Kybernetes28,6/7

792

Kybernetes,Vol. 28 No. 6/7, 1999, pp. 792-809.# MCB University Press, 0368-492X

A sociocybernetic approach toinformation systems

developmentGuohua Bai

Division of Computer and Systems Sciences, LuleaÊ University ofTechnology, LuleaÊ , Sweden, and

Lars-AÊ ke LindbergDivision of Technical Psychology, LuleaÊ University of Technology,

LuleaÊ , Sweden

Keywords Cybernetics, Autopoiesis, Sociocybernetics, Information systems

Abstract A principal problem of information systems (IS) development today is how tointegrate the work activity and social context of users into the IS which is being designed. Thispaper aims to delineate a sociocybernetic approach which enlightens the relationship between ISdesign activity, use activity, and the embedded social context. First and second order cyberneticsare employed as a general skeleton or structure for conducting the sociocybernetic approach.Approaches from social autopoiesis theory and activity theory are introduced respectively toprovide the skeleton of cybernetics with flesh and blood elements of real problems. An emergingconceptual body resulting from an organic integration of the cybernetic structure and social-psychological elements which demonstrates the relationship between IS design activity, useactivity, and embedded social context is presented.

IntroductionThere is increasing dissatisfaction with the usability of information systems(IS) and the traditional design theory and model, such as the life cycle model,which presupposes a static reality and a reliable match between the mentalmodels of users and designers. There are also approaches to resolve theproblem, such as the participatory design approach (Ehn and Kyng, 1987;Bjerknes, 1992; GroÈnlund and Bai, 1993) in Scandinavia in which users'participation in the design process is considered a major way of increasing theusability of IS. However, the participatory approach was initially based onpolitical theory and the unions' democratic movement, and had concentrated onpractical solutions rather than on theoretical understanding of the generalrelationship between design activity, use activity, and embedded social context.

The aim of this paper is, therefore, to focus on a theoretical understanding ofthe relationship between design activity, use activity, and embedded socialcontext from a social-psychological perspective. The structure used to conductthis theoretical approach is based on cybernetics, and the elements that fill inthe structure are identified by integrating social autopoiesis theory (Luhmann,1986) and activity theory (EngestroÈm, 1987; LeonteÁv, 1981).

The current issue and full text archive of this journal is available athttp://www.emerald-library.com

Informationsystems

development

793

Cybernetics approachThe term cybernetics is derived from the Greek word `̀ kubernetes'', meaningsteersman, or the art of steering. It referred principally to the piloting of avessel. Wiener (1948) first introduced the concept in his famous bookCybernetics, or Control and Communication in the Animal and Machine. Thus,the term cybernetics is used to define an area of technical engineering andmachine control, which is referred to later as first order cybernetics.

After the founding of first order cybernetics, scientists from different areasextended the scope of cybernetics beyond feedback control to include manyareas of study, such as the biological and open systems theory (Bertalanffy,1972), information and communication theory (Shannon and Weaver, 1964),self-organisation and regulation (Ashby, 1956), and management andorganisational cybernetics (Beer, 1985). Bateson (1972) uses the single term`̀ cybernetics'' to refer to the aggregate of these areas, and calls cybernetics `̀ thebiggest bite out of the fruit of the tree of knowledge that mankind has taken inthe last 2,000 years (Bateson, 1972, p. 476)''.

Second order cybernetics originated from social and biological studies (VonFoerster, 1984; Luhmann, 1986; Maturana and Varela, 1972) from the early1970s to the 1980s, in reaction to some problems with first order cybernetics inapplying to social studies (Geyer, 1995). Second order cybernetics is thecybernetics of observing rather than observed systems (Von Foerster, 1984).Some concepts of second order cybernetics, such as epistemologicalobservation, self-reference, ontology of systems, and purposeful behaviour, arediscussed in the following.

Approach to observation of realityA problem of first order cybernetics from the social scientists' point of view liesin its principle of objective observation, namely, the properties of an observedbehaviour have nothing to do with the properties of the observers. First ordercybernetics developed in a period when people believed in the possibility ofvalue-free science.

A prerequisite of first order cybernetics is that the lodestar observer is non-intervention, neutral and objective. In this view, observation of the same objectperformed by different observers in different parts of the universe and atdifferent times should always give the same results. Dissimilar results areattributed to human deficiency or deception and will be corrected throughbetter precision and repetition of the observation. When, however, the object isa social system, this basic metaphysical presumption of first order cyberneticsis seriously attacked by constructivism in sociology (Glasersfeld, 1996;Maturana and Varela, 1972; Searle, 1995). As formulated by Bauersfeld (1992):

we cannot `̀ see'' reality `̀ as it is'', there is no direct access. The only way for human beings is toconstruct their own interpretations, their own reality, . . . There is no `̀ copying'',`̀ matching'', `̀ replicating'' or `̀ reflecting'' (as a physical mirror works) of any `̀ external world''(p. 16, bolds original).

Kybernetes28,6/7

794

Many arguments concerning different views of observations andunderstandings of a system may originate in an ontological prejudice, namely,each takes it for granted that systems exist ± such as things exist; therefore, weare all able to observe and talk about the same object, such as a cup. Thus onereason of the different views concerning observations may lie in the differentsettings of the `̀ system''. So, what is a `̀ system''?

Approach to thinking about `̀ system''According to Checkland (1988), a system

is the name of an abstract concept to be used in a conscious process of trying to make sense ofthe world. `̀ System'' is thus not a label word like `̀ cat'' or `̀ table'', words used unreflectively ineveryday living; it is an epistemological device, one to be used consciously in the process oftrying to understand the complexity of perceived reality until such time as it is replaced by amore powerful concept (p. 14).

Senge et al. (1996) defined a system as

a perceived whole whose elements `̀ hang together'' because they continually affect each otherover time and operate toward a common purpose (p. 90).

So a system is not a physical object in the real world, but a subjectively definedconcept for perceiving the world. This implies that it is necessary to define thesystem being observed before any discussion about the observation takesplace.

Under the kernel concept of system, properties of objects perceived in thereal world such as relations, hierarchy, order, purposeful behaviour, variablesand dynamics can all be studied in a coherent framework of a system. We canfurther define different types of systems, such as real systems with entitiesperceived in or inferred from observation and existing independently of anobserver; conceptual systems which essentially are symbolic constructs; andabstracted systems which are abstracted from corresponding reality(Bertalanffy, 1972). According to Checkland's (1981) systems typology, theabsolute minimum number of classes of systems necessary to describe realityis five. They are: natural systems, human activity systems, designed physicalsystems, designed abstract systems, and transcendental systems. Naturalsystems `̀ are systems which could not be other than they are, given a universewhose pattern and laws are not erratic''. The designed physical systems, suchas tools, artefacts, machines, and other material entities, are created by humanbeings with some purposes. Designed abstract systems are various types oftheological, philosophical or knowledge systems created by the human mind.Human activity systems are the link between natural systems and designedsystems. Everything that is carried out by humans is labelled a human activitysystem, such as: materials production, goods transportation, defence, trading,politics, and scientific investigation. Finally, the transcendental systemsrepresent systems beyond current knowledge, or an unknown world.

All those definitions about system are about how to facilitate ourunderstanding of a perceived whole. But by what do the elements `̀ hang

Informationsystems

development

795

together'' as a system or whole as we perceive it? We perceive objects as asystem because they are related to some purposes we are pursuing. This meansthat in order to define a system properly we need to understand what ispurposeful behaviour or goal seeking.

Approach to purposeful behaviourThe idea of purposeful behaviour defined in first order cybernetics is that anobject behaves purposefully if it continues to pursue the same goal by changingits behaviour as its environment changes (Rosenblueth et al., 1943). It is thusrequired that the purposeful systems receive information from the environmentand respond to changes in the environment through feedback. Although werecognise that this feedback process is involved in much purposeful behaviour,this restriction, i.e. available feedback from the reaction of the environment,should not be a necessary requirement for studying all purposeful behaviour,because such transmission of information in many psychological and socialexamples of purposeful behaviour is either not present or is not of concern tothe observers. For example, a person who pursues a promotion or position (apurposeful behaviour) could take many different actions even though theenvironment remains the same or even he or she could not know if theenvironment had changed. Based on this argument, Churchman and Ackoff(1950) developed a category of purposeful behaviour, comprising extensivefunction, intensive function, and purpose:

. A system has an extensive function if it consists of objects thataccomplish certain objectives by displaying relatively invariantbehaviour over a wide range of environments; e.g. an ordinary clock hasan extensive function since it serves a purpose by displaying relativelyinvariant behaviour in all environments.

. A system has intensive function if it consists of objects that accomplishtheir goals by changing their behaviour if the environment changes, butgenerally exhibit only one type of behaviour in any given environment,e.g. light switches, servo-mechanisms such as automatic temperaturecontrol devices, self-aiming guns and missiles. The intensive function iswhat the first order cyberneticians are interested in.

. A system has purpose if it accomplishes its objectives by exhibitingdifferent types of behaviour, even though the environment remainsconstant. Typical examples are artificial intelligence, animal behaviourand human activity systems. This is what the second ordercyberneticians are primarily concerned about.

Approach to self-referenceThe concept of self-reference has dominated the current scene of second ordercybernetics. In a general sense, self-reference is involved in a description thatrefers to something that affects, controls, or has power to modify the form orthe validity of that description. An example is the description of a system by an

Kybernetes28,6/7

796

observer who is part of the system observed. Actually, the idea of self-referenceis related to the concept of feedback. Feedback implies self-reference since thesystem reacts to its own previous action. In this sense reference is the previousstatus of the system itself (self-reference).

Any observation starts from a blind spot which is invisible to the observerwhile he/she observes (Von Foerster, 1984). In first order cybernetic systems,the blind spot of current observation is observed by the future observation, i.e.by continuous feedback or continuous self-reference. Social scientists areobservers of social behaviour, i.e. human beings' self-observation, and theobservation also has blind spots that can only be observed from God'sviewpoint. In order to see blind spots of our own selves, Maturana and Varela(1972) suggest `̀ multi-verses'', not `̀ uni-verses''. Everyone observes from theirlocal point of view, and creates their own world. But a universe can beconstructed when observers co-ordinate, i.e. a universe can be constructed by`̀ multi-verses''. Churchman (1979) proposed a strategy for systems scientists,namely, `̀ sweep-in enemies'' of systems thinking, such as politics, morality,aesthetics, and religion, into systems in order to observe blind spots of systemsscientists themselves.

There is another kind of reference more active than self-reference, especiallyfor human activity systems, namely, external-reference. This indicates to`̀ behaviours'' that a system will purposefully and actively change theenvironment, and by referring to the changed environment the system thenchanges its structure, function, and states in order to be more viable in thechanged environment. As pointed out by Aulin (1982): `̀ Human action is notjust reacting to environment changes ± like adaptation is ± but it entails anactive effort to change the environment for human purpose'' (italic original).

A typical behaviour of this external reference is the activity called `̀ systemsdesign''. The process of design is not only a process of self-reference whichmostly involves such behaviour as adaptiveness, mental model construction, orreflection, but which is rather a process of external reference which involvesdesigning the external world and learning by doing. The external reference ismore significant to our knowledge and understanding of social reality, sinceyou cannot gain systems understanding unless you can take part in changing it(Senge et al., 1996, p. 94).

The cybernetic approach provides a general structure and terminology todescribe a system and its behaviours. To deal with concrete problems of ISdesign, however, the structure and terminology must be specified and filledwith concrete elements; e.g. where does feedback come from in design? What isthe purpose of a design? What are design activity systems? In the following, thetheories of social autopoiesis and activity will be applied to identify thoseconcrete elements.

Social autopoiesis theory approachLuhmann's autopoietic theory is too complex to present adequately here. Forour purposes it is sufficient to present some basic concepts underlying his

Informationsystems

development

797

theory, such as autopoiesis, self-reference, communication and differentiatedfunction. For further discussion of Luhmann's autopoiesis theory see Mingers(1995) and Bailey (1997).

Self-referenceThe notion of self-reference is at the heart of Luhmann's autopoiesis theory. Hefirst proposes a fundamental distinction between system and environment,since one could not talk about a `̀ self'' at all and could not designate a `̀ self'', ifnothing else than this self existed (Luhmann, 1995a, p. 5). Self-referentialsystems can define themselves only with reference to what they are not. Thesystem would never be able to build its own complexity and its own knowledgeif it repeatedly mistook itself for its environment (Luhmann, 1995a).

Based on this distinction of system and environment, Luhmann's term `̀ self-reference'' includes two components: auto-reference and hetero-reference(Bailey, 1997). Thus, the self-reference of an autopoietic system is not confinedonly to reference to the system self (e.g. design is design) which can be calledauto-reference (Bailey, 1997), but also to entities other than the system itself(such as an environment or context). An example is that to understand what isdesign you have to refer to or contrast with what is non-design; i.e. what is use.Thus, a self-referential system cannot exist on its own; it must interact with theenvironment. According to Luhmann (1995b, p. 437), social systems whichinclude societies, interactions, and organisations (Bailey, 1997) are undoubtedlyself-referential systems since we can observe and describe them as systemsonly by acknowledging that they refer to themselves in every operation (Bailey,1997).

According to the concept of self-reference, design activities would not beable to build their own complexity and knowledge without reference to useactivities. Practically, this means that designers must communicate with usersduring the design process. This is an example of hetero-reference. The designprocess also involves auto-reference, i.e. the designers have to refer to or useanother IS, namely, a communication system (CS), in order to communicatewith its context or reference of use (Bai, 1997).

Autopoietic systems`̀ Autopoietic'' is a concept developed in the 1970s by Maturana and Varela(1972) in describing cell systems in which a cell not only creates its own unityas a system and its own structure, but also its own elements. Luhmanngeneralises the concept to cover not only biological systems, but also socialsystems. In Luhmann's own words, autopoiesis: `̀ refers to systems thatreproduce all the elementary components out of which they arise by means of anetwork of these elements themselves''. An autopoietic system isorganisationally closed, but communicatively open. `̀ There is no input and nooutput of unity'' (Luhmann, 1986, p. 174). According to Luhmann (1986; 1995b),social systems are autopoietic systems in which communication is theproduction unit.

Kybernetes28,6/7

798

It would be wrong to consider organisations which conduct IS designactivities as social autopoietic systems, since obviously any IS design will needboth materials, energy, and information from other labour divisions, especiallyfrom users' organisations. So design could not be autopoietic; or we may notcall design activities themselves a social system in Luhmann's term. To beautopoietic, IS design activities must be expanded to include all the necessaryelements embedded in other social labour divisions. At the end of the paper, wepropose this expansion (see Figure 5, to be discussed later).

CommunicationLuhmann boldly characterises social systems, categorised as societies,interactions, and organisations, as autopoietic, but with communication as thebasic unit. Communication includes information, utterance, and understanding(Luhmann, 1989, p. 143). Among the three types of social systems, societiesinclude all communications, in a closed sense. That is, society does notcommunicate with its environment. Interactions take environmentalcommunication into account, and are closed in the sense that theircommunications can only be understood in the context of the interactionsystem. Organisations and groups seem to lie between the relative closure ofsociety and the relative openness of interactions. According to Luhmann (1986)

social systems use communication as their particular mode of autopoietic reproduction. Theirelements are communication which are recursively produced and reproduced by a network ofcommunications and which cannot exist outside of such a network (p. 174).

Even more radically, he states that

for a theory of autopoietic systems, only communication is a serious candidate for the positionof elementary unit of the basic self-referential process of social systems. Only communicationis necessarily and inherently social (Luhmann, 1986, p. 177).

We may question whether communication is the only elementary unit of socialautopoietic systems. However, the crucial role of communication in modernsociety is obvious. In IS design activities, communication between designersand users is very crucial, and therefore, the communication must be supportedby a communication system (CS) properly embedded in the designed IS (Bai,1997).

Differentiated function systemsModern societies are divided into functional subsystems such as laws,economics, science, politics, and education, which Luhmann calls differentiatedfunction systems. A society increasingly contains differentiated functionsystems as the numbers of social labour divisions increase over time.

According to Luhmann, each subsystem is organisationally closed (to unity)and communicatively open (to information). As a consequence of thisorganisational closeness, it is difficult, even impossible (without the help ofGod), to talk about what is rational regarding society as a whole, or what willbenefit society as a whole (Thyssen, 1995). The co-operation of functional

Informationsystems

development

799

subsystems is not based on consensus, but on their mutual benefits and on thecomplementary principle (Bai and Lindberg, 1998). No society can afford tomake co-operation among functional subsystems dependent on consensus(Thyssen, 1995). IS design and IS use are functional subsystems of societies.Neither design organisations nor users' organisations can conduct theiractivities based on their own interest: their activities must benefit bothorganisations.

Critiques of Luhmann's social autopoiesisLuhmann's social autopoietic theory has recently generated both attention andcritiques (Mingers, 1995; Bailey, 1997; Kennealy, 1987). There are two mainproblems regarding Luhmann's social autopoiesis which will be complementedby the activity theory (Davydov et al., 1982; Vygotsky, 1978; LeonteÁv, 1974;EngestroÈm, 1987) in the following. One problem is that it does not provide acomprehensive explanation to the question: what exactly in the society arebeing reproduced if social systems are autopoietic or self-reproducing systems?Or stated another way: what is the proper analysis unit of social systems?There are many possible choices. The unit could be individual (Mingers, 1989;Miller, 1978), situated action (Hutchins, 1994; Suchman, 1987; Lave, 1988),organisations (Beer, 1985), and activity (EngestroÈm, 1987). Luhmann, however,exclusively chooses the communication as the unit of reproduction. ToLuhmann, a social system does not consist of human beings or artefacts. Itconsists of an ongoing stream of communication (Luhmann, 1989). We do notthink that communication is the only unit in a society that is being reproduced.Actually, communication is not a comprehensive unit compared to the unit ofactivity following activity theory. However, we are not forced to abandon theunit of communication, action, individuals, etc. for activity alone; all frames ofreference are fruitful, and none of them is fundamental. To understand ISdesign, we conduct an integration of all these contingent frame works at theend of the paper.

Another problem with Luhmann's theory is the closeness and separation ofdifferentiated function systems. It has been argued that one of the markedfeatures of the modern world has been the extent of its differentiation ordivision of labour, which leads to increased collective productivity. The moredifferentiation there is, the more specialised the roles of the social actors, andtherefore the intellectual and inter-divisional co-operation, communication andcontrol become more important (Corning, 1994). Even though society can beautopoietic, it does not automatically imply that its differentiated functionsystems are also autopoietic, although according to Luhmann, for example, lawis a differentiated function system and an autopoietic, self-referencing systemwhich produces every kind of unity that it requires. Considering law asautopoietic would imply that law alone is able to produce and reproduceautopoietically all legal norms. This undermines the power of social interactionamong other differentiated function systems, such as mass media, economicsand politics. On the operational level, it is true that the law reproduces the

Kybernetes28,6/7

800

elements which produce the laws without exposing the reproduction to theenvironment. There are, however, some vital factors which determine what thelaw is to be. It would be incomplete, even incorrect, if we just saw the operationlevel without looking into its activity level (see `̀ activity theory'' in the nextpart). Seeing the mechanism of producing a law without seeing the embeddedsocial context is like seeing an iceberg as being only that part above thesurface. Let us take an example which concerns the law of weapons control inthe USA. Although the White House has tried for many years to enact aweapons control law, the law has surrendered to the economic interest of theweapons business, in the face of many innocent people and schoolchildrenbeing killed.

With regard to the problem of IS design, it is also clear that designorganisations cannot be autopoietic, as we have pointed out above. In the nextpart, the problem of a more solid unit for social autopoietic reproduction, theproblem of a clarified social context, the problem of interrelation amongdifferentiated function systems, are approached by the activity theory.

Activity theory approachActivity theory originated in the so-called cultural-historical school of Sovietpsychology. It uses the category `̀ activity'' as `̀ a system of its own structure''(LeonteÁv, 1981, p. 46) to approach the relationship between the subjective mindand its social context. Today there is an emerging multi-disciplinary andinternational community united by the central category of activity for studyingvarious forms of human practices; e.g. human-computer interaction (Nardi,1996; Kaptelinin, circa 1996; Kuutti and Favorin, 1993; Kuutti, 1996),development work research (EngestroÈm, 1987; Saarelma and EngestroÈm, 1993),information systems design (Bjerknes, 1992), organisation analysis (Holt andMorris, 1993). Some important characteristics of activity theory, most of whichoriginated in the work of Davydov et al. (1982), Vygotsky (1978), LeonteÁv,(1981), and EngestroÈm (1987), are summarised in the following.

Activity is a necessity of human social lifeLike oxygen and foods to human physical existence, activity is like oxygen andfood to human mind. No one can survive without participating in various socialactivities; e.g. materials production and distribution activities, political anddemocratic activities, scientific and educational activities, religious activities,sports and various entertainment activities. It is `̀ the nonadditive, molar unit oflife'' (LeonteÁv, 1981, p. 46).

But how is activity possible? According to Searle (1995), there are twobiological primitives of human beings: `̀ background capacity'' and `̀ collectiveintentionality'', which enable various social activities. The `̀ backgroundcapacity'' indicates the capacity of each individual being able to follow social orinstitutional rules, norms, laws, etc. As re-formulated by Qvortrup (1996):

. . . Rule-governed structures of human institutions are followed by people not because theyhave explicitly learned and memorised the rules, but because they have developed a set of

Informationsystems

development

801

capacities and abilities that render them at home in the society. These capacities and abilitiesare labelled `̀ background abilities'' (p. 33).

The `̀ collective intentionality'' explains why individuals with their singularintentionality wish to participate in social activities. Searle's (1995) answer is:`̀ in addition to singular intentionality there is also collective intentionality''(p. 23). This collective intentionality is not the sum of singular intentionality,but a new emergent property of collectives. Activities are formed by personssharing a `̀ We intention'' with their `̀ background capacity''.

Activity is an interaction between subjects and their physical, social and culturalenvironmentA subject is a conscious actor or a group of conscious actors. An object is somepart of the real world on which the subject acts. `̀ If I act, there is something infront of me, an object'' (Schwarz 1997, p. 24). LeonteÁv considers the activity asthe `̀ middle link'' in a three-part scheme between subject and object (LeonteÁv1981, p. 46). The object manifests itself only if there is a subjective interaction.There is non-separability between object and subject. Activity theory takessocial, historical, and cultural properties to be as objective as physical andbiological properties, and maintains that consciousness is located in everydayobjective practice: you are what you do. And what you do is firmly andinextricably embedded in the social matrix of which every person is an organicpart (Nardi, 1996). An object (objective) is always held by a subject, a person ora group of persons who is or are engaged in an activity, provides motives forthe activity, and gives the activity specific direction. `̀ Behind the object, therealways stands a need or a desire, to which the activity always answers''(LeonteÁv, 1981, p. 46).

Activity is a reciprocal transformation between subject and objectIn activity, the object is transformed into its subjective form or image (mentalmodels, theories, skills, consciousness, etc.) which is `̀ generalised, verbalised,abbreviated, and most importantly, becomes susceptible to furtherdevelopment that exceeds the possibility of external activity'' (LeonteÁv, 1974, p.18). At the same time, internal processes manifest themselves in externalactions performed by persons, and are converted into objective results andproducts. This reciprocal process, respectively named internalisation andexternalisation (Davydov et al., 1982; Vygotsky, 1978; LeonteÁv, 1974) is shownin Figure 1.

Activity can be analysed in a hierarchical structureLeonteÁv (1974) provided a three-level scheme of activity (Figure 2). Activitiesare realised through actions or clusters of actions that are generated or relatedto the same overall object or motive. Participating in an activity comprisesconscious actions that have immediate, and defined, goals. These actions mustin turn be technically implemented through operations that are well-defined,and habitual routines in accordance with conditions during the operation.

Kybernetes28,6/7

802

The action and operation are dynamically defined: when an action has beenpractised long enough, the action will be collapsed into an operation and newkinds of action will be created with their corresponding new operations. On theother hand, when conditions change, an operation can again `̀ unfold'' andreturn to the level of conscious action (Kuutti, 1996).

Activity is always mediated by artefacts`̀ Mediator objects connect humans not only with objects, but also with otherpeople'' (LeonteÁv, 1974). In particular, mediator objects are understood asobjective transmitters for the internalisation process and externalisationprocess (Figure 3). Artefacts include signs, language, concepts, internalisedmental models and physical tools. Artefacts carry with them a particularculture and history (Kuutti, 1992) and are persistent in structures that stretchacross activities through time and space (Nardi, 1996). So artefacts, which haveso far condensed the human cultural-historical development of activities, play

Internalisation

Human Activity SystemsMan-machine systems

Service systemsInformation design and use Systems

Political Systemsetc.T

he S

ubje

ct W

orld

Externalisation

The

obj

ect W

orld

Designed Abstract System

TheoriesMental ModelsConsciousness

KnowledgeSkill

Natural SystemsAnimals

Plant organsAtomic systems

Crystal system (rocks, seas,..)

Designed Physical Systemstools

machinesframework

cybernetic system

Figure 1.The reciprocaltransformation ofactivity

Activity Motive

Action Goal

Operation Condition

Figure 2.Hierarchical structure ofan activity

Informationsystems

development

803

the role of mediating `̀ subject-object'', and also of mediating the subject with hisor her social context.

Activity is initially social in natureFor much current sociology, in a certain sense only individuals exist: theindividual human being is the ontologically given starting point from whicheverything else must be deduced. Consequently, socialisation, norms, values,and culture are popular means for explaining why individuals unite intosomething called society (Qvortrup, 1996). Activity theory, however, takes therelationship between an individual and his or her social cultures, norms, andvalues as a molar unit of an activity and believes that activity is only developedunder conditions of co-operation and social interaction among people (LeonteÁv1981, p. 55). EngestroÈm (1987) crystallises the social context, such ascommunity, social and cultural rules, norms, and social labour divisions intoone unified system and logical interrelation (Figure 4). The interrelationbetween the social elements will be discussed in terms of contradiction in thefollowing ± another important concept of activity theory.

Tool

Subject ObjectFigure 3.

Tools mediated activity

Instrument

ObjectSubject

Production

Outcome

Consumption

DistributionExchange

Community Division of LaborNorms, RulesFigure 4.

The social activity

Kybernetes28,6/7

804

Activity theory considers contradictions as a basic resource of developmentContradiction is a crucial concept or a general category in activity theory whichdescribes the complementary nature of the oppositions, such as mutuality,interdependence, diversity, and dynamic relationships, that are deemedimmanent in a development process (for detailed discussion, see Bai andLindberg, 1998). Contradictions manifest themselves as problems, ruptures,breakdowns, and clashes. Activity theory sees contradictions as sources ofdevelopment. Activities are virtually always in the process of working throughcontradictions (Kuutti, 1996). In Figure 4, together with the traditionalcontradiction `̀ subject-object'' mediated by artefacts, two new contradictionsare identified after the introduction of community, the contradiction `̀ subject-community'', and the contradiction `̀ object-community''. In a similar manner asthe artefact is introduced as the mediator of the contradiction `̀ subject-object'',EngestroÈm introduces `̀ norms, rules'' as the mediator of the contradiction`̀ subject-community'', and introduces `̀ division of labour'' as the mediator of thecontradiction `̀ object-community''. Based on Marx's terminology of socialproduction, exchange, distribution, and consumption, he further maps thosefour human fundamental activities as four sub-triangles.

According to Figure 4, the development of human production, exchange,distribution, and consumption activities are accordingly driven by fourcontradictions. First, production activity is driven by the contradiction `̀ subject-object''; namely, by using artefacts (`̀ tool'' in the terminology of Vygotsky, firstmediator) the subject works and produces the objects that correspond to thegiven need or an outcome. Second, the exchange activity is driven by thecontradiction `̀ subject-community'', namely, the subject exchanges his labourvalue (exchange value) within the community to obtain his needs (use value)according to the community's rules and social law (second mediator). Third, thedistribution activity is driven by the contradiction `̀ object-community'', namely,the outcome of the object is distributed for social re-production amongmembers (organisations, companies) of the community according to theprinciples of the division of labour (third mediator). Finally, the total socialactivity system (the whole triangle) is driven by a new kind of contradiction:`̀ production-consumption'', namely, by the paradox that we produce outputand, simultaneously, we consume the output in order to re-produce it (self-reference in terms of Luhmann's autopoiesis theory). The contradiction of`̀ production-consumption'' provides an inner and never-ending energy thatdrives an accumulating cycle of consumption and production. `̀ Were it not forthe paradox that consumption necessitates production, and vice versa, activitywould not exist'' (Holt and Morris, 1993, p. 99).

Integration: a sociocybernetic frameworkCybernetic theory provides us with a fundamental structure or `̀ skeleton'' to`̀ hang up'' the `̀ flesh and blood'' elements identified by social autopoietic theoryand activity theory. The aim of this paper is to delineate a sociocyberneticapproach to enlighten the relationship between IS design activity, IS use

Informationsystems

development

805

activity, and embedded social context. The result of the integration of thestructure and the elements is shown in Figure 5.

In Figure 5, a social system as a whole can be viewed as a double-loopsociocybernetic system. The inner loop refers to the activities of socialproduction (IS design), consumption (IS in use), distribution (IS marketing for re-design or for customising) and exchange (IS marketing for end-user) identifiedby the activity theory in Figure 4. The outer loop refers to social-political andsocial-economic control activities. In terms of Luhmann, all those activities aredifferentiated function systems of society as a whole. They are mutuallyreferenced activities. The framework is not intended to contribute an IS designmethodology, but to highlight the broad social context and relationship of socialunits in view of IS design. For methodology based on this general framework,see Bai (1997); Bai and GroÈnlund (1991); GroÈnlund and Bai (1993).

Analysis of the inner loop

. Production-exchange-consumption: as described in the activity theory,social production (IS design) and consumption (IS in use) are the mostbasic human activities. These two comprise a fundamental contradictionwhich drives society forward, namely, the paradox or self-reference thatwe produce products, and, simultaneously, we consume products inorder to re-produce new products. Consumption necessitates production,and vice versa. Basically, production and consumption form a positivefeedback in the sense that the more there is consumed the more theremust be produced. Before a product (an IS) becomes a consumed objectfor individuals or organisations (in use), there is a process of exchange:individuals or organisations buy from the market (exchange value) fortheir needs (use value).

Open for inquiringand learning

Social ValuesHuman rightsPolitical anddemocratic decisions

Materials needs for re-production

Production(IS Design)

Market needs for consumption

Consumption(IS in Use)

Exchange(IS Marketing)

First loop Sociocybernetic system

Social Law,Rules, norm, etc.

Social sensorStatistic bureau

Investigator

Second loop Sociocybernetic system

Distribution(IS Marketing for

re-design, tailoring)

Figure 5.Double loops of

sociocybernetic system

Kybernetes28,6/7

806

. Production-distribution: Some products which are produced in theproduction process will not be consumed directly by members of society,whether individuals or organisations. Some products are sent back, ordistributed to further production as materials, e.g. software or hardwaredistributed to second hand developers for further design for somespecific end-users. The more products are produced, the more materialsneed to be distributed. Thus, production and distribution herebycomprise another positive feedback in the inner loop in Figure 5.

Though the sub-systems of production, consumption, exchange, anddistribution in the inner loop sociocybernetic system are social systems interms of Luhmann, none of them is, however, autopoietic and closed. Each sub-system is not self-referential in the strict sense (auto-reference). Instead, theyare mutually (hetero) referential. Since the inner loop sociocybernetic system asa whole forms a positive feedback loop, it is not a stable system. In the words ofForster

We cannot reach social and political stability for the reason that we continue to makescientific discoveries and to apply them, and thus destroy the arrangements which werebased on more elementary discoveries. If science would discover rather than apply ± if, inother words, men were more interested in knowledge than in power ± mankind would be in afar safer position, and the stability statement talk about would be a possibility. But scienceshows no signs of doing this . . . (Forster, 1951, pp. 90-1).

Analysis of the outer loopThe world, however, is not chaotic either. We see order and well-organised co-operation between differentiated sub-systems, and even between differentpolitical-economic societies or nations. The United Nations and many nationalgovernments can co-ordinate actions in dealing with world matters. Therelative stability of the world can be explained by the outer loop in Figure 5,which applies negative feedback mostly. The function of market/plannedeconomy control is a global adjustment mechanism, such as bank policy,investment policy, and national budgets. Different political-economic systemsmay put different emphasis on this control. From the perspective of outer loopsociocybernetics, the social system as a whole is self-referential and autopoietic.It is self-referential since the laws, norms and rules are created by humanbeings as a whole and are in turn used by human beings to organise humanbeings themselves. It is autopoietic in the sense that all elements inside thesystem are reproduced by the network of the elements themselves. Thefunctions of the social sensors (statistics bureau, reporter, investigator, etc.) areto be responsible for society being observable and for feedback. The never-ending human enquiry and learning will challenge, change, and constructcontemporary values and social laws, rules, norms, and culture as criteria ofsocial goals. The continuous construction of social goals is often manifested aspolitical, democratic or even military, activities. Therefore, the world is notabsolutely, but relatively, stable. When such an unstable situation occurs, theinner loop of the sociocybernetic system will not be stable, namely the social

Informationsystems

development

807

production, consumption, exchange and distribution activities will beinterrupted, destroyed, and reformed. In the case of IS development, this is nodoubt also true since the development of IS cannot be free from the control ofcontemporary social values and social laws.

ConclusionThe double-loop sociocybernetic system in Figure 5 is the result of theintegration from three theoretical disciplines: cybernetics, social autopoiesistheory, and activity theory. It sheds light on the complex social context withinwhich IS development takes place, and provides an epistemologicalunderstanding of the relationship among the elements involved in ISdevelopment. Based on this theoretical framework, one can developmethodologies for the practice of IS development (Bai, 1997; GroÈnlund and Bai,1993), and guide various research activities, such as the socio-technicalapproach (Mumford, 1987; Harker, 1993), and the participatory approach (Ehnand Kyng, 1987; Bjerknes, 1992; Schuler and Namioka, 1993).

References and further reading

Ashby, R. (1956), Introduction to Cybernetics, Methuen, London.

Aulin, A. (1982), The Cybernetic Laws of Social Progress, Pergamon Press, New York, NY.

Bai, G.H. (1997), `̀ Embryonic approach to the development of information systems'', Journal ofStrategic Information Systems, Vol. 6 No. 4, pp. 299-311.

Bai, G.H. and GroÈnlund, AÊ . (1991), `̀ Adaptive decision support systems'', Transactions of the 12thInternational Conference on Decision Support Systems, Chicago, IL, pp. 66-80.

Bai, G.H. and Lindberg, L.-AÊ . (1998), `̀ Dialectical approach to systems development'', SystemsResearch and Behavioral Science, Vol. 15 No. 1, pp. 47-54.

Bailey, K.D. (1997), `̀ The autopoiesis of social systems: assessing Luhmann's theory of self-reference'', Systems Research and Behavioral Science, Vol. 14 No. 2, pp. 83-100.

Bateson, G. (1972), Steps to an Ecology of Mind, Ballantine, New York, NY.

Bauersfeld, H. (1992), `̀ Activity theory and radical constructivism'', Cybernetics and HumanKnowing, Vol. 1 No. 2-3, pp. 15-25.

Beer, S. (1985), Diagnosing the System for Organisations, Wiley, Chichester.

Bertalanffy, L.V. (1972), `̀ The history and status of general systems theory'', in Klir, G.J. (Ed.),Trends in General System Theory, Wiley-Interscience, New York, NY, pp. 407-25.

Bjerknes, G. (1992), `̀ Dialectical reflection in information systems development'', ScandinavianJournal of Information Systems, Vol. 3, pp. 55-77.

Checkland, P. (1981), Systems Thinking, Systems Practice, John Wiley & Sons, Chichester.

Checkland, P. (1988), `̀ Images of systems and the systems image'', General Systems, Yearbook ofthe International Society for the Systems Sciences, Vol. XXXI, pp. 13-16.

Churchman, C.W. (1979), The Systems Approach and its Enemies, Basic Books, New York, NY.

Churchman, C.W. and Ackoff, R.L. (1950), `̀ Purposive behavior and cybernetics'', Social Forces,pp. 32-9.

Corning, P.A. (1994), `̀ Synergy and self-organization in the evolution of complex systems'',Systems Research, Vol. 12 No. 2, pp. 89-121.

Kybernetes28,6/7

808

Davydov, V., Zinchenko, V. and Talyzina, N. (1982), `̀ The problem of activity in the works ofA.N. LeonteÁv'', Soviet Psychology, Vol. 21, pp. 31-42.

Ehn, P. and Kyng, M. (1987), `̀ The collective resource approach to systems design'', in Bjerknes,et al. (Eds), Computer and Democracy: A Scandinavian Challenge, Avebury, Sydney,pp. 17-58.

EngestroÈm, Y. (1987), Learning by Expanding, Orienta Konsultit, Helsinki.

Foerster, H.V. (1984), Observing System, Intersystems Publications, Seaside, CA.

Forster, M.E. (1951), Art for Art's Sake, Harcourt, Brace and World, New York, NY.

Geyer, F. (1995), `̀ The challenge of sociocybernetics'', Kybernetics, Vol. 24 No. 4, pp. 6-32.

Glasersfeld, E.V. (1996), Radical Constructivism: A Way of Knowing and Learning, Falmer Pr.

GroÈnlund, AÊ . and Bai, G.H. (1993), `̀ Participatory information systems'', in Bots, P.W.G., Sol, H.G.and Traunmuller, R. (Eds), Decision Support in Public Administration, (A-26), ElsevierScience Publishers B.V., North Holland, IFIP, pp. 171-98.

Harker, S. (1993), `̀ User participation in prototyping'', Communication of The ACM, Vol. 36 No. 4,p. 77.

Holt, G.R. and Morris, A.W. (1993), `̀ Activity theory and the analysis of organizations'', HumanOrganization, Vol. 1 No. 52, pp. 97-109.

Hutchins, E. (1994), Cognition in the Wild, MIT Press, Cambridge, MA.

Kennealy, P. (1987), `̀ Talking about autopoiesis: order from noise?'', in Teubner, G. (Ed.),Autopoiesis and the Law, de Gruyter, Berlin.

Kuutti, K. (1992), `̀ Identifying potential CSCW applications by means of activity theory concepts:a case example'', CSCW 92 Proceedings, pp. 233-40.

Kuutti, K. (1996), `̀ Activity theory as a potential framework for human-computer interactionresearch'', in Nardi, B.A. (Ed.), Context and Consciousness, The MIT Press, London,pp. 17-44.

Kuutti, K. and Favorin, M. (1993), `̀ Tools for research-simulating learning'', IFIP TransactionsA-35, pp. 41-8.

Lave, J. (1988), Cognition in Practice, Cambridge University Press.

LeonteÁv, A.N. (1974), `̀ The problem of activity in psychology'', Soviet Psychology, Vol. 13 No. 2,pp. 4-33.

Luhmann, N. (1986), `̀ The autopoiesis of social systems'', in Geyer, F. and Van der Zouwen, J.(Eds), Sociocybernetics Paradoxes: Observation, Control and Evolution of Self-steeringSystems, Sage, London, pp. 172-92.

Luhmann, N. (1989), Ecological Communication, Polity Press, Cambridge.

Luhmann, N. (1995a), `̀ Why systems theory?'', Cybernetic & Human Knowing, Vol. 3 No. 2,pp. 3-10.

Luhmann, N. (1995b), Social Systems, Stanford University Press, Stanford.

Maturana, H. and Varela, F. (1972), Autopoiesis and Cognition, D. Reidel Publ. Company, Boston,MA.

Miller, J.G. (1978), Living Systems, McGraw-Hill, New York, NY.

Miller, M. (1994), `̀ Intersystemic discourse and co-ordinated dissent: a critique of Luhmann'sconcept of ecological communication'', Theory, Culture, and Society, Vol. 11, pp. 101-21.

Mingers, J. (1989), `̀ An introduction to autopoiesis: implication and applications'', SystemsPractice, Vol. 2, pp. 159-80.

Mingers, J. (1995), Self-producing Systems: Implications and Applications of Autopoiesis, PlenumPress, New York, NY.

Informationsystems

development

809

Mumford, E. (1987), `̀ Sociotechnical systems design ± evolving theory and practice'', in Bjerkneset al. (Eds), Computer and Democracy: A Scandinavian Challenge, Avebury, Sydney,pp. 58-97.

Nardi, B.A. (1996), `̀ Studying context: a comparison of activity theory, situated action model, anddistributed cognition'', in Nardi, B.A. (Ed.), Context and Consciousness, The MIT Press,London, pp. 69-102.

Qvortrup, L. (1996), `̀ How is society possible? The epistemology of social constructivism. Acomment on John R. Searle's The Construction of Social Reality'', Cybernetic & HumanKnowing, Vol. 3 No. 4, pp. 27-40.

Robert, L. (1978), The Rise of Systems Theory, John Wiley & Sons, New York, NY.

Rosenblueth, A., Wiener, N. and Bigelow, J. (1943), `̀ Behavior, purposeful and technology'',Philosophy and Science, Vol. 11, pp. 18-24.

Saarelma, O. and EngestroÈm, Y. (1993), New Work ± New Tools: Information SystemsDevelopment in the Context of Developmental Work Research, University of California, SanDiego, CA.

Schuler, D. and Namioka, A. (1993), Participatory Design: Principles and Practices, LaurenceErlbaum, Hillsdale, NJ.

Schwarz, E. (1997), `̀ Toward a holistic cybernetics: from science through epistemology to being'',Cybernetics & Human Knowing, Vol. 4 No. 1, pp. 17-49.

Searle, J. (1995), The Construction of Social Reality, The Free Press, New York, NY.

Senge, P. et al. (1996), The Fifth Discipline, Nicholas Brealey, London.

Shannon, C.E. and Weaver, W. (1964), Mathematical Theory of Communication, Urbana, IL.

Suchman, L. (1987), Plans and Situated Actions, Cambridge University Press, Cambridge.

Thyssen, O. (1995), `̀ Some basic notions in the systems theory of Niklas Luhmann'', Cyberneticsand Human Knowing, Vol. 3 No. 2, pp. 13-22.

Vygotsky, L.S. (1978), Mind and Society, Harvard University Press, Cambridge, MA.

Wertsch, J.V. (Ed.) (1981), The Concept of Activity in Soviet Psychology, Sharpe Inc., Armonk, NewYork, NY, pp. 37-71.

Wiener, N. (1948), Cybernetics, or Control and Communication in the Animal and the Machine,MIT Press, Cambridge, MA.