The ‘Post-Institutional’ Perspective:

Society as an emerging system with dynamically changing boundaries


Loet Leydesdorff

Science and Technology Dynamics

Department of Communication Studies

Oude Hoogstraat 24

1012 CE  Amsterdam






One does not sufficiently appreciate the relation between a social system and its environment by using the metaphor of autopoiesis of an observable system.  A social system cannot be considered as a thing separate from a context, whether temporal, spatial, or in terms of relations.  Reflexive actors, however, are able to specify hypotheses concerning social systems on the basis of observations, while they operate as participants from within the system.  Structures, functions, and boundaries then remain expectations.  The reflexive specification of expectations can make a difference for the system of expectations.  A social mechanism for updating expectations, and thus for changing social systems reflexively, can be proposed.  From this perspective, implications for changing the boundaries of the system by sociological theorizing and research are suggested.




Luhmann (1984; 1986) noted that society should not be considered as a living system. Thus, the biological metaphor in which functionality is geared to reproduction and survival in a given (‘natural’) environment is inadequate (cf. Maturana 1978).[1]  The social system does not have to be physically or biologically integrated; it remains an uncertain expectation of order among reflexive people (‘consciousness systems’).


What may functional differentiation mean if uncertainty prevails at the level of the (super‑)system?  In his study Die Wissenschaft der Gesellschaft (1990), Luhmann addressed this consequence of the sociological thesis of functional differentiation.  For example, he noted that scientific communication is internally differentiated in terms of experience and action.  Thus, one is able to raise the question of whether functionality in the differentiation is recursive.  Luhmann, then, made a decisive move beyond structural functionalism and neo-institutionalism by stating that ‘the differentiation of society also changes the social system in which it occurs’ (ibid., at p. 340).  He called for ‘an accordingly complex systems-theoretical arrangement’. 


In this paper, I introduce the problem of further differentiation of functionally coded communications using the formal perspective of second-order systems theory.  In subsequent sections, I shall indicate the fruitfulness of my hypothesis for sociological theorizing and research concerning the topic of this special issue: the social reflection of system’s boundaries by new social movements.



Differentiation and Integration


Traditionally, structural functionalism has used the model of a dialectic between functional differentiation and institutional integration at the system level.  Action is then taken as the integrating category, but action is not considered as itself constructed (Parsons 1937; Münch 1982/1988).  Parsons (e.g., 1968) and Luhmann (e.g., 1977 and 1984) studied this phenomenon under the heading of ‘interpenetration’.[2]  Giddens (1984) has called this dialectic operation ‘the duality of structure’, but the focus in his sociology has remained firmly on action as the system of reference.[3]


Institutions have been studied as aggregates of action, in neo-institutionalism and social network analysis (e.g., Burt 1982).  However, human actions are also interactive and reflexive.  These communicative dynamics lead to non-linear terms in the further development of the system that cannot be captured by an aggregation model (Leydesdorff 1993a).  When the aggregation rules themselves are changing, the institutional distribution of action is only a first-order observation: this data does not yet reveal the dynamics.


From a dynamic perspective, institutions can be considered as the fingerprints of the communications which have served us hitherto.  They carry the reproduction of the social system, while the institutions are themselves carried by reflexive actors who in turn are expected to change them, for example, by exploring new means of communication.  Other aggregations of action among functionally differentiated communication systems are possible when the interactions among the actors and accross institutional boundaries become recurrent. 


Each aggregation rule can be considered as a specific selection of an (hypothesized) aggregating super-system.  Selection operates on functionality; functionality for the next-order system can be provided by different structural arrangements.  Although generated by aggregation, the higher-level system cannot always be reduced to an aggregate of lower-level ones.  In general, aggregation is dominant in a stratified system, while interaction prevails in a functionally differentiated one (Leydesdorff 1994).  However, the functionally differentiated system remains evolutionarily constrained by its institutional history.


In Luhmann’s sociology, the function systems select upon each other using their internal codifications.  This induces aggregation of the variation for functional instead of institutional reasons.  If the function of these aggregations is no longer necessarily the institutional one of adaptive reproduction-that is, undifferentiated action-we have to specify the mechanism of the production of these reflexive layers and their social functions.



Inter-system Dependencies


In a fully differentiated system, the question of the consequences of further differentiation for the reproduction leads one to focus on inter-system dependencies.  Can a differentiated system perhaps be integrated at the interfaces without assuming de-differentiation by action (or in terms of its institutional aggregates)?  In my opinion, interactions among inter-system dependencies provide us with a mechanism for integration which can analytically be distinguished from institutional integration.


Let me use pictures to illustrate this alternative option for integrating differentiated sub-systems.  Figure One shows a Venn-diagram representating three differentiated sub-systems of a single system.  The three sub-systems are here considered as specifications (ij) of a common system (i). Each of the sub-systems has a different value in a second (e.g., functional) dimension (j).  In other words: the uncertainty i that is communicated is coded with a meaning j that is specific for the respective sub-system.


1Figure 1

Three functionally differentiated subsystems with a common intersection

2Figure 2

Three differentiated sub-systems without a communality

The intersection in Figure One can be associated with Max Weber’s conceptualization of ‘culture’ as a meeting place between otherwise incompatible value-orientations (Leydesdorff 1993b).  It is a locus of both interpenetration and integration.  Note that the picture should be thought dynamically as an interaction among spirals of communication.


Figure Two provides a different conceptualization of inter-system dependencies.  The systems have grown so far apart that the communality in the intersection has been dissolved.  This system can no longer be considered as emerging from a shared origin; it has grown constitutionally complex.  In this case, the integrating mechanisms cannot be identified in terms of an undifferentiated (‘common’) communication (i).  The intersections are themselves distributed as interfaces among various systems, and they are expected to contain more complexity than the functionally differentiated systems, since they communicate in terms of two codes (j and k).  It is a form of interprenetation, but not yet integration.


Since the meaning of the information changes when the latter is passed from one codification to another, communications at these interfaces can be considered as translations.[4]  This complex dynamics, however, can still be integrated dynamically, as indicated in Figure Three.  While integration was based on a common carrier in the case of Figure One, it has to be constructed dynamically, that is, as a series of translations at the interfaces, in Figure Three.



3Figure 3

Integration at the hyper-cyclic level

4Figure 4

Hyper-cyclic closure of commun­icat­ion in the case of four degrees of freedom



Such a series of translations can be considered as a hyper-cycle constituted as a new communication system on top of the underlying ones. Note that this integrating communication is internally differentiated.  It is not emerging from a differentiation in an underlying communality, but its emergence is based on interactions among function systems.  If one mentally takes the hyper-cycle out of the basic plane, the emerging structure can be conceptualized as a tetrahedron (Figure Four).  The emerging order is different from the underlying ones: the hyper-cycle adds to the complexity of the system.


This form of integration does not secure reproduction, since it remains distributed and thus fragmented.  Because of its failure to integrate fully, the emerging system of communications is expected to remain fractal in its dimensionality.  Fractal, fragile, fragmented, and failure all share a common root (of the Latin verb frangere): no communality is left, not even in the integration.  We may wish to call it a sub-integration.  This system is never integrated and immediately observable; yet, it can be considered in terms of a distribution of translations.



Fragmentation and Translation


A hyper-cyclic communication system is expected to fail to be integrated in an unambiguously observable mode.  Note that there is no evolutionary need for identifiable integration because the system is reproduced in a distributed mode.  Given a distribution, uncertainty prevails.[5]  Different observers are then expected to have different perspectives. 


Hyper-cyclic integration by means of a series of translations transforms the underlying (institutionally based) regime of functional differentiation into a regime of translations.  In a regime of translations, interactions tend to disorganize institutional boundaries (Turpin and Garrett-Jones 1997).  The sysem is ‘on the move’: the participant’s role is increasingly different from an observer’s, since the latter has to stabilize a perspective reflexively.


How does translation among function systems operate?  First, each of the function systems translates in terms of its code.  For example-following Luhmann-the science system uses resources for translation in terms of the truth-value of the communications.[6]  Differentiated sub-systems communicate in terms of different codes, like true/false, power, or economic value.  Thus, what a communication means is expected to differ in various sub-systems.  Translation is an asymmetrical operation.  At the interfaces a communication in one medium is translated into a communication in another.  For example, ij can be translated into ik.  In other words: the meaning of the information changes. 


Such a transformation can also be considered as a reflection: an input (ij) is transformed into an output (ik).  Biological systems already contain many such interfaces.  For example, at a synapse the signal from the nerve is translated into a signal that can be made relevant for the muscle.  The i of the signal may vary for substantive reasons.  The code j of the neural system is changed into a k that provides it selectively with meaning for the muscle.  Note that the variation in the signal (i) remains structurally coupled in a second dimension by being selected in different systems, but the transformation requires an operational coupling (between j and k) at the synapse.  (I return to the distinction between structural and operational coupling in a later section.)


If a muscle is denervated, it may begin to exhibit uncontrolled contractions.  Selections can also be considered as negative feedbacks.  The feedbacks of various function systems are expected to be different, and therefore each recursion of the selection introduces another sub-dynamics.  Three sub-dynamics are sufficient, in principle, for constituting a complex system (Figure Two): the two others provide both a variation and a selecting instance for each of them.  Two sub-dynamics, however, can be balanced in an observable stabilization (like ‘action’ and ‘re-action’).  Using a third sub-dynamics, some stabilizations can be entrained in a globalization (Figure Three; cf. Kampmann et al. 1994).


In summary, a communication system is sufficiently complex for a translation at the interface between two sub-dynamics if the uncertainty contains three degrees of freedom (i, j, and k).  In the first dimension, the uncertainty is generated as variation, in the second it is specifically (e.g., functionally) selected as meaningful, and in a third it can be translated and thus provided with another or new meaning.  By this last operation, the information is stabilized, since the original signal is re-written by the receiver in terms of its own code.  This is done locally, that is, at the locus of reception.





The distribution of the local translations provides us with a fourth degree of freedom in the uncertainty.  A next-order or hyper-reflexive system is expected to operate in four dimensions of uncertainty.  The additional degree of freedom enables it to change the aperture of its reflection.  For example, a hyper-reflexive observer-who knows herself to be one among a distribution of observers-is able to change the distinction between the information of a message and the message with hindsight (Luhmann 1997a). 


Note that this potential for changing the aperture of the reflection at the social level is precisely the arrangement that Luhmann indicated in the passage quoted above: the spread of reflexivity in modern society has recursively changed the differentiation of social communication (cf. Giddens 1990). In addition to functionally differentiated sub-systems, the system transforms inter-system dependencies into systems of translations that are able to reintegrate the system, by using a distribution of locally specific perspectives.


The institution can then be considered as a stabilization of local integration.  Other aggregations of action become possible with increasing reflexivity about different functions at the level of the social network among the actors: the uncertain functionality of the communication puts evolutionary pressure on institutional structures. 


Figure Five illustrates how sometimes even a new sub-system can emerge in the interface between two systems that were originally differentiated. If the translation in the intersection becomes recurrent, the interaction may increasingly develop systemic properties.  Using a biological metaphor, one may think of sexual reproduction.  Society, however, has the option to develop systems of translation that exhibit recurrent systemness without directly observable stabilization. 



5Figure 5

Emergence of systemness from interaction

Algorithmic approaches enable us to visualize series of geometrical representations (that is, stabilizations) without becoming confused about what is an observable instantiation (at one moment in time), an observable trajectory (over time), or a globalized regime (in ‘time-space’; Giddens 1984).  For example, ongoing translations between national systems have developed measurable systemness at the European level in some dimensions, but not in others (Leydesdorff 1992).[7] 


Translation can only be achieved between sufficiently different systems.  Sub-integration has in common with integration that it presupposes differentiation.  At the level of the complex social system, differentiation and translation are structurally coupled as a second (j) and third dimension (k) of the uncertainty.  Translation can have a function in informing the (sub‑)systems of communication, but not necessarily in their institutional reproduction.  A ‘niche’ of communication does not yet have to be institutionalized (Tong 1996): it is based on asymmetrical ‘code-sharing’ in a translation system.  Adaptation of the stabilizing institutional arrangements can thus be considered as another sub-dynamic of the complex system.  The institutional layer provides the retention mechanism of processes of social learning, while institutions remain the social dimension of function systems (cf. Luhmann 1984).


In summary: distributed action (by different observers) generates uncertainty in a first dimension.  This variation can be selected (‘provided with meaning’) in a second dimension, and then this selection can recursively be selected for translation using a third degree of freedom in the communication.  If a translation is locally understood, the initial action is provisionally stabilized since the signal can be re-written.  In a fourth dimension (of the uncertainty) some of these stabilizations are recursively selected for globalization.  In this context, globalization means sub-integration of the system at a next level through a series of translations.  The next-level (super‑)system is perceived as global from the perspective of each participant/observer.



Emergence as a systemic property


The order of this global hyper-cycle is continuously under construction, and therefore remains necessarily emergent.  The observer participates in the reconstruction by being reflexive.  Note the difference between hyper-cyclic sub-integration and integration at the institutional level: the latter is historically embedded.  Sub-integration, however, remains sub-symbolic since it fails to be.  One is able to provide it with meaning from a locally reflexive context.  Each local ‘instantiation’ then remains only one example from among a variety of possible reflections.


All (functionally differentiated) sub-systems can recursively use their system of reference as a code for sub-integration.  In a regime of translations, integration no longer takes place at a center, as in a stratified society, but in a distributed mode (cf. Gibbons et al. 1994).  The social system is implied, but only its instantiations and trajectories can be observed.  Thus, one should never reify the social system, e.g., by attributing it a specific identity.


In other words, social systems do not exist in the strong sense of the Latin verb esse.  One is in need of a different ontology for understanding social systems as distributions of expectations: peaks in the distributions can be considered as identities from a reflexive perspective, and hypotheses concerning social systems can be formulated on this basis.[8]  The thesis of functional differentiation is such a hypothesis.  Hypotheses direct the observations, and thus function as heuristics in scientific research.  However, the social system is then no longer empiristically available for observation in terms of action or at the institutional level, as in structural functionalism or neo-institutionalism.


The consequences of this epistemological shift pervade the logic of reasoning. In second-order systems theory, statements using predicates can no longer be made without generating paradoxes (Luhmann 1990).[9]  Consequently, the purpose of a single sociology tends to become self-defeating.  One can only make statements of the kind that ‘society can be considered as ...’, because such a formulation accounts reflexively for the contextual nature of the very statement.  (Yet, the context may be codified into one or another sociology; cf. Leydesdorff 1997a.)  If a system can no longer be identified, then its boundaries can no longer be considered as ‘given’. They, too, remain expectations that can be deconstructed reflexively.  However, before turning to this consequence, let me provide the reader with a metaphor for a system of translations.


Imagine a system of translations among natural languages such as may occur at a meeting of a European organization in Brussels. In each of the boxes for interpreters, two or more European languages are simultaneously translated.  The system of translations at the level of society, however, is decreasingly based on national differentiation.  National differentiation has been correlated to differences among national languages, while functional differentiation is associated with differences among codes in the communication.  The ‘languages’ in functional translations are no longer ‘natural’ and codified dictionaries will therefore not be available: one expects that the words change in meaning by being translated into a different context, and this change of meaning potentially feeds back on the original categories.  Adjustments are made with hindsight and according to the functional logic of an emerging system.


A Tower of Babel-like confusion of tongues becomes the normal state of this system, and reflexively all sub-systems have continuously to update their dictionaries.  The system of translations has gained an independent function among the function systems as it sub-integrates the system.  Fragmentation and poly-contexturality are constitutive for this post-modern regime.



The post-institutional perspective


The regime of translations is already an everyday experience in domains of society other than science.  The European Union, for example, cannot be considered as a union like the United States; it is a system of translations which is gaining an autonomous momentum on the basis of interactions among the national systems on which it rests.  The progression may have been stronger in some areas of policy-making than in others.


The example illustrates the capacity of the social system to develop new systemic layers internally on the basis of recurrent interactions (as also illustrated in Figure Five above).  The new layers can be expected to have different functions as they become increasingly codified.  The codification, however, is no longer unambiguous: it is an dynamic expectation in a variety of contexts.  Therefore, codification can itself become distributed and poly-contextural.  Each dimension is binary in the extremes, but it extends as a degree of freedom that may contain shades of grey.


Let us now specify the model more fully.  First, remember that selection is a negative feedback.  If ten percent is selected, ninety percent is discarded.  Thus, if the distribution is normal to begin with, we may obtain a skewed distribution after selection.  However, the result of two negative operations upon each other contains a positive term: some selections are selected for observable stabilization.  In general, this positive result is at an order of magnitude smaller than the variation upon which it operates because there are two selections involved: translation is a form of communication, but it is a highly specific communication.


The sign is expected to alternate between the even and the odd dimensions of the information because of the introduction of a minus-sign by each recursion of the selection.  At first, selection is a negative feedback on the variation, as in the case of market clearing.  The positive feedforward in a next selection allows for stability and thus for observable trajectories.  On the basis of a third layer of selections, the complex system incurs as a pending regime on its underlying sub-systems.  The sub-dynamics inhibit one another as negative feedbacks: noise is continuously filtered out. 


The stabilizations can be considered as the institutions on which the regime of translation builds new functions.  The institutions have been selected socially because of their service in the reproduction of the communication.  However, one can expect them to be changed, since they are entrained in the further developments like the trajectories of a regime.


I follow Luhmann in defining the social system as the system of inter-human communication, while, in my opinion, other forms of (irreflexive) communication are also possible (Latour 1987).  Human communication is evolutionarily based on the possibility of distinguishing between meaningful information and uncertainty.  The social system provides us with codes that can be processed reflexively, so that an uncertainty can be reconstructed and then communicated while suppressing the noise.  Precisely because we are able to communicate our reflections, we do no longer need to follow our instincts.  Hitherto, human agency has been the sole source of variation at this level of reflexive discourse (cf. Leydesdorff 1994).


Code is selective at the network level: some actions are provided with meaning by the social system, while others are not.  This may be different from various perspectives as the code becomes differentiated during cultural evolution.  Different meanings can be translated into each other by using reflexive agency at distributed nodes of the network.  In summary: agency generates the variation; constructed (and therefore uncertain) codes at the network level select specific actions as communications; and the distribution adapts itself to interactions among codes in the institutional layer.



Structural, operational, and loose coupling


Let me use the model of a standard social science design for formalizing the increase of complexity in terms of distinguishable forms of coupling between systems.  In a social science methodology, like in SPSS, the rows of the matrix are considered as actions attributed to the actors.  The columns represent the variables or communications among the actors. 


The two dimensions of this matrix (e.g., i and j) determine each other in the co-variation.  However, the dynamics of each system, that is, the action system and the communication network, are different because they refer to the previous states of these systems in terms of the total of their respective variations.  The total variation is by definition equal to the co-variation and the remaining variation.  (The two systems condition each other in the remaining variation.)  Giddens (1984), for example, denotes this relation between conditioning and determining in the reproduction with the expression of ‘enabling and constraining of action by structure’.


Two dimensions (in this case, action and communication) coupled like the rows and columns of a matrix are said to be structurally coupled: operation in one dimension necessarily has an effect on the other which constitutes its environment.  (In terms of the matrix representation, a cell value is changed that affects a row and a column locally.)  Over time, structurally coupled systems ‘mutually shape’ each other or ‘co-evolve’ (Leydesdorff 1994). 


As noted, communication between a sender and a receiver requires transmission of the signal by the network between them, and thus communicating systems (i.e., actors) are operationally coupled by the network between them.  First, the sender operates and thus disturbs the network (i), then the network operates in terms of transmitting the message (j), and only on the basis of this first selection, the receiver reconstructs using his/her interface (jk) with the network as a second structural coupling.  The recursive selections, however, operate in substantively different layers.


Codification of the communication (along the column vectors of the communications), self-organizes the system by adding another (reflexive) selection when grouping the communications.  This grouping along the column dimension of the matrix (j’ or l) can become functional in sorting various types of communication apart.  Using this additional interface, the complex system is able to self-organize the (sub‑)systems in terms of their operational and structural couplings.  The distribution of previous reflections of interactive events is then reconstructed, yet hyper-reflexively. 


In a biological system this degree of freedom is functionally geared towards survival and reproduction (Maturana 1978).  However, a social system of inter-human communication contains one more degree of freedom, since the participants are able to reflect on what they observe and the reflections can also be communicated.  Consequently, this system is only loosely coupled in principle: it has an internal degree of freedom for adjusting its own further development (cf. Simon 1969).  While human cognition is biologically constrained, social learning can develop itself as a cultural evolution on top of the biological evolution.


In other words, a self-organizing system feeds back on its instantiations and their trajectories, and it is able to reorganize its past with hindsight and selectively in terms of new recombinations.  Thus, a social system is able to learn in terms of institutional rearrangements (Etzkowitz & Leydesdorff 1997).  The recombinations are dependent on the system’s capacity to adjust to recurrent translations at the relevant interfaces.  However, there is no center of control, since the actions on which the system rests are distributed.  Control emerges; the system can only sub-control with hindsight, that is, based on suppression by selection.


The system has to select continuously because it is disturbed by a multitude of lower-level interactions.  If the channels of communication were to become overburdened, integration might fail, and the system would experience crises in its reproduction.  The hyper-cyclic social system, however, can increase its complexity without being delimited by biological constraints, since it couples only operationally and not structurally to the biological layer.  The psychological level or-as Luhmann has called it-‘human consciousness’ has to operate in order to make biological issues (like ‘bio-diversity’) culturally relevant, and vice versa.



‘All that is solid melts into air’


This quote from Marx (Communist Manifesto, 1848) illustrates that Marx as a philosopher had already grasped this evolutionary dynamics of modern society (Berman 1982).  While Ricardo and Malthus had a notion of naturally given limitations, Marx entertained a dialectic notion of the relation between nature and culture.  In his opinion, the dynamics of society would dialectically be confined because this evolution would eventually self-generate its own crises.  The ensuing tensions (between Capital and Labour) would have to be resolved in a single, global solution.


In social-systems theory one should radically reflect on Luhmann’s (1984) conclusion that society is not a living system.  Because society is not biological, it is not inherently limited in its extension, and it is not necessarily confined by its natural environment.[10]  Society is not expected to die (but it might have to stop if there were no reflexive carriers, i.e., human beings, left).  Thus, society needs no resolution; on the contrary, it needs further differentiation in order to be able to operate in terms of recurrent communications. 


From each perspective (that is, ‘bounded rationality’), other perspectives are perceived as not yet sufficiently reflected environments, since the boundaries of a rationality can only be perceived from another perspective.  Thus, environments challenge the human carriers to explore the possibilities of increasing society’s internal complexity.  I have specified above how this may happen at the (super‑)systemic level, that is, ‘beyond intent and consciousness’ by distributions operating upon one another.  The boundaries of society-whether conceptualized in terms of ‘risks,’ ‘ecology,’ or ‘exclusion’-are then no longer considered as given, but as constructed.  The reflexive analysis is pursued from a post-institutional perspective.  Although not given, boundaries are expected by reflexive and distributed actors. 


What does it mean, for example, to consider ‘exclusion’ as a boundary of functional differentiation from within the social system (Luhmann 1997b)?  In my opinion, the awareness of ‘exclusion’ can be considered as an irritation at certain places within the social system.  If codified in interactions among people, it may motivate the rise of a social movement.  Analogously, one can argue that the ‘ecology’ and ‘risks’ are irritations within the system.  Social movements feed back on the social system by challenging its capacity to handle complexity. 


The ‘new social movements’ can thus be considered as internally produced irritations of the system that call for a revolutionary change by taking into account another, not yet sufficiently reflected dimension of the system’s evolution.  As noted, the social system can cope with internally generated irritations by increasing its complexity.  Note that this has already happened to a large extent in the case of the ‘old social movements.’  In this sense, the new social movements replace the traditional opposition by Labour.


‘Labour’ has been accommodated in the system by mechanisms of compromise contained in the creation of the welfare state.  Similarly, the ‘new social movements’ appeal to new dimensions of differentiation in social organization that have not yet been sufficiently specified.  If these emerging dimensions were codified, it could mean that the corresponding social movement would have fulfilled its transformative function.  The problems that generate the tensions are then not resolved, but sub-integrated like issues that provide a selective constraint to the further dynamics of the social system.



Theoretical Perspectives


One of the functions of ‘social movements’ is to put issues on the agenda for codification.  However, social movements are not the only mechanism that is functional for this purpose.  Science has also a function here: social movements generate, among other things, the political pressure to legitimate the further development of the system, for example by making resources available for the sciences to reconstruct the issues thus placed on the agenda.


What can systems theory teach us with respect to the dynamics that one may expect in relation to these social challenges to the social system?  The systems theoretician in this case is not able to forecast the new dimension substantively because the code for the failing dimensions of the system is found at the level of the social system and not at a formal level.  At the substantive level, the qualitative sociologist has to step in for the development of a heuristic theory.  However, the expected dynamics of this ‘infra-reflexivity’ (Latour 1988)-that is, the capacity to reflect on the system’s operation both from within the system and as an external observer-can be specified using a formal perspective.


The urge to new code is social because the subjects (‘consciousness systems’) know themselves to be limited, and therefore unable to carry responsibility for the emerging limitations to the social system.  The perceived limitations negatively indicate the need for new code.  Because of the substantive nature of this quest for code, the issue obtains the character of an essential question: how can a system that is historically contingent be understood as not clearly limited in time, space, or in terms of participation?  In which respects does such a system differ from a God-like being?


Note that the problem is somewhat analogous to Hegel’s quest for the relation between the Subjective, Objective, and Absolute Spirit: how does the negation of a negation lead to a positive affirmation?  In our models, however, there is no assumption left of an eschatological reconciliation.  The social system is a complex dynamics that can further complicate its own internal structure by self-organizing additional dimensions, that is, codes to the communication.  When the system changes qualitatively in terms of its dimensions, all bets are off and earlier theories tend to become obsolete.


What does this mean in terms of relevant social science research?  In my opinion, it changes the relevant research questions like a paradigm shift: not the number of observable people or their behaviour, but the assessment of the complexity in communications becomes crucial for the carrying capacity of social systems.  In the above metaphor of a matrix, the emphasis shifts from the rows which represent actors and their actions to the columns that represent the communications among them.  The question about the nature of boundaries can then be translated into a researchable question about the process of exploring new codifications.  The role of new social movements can be appreciated from this perspective.


In my opinion, the inter-system dependency of technological innovation can be used to develop a model for studying the effects of historically emerging constraints.  First, the laboratory has allowed for the discursive reconstruction of what was considered as ‘naturally’ given, and thereafter for its reconstruction in social practice (Latour 1983).  Second, the systematic institutionalization of an interface between science and the economy since the scientific-technical revolution of the period 1870-1910 (e.g., Braverman 1974; Noble 1977) has created a full blown inter-system dependency that can be analyzed sociologically (Leydesdorff 1997b). 


The study of the resulting university-industry-government relations has, among other things, taught us that the relevant codes have become internally differentiated during the process.  Given innovation in a knowledge-based economy, product competition has become a relevant criterion in addition to price competition.  Truth-finding has been supplemented with puzzle-solving as an heuristic in the sciences.  Power has to be offset against functionality of control given a state bureaucracy.  In other words, the functional codes have become internally differentiated in order to cope with complexities in processes of translation at the interfaces.


The social system, then, can no longer be characterized only in terms of its functional differentiation.  The categories have become fluid: translations (at the hyper-cyclic level) entrain the further development of functional codes.  However, such a system cannot be observed directly, while it remains emergent.  It can be specified only in terms of expectations.  The position of each observing discourse provides a window on the complex dynamics, but there is no longer the expectation of any meta-position.


If we turn to the tetrahedron of Figure Four, each of the discourses can achieve the top position, and therefore claim a priority from its own perspective.  This corresponds with the distributedness of control in a pluriform society.  No objective criteria are given that allow us to escape from this post-modern dilemma.  However, one expects that all reflexive discourses have to be updated.  One means of stimulating them to update is to confront them with the unexpected results of complex interactions that potentially have generated new dimensions of communication. 


The emerging dimensions allow for new social movements, since other limits to the system’s development can be perceived from each new perspective.  Reflexively, the interactive possibilities of translation have a function by enabling us to analyze a post-modern society in terms of intertextualities among scientific disciplines.  For example, by integrating other forms of sociological analysis (e.g., symbolic interactionism with a focus on ‘meaning’) and the systems-theoretical approach, Luhmann has been able to address essential questions like those raised in this theme issue. 


Reconstruction by translation can thus be specified as an internal mechanism for the expansion of a non-reified social system.  This system continues to expand since the carriers cannot escape from their reflexivity.  This resource is used in a cultural evolution, and therefore the social system is no longer physically or biologically constrained.  We as carriers are constrained by our physical existence and our biological survival, on the one side, and by social communication, on the other.  Society is only constrained by our mediating capacity, and the other way round: society puzzles us because sociology has made us aware of our ambivalent position in it as both observers and participants.


While one is not able to delineate the social system and its boundaries at any moment in time, they remain historical.  The social system is expected to survive us, since a non-biological system cannot physically die.  The new evolutionary theorizing thus contributes to the secularization of essential questions as a predicament: the quest for the missing dimensions can be made a subject for sociology.





*          I am grateful for useful comments of Rudolf Stichweh and Gerald Wagner.



Berman, Marshall (1982): All That is Solid Melts into Air: the Experience of Modernity.  New York: Simon and Schuster.


Braverman, Harry (1974): Labor and Monopoly Capital. New York/London: Monthly Review Press.


Brown, George Spencer (1969): Laws of Form. London, Allen and Unwin.


Burt, Ronald S. (1982): Toward a Structural Theory of Action.  New York, etc.: Academic Press.


Henry Etzkowitz & Loet Leydesdorff (Eds.), (1997): Universities and the Global Knowledge Economy: A Triple Helix of University-Industry-Government Relations. London: Cassell Academic.


Gibbons, Michael, Camille Limoges, Helga Nowotny, Simon Schwartzman, Peter Scott, and Martin Trow (1994): The new production of knowledge: the dynamics of science and research in contemporary societies. London, etc., Sage.


Giddens, Anthony (1979): Central Problems in Social Theory. London: Macmillan.


Giddens, Anthony (1984): The Constitution of Society. Cambridge: Polity Press.


Giddens, Anthony (1990): The Consequences of Modernity. Cambridge: Polity Press.


Habermas, Jürgen (1987): Excursus on Luhmann’s Appropriation of the Philosophy of the Subject through Systems Theory. Pp. 368-85 in: The Philosophical Discourse of Modernity: Twelve Lectures. Cambridge, MA: MIT Press.


Kampmann, Christian, Christian Haxholdt, Erik Mosekilde, and John D. Sterman (1994): Entrainment in a Disaggregated Long-Wave Model. Pp. 109-24 in: Loet Leydesdorff and Peter Van den Besselaar (1994).  Evolutionary Economies and Chaos Theory: New directions in technology studies. London: Pinter.


Latour, Bruno (1983): Give Me a Laboratory and I will Raise the World. Pp. 141-70 in: Karin D. Knorr-Cetina and Michael J. Mulkay (Eds.), Science Observed. London: Sage.


Latour, Bruno (1987): Science in Action. Milton Keynes: Open University Press.


Leydesdorff, Loet (1992): The Impact of EC Science Policies on the Transnational Publication System. Technology Analysis and Strategic Management 4, 279-98.


Leydesdorff, Loet (1993a): ‘Structure’/’Action’ Contingencies and the Model of Parallel Distributed Processing. Journal for the Theory of Social Behaviour 23, 47-77.


Leydesdorff, Loet (1993b): Why the Statement ‘Plasma-membrane Transport is Rate-limiting for its Metabolism in Rat-liver Parenchymal Cells’ Cannot Meet the Public. Public Understanding of Science 2, 351-364.


Leydesdorff, Loet (1994): The Evolution of Communication Systems. Int. J. Systems Research and Information Science 6, 219-30.


Leydesdorff, Loet (1995): The Challenge of Scientometrics: The development, measurement and self-organization of scientific communications. Leiden: DSWO Press, Leiden University.


Leydesdorff, Loet (1996): Luhmann’s Sociological Theory: Its Operationalization and Future Perspectives. Social Science Information 35, 283-306.


Leydesdorff, Loet (1997a): The Non-linear Dynamics of Sociological Reflections,’ International Sociology 12, no. 1, 25-45.


Leydesdorff, Loet (1997b): The New Communication Regime of University-Industry-Government Relations. In: Etzkowitz & Leydesdorff (1997), 106-17.


Leydesdorff, Loet & Van den Besselaar, Peter (1997): Scientometrics and Communication Theory: Towards Theoretically Informed Indicators. Scientometrics 38, 155-74.


Luhmann, Niklas (1977): Interpenetration - Zum Verhältnis personaler und sozialer Systeme. Zeitschrift für Soziologie 6, 62‑76.


Luhmann, Niklas (1984): Soziale Systeme. Grundrisz einer allgemeinen Theorie. Frankfurt a. M.: Suhrkamp.


Luhmann, Niklas (1986): The autopoiesis of social systems. Pp. 172-92 in: F. Geyer and J. van der Zouwen (Eds.), Sociocybernetic Paradoxes. London: Sage.


Luhmann, Niklas (1990): Die Wissenschaft der Gesellschaft. Frankfurt a.M.: Suhrkamp.


Luhmann, Niklas (1997a): Die Gesellschaft der Gesellschaft. Frankfurt a.M.: Suhrkamp.


Luhmann, Niklas (1997b): Globalization of World Society: How to Conceive of Modern Society?  International Review of Sociology (1997, Nr. 1).


Maturana, Humberto R. (1978): Biology of Language: The Epistemology of Reality. Pp. 27-63 in: G. A. Miller and E. Lenneberg (Eds.), Psychology and Biology of Language and Thought. Essays in Honor of Eric Lenneberg.  New York: Academic Press.


Maturana, Humberto R. (1991): Response to Jim Birch, J. of Family Therapy 13, 375-93.


Mingers, John (1995): Self-Producing Systems: Implications and Applications of Autopoiesis. New York/London: Plenum.


Münch, Richard (1982): Theorie des Handelns. Frankfurt a.M.: Suhrkamp; [Engl. translation: Understanding Modernity. London/New York: Routledge, 1988].


Noble, David (1977): America by Design.  New York: Knopf.


Parsons, Talcott (1937): The Structure of Social Action. Glencoe, IL.: The Free Press.


Parsons, Talcott (1968): Interaction: I. Social Interaction.  Pp. 429-41 in: The International Encyclopedia of the Social Sciences, Vol. 7.  New York: McGraw-Hill.



Simon, Herbert A. (1969): The Sciences of the Artificial. Cambridge, MA.: MIT Press.


Tong, Jian (1996) ‘Reflections on Human Capital Theory and Niche Theory of Evolutionary Economics.’ Paper presented at the Triple Helix Conference, Amsterdam, January 1996.


Turpin, Tim & Garrett-Jones, Sam (1997): Innovation Networks in Australia and China. In: Etzkowitz & Leydesdorff (1997), pp. 21-32.





[1].         Whether an autopoietic system should be identified with a ‘living' system has been debated in the literature.  See, e.g., Mingers (1995) for a review.  More recently, Maturana (1991, at p. 376) stated:  ‘Yet ..., it could have been proper to claim that all autopoietic systems, regardless of the space in which they occur, are living systems.'

[2].         For example, Parsons (1968, at p. 473) formulated: ‘The phenomenon that cultural norms are internalized to personalities and institutionalized in collectivities is a case of the interpenetration of subsystems of action, in this case social system, cultural system and personality (...).  Here the critical proposition is that institutionalized normative culture is an essential part of all stable systems of social interaction.  Therefore, the social system and the culture must be integrated in specific ways of their interpenetration.'

[3].         Giddens (1984) allowed for iuxtaposing systems of reference, but the reflection on the ‘duality of structure' as an operation was not elaborated (cf. Leydesdorff 1993a).

[4].         I use the term ‘information' here as equivalent to Shannon's concept of probabilistic entropy.  Whenever I speak of information as meaningful information, I will consider such information as structurally coupled to an observer (or a distribution of observers) that has provided the information with meaning (cf. Leydesdorff 1996).

[5].         Identity can be considered as the special case that the distribution has vanished because the probability has reached the value of one at a single locus.

[6].         Luhmann has argued that a functional sub-system can communicate only in terms of a binary dichotomy like ‘true' and ‘false'.  In general, a dichotomy spans a dimension which allows for shades of grey.  In my opinion, a code can be considered as a meaning that can be attached to the information (‘uncertainty') in a second dimension (Leydesdorff 1996).

[7].         A statistical test can be developed for distinguishing whether the history of the composing elements or the assumption of systemness in their interaction provides a better prediction for the next stage of the expected information content of the distribution (Leydesdorff 1995).

[8].         See note 5.

[9].         For example, at p. 355: ‘Die Ausdifferenzierung der Wissenschaft führt mit anderen Worten, zu einem Doppelzugriff der Gesellschaft auf Wissenschaft - nämlich von aussen und von innen, über andere Funktionssysteme und über die Sonderautopoiesis des Wissenschaftssystems selber.'

[10].        Habermas (1987) has argued that Luhmann's sociology can be considered as a meta-biology.  In my opinion, Habermas has failed to understand the epistemological order: insights in the evolution of communication networks feed back on the self-understanding of the biological sciences that as discourses have always (yet unreflexively) transcended biological domains.