synthetically indeterministic, historydependent, analytically indeterministic and an.u,tically unpredictable. But so are socialsystems in general!

Nevertheless, multi-actor simulation can help us, facilitators and participants,understand the processes of structural coupling within and between relative autonomoussocial systems. Free form games are especially appropriate in this regard (Klabbers,1991). They can show facilitators as weB as participants, configuration theory, i.e.,cognitive-/socio-structural distinctions, 'at work' . Games can teach us the boundaries ofsteering and self-steering in social systems.

Those who participate in free-form games willlearn to understand the meaningsof steering and self-steering. I see them primarily as a switching of perspective fromoutsider/spectator to insider/participant and vice versa. This switching has a cognitive,social, ethical as weB as an affective meaning. In this sense free-form games can be seenas support systems for improving competence in (self-)steering with regard to socialsystems, because they enable one to frame and re-frame the multiple realities embeddedin games, that is, they help us cope with numerous worlds of meaning.

221

Part IV

CONFERENCE IMPRESSIONS

Chapter 19

A STRANGER IN PARADISE?

AUTOPOIESIS, CONFIGURATION AND SOCIETAL STEERING

H. Mastik

1. Introduction

In 1908 Georg Simmel described one of his so called social types 'Der Fremde' asfollows: 'The stranger will (...) not be considered here in the usual sense of the term,as the wanderer who comes today and goes tomorrow, but rather as the man whocomes today and stays tomorrow - the potential wanderer so to speak, who, althoughhe has gone no further, has not quite got over the freedom of coming and going. Heis fixed within a certain spatial circle - or within a group whose boundaries areanalogous to spatial boundaries - but his position within it is fundamentally affectedby the fact that he does not belong in it initially and that he brings qualities into itthat are not, and cannot be, indigenous to it." (Levine, 1971, p.143). This descriptionseems to apply both to my experience at the conference as to its subject matter. Toput it another way: I could not help thinking of a traditional anglo-saxon superstitiousrhyme that tells brides to wear: 'something old, something new, something borrowed,something blue' on their weddingday.

When the foregoing may convey a negative impression I hasten to correct that.Both the lectures and the discussions were fascinating and inspiring and I'll try toanalyze some of the issues from the standpoint of Simmels stranger rather than thatof the superstitious bride to be, although the rhyme holds great wisdom. Undertakingthis activity I draw upon my understanding of cultural sociology and sociology ofknowledge. This detour however may well lead up to the conclusion that within thecommunity of public administration and policy science I take the stand of apermanent stranger whose position "(...) stands out more sharply if, instead of leavingthe place of his activity, he settles down there." (Levine, 1971, p.144).

2. Societal steering and the probability of directing innovation and change

The attractiveness of both configuration and autopoiesis theory to students of publicadministration and policy science holds the danger of scientific suicide. Boththeoretical perspectives point very strongly at the self referential and self reproductive

225

properties of social realities and take into account their simultaneity, multiplicity andautonomy. Configurations and systems lead their own lives according to their ownimmanent 'laws of nature'. Both theoretical perspectives do away with causality andlinearity as applicable concepts to describe let alone to predict the occurence ofphenomena. Rational planning and societal steering in the sense that means aredirected to attainment of goals is sequentially intervening in synchronic processes andtherefore outcomes are unpredictable in principle. Theoretical notions of societalsteering that focus on policy implementation by control and regulation can be putback to storage next to concepts and beliefs on the potential of individual creativityfor bringing about any intended and/or desired innovation. Both the state and thewizard's pupil are confronted with the same problem: they cannot master their owncreations. To stop here however tempting would be cynical, even cowardly.

Further scrutiny of autopoiesis and configuration theory leads to some veryfruitful insights into what exactly could be the realm of societal steering because bothapproaches very clearly point out the boundaries and constraints of directedinterference.

3. Theoretical pluralism and the historical relativity of knowledge

Students of sociology in the late seventies have become very familiar with discussionson the so called theoretical pluralism of the social sciences and witnessed manyfruitless debates between adapts of different schools of thought. What happenedlooking back, was that either discussions focused on ontological and epistemologicalquestions thus moving away from empirical questions in philosophical regression, ordilemmas were ignored completely by sociologists who claimed their activities to bescientific merely because they were using 'scientific' research methods. This did nothelp to enlarge the esteem of the discipline: sociologists seemed to either generatefutile knowledge, rephrasing in difficult terminology what everybody already knew, orto generate no practical knowledge at all.

Looking at the history of the discipline one notices that the same type of debatereturns as conjunctural cycles. An interested observer of the debates over time canhardly escape from the impression that some ideas and notions have been brought upbefore, be it with other words and in a different language. Although it is tempting itdoes not seem fruitful to prove that point. Such a venture may well be termed as 'artfor the sake of art', and will not bring a useful synthesis to sociological thought.Observing the recursiveness of notions and ideas in sociology does not mean that noknowledge has been added since Comte, Durkheim and Weber.

Meanwhile, all concepts, models and theories, produced by the (social) sciences,have been out there, in the real world. No longer tied to their originators, they startliving their own lives in a universe that is continuously expanding by realityconstructing and surplus generating processes. Disciplinary attempts to create order inorder to influence the direction of developments have become part of this expandinguniverse. Bodies of knowledge themselves, can be characterized as autopoieticsystems. Scientific controversies have their own life cycles and yet they never die.Both verification and falsification of concepts and theories are principally andultimately impossible. Whatever may be true, depends on time and place andperceptions and interpretations of individuals, configurations, or systems. Whatever is

226

going to work, depends on the complex multiple and mutual interactions of allvariables simultaneously in time. Anticipating outcomes of directed interferencebecomes extremely complicated because of the fluidness of this dynamic process.Accepting insights of autopoiesis and configuration theory means that successfulsocietal steering is only possible if it results in changes in the perceived interests andinterpretations of actors, configurations and systems.

Simmel points at the tragedy of culture. He argues that the relation betweenobjective and subjective culture is declining, loosening. Eventually this will lead totheir definite separation. In his opinion change can only result from interaction ofsubjective and objective culture. However, subjective culture is developing away formobjective culture towards infinite regression in subjectivity. As a consequenceindividuals are no longer encompassed by (included in) the public realm. Objectiveculture is thus developing and expanding more and more according to its ownintrinsic logic (Simmel, 1957). In this process change emerges increasingly by chance,unintended and possibly undesired. Such a society becomes more and more resistantto directed interference.

Today evidence seems to prove Simmel right. Objective culture has developedinto an ever expanding universe of freefloating artifacts that, because of theirprincipal multi interpretability form a reservoir of potential meanings from which anyuser can choose according to his personal needs and purposes. Not only beauty, butalso meaning seems to exist in the eye of the beholder. Moreover the number ofpossible unintended outcomes grows with a tremendous multiplying factor due to thepotential of the 'internal chemistry' within the expanding cultural universe. All kindsof 'Wahlverwandschaften', collisions and coincidents generate new realities apart fromthe permanent (re)construction of meaning by actors that are also reinforcing theexpansion. The strange paradox of modernization becomes apparent: the moreknowledge is generated by 'rational sciences' the less applicable it will be for societalsteering purposes. Or in other words, the larger the amount of objectively availablepossibilities to create, arrange and control society, the larger the probability that merechance will shape the future of mankind.

4. The position and situation of the social sciences

Following Weber, Mannheim develops the concepts substantial and functionalrationality; substantial rationality referring to the normative, qualitative aspects ofquestions and answers. Very optimistically he suggested a role for the 'freefloatingintelligence' to generate substantially rational knowledge because of its presumed lackof self interest (Mannheim, 1951).

To a certain degree autopoiesis theory might very well be used to explain thedevelopment of schools in scientific thought. Instead of being unselfish and withoutself-interest scientific disciplines and schools themselves have constructed their ownrationalities according to interests in the 'arena of truth', that are reproducingthemselves and generating their own surplus. Many discussions between scientists whoare adherents of differing perspectives within the same discipline can be characterizedas selective rephrasing of elements of the others contribution in terms of ones ownschool of thought. The social sciences are especially vulnerable. Concepts havebecome relative autonomous from their original inventors, the same words may be

227

referring to different realities. Establishing intersubjectivity (are we all talking aboutthe same things?) is further complicated by the fact that languages are 'living'symbolic universes that are closely related to the societies they express and reinforce.The sort of scientific pig English that emerges at encounters of non native speakingscholars further complicates matters, as Koppen observed at the conference underscrutiny here.

5. Good and bad theory and the scientific method

Having decided to accept theoretical pluralism as a natural fact, produced andreproduced by the very processes and mechanisms that are studied by autopoiesis andconfiguration theorists, does not necessarily lead to the conclusion that every theory isas valuable as the next one. Distinguishing between 'good' and 'bad' theory, or even'good' or 'better' theory however, becomes extremely difficult when we take intoaccount that the appearance of any theory at any time is the beginning of theirautonomous existence, influencing form and contents of historical social reality and itstaken for granted character vis a vis individuals and groups, including scientists andpolicymakers. Consequently the activity undertaken here cannot result in an embraceof either autopoiesis or configuration theory but should focus on their potential forfinding answers to questions that confront scientists and policymakers. Attempts toapply models from the natural sciences to the social and cultural sciences as such, arenot sufficient. The history of the social sciences shows that this is a tradition that goesback to the founders of positivistic sociology and ever since can be seen as attemptsto enhance the status of the disciplines. But, as Kickert pointed out "a hard sciencemodel cannot be blamed for not fitting a soft science reality". Moreover, as Giddensstated, the history of the social sciences also shows "(...) it wrong to slant social theorytoo unequivocally towards abstract and highly generalized questions of epistemology,as if any significant developments in social science had to await a clear cut solution tothese." (Giddens, 1984, p.xviii).

6. The unsolvability of problems

At the conference part of the discussion focussed on ontological and epistemologicalproblems, and rightfully so because these problems did arise (again). Recognizingtheir existence is both necessary and sufficient but also leads inevitably to theconclusion (again) that they are immanently unsolvable. As Van Dongen pointed out,some problems can be pointed at and should be clarified only to establish as a fact(again) that they cannot be solved.

This observation however, also holds on the disciplinary level both theoreticaland empirical. The notion that we should learn to recognize, distinguish anddetermine that some problems may well be conceivable and are conceived differentlyby different stakeholders, but cannot be solved, is very important for the policysciences. Thus we can return to the core issue of the conference: what do autopoiesisand configuration theory teach us that is relevant and useful for the development of adynamic theory of societal steering?

228

7. Confrontation, reflection and communication, autopoiesis and configurationtheory

The positive contribution that can be derived from both theoretical perspectives liesin their negative evaluation of the possibilities for planned goals attainment. In otherwords by pointing out clearly the limits to steering. Analyzing the very processes thatexist and shape the social world as it develops in its own right, both perspectivesenlighten the possibilities for directed intervention. Autopoiesis theory points out themechanisms of 'coupling' and 'binding' as strategical change agents, stressing theimportance of communication and self referentiality. It also offers the valuable insightthat crisis situations, 'soft spots', offer tremendous possibilities and opportunities fordirected interventions. Following Mead, configurationists develop the concepts of'play' and 'game' to either carry change (game) or to be changed (play). Also theobservation that actors are partially and multiple included offers possibilities fordirected change. Actors are already familiar with more than one reality, articulationof fixations through reflections generate space for new internalizations, andconsequently a change of attitudes.

To my understanding the debate between autopoiesis and configuration theoristsat the conference revealed their underlying common interests and intentions toexpose naive conceptions of societal steering. Both approaches focus on self referenti­ality and reflection through recursive processes in a dynamic way, however autopoiesistheory is more of an abstract, reified nature, reducing people to 'mere psychicartifacts'. Social reality is not explained in terms of repeated sameness, but in termsof setf(re)production either through selfreferential rationality generating surplus(autopoiesis), or through ongoing (re)construction of social (symbolic) reality throughinteraction, allowing for multiple realities and thus different possible futures(configuration). Autopoiesis theory stresses the 'hardness' of the intrinsic rationality of(sub)systems, while configuration theory points out the social (re)construction ofrationality. Both approaches dismiss the possibility of knowing 'ideal truth' andformulate a negative opinion on the role of the state, i.c. government. Teubner holdsthat the state is 'living on the surplus created by societal subsystems' and as suchparasiting according to its own rationality. Van Dongen, slightly more optimistic,limits the realm of the state to the role of 'safeguard' of the 'ethical space' in whichactors articulate and negotiate 'possible futures'.

Crucial to both approaches are the notion of closure of systems, or actors with ashared definition of reality and the notion of language, the carrier of selfreferentialmeaning. Closure means that the environment is perceived, defined and interpreted inthe systems' or configurations' own terms. Thus irrelevant signals and information willbe ignored, or better, cannot be perceived at all. Development and growth merelyamount to expansion through accommodation within systems or configurations. As aresult it is impossible to bring about innovative change through outside intervention,without a translation in their 'language'.

Consequently it becomes clear that both approaches advocate solutions throughlanguage. Articulation of meaning, on the verge of the system or configuration incontact with others, is a necessary step to 'plough the soil' for whatever change needsto be established, from whatever interest or perspective. Van Twist and Schaapdiscuss Luhmanns communication theory, pointing at ways to influence or even todirect interdependent subsystems. Configurationists point at several methods of

229

intervention, all of them referring to the need for 'thirdness' (outsider in or insiderout) to (help) objectify the constructiveness of a specific definition of reality, thusexposing its 'taken for granted' character in order to reflect and to free energy forcreating other (new) realities.

8. Agenda suggestions for policymakers and policy scientists

The agenda of policy scientists and policymakers who are inspired by autopoiesis andconfiguration theory will look very different from that of practitioners who get theirinspiration from linear and causal sources. Policy formulation and policyimplementation can no longer be distinguished as fairly separate fases in the policyprocess. The polycentric and network approach for the analysis of empirical policyprocesses that were presented at the conference make this very clear. Interests anddefinitions of the situation change over time, learning processes tend to have aninbuilt time lag. Societal steering if at all possible, can only be effective if the timingis right, momentum is sufficient and if actions are directed at the process through'fine tuning' rather than through traditional instruments. Like a football player needsto have an 'eye for the ball' the policymaker needs to have intuition for the 'oppor­tunity'. Shakespeare already said: 'there is a tide in the events of man, which taken atthe flood, leads on to fortune'.

The policy scientist should focus on the debate between stakeholders in thepolicy negotiating process (Majone, 1989). This is extremely relevant from anormative point of view. If it is agreed upon that part of the 'game' in the policyarena is persuasive argumentation it is essential in a democratic society thatunderlying norms and values be exposed or even demasked.

The above leads to a final remark about the potential of both theories. Societalsteering is ultimately aimed at establishing satisfactory and just arrangements for coexistence of all inhabitants of a society. Political intervention is normative perdefinition. The cynicism with which autopoiesis theorists view the role of the stateshould not lead to the conclusion that nothing 'good' can ever come from a parasite.Further analysis of change potential of systems should take into account a stipulationof underlying norms and values. In a slightly different way the same holds for the'safeguard' role of the state. The notion of un-values presupposes dichotomization ofvalues. To distinguish an un-value implies a value orientation. In my opinion thisoverlooks the enormous variety in value orientations and value prioritizations ofactors, groups and systems. As In 't Veld points out, from a democratic point of viewthere will always be the need of societal intervention in certain policy domains.

9. Ascetic opportunism, a post modern approach

The foregoing cannot pay enough tribute to the richess of the issues at stake. Theabove impression necessarily only scratches the surface and does not pretend to domore than that. The basic argument I want to make though is, that constantrevitalization and re-examination of ideas that came up earlier in history is needed.For reasons of modesty as well as for improving our understanding of the socialworld. From a post modernist point of view 'new' theories thus rather refer to the

230

fusion of theoretical fragments and re-application of earlier concepts in such a waythat questions and answers are rephrased and reframed. Description, explanation andprescription are contextual. In my opinion the value added in the autopoiesis andconfiguration conference I attended consists of the 'mediated communication'between the two perspectives. To avoid some of the mechanisms described by bothperspectives (closure, fixation) communication between schools of thought andscientific communities may well lead to 'new' insights based on 'old' ones.Consequently we can set the agenda of conference managers. To name but a few,confrontation of concepts from chaos theory, notions of crisis and coincidence,probability, chance and serendipity could possibly contribute to the furtherdevelopment of a dynamic steering theory.

In order to be able to decide which confrontations are relevant I suggest the useof ascetic opportunism. Accepting the historical relativity of knowledge as well as theimmanent unsolvability of problems, imposes constraints on our intention to useknowledge for arranging innovation and change in the real world. If different theoriesoffer plausible explanations for phenomena, and scientific methodology doesn't offerobjective criteria to distinguish the most valid explanation we should undertake thediscussion on how to choose relevant theoretical perspectives for our purpose.Following Weber and Schutz I propose to theorize in an 'ascetic opportunistic' way.Ascetic opportunism refers to the need for integration of concepts from differentdisciplines for the multidisciplinary art and science of public administration and policyscience. It offers both substantial and functional criteria to choose and fuse relevanttheoretical approaches. The adjective 'ascetic' refers to the need for explicitarticulation of value orientation as an essential obligation of any scientific effort, bothas a means to distinguish between theoretical notions as well as to evaluate thesenotions with regard to others (Zijderveld, 1979). Opportunism, not meaning vulgareclecticism but referring to the practical value of theories for certain levels, domains,situations or processes in the social world. Instead of disputing on the validity ofconcepts (scientific method) I advocate to regard concepts with regard to theiradequacy to describe and explain complex phenomena (Schutz, 1976). Adequateconcepts are reflections on the taken for granted images of actors, acquired through'thirdness', but meaningfully congruent with those images. This leads me back toSimmel's stranger, who embodies a very special type of 'thirdness'. Policy scientistsneed to be craftsmen and artists, but most important of all, they should remainstrangers. Going on Simmel's definition the stranger is in the best position. Situatedon the very edge of 'black holes' he is close enough to observe them clearly and toofar away to run the risk of being swallowed. This position allows for original thinking,as well as for commitment to the community, because he belongs there ('settlesdown'). This distinguishes him from the mere visitor, whose position can easilydegrade towards vulgar opportunism. The strangers domain is a safeguarded 'creativespace' for playing with possible theories.

231

Bibliography

Ackoff, R. (1974), Redesigning the Future, New York: John Wiley.Adewoye, O. (1986), Legal Practice in Ibadan, 1904-1960, 24 Journal of Legal Pluralism 57.Aiken, MJ. Hage (1968), Organizational interdependence and intra-organizational stmcture, American

Sociological review, vol. 33, pp. 912-30.Aldrich, H.DA. Whetten (1991), Organization-sets, action-sets and networks: making the most out of

simplicity, in: Nystrom/Starbuck.Allison, G.T. (1971), Essence of Decision. Explaining the Cuban Missile Crisis, Boston: Little, Brown and

Company.Allott, A. (1960), Essays in African Law, London: Butterworths.Argyris, c., DA. Schon (1978), Organizational Learning: A Theory of Action Perspective, Reading:

Addison-Wesley Publishing Company.Arthurs, H.W. (1985), Without the Law: Administrative Justice and Legal Pluralism in the Mid

19th-Century England, Toronto: University of Toronto Press.Ashby, W.R. (1952), Design for a Brain, England: Chapman & Hall.Arrow, KJ. (1963), Social choice and Individual Values, 2nd edition, New YorkAshby, W.R. (1956), Introduction to Cybernetics, London: Chapman & Hall.Auerbach, J.S. (1983), Justice Withow Law?, New York: Oxford University Press.Bachelard, G. (1983), La Philosophie du Non, Paris: PUF (9th Edition).Bachrach, P, M.S. Baratz (1970), Power and Poverty: Theory and Practice, New York: Oxford Univer-

sity Press.Bacow, L.S., M. Wheeler (1985), Environmental Dispute Resolution, New York, Plenum Press.Baecker, D. 1988), Infomlation WId Risiko in der Marktwirtschaft, Frankfurt: Suhrkamp.Bakker, R. (1981), Wijsgerige anlropologie van de twinligste eeuw, Assen: Van Gorcum.Bakker, R. and J. de Graaf (1979), Wijsgerige ethiek van de twintigste eeuw, Utrecht: Bijleveld.Ballmer, T., E. von Weizsacker (1974), Biogenese und Selbstorganisation, in E. von Weizsacker (ed.),

Offene Systeme I: Beitriige zur Zeilstmktur von Infonnation, Entropie und Evolution, Stuttgart: Klett.Ballweg, O. (1972), Interdisziplinare Forschung als Sophiatrie, 2 Jahrbuch Jilr Rechtssoziologie und

Rechtstheorie pp.578-582.Bardach, E., RA. Kagan (1982), Going by the Book. The Problem of Regulatory Unreasonableness,

Philadelphia: Temple University Press.Beer, S. (1980), Preface, in: Maturana, Varela 1980.Behn, R.D. (1978), How to Terminate a Public Policy; A Dozen Hints for a Would-Be Terminator,

Policy Analysis, 4, pp.393-413.Bendix, R. (1984), Force, Fate and Freedom. On Historical Sociology, Berkeley: University of California

Press.Benseler, F., P.M. Heijl, W.K Kock (eds.) (1980), Autopoiesis, communication and society, Frank-

furt/New York: Campus Verlag.Benson, J.K (1978), The interorganizationalnetwork as a political economy, in: Karpik (ed.).Benson, J.K (1982), A framework for policy analysis, in: Rogers and Whetten.Bentsi-Enchill, K (1969), The Colonial Heritage of Legal Pluralism, 1 Zambia Law Journal 1.Bercusson, B. (1987), Juridification and Disorder, in G. Teubner (ed.), Juridijication of Social Spheres.

Berlin: de Gruyter, pp.49-90.Berger, J. (1987), Autopoiesis : Wie 'systemisch' ist die Theorie sozialer Systeme? In: H. Haferkamp, M.

Schmidt (Hrsg.), Sinn, Kommunikation und soziale Differenzierung, Frankfurt am Main: Suhrkamp.Berger, P.,S. Pullberg (1966), Reilication and the Sociological Critique of Consciousness, New Left

Review, nr.35 pp.56-71.Berger, P.,T. Luckmann (1971), the Social Construction of Reality, Penguin Books.Berger S.D.(ed.) (1981), Organizing illlerests in Western Europe: pluralism, corporatism and the

transfonnation ofpolitics, Cambridge: Cambridge University Press.Berman, P. (1978), The study of macro- and micro-implementation, in: Public Policy 2, pp.I77-179.

233

Bermbach U. (1984), Politische Theoriengeschichte, Probleme einer Teildisziplin der Politischen JVissenschaft, (Sonderheft), pp.99-125, Opladen: Westdeutscher Verlag.

Bertels, K (1973), Geschiedenis tussen stmktuur en evenement, Amsterdam: Wetenschappelijke UitgeverijB.V.

Berting, J. (1981), Structures, actors and choices. In: Klabbers, J., et al. (eds.) Simulation-gaming: On theImprovement of Competence in Dealing with Complexity, Uncertainty and Value Conflicts, Oxford:Pergamon.

Bingham, G. (1985), Resolving Environmental Disputes: A Decade of Experience, Washington D.C.:Conservation Foundation.

Blake, J., K Davis (1968), On norms and values, in: RA. Manners and D. Kaplan (eds), Theory inAnthropology: Source Book, pp.465-472, Chicago: Adline.

Boer, Th. de (1976), Tussen filosojie en profetie, Baarn: AMBO.Boer, Th. de (1980), Grondslagen van een kritische psychologie, Baarn: AMBO.Bohne, E. (1981), Der infonnale Rechtsstaat, Berlin: Duncker & Humblot.Boissevain J., Mitchell J.e. (1973), Network analysis, studies in human interaction, The Hague/Paris:

Mouton.Bolk, H. (1989), Organizing, Changing, Simulating, An Organizing Perspective for Management, Consulting

and Science Policy, Delft: Eburon.Boulding, KE. (1981), Foreword, in: Zeleny 1981.Bovens, MA.P. (1990), VeranlWoordelijkheid en organisatie, Zwolle: Tjeenk Willink.Braudel, F. (1966), La Meditmanee et Ie Monde Medite"aneen a I'Epoque de Philippe II, Seconde

edition revue et augmentee, Paris: Librairie Armand Colin.Bredemeier, M.E., Greenblat, C.S., The Educational Effectiveness of Simulation Games: A Synthesis of

Recent Findings, in: Simulation & Games: An International Journal (1981), 12 (3).Bressers, J.ThA., PJ. Klok (1987), Grondslagen voor een instrumententheorie, in: Beleidswetenschap

1987, nr.1, Alphen aan den Rijn: Samsom H.D. Tjeenk Will ink.Broekstra, G. (1984), MAMA: MAnagament by MAtching, a Consistency Model for organizational

assessment and change, in R.Trappl (ed.), Cybernetics and Systems Research, pp.413-420, ElsevierScience Publishers BV (North Holland).

Broekstra, G. (1986a), Organizational humanity and architecture: duality and complementarity ofPAPA-logic and MAMA-logic in managerial conceptualizations of change, Cybernetics and Systems:An International Journal, 17, pp.13-41.

Broekstra, G. (1986b), The paradigm-shift from architectured inhumanity to humanized architectures incorporate reality, in R.Trappl (ed.), Cybernetics and Systems, pp.301-308, Dordrecht: ReidelPublishing Company.

Broekstra, G. (1986c), A soft systems methodology for organizational assessment and change from aholistic perspective, in John A. Dillon (ed.), Proc. International Conference on Mental Images,Values, & Reality, pp. J-57-69, Philadelphia: Soc. General Systems Research.

Broekstra, G. (1989), Creating Intelligent Organizations (in Dutch), Delft: Eburon.Broekstra, G. (1990), Holomanagement: the 21st century visionary leadership style for creating intelligent

organizations, in R.Trappl (ed.), Cybernetics and Systems, pp.593-600, Singapore: World Scientific.Broekstra, G. (1991a), Organizational closure and the quantum view of organizations. In McJackson

(ed.), Systems Thinking in Europe, New York: Plenum Press. To appear.Broekstra, G. (1991b), Effective Management = Holomanagement, in Proc. 35th Annual Meeting ISSS,

Sweden. To appear.Broekstra, G. (1991c), Chaos, the fifth environment, and the revolution of interorganizational

cooperation, in De Zeeuw (ed.), Problems of Support, Survival & Culture, Amsterdam.Brown, J.S., Collins, A., and Duguid, P., Situated Cognition and the Culture of Learning, In: Educational

Researcher, 1989, 18(1).Bruha, T. (1986), Rechtsangleichung in der Europiiischen Wirtschaftsgemeinschaft: Deregulierung durch

'Neue Strategie?', 46 Zeitschrift fur ausliindisches und offentliches Recht pp.1-33.Bruijn, JA. de (1990), The Ministry of Economic Affairs and Economic Subsidies, The Hague: Vuga

(in Dutch).Bruijn, JA. de, E.F. ten Heuvelhof (1991), Instmments for Government Action. About Complex

Networks and a Second Generation Policy Instntments, Leyden: Stenfert Kroese (in Dutch).

234

Budde, A, J. Child, A Francis, A Kieser (1982), Corporate Goals, Managerial Objectives, andOrganizational Structures in British and West German Companies, 3 Organization Studies pp.I-32.

Burggraeve, R. (1977), Ethische grondslagen van een menswaardige samenleving. De bijdrage van Em­manuel Levinas in: Ethische vragen voor onze tijd, Antwcrpen: De Nederlandse Boekhandel.

Burrell, G., G. Morgan (1979), Sociological paradigms and organizational analysis, Aldershot: Gower.Cangelos~ V.E., W.R. Dill (1965), Organizational Learning: Observations towards a theory, in Admini-

strative Science Quarterly, 10, pp. 175-203.Cawson, A (1986), Corporatism and political theory, Oxford: Basil Blackwell.Checkland, P. (1985), Systems Thinking. Systems Practice, Chichester: Wiley & Sons.Cobb, R,W., e.D. Elder (1972/1983), Participation in American Politics: The Dynaics of Agenda-Buil­

ding, Baltimore: The Johns Hopkins University Press.Cohen, J., A Arato (1990), Politics and the reconstruction of the concept of civil society. In: A Honneth,

Th.McCarthy, C. Offe, A Wellmer (Hrsg.), Zwischenbetrachtungen; im Prozess der Aujkliirung,Frankfurt am Main: Suhrkamp.

Cooper, R. (1989), Modernism, post modernism and organizational analysis 3: The contribution ofJacques Derrida, Organization Studies, pp.479-502.

Crozier, M., E. Friedberg (1980), Actors and systems; the politics oj collective action, Chicago/London:University of Chicago Press.

Cummings, T., E. Huse (1989), Organization Development and Change (4th Ed.), St Paul: West.Cyert, R., J. March (1963), A Behavioral Theory of the Finn, Englewood Cliffs: Prentice Hall, Inc.Daintith, T. (1986), The Design and Performance of Long-Term Contracts, in T.e. Daintith and

G. Teubner (eds.), Contract and Organisation, Berlin: de Gruyter.Davis, GA. (1973), Psychology ojproblem solving: Theory and practice, New York: Basic Books.Dery, D. (1984), Problem definition in policy analysis, Kansas: University Press of Kansas.Diekmann, A. (1980), Die BeJolgung von Gesetzen: Empirische Untersuchungen zu einer

rechtssoziologichen Theorie, Berlin: Duncker and Humblot.Dijk, N. van (1989). Een methodische strategie van organisatie-verandering. (in Dutch) (A methodical

strategy of organizational change), Delft: Eburon.Doelen, FJ.C. van der (1989), Policy instntments and energy conservation: TIle implementation and

effectiveness of infonnation campaigns and subsidies directed at industrial energy conservation Jrom1977 to 1987, Enschede: TU Twente (in Dutch).

Doise, W., G. Mugny (1984), The social development oj the Intel/ect, Oxford: Pergamon.Dongen, HJ. van (1979), Technology assessment en sociale rationaliteit, inaugural, Delft: Interuniversitair

Instituut Bedrijfskunde.Dongen, HJ. van (1982), Technologie en Sociale Verandering in J. Berting en T. Huppes: Naar een in

fonnatiemaatschappij, Leiden: Stenfert Kroese.Dongen, HJ. van (1983), Regeren is achterom zien, methodische kanttekeningen bij het nadenken over

de staat, in: R.J. in 't Veld en Tj. Koning (red.), Toekomstdenken en openbaar bestuur, 's­Gravenhage: Staatsuitgeverij.

Dongen, HJ. van (1988). TIle social Construction oj Futures. Internal report, Rotterdam: EUR/RS.Dongen, H,J. van, A. Maas, H. Bolk, T. de Graaf, J. Pavel (1990), Netwerking en netwerkvorming

faciliteren. (in Dutch) (Facilitating networking and the formation of networks: informing andadvising medium and small-scali industry). Internal report, Leiden: Intervisie.

Dongen, HJ. van (1991), Organizing Realities: Rethinking organizing and social integration. Unpubl. doc.Drupsteen, T.G. (1985), Een recht-gestuurd milieu?, in MA.P. Bovens and W,J. Witteveen, eds. Het

Schip van Staat. Beschouwingen over Recht, Staat en Sturing, Zwollc: Tjeenk Willink.Duncker, K., On problem solving, In: Psychological Monographs, 1945, no. 270.Duyndam, J. (1984), De meervoudigheid van de mens als voorwaarde van ethiek, Delft: Eburon.Easterbrook, F.H., D.R. Fischel (1982), Antitrust Suits by Targets of Tender Offers, 80 in: Michigan

Law Review pp.1155ff.Edelman, L.B. (1990), Legal Environments and Organizational Governance: The Expansion of Due

Process in the American Workplace, 95 American Journal of Sociology pp.I401-1440.Edwards, A.R. (1990), Planning betwist; communicatieve strategieen van boeren en natuurbeschermers in de

ruilverkaveling Wommeis, Utrecht: Jan van Arkel.

235

Eemeren, F.H. van, R. Grootendorst (1984), Speech Acts in Argumentative Discussions, Dordrecht: ForisPublications

Ehrlich, E. (1913), Grundlegung der Soziologie des Reclus, Reprint 1967, Berlin: Duncker and Humblot.Eichmann, R. (1989), Systemische Diskurse: zur produktiven Nutzung von Dissens. In: M. Glagow, H.

Willke, H. Wiesenthal (Hrsg.), Gesellschaftli.he Steuerungsrationalitiit und partikulareHandlungsstrategien, Pfaffenweiler: Centaurus-Verlagsgesellschaft.

Eigen, M. (1987), Stufen zum Leben: Die fmhe Evolution im Visier der Molekularbiologie, Miinchen:Piper.

Eigen, M., P. Schuster (1979), The Hypercycle: A Principle of Natural Self-Organisation, Berlin: Springer.Eigen, M., Winkler, R. (1987), Het Spel, Bert Bakker, Amsterdam.Elias, N. (1939/1982), Uber den Prozess der Zivilisation. Basel: Haus zum Falken.Elias, N. (1969), Uber den Prozess der Zivilisation, Bern: Franke Verlag.Ellwein, T., JJ. Hesse, R. Mayntz, F.W. Scharpf (1987), Jahrbuch zur Staats- und Verwaltungswissenschaft,

Band 1/1987, Baden-Baden: Nomos Verlagsgesellschaft.Elmore, R.F. (1985), Forward and backward mapping: reversible logic in the analysis ofpublic policy, in:

Hanf and Toonen, 1985.Etheredge, L.S. (1981),Governmental Learning: an overview, in S.L. Long (ed.), Handbook of Political

Behavior, New York: Plenum Press.Evan, M. (1972), An organization-set model of interorganizational relations, F. Tuite, M. Radnor and

R.K Chisholm (eds.), Interorganizational decision making, Chicago: Aldine-Atherton Publishing Co.Evans, WA., Sculli, D., An Evaluation of Business Games with respect to the Development of

Managerial Talent, in: Sagset Journal, 1984, 14 (1).Ferner, A (1988), Governments, Managers and Industrial Relations, Oxford: Basil Blackwell.Fisher, R., W. Ury (1981), Getting To Yes. Negotiating Agreement Without Giving In, Boston: Houghton

Mifflin.Foerster, H., von (1984) Principles of Self-Organization in a Socio-Managerial Context. In: Ulrich, R,

and Probst, GJ.B. (eds.), Self-Organization and Management of Social Systems, Berlin: Springer­Verlag.

Foulkes, F.K (1980), Personnel Policies in Large Nonunion Enterprises, Englewood Cliffs: Prentice Hall.Galjaard, J.H. (1979), Infonnatisering, paradox van organisatietecllllologie, Delft: Eburon.Gawthrop, L.c. (1984), Public Sector Managemelll, Systems, and Ethics, Bloomington: Indiana University.Gergen, KJ. (1985), The Social Constructionist Movement in Modern Psychology, American Psychologist,

vol.40, pp. 266-275.Gergen, KJ. (1990), Het binnenste buiten. Organisatietheorie in de postmoderne tijd, Filosofie in Bedrijf

vol.2 or.l, pp.21-30.Geyer, F., J. van der Zouwen (eds.)(1986), Sociocybernetic Paradoxes, London: Sage.Giddens, A (1979), Central Problems in Social theory, London: University of California Press.Giddens, A (1984), The Constitution of Society, Cambridge: Polity Press.Giddens, A, J. Turner (1987), Social Theory Today, California: Stanford University Press.Giddens, A (1990), 17Je Consequences of Modernity, Cambridge: Polity Press.Gleick, J. (1988), Chaos, making a new science, Harmondsworth: Penguin.Godfroy, AJA. (1981), Netwerken van organisaties: strategieen, spelen, structuren, 's-Gravenhage: Vuga.Godfroy, AJA. (1989), Netwerken van organisaties in: Handboek Organisaties, Alphen aan den Rijn:

Samsom.Goguen, JA. and FJ. Varela (1973/1979), Systems and Distinctions, Duality and Complementarity, Inter

national Journal of General Systems, vol.5, pp.31-43.Goldstein, J. (1986). A far-from equilibrium approach to resistance to change. Organizational Dynamics,

16-26.Gotthold, J., R. Vieth (1982), Erfiillung von 6ffentlichen Aufgaben durch Verhandlungen mit Privaten

im Bereich der Wettbewerbspolitik, in V. Gessner & G. Winter (eds.), Rechtsfonnen derVeif1echtung von Staat und Wirtschaft, Opladen: Westdeutscher Verlag.

Goud, J.F., Wat men van zichzelf eist, eist men van een heilige. Ter Herkenning, lIe jrg. nrs. 2, 3 en 5,Den Haag: Boekencentrum.

Graaf, J. de. (1972), Elementair begrip van de ethiek, Haarlem: De Erven F. Bohn.

236

Gunsteren, H.R. (1985), Het leervermogen van de overheid, in: MA.P. Bovens en WJ. Witteveen, Hetschip van staat, Zwolle: W.EJ. Tjeenk Willink.

Gunsteren, H.R. van (1985), Het begrip sturing: schetsen ten behoeve van onderzoek, in MA.P. Bovensand WJ. Witteveen, eds. Het Schip van Staat. Beschouwingen over Recht, Staat en Sturing, Zwolle:Tjeenk Willink.

Habermas, J., N. Luhmann (1971), Theorie der Gesellschaft oder Sozialtechnologie?, Frankfurt am Main:Suhrkamp.

Habermas, J. (1981), Theorie des kommunikativen Handelns, lund ll, Frankfurt am Main: Suhrkamp.Habermas, J. (1985), Der philosophische Diskurs der Moderne, Frankfurt am Main: Suhrkamp.Hall, R.H., J.P. Clark, P.c. Giordano, P.V. Johnson, M. van Roekel (1978), Interorganizational

coordination in the delivery of human services in: Karpik (ed.).Hanf, K, F. Scharpf (eds)(1978), Interorganizational Policy-making, Limits to Central Coordination and

Control, London: Sage.Hanf, K, ThAJ. Toonen (1985), Policy implementation in federal and unitary systems, Dordrecht/­

Boston/Lancaster: Martinus Nijhoff.Harrington, C. (1985), Shadow Justice: The Ideology and Institutionalization of Alternatives to Courts,

Westport: Greenwood Press.Hart, P. 't, JA.M. Hufen, MJ. van Duin (1988), De lerende overheid: Mogelijkheden en grenzen van

een modieuze metafoor, in Beleid en Maatschappij, 2, pp. 83-102.Harter, PJ. (1982), Negotiating Regulations: A Cure for Malaise, 71 Georgetown Law Journal pp.1-118.Hedberg, B. (1981),How Organization Learn and Unlearn, in P.c. Nustrom en W.H. Starbuck (eds.).

Handbook of Organizational Design, New York: Oxford University Press. pp. 3-27.Heidorn, J. (1982), Legitimitiit lind Regierbarkeit, Berlin: Duncker & Humblot.Helmers, H.M., RJ. Mokken, RC. Plijter, F.N. Stokman (1975), Graven naar macht, Amsterdam: Van

Gennep.Henry, S. (1983), Private Justice, Boston: Routledge and Kegan Paul.Henry, S. (1987), The Construction and Deconstruction of Social Control: Thoughts on the Discursive

Production of State Law and Private Justice, in J. Lowmanm, R. Menzies and T. Palys (eds.),Transcarceration, Farnborough: Gower.

Herbst, Ph.G. (1976), Alternatives to Hierarchies, Leiden: Nijhof.Heylighen, F. (1990), A new transdisciplinary paradigm for the study of complexsystems. In: Heylighen,

F. Rosseel, E. and Demeyere, F., (eds.) Self-steering and cognition in complex systems: Towards anew cybernetics, London: Gordon and Breach.

Heyting, G. (1966), Intuitionism: An Introduction, Studies in Logic and Mathematics, Amsterdam: NorthHolland.

Hinich. MJ., (1991) A spatial theoretic approach to environmental politics, in: Kraan, DJ. and RJ. In'tVeld (eds.), Environmental Protection: Public or Private Choice, Dordrecht: Kluwer AcademicPublishers.

Hoed, P. den et al. (1983), Planning As Enterprise, Scientific Council for Government Policy, The Hague(in Dutch).

Hoffmann-Riem, W. (1989), Konfliktmiltler in Verwaltungsverhandlungen, Heidelberg: C.F. Miiller.Hogwood, B.W., B.G. Peters (1982), The Dynamics of Policy Change: Policy Succession, Policy

Sciences, 14, pp.225-245.Hogwood, B.W., B. Guy Peters (1985), The Pathology of Public Policy, Oxford: Clarendon Press.Hood, c.c. (1983), The Tools of Govemment, London: Chatham House Publishers.Hoogerwerf, A, (1984), Beleid berust op veronderstellingen, in: Acta Politica 1984 orA, ppA93-531, Den

Haag: VNG.Hoppe, R (1989), Het beleidsprobleem geproblematiseerd, Muiderberg: Coutinho.Hucke, J. (1978), Bargaining in Regulative Policy Implementation: the case of air and water pollution

control, 4 Environmental Policy and Law, pp.109-115.Hufen, JA.M. (1990), Instrumenten in het technologiebeleid, Groningen: Wolters Noordhoff (In Dutch).Hufen, JA.M., A.B. Ringeling (1990), Beleidsnetwerken, Den Haag: Vuga.Hutter, M. (1989), Die Produktion von Recht. Eine selbstreferentielle Theorie der Wirtschaft und der Fall

des Arzneimittelpatentrechts, Tiibingen: Mohr and Siebeck.

237

Jantsch, E. (1980), The self-organizing universe, Oxford: Pergamon Press.Jessop, B. (1990), State theory, Cambridge: Polity Press.Joerges, C. (1981), Verbraucherschutz als Rechtsproblem, Heidelberg: Recht und Wirtschaft.Jonas, F. (1969), Geschichte der Soziologie, Hamburg: Rowohll.Karpik, L. (ed.), (1978), Organization and environment, London: Sage.Katz, D., H. Kahn (1978), The social Psychology of Organizations, New York: Wiley.Kaufman, A, EJ Mestmacker, H.F. Zacher (1988), Rechtsstaat und Menschenwiirde, Festschrift fiir Wer

ner Maihofer zum 70. Geburtstag, Frankfurt am Main: Vottorio Klosterman.Kaufmann, EX. (1988), Steuerung wohlfahrtsstaatlicher Ablaufe durch Recht, in D. Grimm and W.

Maihofer (eds.), Gesetzgebungstheorie und Rechtspolitik, Opladen: Westdeutscher Verlag, pp.65-108.Kaufmann, H. (1976), Are Govemment Organizations Immortal, Washington: Brookings Institution.Kaupen, W. (1975), Die Grenzen des Rechts als Mittel gesellschaftlicher Gestaltung, in Institut fiir

Gesellschaftspolitik, Wien (ed.), Recht und Politik, pp.33-42.Kee, B. (1988/1989), Organisatie: een gekwalificeerd geheel van personen, in Wijsgerig perspectief,

2ge jrg. Of. 5, Meppel: Boom.Kee, B. (1990), Theorie en praktijk, paper, Amsterdam: VUAKeulen, S. van (1989), De dubbelzinnigheid van het spel (in Dutch) (Equivocality of game), in: G.

Vandenakker e.a. (ed.). Filosojiedag Utrecht 1989 (Utrecht day of philosophy), Delft: Eburon.Kickert, W.J.M. (1980), Organization of Decision-making, Amsterdam: North Holland.Kickert, W.J.M. (1991), Steering at a Distance, a New Paradigm of Governance in Dutch Higher Educa­

tion, to be published in Govemance.Kingdon, J.W. (1984), Agenda's, Aleemaeives, and Public Policies, Boston: Little, Brown and Compagnies.Kirsch, G. (1988), Die Deregulierungsdebatte: Anmerkungen zu einem bornierten Streit, in Th. Schmid

(cd.), Entstaatlichung: Neue Perspekeiven auf das Gemeinwesen, Berlin.Klabbers, J. (1975), Interactive simulation: On-line interaction between man and machine for the study

and management of social systems, In: 1nfonnatie, 17 (10).Klabbers, J., P. van der Hijden, Hoefnagels, K., Truin, G. (1980), Development of an interactive simula­

tion game: a caste study of dentist, In: Simulation & Games: An Intemational Journal, 33 (3).Klabbers, J. (1986), Improvement of (sclf-)steering through support systems. In: Geyer, F., and van der

Zouwen, J., (eds.) Sociocybemetic Paradoxes, London: Sage.Klabbers, J., Scheper, W. Takkenberg, C. and Crookall, D., (cds.) (1989), Simulation-Gaming: On the

Improvement of Competence in Dealing wieh Complexity, Uncertainty and Value Conflicts, Oxford:Pergamon Press.

Klabbers, J., B. van der Waals (1989), From Rigid-rule to free-form games: Observations on the role ofrules. In: Klabbers, J. et al. (cds.) Simulaeion Gaming: On the Improvement of Competence inDealing with Complexity, Uncertaincy and Value Conflices, Oxford: Pergamon.

Klabbers, J. (1989), Methodological aspects for designing learning environments (in Dutch). In: Klep, J.and Kommers, P. (cds.), Courseware en Leerplan-oncwikkeling Didactische systeemanalyse, Enschede:SLO.

Klabbers, J. (1991), The fimo managers, Center for Policy Analysis and Advice, Nijmegen: CBA/N.Klijn, E.H., G.R. Teisman (1990), The power of dispersion: effective policy making in networks, con­

ference paper November 21th 1990, Rotterdam: Departement of Public Administration ErasmusUniversity.

Klijn, E.H. (1990), De vissen en het aas; de toepassing van twee financieringsinstrumenten in devolkshuisvesting in: Hufen, JA.M., AB. Ringeling, Beleidsnecwerken; Overheids- semioverheids- enparticu/ire organisaties in wisselwerking, 's-Gravenhage: Vuga.

Knaap, P. v.d. (1991), Leertheorieen, Rotterdam: Risbo.Koppenjan, J.F.M., AB. Ringeling, R.HA. te Velde (1987), Beleidsvorming in Nederland, 's-Gravenhage:

Vuga.Koppenjan, J.F.M. (1991), Falen en leren rond de paspoortaffaire: de hardleersheid van een ministerie

geanalyseerd, in Beleid en Maatschappij, 1, pp. 20-30.Kratky, K.W., F. Wallner (cds.) (1990), Gnmdprinzipien der Se/bstorganization, Darmstadt:

Wissenschaftliche Buch Geschiift.Krohn, W., G. Kiippers (Hrsg.), Emergenz wzd Selbstorganisaeion, 1990 (in print).

238

Kunneman, H. (1986), De waarheidstrechter, Meppel: Boom.Kunneman, H. (1990), De grenzen van de henneneutiek. In: GA.M. Widdershoven en Th. de Boer (red.).

Hermeneutiek in discussie, Delft: EburonLaat, W. de (1983), Vragen naar de onbekende weg. (in Dutch) (Asking for the untrodden Path. A Social

Research Project into the topic of Expertise in the Framework of a Theory about Organizing).Ph.D. Thesis, Delft: Eburon.

Laat, WA.M., de (1984), In tennen van organiseren, over het raakvlak van sociale psychologie,organisatietheorie en bedrijfskunde, een inleiding, Delft: Eburon.

Laat, WA.M. de (1985), Vragen naar de onbekende we& een sociaal wetenschappelijk onderzoek naardeskundigheid in het kader van een theorie over organiseren, Delft: Eburon, handelseditie.

Lammers (1987), Organisaties vergelijkenderwijs, Utrecht: Spectrum.Laszlo, E. (1986), Systems and Societies, The Basic Cybernetics of Social Evolution, in Geyer and Van

dec Zouwen, op.cit.Laumann, E.O., F.D. Pappi (1976), Networks of collective action, New York: Academic Press.Leeuw, F.L. (1988), Overheidsbeleid en Sociaal- en Geesteswetenschappelijk onderzoek; een analyse van

beleidstheorieen, Zoetermeer: Ministery for Education and Sciences.Lehmbruch, G., P.C. Schmitter (ed.), (1982), Pattems of corporatist policy-making, London: Sage.Lehning, P.B., M.P.CM. van Schendelen (eds.) (1981), Actualiteit van politieke filosofie, Amsterdam:

Intermediair bibliotheek.Leon, P. de (1978), Policy Termination, Policy Analysis, 3, pp.369-393.Levinas, E. (1971), Het menselijk gelaat, Bilthoven: Ambo.Levinas, E. (1987), De IOtaliteit en het Oneindige (oorspr. 1961), Baarn: Ambo.Levinas, E. (1989), De tijd en de ander, Baarn: Ambo.Levine D.N. (ed.) (1971) Georg Simmel on Individuality and Social FomlS, Chicago: University of

Chicago Press.Levy, A., U. Merry (1986), Organizational Transfonnation, New York: Praeger.Lewicki, RJ., JA. Litterer (1985), Negotiation, Homewood Illinois: Irwin.Linstone, HA. (1984), Multiple Perspectives for Decision Making, New York/Amsterdam/Oxford: North

Holland.Luhmann, N. (1970), Oeffcntliche Meinung. In: Politische Vierteljahresschrift 11, pp.2-28.Luhmann, N. (1981), Politische 771eorie im Wohlfahrtsstaat, Miinchen: Giinther Olzog Verlag.Luhmann, N. (1982a), The differentiation of society, New York: Columbia University Press.Luhmann, N. (1982b), Autopoiesis, Handlung und kommunikative Verstiindigung, in: Zeitschrift fUr

Soziologie 1982, nr.4 p.366-379, Stuttgart: F. Enke Verlag.Luhmann, N. (1984), Soziale Systeme: Grnndriss einer allgemeinen 77leorie. Frankfurt am Main: Grnndriss

einer allgemeinen 77leorie, Frankfurt am Main: Suhrkamp.Luhmann, N. (1984b), Staat und Politik. Zur Semantik der Selbstbeschreibung politischer Systcme, in:

Bermbach 1984 .Luhmann, N. (1984c), Die Wirtschaft der Gesellschaft als autopoietisches System, in: Zeitschrift fUr

Soziologie 1984, nr.4 p.308-327: Stuttgart: F. Enke Verlag.Luhmann., N. (1986), Die soziologische Beobachtllng des ReclUS, Frankfurt am Main: Alfred Metzner

Verlag.Luhmann, N. (1986a), The Autopoiesis of Social Systems, in F. Geyer and J. van der Zouwcn, eds.

Sociocybemetic Paradoxes: ObselVation, Control and Evolution of Self-Steering Systems, London:Sage.

Luhmann, N. (1986b), The Theory of Social Systems and Its Epistemology: Reply to Danilo Zolo'sCritical Comments, 16 Philosophy of the Social Sciences, pp.129-134.

Luhmann, N. (1988a), Closure and openness: on reality in the world of law, in: Teubner 1988a.Luhmann, N. (1988b), Die Wirtschaft der Gesellschaft, Frankfurt am Main: Suhrkamp.Luhmann, N. (1986/1989), Ecological Commllnication, London: Polity Press.Luhmann, N. (1989), Wirtschaft lind Recht: Probleme strnklllreller Koppillng, Bielefeld: Manuskript.Luhmann, N. (1990), The Coding of the Legal System, in Gunther Teubner, ed. Die Wissenschaft der

Gesellschaft, Frankfurt: Suhrkamp.Maas, AJJA. (1988), Ongedefinieerde rnimten. (In English: Undefined Social Spaces), Delft: Eburon

239

Macaulay, S. (1963), Non-Contractual Relations in Business: A Preliminary Study, 28 American Socio­logical Review pp.55-67.

Macaulay, S. (1986), Private Government, in L. Lipson & S. Wheeler (eds.), Law and the SocialSciences, New York: Russell Sage.

Majone, G. (1989), Evidence, Argument, & Persuasion in the Policy Process, New Haven: Yale UniversityPress.

Mandell, M.P. (1989), Network Management: Strategic Behavior in the Public Sector, in: Robert W.Gage and Myrna P. Mandell (eds), Strategies for Managing Intergovernmental Policies and Networks,New York: Praeger.

Mannheim K. (1951), Man and Society in an Age of Reconstruction, London: Routledge Kegan Paul Ltd.March, J.G.M., J.P. Olsen (1976), Ambiguity and Choice, Bergen: Universitets forlaget.Maturana, H.R., Fl. Varela (1972/1980), Autopoiesis and Cognition, the Realization of the Living,

Dordrecht: Reidel.Maturana, H.R. (1980), Man and society, in: Benseler c.s. 1980.Maturana, H.R., Fl. Varela (1980), Autopoiesis and Cognition, the Realisation of the Living, Boston:

Reidel.Maturana, H.R. (1981), Autopoiesis, in Zeleny op.cil.Maturana, H.R., Fl. Varela (1987), The tree of knowledge, Boston: New Science Library.Maturana, H.R., Fl. Varela (1987), Der Baum der Erkenntnis, Miinchen: Scherz.Mayntz, R. (1987), Politische Steuerung und gesellschaftliche Steuerungsprobleme - Anmerkungen zu

einem theoretischen Paradigma, in T. Ellwein et aI., eds. Jahrbuch zur Staats-undVerwaltungswissenschaft, Band 1, Baden-Baden: Nomos.

McKelvey, B. and H. Aldrich (1983), Populations, Natural Selection and Applied Organizational Science,Administrative Science Quarterly 28, pp.101-128.

Meer, F.B.L. van der (1983), Organisatie als een spel: sociale simulatie als methode in onderzoek naOT organiseren (in Dutch) (Organization as a Game: Social Simulation as a method for Research intoOrganizing), Enschede.

Meer, F.B.L., van der (1987), Organisatieprocessen: een onderzoek met behulp van sociale simulatie, in:M&O, nr.6, pp.456-471.

Meltsner, Al. and C. Bellevita (1983), TIle Policy Organization, Beverly Hills: Sage.Merry, S.E. (1988), Legal Pluralism, 22 Law & Society Review, pp.869-901.Miles, R.E. and e.e. Snow (1978), Organizational Strategy, Structure, and Process, Tokyo: McGraw-Hill

Kogakusha, Ltd.Milgram, S. (1974), Obedience to authority, New York: Harper & Row.Miller, D., P.H. Friesen (1984), Organizations: A Quantum View. New Jersey: Prentice-Hall,Inc.Milward, H.B., G.L. Wamsley (1985), Policy subsystems, networks and the tools ofpublic management, in:

Hanf and Toonen.Mingers, J. (1989), An introduction to autopoiesis- implications and applications. Systems Practice, 2,

pp.159-18O. For the debate between Mingers and Robb, consult: Research notes, Systems Practice,2(1989), pp.343-360.

Mintzberg, H. (1983), Structures in jives; designing effective organizations, Englewood Cliffs: Prentice Hall.Mintzberg, H. (1989), Mintzberg on Management: Inside our Strange World of Organizations. New York:

The Free Press.Mitchel, J. C. (1%9), The concepts and use of social networks, J.e. Mitchell (ed.), Social networks in

Urban situations, Manchester: Manchester University Press.Mitnick, B.M. (1980), The Political Economy of Regulation, New York: Colombia University Press.Mnookin, R.H., L. Kornhauser (1978), Bargaining in the Shadow of the Law: The Case of Divorce, 88

Yale Law Journal pp.950-997.Moore, 0., A. Anderson (1975), Some principles for the Design of Clarifying Educational Environments,

in: Greenblat,e.S., andDuke, R. (eds.) Gaming Simulation, New York: Halsted.Moore, S.F. (1986), Legal Systems of the World: An Introductory Guide to Classifications, Typological

Interpretations, and Bibliographic Resources, in: L. Lipson and S. Wheeler (eds.), Law and theSocial Sciences, New York: Russell Sage Foundation.

Morgan, G. (1986), Images of Organization, Beverly Hills/London/California: Sage Publications, Inc.

240

Nader, L. (1984), The Recurrent Dialectic Between Legality and Its Alternatives, 132 University ofPennsylvania Law Review p.621.

Nadler, D., M. Tushman (1988). Strategic Organization Design: concepts, tools and processes. Glenview, IU:Scott, Foresman.

Nahamowitz, P. (1990), Interventionistisches Recht als Steuerungskonzept, Jahresschrift fUr Rechtspolitologie pp.7-36.

Negandhi, A.R. (ed.), (1975), Interorganization theory, Kent.Nozick, R. (1974), Anarchy, State and Utopia, Oxford: Basil Blackwell.Nystrom, P.C., W.H. Starbuck (ed.), (1981), Handbook of organizational design, Oxford: Oxford

University Press.Offe, C. (1990), Sozialwissenschaftliche Aspekte der Kontroverse tiber Regulierung und Deregulierung,

in 11. Hesse and C. Zopel (eds.), Der Staat der Zukunft, Baden-Baden: Nomos.Okoth-Ogendo, H.W.a. (1979), The Imposition of Property Law in Kenya, in S.B. Burman and B.E.

Harrell-Bond (eds.), The Imposition of Law, New York: Academic Press.Opp, K.D. (1973), Soziologie im Recht, Reinbek: Rowohlt.Ornstein, D.S. (1989), Ergodic Theory, Randomness, and "Chaos ", In: Science, 243 (1).Ostrom, V. (1985), Muitiorganizational arrangements in the govemance of unitary and federal systems in:

Hanf and Toonen.Ozawa, c.P., L. Susskind (1985), Mediating Science-Intensive Policy Disputes, 5 Journal of Policy

Analysis and Management pp.23-39.Pavel, J. (1990), Netwerkvonning en netwerking in communities c.q. culturen. (in Dutch) (The Formation of

networks and networking in communities and cultures. A Research study into criteria of ongoingprocesses of interaction). Thesis, Rotterdam: EUR/RSM.

Peirce, Ch. (1940), The Principles of Phenemenology, in: J. Buchler (1940), Philosophical writings ofPeirce, New York: Dover Publishers.

Polanyi, M. (1958), Personal Knowledge: Toward a Post-Critical Philosophy, Chicago: Chicago UniversityPress.

Posner, RA. (1987), The Decline of Law as an Autonomous Discipline: 1982-1987, 100 Harvard LawReview pp.761-780.

Prawat, R.S. (1989), Promoting Access to Knowledge, Strategy, and Disposition in Students: A ResearchSynthesis, In: Educational Research, 59 (1).

Pressman, J.L., A. Wildavsky (1973), Implementation: How Great Expectations in Washington are Dashedin Oakland, Berkeley: University of California Press.

Pressman, J.L. (1975), Federal programms and city politics, Berkeley: University of Columbia Press.Pressman, J.L., A. Wildavsky, (1983), Implementation, Berkeley (3rd edition): University of California

Press.Pressman, J.L., A. Wildavsky (1984), Implementation, Berkeley: University of California Press.Prigogine, I., I. Stengers (1985), Order out of chaos, New York: Bantam.Prins, H. (1980), Offenders, Deviants or Patients? An Introduction to the Study of Socio-Forensic Problems,

London: Tavistock.Raiffa, H. (1982), The Art and Science of Negotiation: How to Resolve Conflicts and Get the Best out of

Bargaining, Cambridge MA: Harvard University Press.Raiser, T. (1990), Uber die Beziehungen zwischen Rechtssoziologie und Rechtsdogmatik, in W.Hoff­

mann-Riem, KA. Mollnau and H. Rottleuthner (eds.), Rechtssoziologie in der DeutschenDemokratischen Republik und in der Bundesrepublik Deutschland, Baden-Baden: Nomos.

Ramanadham, V.V. (1984), The Nan/re of Public Enterprise, London: Croom Helm.Ranger, T. (1983), The Invention of Tradition in Colonial Africa, in E. Hobsbawm and T. Ranger

(eds.), The Invention of Tradition, Cambridge: Cambridge University Press.Ravetz, J. (1990), Knowledge in an Uncertain World, New Scientist, 22 September 1990, p.18.Rawls, J. (1990), A Theory of Justice, Oxford: Oxford University Press (10th Edition).Rehbinder, E. (1982), Vertragsgestaltung, Frankfurt: Metzner.Reich, R.B. (1990), The Power of Public Ideas, Cambridge: Harvard University PressRein, M. (1976), Social Science and Public Policy, New York: Pinguin Education.Rescher, N., R. Brandom (1980), The Logic of Inconsistency, A Study in Non-Standard Possible World

Semantics and Ontology, Oxford: Basil Blackwell.

241

Richardson, G. et.al. (1983), Policing Pollution: A Study of Regulation and Enforcement, Oxford: OxfordUniversity Press.

Ringeling, A.B. (1989), Politici en de overheidsorganisatie, in Bestuur, 4, pp. 111-115.Rogers, D.L., DA. Whetten (eds.), (1982), Interorganizational coordination: theory, research, and

implementation, Iowa: Iowa State University Press.Ronge, V. (1980), Am Staat vorbei, Frankfurt a.M.: Campus.Rosanvallon, P. (1988), The Decline of Social Visibility, in J. Keane (ed.), Civil Society and the State,

London: Verso.Rosenthal, U. (1989), Van begin tot einde: parlementaire enquetes en hun latente gevolgen, in Acta

Politica, pp. 257-271.Rosewitz, B., U. Schimank (1988), Verselbstiindigung und politische Steuerbarkeit gesellschaftlicher

Tei/systeme, Frankfurt/New York: Campus Verlag GmbH.Rottleuthner, H. (1987), Einfiihnmg in die Rechtssoziologie, Darmstadt: Wissenschaftliche

Buchgesellschaft.Rottleuthner, H. (1989), The Limits of Law - The Myth of a Regulatory Crisis, 17 Intemational Joumal

of the Sociology of Law pp.273-285.Rottleuthner, H. (1991), Grenzen rechtlicher Steuerung, in P. Koller and O. Weinberger (eds.), Grund

lagen der Rechtspolitik, Wiesbaden: Steiner. (forthcoming).Ruebens, M. (1990), Sociologie van het alledaagse leven; een grondslagendebat tussen handelingstheorie en

systeemtheorie, Leuven/Amersfoort: Acco.Sabatier, PA. (1988), An Advocacy Coalition Framework of Policy Change and the Role of Policy­

oriented Learning Therein, in Policy Sciences, 21, pp. 129-168.Salamon, L.M. (ed.) (1989), Beyond Privatization, The Tools of Government Action, Washington: Institute

for Fiscal Studies.Santos, Boaventura De Sousa (1987), Law: A Map of Misreading; Toward a Postmodern Conception of

Law, 14 Journal of Law and Society pp.279-302.Schaap, L., MJ.W. van Twist, R.J. in 't Veld (1990), De eigenwijze samenleving, sturing van gesloten

systemen, in Bestuur, lO, pp. 263-267.Schaap, L., M.J.W. van Twist (1990), Integration measures: an autopoietic approach and some implica

tions, paper for the lIAS-conference, Madrid November 1990.Scharpf, F.W., B. Reissert, F. Schnabel (1976), PolitikverOechtung; theorie und empiJie des kooperativen

federalismus in der Bundesrepublik, Regensburg: Scriptor Verlag.Scharpf, F.W., (1978), Interorganizational policy studies: issues, concepts and perspectives in: Hanf and

Scharpf.Scharpf, F.W., B. Reissert, F. Schnabel (1978), Policy effectiveness and conflict avoidance in

intergovernmental policy fonnation in: Hanf and Scharpf.Scharpf, F.W. (1987), Grenzen der institutioncllen Reform, 1 Jahrbuch zur Staats- und Verwaltungswissen

schaft pp.ll1-151.Scharpf, F.W. (1988), Verhandlungssysteme, VerteilungskonOikte und Pathologien der politischen

Steuerung, in Manfred G. Schmidt, cd. Staatstiitigkeit. International und historisch vergleichendeAnalysen, Opladen: Westdeutscher Verlag.

Scharpf, F. (1989), Politische Steuerung und politische Institutionen, 30 Politische Vierteljahresschriftpp.lO-21.

Schattschneider, E.E. (1960), The Semi-sovereign People. A Realist's View of Democracy in America, NewYork: Holt, Rinehart and Winston.

Schmidt, M.G. (Hrsg) (1980), Staatstiitigkeit. International und historisch vergleichende Analysen, Opladen:Westdeutscher Verlag.

Schmitter, P.C, G. Lehmbruch (ed.), (1979), Trends toward corporatist intennediation, London: Sage.Schneider, V. (1988), Politiknetzwerke der Chemikalienkontrol/e; einanalyse einer transnationalenpolitikent

wicklung, Berlin/New York: Walter de Gruyter.Schon, D. (1983), The reflective practioner; how professionals think in action, New York: Basic Books.Schon, D. (1987), Educating the Reflective Practitioner, San Francisco: Jossey-Bass.Schumann, K.F. (1977), Der Handel mit Gerechtigkeit, Frankfurt: Suhrkamp.Schutz A. (1976), The Phenomenology of the Social World, London: Heineman Educational BooksSearle, J.R. (1%9), Speech Acts, Cambridge: Cambridge University Press.

242

Selznick, P. (1969), Law, Society and Industrial Justice, New York: Russell Sage.Shubik, M. (1983), Gaming: A state-of-the-art survey. In: Stahl, I., (ed.) Operational Gaming, Oxford:

Pergamon.Simmel G. (1957), Vom Wesen der Kultur, in Briicke und Tor, Stuttgart: K.F. Koehler VerlagSimon, HA. (1969), The Sciences of the Anificial, Cambridge MA: MIT Press.Smircich, L., Ch. Stubbart (1985), Strategic Management in an Enacted World, Academy of Management

Review vol.W pp.724-36Smircich, L., Ch. Stubbart (1989), Do Organizations 'Enact' Their Environments, in Creative organisation

theory: A Resource book, Newbury Park: SAGE.Snellen, I.Th.M. (1987), Boeiend en geboeid, ambitities en ambivalenties in de bestuurskunde, Alphen aan

den Rijn: Samsom H.D. Tjeenk Willink.Snyder, F. (1981a), Capitalism and Legal Change: An African Transformation, New York: Academic

Press.Snyder, F. (1981b), Colonialism and Legal Form: The Creation of 'Customary Law' in Senegal', 19

Journal of Legal Pluralism and Unofficial Law pp.49-90.Spencer Brown, G. (1972), Laws of Form, New York: Julian Press.Spencer Brown, G. (pseudonym J. Keys, 1972), Only Two can play this Game, New York: Julian Press.Starr, J., J. Collier (1989), History and Power in the Study of Law: New Directions in Legal Anthropology,

Ithaca: Cornell University Press.Stigler, GJ., G.S. Becker (1977), De gustibus non est disputandum, 67 American Economic Review

pp.76-90.Stout, H., J. Stoop (1991), Autopoiese aan de Maas, ofwei: Instrumentele Wetgeving Reddeloos

Verdronken, 6 RegelMaat, pp.5-9.Strasser, S. (1981), De burger voorbij, Baarn: Ambo.Susskind, L., G. McMahon (1985), The Theory and Practice of Negotiated Rulemaking, 3 Yale Journal of

Regulation, pp.133-165.Susskind, L., J. Cruikshank (1987), Breaking 17le Impasse: Consensual Approaches to Resolving Public

Disputes, New York: Basic Books.Swaan, A. de (1988), In care of the state, Cambridge/Oxford: Polity Press.Teubner, G.C.M. (1978), Organisationsdemokratie und Verbandsve!fassllllg, Tiibingen: Mohr and Siebeck.Teubner, G.C.M. (1983), Substantive and reflexive elements in modern Jaw, in: Law and socieryview 1983,

nr.Z.Teubner, G.C.M. (1984), Autopoiesis in law and society: a rejoinder to Blankenburg, in: Law and society

review 1984, nr.2 p.291-301.Teubner, G.C.M. (1984), After Legal Instrumentalism? Strategic Models of Post-regulatory Law, 12

International Journal of the Sociology of Law, pp.375-400.Teubner, G.C.M., H. Willke (1984), Kontext und Autonomie. Gesellschaftliche Selbststeuerung durch

reflexives Recht, In: Zeitschrift fUr Rechtssoziologie 5, pp.4-35.Teubner, G.C.M. (1987), Juridification of Social Spheres: A Comparative Analysis in the Areas of Labor,

Corporate, Antitrust, and Social Welfare Law, Berlin/New York: de Gruytcr.Teubner, G.C.M. (eds) (1988), AlIlopoietic Law, a New Approach to Law and Society, Berlin: De Gruyter.Teubner, G.C.M. (1988a), Enterprise Corporatism: New Industrial Policy and the 'Essence' of the Legal

Person, 36, in: The American Journal of Comparative Law pp.130-155.Teubner, G.C.M. (1988b), Napoleons verlorener Code, Eigendynamik des Rechts als politisches Problem,

in: Kaufman c.s. 1988.Teubner, G.C.M. (1988b), Hypercycle in Law and Organization: The Relationship between Self­

Observation, Self-Constitution and Autopoiesis, in: European Yearbook in the Sociology of Lawpp.43-70.

Teubner, G.C.M. (1989), How the Law Thinks, Towards a Contructivist Epistomology of Law, in: Lawand Society Review, 23, 5, pp.727-759.

Teubner, G.C.M. (1989a), Recht als alllopoietisches System, Frankfurt am Main: Suhrkamp.Teubner, G.C.M. (1989b), "And God Laughed.. .": Indeterminacy, Self-reference and Paradox in Law, in

Christian Joerges and David Trubek, eds. Critical Legal Thought: An American-German Debate,Baden-Baden: Nomos.

243

Teubner ed. (1990) State, Law, Economy as Autopoietic Systems: Regulation and Autonomy in a NewPerspective, Berlin: de Gruyter.

Teubner, G.C.M. (1990a), Die vielkapfige Hydra: Netzwerke als kollektive Akteure hahere Ordnung in:Krohn 1990 (in print).

Teubner, G.C.M. (1991a), 1st das Recht auf Konsens angewiesen?, in HJ. Giegel (ed.), Kommunikationund Konsens, Frankfurt: Suhrkamp (forthcoming).

Teubner, G.C.M. (1991b), Law as an Autopoietic System, London: Blackwell. (forthcoming).Teubner, G. (1991c), Regulatory Law: Chronicle of a Death Foretold, in 1. Lenoble (ed.) Law and the

Crisis of the Welfare State, Bruxelles: Scientia (forthcoming).Toffler, A (1968), Future shock, New York: Random House.Twist, MJ.W. (1991), Organizeuren: configupraatjes en autopoiezie, Rotterdam: Risbo.Uribe, R.B. (1981), Modeling Autopoiesis, in Zeleny, op.cil.Varela, F.G., H.R. Maturana and R. Uribe (1974), Autopoiesis, The Organization of Living Systems, Its

Characterization and a Model, Biosystems, 5, pp.187-1%.Varela, FJ. (1978), On being autonomous: the lessons of natural history for systems theory. In GJ.K1ir

(ed.), Applied General Systems Research: Recent Developments and Trends, pp.77-84, New York:Plenum Press.

Varela, FJ. (1979), Principles of Biological Autonomy. New York: Elsevier North Holland.Varela, FJ. (1981), Describing the Logic of Living, in: Zeleny.Veld, R.I. in 't (1985), Het leervermogen van de overheid, in MA.P. Bovens, W. Derksen, WJ.

Witteveen, (eds.), Stun'ng van de samenleving. Zwolle: WEJ. Tjeenk Willink.Veld, R.I. in 't (1989), De verguisde staat, 's-Gravenhage: Vuga.Veld, R.I. in 't (1991), Roadpricing, a logical failure, in: Kraan, DJ. and R.I. in't Veld, Environmental

Protection: public or private choice, Dordrecht: K1uwer Academic Publishers.Velde, E. van de (1977), Rede en Vrede, Kampen: 1.H. Kok.Ven, AH. van der, D.C. Emett, R. Koening jr. (1975), Frameworks for interorganizational analysis, in:

Negandhi.Voogt, AA. (1990), Managen in een meervoudige COil/ext, naar een methode voor het ontwikkelen en

veranderen van sociaal cognitieve configuraties (In Dutch) (Managing in multiple Contexts. AMethod for developing and changing social-cognitive Configurations), Delft: Eburon.

Vorliinder, K (1966), Philosophie der Neuzeit: Geschichte der Philosophie W, Hamburg: Rowohll.Vught, FA. van (1984), Experimentele beleidsplanning: bestuurskundige expeditites in de jungle van het

planningsdenken. 's-Gravenhage: Vuga.VUght, FA. van (1986), Op zoek naar een interactionistische sturingsconceptie, In: MA.P. Bovens, W.

Derksen en WJ. Witteveen (red.), Conferentie over "het schip van staat", Zwolle: W.EJ. TjeenkWillink.

Wallace, AF.C. (1961), Culture and Personality, New York: Random House.Wamsley, G.L. (1985), Policy subsystems as a unit of analysis in implementation studies: a struggle for

theoretical synthesis, in: Hanf and Toonen.Warren, R.L., A.F. Burgunder, 1.W. Newton, S.M. Rose (1975), The interaction of community

decisionorganizations: some conceptual considerations and empirical findings In: Negandhi.Weber, M. (1978), Economy and Society, Berkeley: University of California Press.Weick, KE. (1976), Educational organizations as loosely coupled systems, in Administrative Science

Quarterly, 21, pp. 1-19.Weick, KE. (1979), The Social Psychology of Organizing. Reading, Massachusets: Addison-Wesley.Weick, KE. (1985), The significance of corporate culture. In PJ.Frost (ed.), Organizational Culture, Sage

Publications.Weick, KE. (1987), Perspectives on Actions in Organizations, in 1.W. Lorsch (Ed.): Handbook of

organizational Behavior, Englewood Cliffs, NJ.: Prentice-Hall.Westin, A.F., A.G. Feliu (1988), Resolving Employment Disputes without Litigation, Washington: Bureau

of National Affairs.Whetten, DA. (1982), Issues in conducting research in:Rogers and Whetten.White, P.E., S. Levine, GJ. Vlasak (1975), Exchange as a conceptual framework for understanding

interorganizational relationships: application 10 nonprofitorganizations in: Negandhi.Wildavsky, A (1979), The Art and Craft of Policy Analysis, Macmillam

244

Wiener, N. (1948), Cybernetics, New York: J. Wiley.Williamson, O. (1985), The Economic Institutions of Capitalism. Finns, Markets, Relational Contracting,

New York: Free Press.Willke, H. (1983), Entzauberung des Stoates. Uber/egungen zu einer sozietalen Steuerungstheorie, Konigste

in: Athenaum Verlag.Willke, H. (1987), Entzauberung des Staates. Grundlinien einer systemtheoretischen Argumentation., in:

Ellwein c.s. 1987, pp.285-308.Winter, G. (1990), Freisetzendes und eingrenzendes Recht, in W. Hoffmann-Riem, KA. Mollnau and H.

Rottleutner (eds.), Rechtssoziologie in der Deutschen Demokratischen Republik und in derBundesrepublik Deutschland, Baden-Baden: Nomos.

Wolfe, J. (1985), The Teaching Effectiveness of Games in Collegiate Business Course: a 1973-1983Update. In: Simulation & Games: An Intemational Journal, 16 (3).

Woodside, K. (1986), Policy Instruments and the Study of Public Policy, in: Canadian Journal of PoliticalScience, or. 4, pp. 775-793.

Zeleny, M. (1977), Self-Organization of Living Systems, a Formal Model of Autopoiesis, InternationalJournal General Systems, 4, pp.13-28.

Zeleny, M. (1980) (ed.). Autopoiesis, Dissipative Structures and Spontaneous Social Orders. AAASSelected Symposium 55, Boulder: Westview Press.

Zeleny, M. (eds.) (1981), Autopoiesis, A Theory of Living Organization, New York: North Holland.Zeleny, M. (1981), Autogenesis, in M. Zeleny (ed.), Autopoiesis: A Theory of Living Organization, New

York: Elsevier.Zijderveld, AC. (1971), Over de relativiteit van kennis en werkelijkl1eid, Meppel: Boom.Zijderveld, AC. (1973), De theorie van l1et symbolisch interaktionisme, Meppel: Boom.Zijderveld AC. (1979), Over waardengebondenheid, in: Amsterdams Sociologisch Tijdschrift, no.2

245

Subject Index

acausal parallel processing 14, 129, 133aggregares 61,64assembling 26,69,71,72,74-77- cognitive assembling 75-77- social assembling 26,77autonomy 32,43, 101, 103, 113, 114, 116, 117,

124, 136, 193, 195, 196, 199, 200, 204,205,226

autopoiesis 3,4,13,14,16,31-34,36,40,41,43,49,113-117,123,127,128,132,133,134,135, 143-146, 149-152, 158, 159, 163,166,167,171,176-181,184,187,193-205,207,219,221,225-230

- social autopoiesis 113autopoiesis theory 13,40,178-181,229bargaining 104, 106, 107, 130, 138, 140, 144,

147-149,168,205boundary 49,50, 114, 118, 127calculus 129catalytically 134, 136centering 55,56-59,62,65change (social) 24,25,28chaos 123,127,196,202,203,214,219,231circularity 34,49,157,198,200c10sedness 201closure 34,39,113-124,133,134,144-152,154,

156, 157, 163-165, 170, 177, 178, 180,193,195,198-203,207,208,229,230

- autopoietic closure 145- operational closure 152- organizational closure 115, 123coalition 107,116,119,173,174,189communication 13,22,34-36,39-42,58,60,61,

81, 102, 107, 108, 124, 133, 136, 147,148-150,153,173,177-179,183-186,188,189, 191, 193, 198,208,209,214,218,219,229,230

complexity 8, 11, 12,36,37,42,63,64,77,99,118, 144-146, 170, 184, 185, 191, 204,208,210,219,220

configuration 3, 4, 13, 15, 16, 19, 20, 20-23,23-28, 52, 54, 60-62,64, 65, 67-70, 74,77,79-85,87,88,91-94,96,97,99, 102-­104, 110, 113, 116, 119, 120, 121-124,158, 164,201,207, 208,217,221,225­-230

configuration theory 13, 97, 221confrontation 25-28, 69, 75, 76, 92, 164, 169,

170,178,180,229,231consciousness 13, 34-37, 42, 84, 93, 124, 133,

189consistency model 116-118

construction 22, 41, 47, 52-54,60-62,67, 68, 72,82,83,198,208,214,219,227,229- social construction 219- reconstruction 186context 9, 12, 16, 20, 21,26-28, 37, 47, 52,57

-60,62,65,67-72,77,80-83,87,88,115,121, 131, 140, 145, 152, 154, 163-165,168, 187, 188, 200, 208-211,213, 215,217,218,220,221

coupling 13-15, 32, 33, 42, 69-72, 74-77, 106,116, 117, 128, 130, 132-135, 137, 139,140, 141, 152, 180, 181, 188, 190, 199,208,219,221,229

- structural coupling 13,14,32,128,132-135crisis 6-9, 11, 129, 132, 145, 180, 193,200,220,

229,231cycle 9, II, 15,42, 120, 121, 136,217decision-making 12, 54, 100-102, 107-110, 139,

140,205dependency 71,74,102,103,177differentiatedness 49-51,53differentiation 31,36-38,61,119,177,184,185,

187,189-191- functional differentiation 36, 37discourse 129, 132-134,136-139,141, 144, 145,

148,149,178,179,186,189,198,199distinction 14, 15,27,34,37,38,41,48-52,57,

60, 67, 88, 92, 114, 115, 118, 124, 133,135, 151, 152, 155, 160, 197,201,202,204,207,209,212,213,219

dynamic conservatism 12,203dynamics 4, 7-9, 12, 13, 15, 19,21, 24, 56, 60,

69, 76, 81, 82, 104, 135, 150, 186, 199,213,220

ecological recursiveness 14, 128, 136effectiveness 3,5-11, 14,39, 117, 119-121,127,

128, 143, 147, 151, 163, 176, 179, 186,187

entities 14, 1,23,51, 122, 175,204epistemology 79,80,84, 198,220,228equilibrium 83,123,196,201,219ethical space 15, 16, 26, 53, 54, 88, 89, 91, 97,

229facilitating 26,28,59,63,65,69,72feedback 116,172,175-180,218fixation 27,28,59,63,70-72,74-77,230cognitive fixation 75,76game 15, 16,52,56,57,59,62,71,72,74,76,

109, 132, 147, 210, 215, 217, 218,229,230

gaming 57,62,65,207-210,215-221government 8, 12, 19,21,25, 38, 39,99, 100,

247

101, 102, 106-109, 138, 143, 148, 149,151, 155-157, 161, 163-170, 172, 176,193,199,203-205,229

identity 37,42,113,115-117,120,137, lSI, 152,154-160, 163-165, 168, 174, 186, 197,200,205,207

impenetrability 12,157,160incentives 143,146,157,159,164,168,176inclusion 13, 16,23-25,51,62,69,74, 75,77,

158information 6, 35, 36, 38,40,42,60,61,72-74,

101, 102, 104, 108, 116, 117, 122, 123,129, 134, 146, 147, 157, 165-167, 172,173, 175, 176, 179, 180, 184, 188, 198,204,209,210,214,218,220,229

instrument 38,40,128,132,143,157,161-169instrumental 21,26,71,162-166,169, 170,174,

175,189,203,220integration 15,20,28,47,49, SO, 52-54,58,59,

61,67,87-91,93,96,97, 119, 121, 185,187,207,208,221,231

- social integration 52,53,88,91,97,207interaction 9, 13-15, 19, 20-28, 32, 33, 40, SO,

53,57,58,60,61,63,67-72,74-77,81­-85,88-91,93-97,102-104,104,106,108,110, 130, 131, 144, 146, 149, 158, 165,172, 178, 179, 183, 188, 189, 195, 200,203,208,213,221,227,229

- ongoing interaction 22,24-26,28,33,68, 103interference 42,43,133,149,165,166,226,227internalization 9, 14interorganizational 102,173,174,193,199interpretation 13, 22, 26, 35, 36, 40, 56, 69, 73,

79, 80, 82, 84, 88-92, 96, 116, 128, 129,130-132, 136, 145, 149, 154, 157, 160,165, 176, 190, 193, 196, 198, 200, 201,203

intervention 20,25-28,56,59,71,72,88,91,97,104,116,130,143-145,190,229,230

knowledge 12, 16,20,22,24,40,48-50,56,57,65,67-72,74-77,79,80,81-84,100,103,109, 135, 144, 169, 172, 173, 176, 177,180, 187, 188, 196, 207, 208, 210-215,218-221,225-227,231

law 4-9,11, 12,14-16,31,37,48, 114, 127-130,130-132,133-141,143-147,149,177,184,187,189,193,198,199,211

layers of history 160learning 5-16, 19,21,22,26, 51, 53, 54,56-60,

63,72,95, 119, 141, 145, 163, 171, 172,173-181,189,207-213,218-221,230

- deutero learning 175learning theories 172,173,180,181legal norms 127-130,132,133,135-137,1451ifewor1d 184,186-191

248

linkage 14,128,133,134,136-138,139linkage mechanisms 133- ultracyclical linkage 134maintenance 26,34,54,63,71, liS, 117,122,

159,196,198,200,201,203,204management 10,26,27,28,64,68,69,72,77,

84, 88, 91, 92, 97, 102, 104, 110, 115,117-120, 122, 131, 145, 159, 165, 175,176,179,216-218

- network management 74, 110,165managing 26,54,59,67-69,71,72,77,91,165meaning 4, 22,23, 34-36, 38, 40-42, 49, 51-53,

57,58,60,75-77,80,82,88-93,96, 151,152-160, 164, 170, 176, 177, 184, 187,198, 207, 209, 210, 215, 218-220, 221,227,229,231

media 137, ISS, 184, 186, 188, 190mediation 134,143,144,147-150mediator 144, 148, 150metaphorization 52,54,56,57-59,62-65metaphors 55,57,58,63,151,173,193,199negotiation 51,54,63,68,81,82,84, 105, 147

-149network 39, 40, 43, 73, 74, 99-104, 104, 105,

106, 106, 106-110, 113, liS, 116, 119,120, 122, 129, 138, 157, 165, 174, 186,188, 194, 195, 198, 199, 205, 210, 218,230

norms 4,6,7,9, 10, 14,25,26,28,52,53,71,127-130, 132-137, 140, 145, 146, 152,153, 156, 158, 159, 164, 172, 173, 209,215,230

organlZlng 13, 15, 16, 19, 23, 27, 33, 55, 57,59-63,65, 67-72,75, 81-84, 88, 90, 91,93, 94, 95, 105, 113, 195-198,201,210,218-220

'other' (the) 16,25-28,89-96perceptions 39,172,176,179,180,226play IS, 16,36, 52, 57, 59, 62, 71,74,76, 87,

108, 109, 123, 144, 147, ISS, 157, 167,170,200,207,229

pluralism 134,135,137-139,226,228policy accumulation 6,8,9, 11, 14- law of policy accumulation 6politics 14, IS, 36-38, 41, 109, 127, 128, 134,

136, 137, 139, 140, 146, 147, 149, 177,184,185,199,203,204

post-modem 135rationality 39,81, 127,154, 178, 183, 186, 189,

209,220,227,229reality 6, 8, 11, 13, 16, 20-28, 36, 37, 39, 47-49,

51-54, 56, 57, 59, 60, 62, 64, 65, 67,68-77, 79-85, 87-92, 96, 97, 103, 106,107, liS, 118, 122, 129, 132, 139, 144,145, 149, 174, 177, 198,202-204,207--

212,214-216,218-221,226,228,229,230- definitions of reality 21,22,24,25reflexivity 4,5,9, 11,12,14,51-54,57,58,186,

219reframing 15,27,58,59, 105,regulation 3,5,8-10,16,128,129,131,132,133,

134, 137-140, 143-147, 149, 157, 162,164,196,204,226

reification 21,22,59,79,80,82,84,162safeguarding 22,26,54,68,91selective activation 19, 107self-description 38,41,176self-maintaining 33self-organizing 218self-referential 178,188,202,212self-reflection 41,176self-steering 42, 207simulation 7,27,195,207-210,216,219,221stability 103,106, 133, 196, 197,200-203,205steering 3-5,7, 8, II, 13-17, 19,20,22,24-28,

31, 33, 34, 38-43, 47, 54, 68, 69, 71, 72,87-89,91,92,94, 96, 97, 99, 100, 101­105, 107-109, 115, 117, 123, 127, 128,143, 145-148, 150-152, 154-161, 163-168,170-172,174, 176-180,183-185,187-191,193, 199, 204, 207-210,217, 219, 221,225-231

- governmental steering 155, 156- political steering 41- societal steering 184, 225, 230steering ~ movenwnt 154steering theory 87, 97systems 3~, 9, 11-16,31,32,32-38,40-43,48,

62,68-72, 103, 104, 105, 113, 114, 115-­117, 119-124, 128, 129, 131-134, 136,137, 139-141, 144, 145, 149, 150, 151-­157, 159, 160, 163, 171-173, 175-181,183, 184, 1, 184-191, 193-195, 196-200,202, 204, 205, 207-215, 219-221, 226,227,229,230

- autopoietic systems 32, 177, 198- living systems 32,34, 177- policy systems 16- psychic systems 34, 35- social systems 12, 13,34-37,184,208,210,221'third' (a) 60, 82, 94'third' (the) 87-97thirdn~ 48,49,52-54,57,58,229,231triangularity 50-53utterance 35,36,42,198value 9, 10, 17,26,31,50,52,53,73,74, 88,

108, 119, 120, 128, 129, 136, 137, 147,151, 165, 168, 180,200, 217,218,220,230,231

249

Authors Index

Aldrich, H.E., 102, 103, 106,204Allison, G.T., 38,40,175Argyris, C., 172, 179Ashby, W.R., 113,204Bachelard, G., 47,53Beer, S., 31,114Bellevita, C., 165, 167Benson, J.K., 102,103,Berger, J., 84,187Blake, J., 52, 53Boer, Th. de, 92, 93Bolk, H., 15,20,21, 23, 23, 24, 25-27,58-60,

81-83,91,92,94,Braudel, F., 151,153Bredemeier, M.E., 217,221Broekstra, G., 113,116,118,119,122-124Brown, J.S.,220, 221Bruha, T., 139,141Bruijn, J.A. de, 161-163,165Child, J., 50,57, 120Crozier, M., 102, 106, 108Cruikshank, J., 146,147Davis, K., 52,53,214Diekmann, A., 129, 133Dijk,N. van, 20-22,27,58,59,80DODgen, H.J. van, 13,20,26,47,54,57,60,62,

63,65,72-74,74-77,87-90,92, 96, 97,116,158,207,208,212,219,228,229

Edelman, L.B., 137,138Edwards, A.R., 183,188,189Eichmann, R., 188,190Eigen, M., 136,210,219Elder, C.D., 55,175Evans, W.A., 217,221Foerster, H. von, 211-213Friedberg, E., 102, 106Friesen, P.H., 121-123Giddens, A., 47, 104, 183, 184, 187,228Godfroy, A.J.A., 103,108Goguen, J.A., 50,51,196Graaf, T. de, 93,96Greenblat, C.S., 217,221Habermas, J., 80,183,184-191,198Henry, S., 121,138Herbst, Ph.G., 47-50,52Heuvelhof, E.F. ten, 161,165Heylighen, F., 208,210Hogwood, B.W., 161,204Hood, C.C., 38, 162Hufen, J.A.M., 19,39,163,171,173,176Janisch, E., 33,196,197,201Jessop, B., 193, 199,205

Kaufmann, A., 129,137,203,204Keulen, S. van, 56,57Kickert, W.J.M., 178,193,204,205,228Kingdon, J.W., 175,180Klabbers, J., 207,209,212-217,221Klijn, E.H., 38,39,99, 106Koppenjan, J.F.M., 106,171,179Kunneman, H., 187,189Laat, W.A.M. de, 20-22,26,61, 81, 84, 85, 89,

91Laszlo, E., 197,201Levinas, E., 49,54,88-93,95-97Levine, D.N., 103,225Levy, A., 27,58,59,162,168Luhmann, N., 13, 14,31,34-36,36-38,40,41,

43, 114, 131, 133, 136, 144, 145, 149,153, 183, 184-188, 190, 191, 193, 197-­199,201,202,214

Maas, A.J.J.A., 21-24,26,27,52,55,57,80,81,83,88,90-92,94,96,104

Macaulay, S., 135,138Majone, G., 148,230March, J.G.M., 107,175Maturana, H.R., 32-34, 42, 43, 113-117, 133,

177,193-197,199,208Mayntz 35,134,137,152Meer, F.B.L. van der, 20,24,25,60Meltsner, A.J., 165,167Merry, D., 27,58,59,127,128Miles, R.E., 120, 121Milgram, S., 94-97Miller, D., 121-123Mingers, J., 114,115,117Mintzberg, H., 106, 121Moore, 0., 128,218Morgan, G., 31,34,104,117,163,193,199,200,

203Nozick, R., 53,89Olsen, J.P., 107,175Opp, K.D., 129,133Peirce, Ch., 49,50Peters, B.G., 161,204Pressman, J.L., 99,106,127,171Prigogine,1., 123,196,202,209,219Ranger, T., 127,128Rawls, J., 53,89Ringeling, A.B.,19, 39,176Rosewitz, B., 152, 153ROllleuthner, H., 127-129,133,145Sabatier, P.A., 172-175Schaap, L., 13,31,152,177,229Scharpf, F.W., 19,33,35,39,102,103,105,107,

251

128,134,137,145,205Sculli, D., 218,221Selmick, P., 137,138Simmel, G., 225,227,231Simon, H.A., 208,209Snellen, I.Th.M., 19,99,151,209Snow, C.C., 120,121Spencer Brown, G., 47,48,50Stengers, I., 196,202,209,219Strasser, S., 92,93Susskind, L., 146-148Teisman, G.R., 38,39,99Termeer, C.l.A.M., 15, 19, 81, 82, 88-89, 103,

104Teubner, G.C.M., 13-15,31,33,34,41,42,43,

127, 133, 134, 137, 141, 144-146, 149,152, 158, 159, 190, 193, 198,229

Twist, M.l.W. van, 13-15,19,31,81,82,88-90,103,104,152,177,229

Uribe, R.B., 113,193-195Varela, F.l., 32-34, 42, 43, 50, 51, 113-117,

121-124,133,177,193-197,199,201,208

252

Veld, R.I. in 't, 3,4, 7, 19,22, 145, 172, 177,230

Voogt, A.A., 15,20-22,24-26,27,28,52,57,67,72-77,80,83,90,91,94,103

Vught, F.A. van, 19,172Weber, M., 129,131,132,226,227,231Weick, K.E., 15,19,22,23,51,52,84, 106, 120,

177,208Whetten, D.A., 103, 1M, 106Wildavsky,A., 7,99,106,127,171Willke, H., 34,37,41,42,43,152Winkler, R., 210,219Wolfe, l., 217,221Zeleny, M., 114,136,195,196

About the Authors

Prof. dr. R.I. in 't Veld (1942) studied Law at Leiden University. He is former director-general forHigher Education and Scientific Research of the Dutch Ministery of Education. Presently he is Head ofthe Department of Public Administration of Erasmus University Rotterdam and Dean of the NetherlandsSchool of Government in The Hague. Recent publications: The Decried State 1989 (in Dutch),Environmental Protection: Public or Private Choice, 1991 (with D.I Kraan (eds.), Evaluation in HigherEducation, 1991 (with U. DahllOf, M. Shattock, A. Staropoli).

Drs. L. Schaap (1962) studied Public Sciences at Groningen University and is now research trainee at theDepartment of Public Administration of Erasmus University Rotterdam. His main research interestconcerns intergovernmental relations between local and central government, especially fmancial relations,steering and autopoiesis theory. Publications: The selfish society; steering and selfreferential systems, in:Bestuur, November 1990 (with M.I.W. van Twist and R.I. in't Veld; in Dutch), lntegration measures: anautopoietic approach and some implications, paper for the IIAS-conference, Madrid November 1990(with M.I.W. van Twist), Steering and communication, paper for the Politicologenetmaal 1991, Rotter­dam 1991 (with F.B. van der Meer and M.I.W. van Twist; in Dutch).

Ir. C.I.A.M. Termeer (1960) studied Agricultural Sciences at the Agricultural University of Wageningen.Since 1987 she is affiliated to the Department of Public Administration of Erasmus UniversityRotterdam. She works on a dissertation on the manure problem and changing configurations in theagricultural world. Publications: Dynamics in the Agricultural Policy Network, in: Hufen, I.A.M. and A.B.Ringeling (eds.), Policy Networks, 1990 (in Dutch), Dutch Manure Policy: the Lack of economicinstruments, in: Kraan, D.J. and R.I. in 't Veld (eds.) (with. F.Dietz), Environmental protection, public orprivate choice, 1991, The configuration approach, a procestheory for steering problems, in: Beleid enMaatschappij, July/August, 1991 (with M.J.W. van Twist; in Dutch).

Drs. M.J.W. van Twist (1966) studied Public Administration and Business Administration. Since 1989 heis a research trainee at the Department of Public Administration of Erasmus University Rotterdam. Hiscurrent research is concentrated on applying Autopoiesis and Configuration Theory to steering problemsregarding international environmental problems. Publications: The selfish society; steering and selfrefere­ntial systems, in: Bestuur, November 1990 (with L. Schaap and R.J. in 't Veld; in Dutch), lntegrationmeasures: an autopoietic approach and some applications, paper for the IIAS-conference, Madrid,November 1990 (with L. Schaap), The configuration approach, a procestheory for steering problems, in:Beleid en Maatschappij, July/August, 1991 (with C.J.A.M.Termeer; in Dutch).

Prof. dr. H.J. van Dongen (1936) studied clinical and industrial psychology at Leiden University. InTilburg he received his PhD in Social Psychology. He worked for eleven years in industry, first as ageneral industrial psychologist and after that in data-theory and organizational development. He wasmember of several governmental or semi-governmental comrnillees and has extensive consultancyexperience. He developed a research program on socio-cognitive configuration and social integration. Hepublished several books and articles on technology and impact, strategies for change and socialintegration.

Dr. A.J.J.A. Maas (1952) is associate professor at the Faculty of Business Administration of theRotterdam School of Management of Erasmus University Rotterdam. His scientific interest concernsorganizational theory, organizational development (OD), organizational change and methodology.Publications: Disturbed Managing, 1981 (with Brevoord and Van Dongen; in Dutch); Ministries understress, 1981 (with J. Kooiman; in Dutch), Undefined Social Spaces, 1988 (in Dutch).

Dr. A.A. Voogt (1938) is member of the board of directors of Incotrans BV. His scientific interestconcerns management and organization. He published Managing in multiple Contexts, 1990 (in Dutch).

Prof. dr. B. Kee (1942) is affiliated to Erasmus University ROllerdam and to Free University Amsterdam.

253

His scientific interest concerns the philosophy of management and organization. Recent publications:Strategic management; to create or to survive, in: M&O Tijdschrift voor organisatiekunde en sociaalbeleid, nr.3, 1988 (in Dutch), Manager and Organization: Egg or Chick, in: Human Systems Manage­ment, nr.8, 1989, Chaos as a path to order? Management and Chaos Theory, in: Filosofie in Bedrijf, nr.2,1990 (in Dutch).

Drs. N.L. Rade is lecturer Organization Sciences at the Department of Business Administration at theundergraduate school of polytechnics in Rijswijk.

Drs. E.H. Klijn (1958) is affiliated to the Department of Public Administration of Erasmus UniversityRotterdam. From 1984 to 1989 he worked as a researcher at the RIW- Research Institute for Housingand Urban Renewal of the Faculty of Architecture, Housing, Urban Design and Planning of TechnicalUniversity Delft. His scientific interest concerns policy making in complex networks, especially in thefield of housing policy. Publications: Post-War Housing investigated, 1987 (with H.M. Koolma, in Dutch);From a nobel mandate towards a silent redistribution: distribution, steering and politicalizing in housingpolicy, in: A.B. Ringeling, I.Th.M. Snellen, Government: on retreat or towards a new profile?, 1989 (withH.M. Koolma, in Dutch), The fish and the bait: the implementation of two financial instruments inhousing in: J.A.M. Hufen, A.B. Ringeling, Policy Networks, 1990 (in Dutch).

Drs.ing. G.R. Teisman (1956) is affiliated to the Department of Public Administration of ErasmusUniversity Rotterdam. His research field is physical planning, infrastructure and public works. Hisscientific interest concerns inter-organizational decision making in policy-arenas. Publications: Projectdevelopment as 'konzertierte Aktion' in: J.A.M. Hufen, A.B. Ringeling (eds.), Policy networks, 1990, (inDutch). Investments in infrastructural projects: a bargaining and decision-making process withparticipation of governmental agencies, public enterprises and private companies in: L.J. Roborgh, R.R.Stough & Th.A.J. Toonen (eds.), Public Infrastructure redefined, 1988 and To decide on city infrastruc­ture; an inquiry into interorganizational decision-making processes, 1991 (with R.J. in 't Veld; in Dutch).

Prof. dr. G. Broekstra (1941) is affiliated to the Rotterdam School of Management of ErasmusUniversity Rotterdam. His scientific interest concerns systems methodology and holomanagement. Recentpublications: Holomanagement: The 21st Century Visionary Leadership Style for Creating IntelligentOrganizations, in: R. Trappl (ed.), Cybernetics and Systems London, 1990, Parts and Wholes inManagement and Organization, in: G Lasker (ed.), Proc. Int. Conference on Systems Research,Informatics, and Cybernetics, Baden-Baden, 1990, Organizational closure and the quatum view oforganizations, in: McJackson (ed.) Systems thinking in Europe, New York, 1991.

Prof. dr. G.C.M. Teubner (1944) is Professor of Law at Bremen University and the European UniversityInstitute in Florence. He was visiting professor in Berkeley (198011981), Ann Arbor (1987), and Stanford(1988). His scientific interest concerns theoretical sociology of law, private law theory and comparativesocietal law. His recent publications include: Law as an Autopoietic System, Frankfurt am Main, 1989 (inGerman, English version forthcoming), State, Law, Economy as Autopoietic Systems: Regulation andAutonomy in a New Perspective, Berlin, 1990, Regulatory Law: Chronicle of a Death Foretold, in:Lenoble (ed.) Law and the Crisis of the Welfare State, Bruxelles, 1991.

Mr. I.J. Koppen (1957) is trained as an environmental lawyer, both in the United States (BA 1981,Honors College, University of Oregon) and in the Netherlands (1985, University of Amsterdam). Since1986 she is affiliated to the European University Institute in Florence, presently as Executive Coordinatorof the European Policy Unit. Fullbright Visiting Scholar at the Harvard-MIT program on Negotiation(1988-1989). Expertise especially in two areas: Negotiations and EC environmental law and policy. Aboutto finish her PhD thesis on 'Negotiating Waste Reduction.' Publications: The Single European Act asconstitution of the environmental policy of the European Community, SEW, Tijdschrift voor Europees enEconomisch Recht, 1988 (in Dutch), Environmental Impact Assessment and Negotiation, in:Environmental Impact Assessment Manual, ERVET, Bologna, 1989 (in Italian), Environmental Rights,in: A. Cassesse et al.(eds.) Human Rights and the European Community; the Substantive Law, 1991(with K. Ladeur).

254

Prof.Mr. Dr. I.Th.M. Snellen (1933) holds a degree in Law and in Political and Social Sciences of theUniversity of Amsterdam and is now professor in Public Administration at Erasmus UniversityRotterdam. He is specialized in informatization in Public Administration. Publications: Approaches toStrategy Formulation, 1975, Limits of Government: Dutch Experiences, 1985 and Expert Systems inPublic Administration: Evolving Practices and Norms, 1989 (with W.B.H.I. van de Donk and I.P.Baquiast (eds.».

Mr. dr. I.A. de Bruijn (1962) is affiliated to the Department of Public Administration of ErasmusUniversity Rotterdam. He studied Political Science and Law at Leiden University. In 1990 he received hisPhD in Public Administration. His scientific interest concerns management in public organizations andpolicy instruments. Publications: The Ministry of Economic Affairs and Economic Subsidies 1990 (inDutch), Steering Instruments for government; About complex networks and second generation steeringinstruments, 1991 (with E.F. ten Heuvelhof; in Dutch).

Mr. dr. E.F. ten Heuvelhof (1954) is associate professor at the Department of Public Administration ofErasmus University Rotterdam. His scientific interest concerns environmental policy, physical planningand tools of government. Publications: To an empirical policy theory, 1982 (in Dutch), SteeringInstruments for government: About complex networks and second generation steering instruments, 1991(with I.A. de Bruijn; in Dutch) and Integrated environmental and physical planning in urban renewal,1988 (in Dutch).

Drs. I.F.M. Koppenjan (1955) is affiliated to the Department of Public Administration of ErasmusUniversity Rotterdam. He is specialized in policy management and policy oriented learning. He preparesa PhD thesis on public policy making in the field of Home Administration. His main publications are:Public policy making in the Netherlands, 1987 (with A.B. Ringeling and R.H.A. te Velde; in Dutch).Together with prof. dr. A.B. Ringeling he wrote the report Making decisions about the new DutchPassport, 1988 (in Dutch) on behalf of the parliamentary inquiry commission which investigated thefailure to introduce the new passport-system.

Dr. I.A.M. Hufen (1959) is affiliated to the Department of Public Administration at Leiden University.His scientific interest concerns policy instruments, implementation research and interorganizationalanalysis. Empirical research in the last years was directed at technology policy, economic policy andenvironmental policy. Publications: Instruments in technology policy, 1990 (in Dutch), Policy Networks,1990 (with A.B. Ringeling (eds.); in Dutch).

Dr. A.R. Edwards (1950) is affiliated to the Department of Public Administration of Erasmus UniversityRotterdam. His scientific interest concerns communication in the public sector. Publications: The study ofPublic Administration and its integrating functions, in: Bestuurswetenschappen, nr.6, 1986 (in Dutch),Regulating room for decision-making (with H. ten Holt; in Dutch), 1989, Planning in dispute; communi­cative strategies of farmers and environmentalists in the Wommels land development project, 1990 (inDutch).

Prof. dr. W.I.M. Kickert (1950) is affiliated to the Department of Public Administration of ErasmusUniversity Rotterdam. His scientific interest concerns management and organization of the publicadministration and steering in complex networks. Publications: Organisation of decision-making, 1980,Planning within bounds, 1985 (in Dutch), Government planning, 1986 (in Dutch).

Prof. dr. I.H.G. K1abbers (1938) is affiliated to the Department of Public Administration of ErasmusUniversity Rotterdam. His scientific interest concerns (Self-) steering in social systems and multi-actorsimulation. Publications: Instruments for Planning and Policy Formation: some methodological considera­tions, in; Simulation & Games, vol. 16(2), 1985, Design Methods for Improvement of Self-Steering ofSocial Systems, in: Heylighen, F. et aJ.(eds.) Self-Steering and Cognition in Complex Systems: Towards aNew Cybernetics, 1990 (with W. Scheper), Problem Setting Through Gaming. In: Oosterhuis, I. etal.(eds.): Problem Solving with Cases. Educational Technology Publication, 1991.

255

Drs. H. Mastik (1943) studied sociology and education science at Erasmus University Rotterdam. After acareer outside the academic world, she joined the Department of Public Administration of ErasmusUniversity Rotterdam in 1989, teaching management and organization, and skillbuilding throughexperiential learning. She developed simulation games for educational, research and consultancypurposes. Since 1990 she is Managing Director of ACES, Academic Center for Experiments inSimulation, an initiative of the Department of Public Administration. Major research interest: theoreticaland methodological aspects of simulation and 'gaming' for public management and policy analysis.

256