Practices of Stigmergy in the Building Process

Practices of Stigmergy in the Building Process

Lars Rune ChristensenAalborgUniversity Copenhagen,Department of Learning andPhilosophy, Techno-antropology ResearchGroup, A. C. Meyers Vænge 15, DK-2450 Copenhagen S, Denmark (E-mail: [email protected])

Abstract. Actors coordinate their cooperative efforts by acting on the evidence of workpreviously accomplished. Based on a field study this article introduces the concept of stigmergyto the analysis of coordinative practices in the building process. It distinguishes between practicesof stigmergy, articulation work and awareness practices. Stigmergy is understood as coordinationachieved by acting directly on the evidence of work previously accomplished by others. The articleprovides descriptions of stigmergy in the building process i.e. in design as well as constructionwork. It seeks to (1) introduce the concept of stigmergy to CSCW, (2) to delimit this concept inregard to other concepts of coordination such as articulation work and awareness and (3) to providedescriptions of practices of stigmergy in the building process and, in this capacity, to help explainhow complex large-scale cooperative work is coordinated.

Key words: coordination, building process, stigmergy, articulation work, awareness

1. Introduction1

Designing and constructing a large building is a highly complex cooperativework endeavour constituted by numerous distributed and interdependent taskscarried out by a diverse network of actors over a period of several years. For eachunique major building project a network of actors is created or configured. Thenetwork is a diverse ensemble drawn from many different professions, workingfor many different companies. Some are with the project from start to finish, forinstance the client and the main architects, while others such as the generalcontractor and the various subcontractors, including building engineers, concretespecialists, carpenters, plumbers, electricians, painters, roofing specialists,ventilation specialists etc. are associated with the project only for the durationof their allotted tasks. Specifically, the network of actors are involved ininterdependent tasks pertaining to design and construction such as ensuring thatthe design of the concrete structure of the building is aligned with the design ofthe building services2 or, in the construction phase, that the construction of thesteel skeleton of walls is aligned with the wiring of electricity—one large

1 This article draws on findings and analyses presented in articles published over the last couple of years(i.e. Christensen 2007; Christensen 2008). Furthermore, it overlaps somewhat with findings presented inChristensen (2013).2 Building services include e.g. electricity, ventilation, sanitation, heating and cooling and etc.

Computer Supported Cooperative Work (2014) 23:1–19 © Springer Science+BusinessDOI 10.1007/s10606-012-9181-3 Media Dordrecht 2012

building project has tens of thousands of tasks and sub-tasks that must becoordinated. The question is, then, how is this done?

In this study, this question has been explored through ethnographic fieldworkin an architectural office and on building sites. Initially, at the onset of thefieldwork upon which this paper is based, the focus was on how articulation work(Schmidt and Bannon 1992; Strauss 1985) was performed by the different actorsinvolved in the design and construction of the building, how key coordinativeartefacts (e.g. time schedules and Gantt charts) were instrumental in theperformance of articulation work. However, during the fieldwork another modeof coordination became apparent, one that cannot tenably be described asarticulation work, and which is presented in this article as practices of stigmergy.The notion of stigmergy refers to the phenomenon that distributed cooperativework tasks may be coordinated by virtue of actors simply acting on the physicalevidence of work previously accomplished by others. For example, in ventilationdesign the engineer acts on the models of the building previous made by thearchitects and in the process integrates his work with theirs. Another example isthe construction of interiors walls where an electrician acts on the physicalmanifestation of the work of the carpenters in order to align his work with theirs.This is stigmergy (Grassé 1959). Practices of stigmergy in the building processare the subject matter of this paper.

It is important to realise from the onset that stigmergy, as a mode ofcoordination, does not stand alone in the building process, rather it iscomplementary to other coordinative practices such as various kinds ofarticulation work. Stigmergy is almost always, at least in the building process,performed in conjunction with articulation work in meetings, emails, phone calls,and etc. (we shall return to this below).

What kind of concept stigmergy is, and how it is different fromarticulation work, or awareness practices for that matter, is one of the keyquestions addressed in this article. Carefully delimitating the concepts ofstigmergy (Grassé 1959), articulation work (Schmidt and Bannon 1992;Strauss 1985) and awareness (Heath et al. 2002) will not only clarify themerits of the concept of stigmergy, as an added windfall it will also clarifythe concept of articulation work and awareness. Especially the latter conceptof awareness may benefit from clarification or delimitation, as it is used in amultitude of ways within CSCW (Schmidt 2002a). The main contribution ofthis paper, then, is (1) introducing the concept of stigmergy to CSCW, (2)delimiting the concepts of stigmergy, articulation work, and awareness, andlast but not least (3) providing descriptions of practices of stigmergy in thebuilding process and, in this capacity, help explain how complex large-scalecooperative work is coordinated.

We will proceed by first introducing the concept of stigmergy in relation to theconcepts of articulation work and awareness. Secondly, we will account for themethods of the study and give a short introduction to the building process. Third,

2 Lars R. Christensen

we will present an analysis of practices of stigmergy in the building process.Finally, a conclusion and the perspectives of the study will be drawn up.

2. On the concept of stigmergy

One of the main research issues in CSCW is the understanding of howcooperative work is coordinated. This issue has often been cast as a question ofexploring how articulation work is practiced or how coordination is achievedthrough awareness practices.

With regard to coordination achieved through articulation work (Strauss 1985),a series of focused, in-depth field studies have been undertaken with the specificpurpose of investigating how the distributed activities of cooperative workarrangements are articulated and, in particular, how prescribed artifacts aredevised, appropriated and used for these purposes (e.g. Bardram and Bossen 2005;Carstensen and Sørensen 1996; Schmidt and Bannon 1992; Schmidt and Simone 1996;Schmidt and Wagner 2004).

Turning to the notion of awareness. The idea of awareness, at least in CSCW,originally emerged in a number of work place studies not least by Heath and Luff(1992; 1996) of Line Control Rooms on the London Underground as well as thestudies of air traffic control work by the Lancaster group (Harper andHughes 1993; Harper et al. 1989a; b). In these studies it was noted howcollaborative activity in complex organizational environments rests on theindividuals’ abilities to create awareness through bodily conduct whilstengaged in their respective activities. That is, it was described how actorsproduce awareness by rendering a feature of their conduct or a feature in theenvironment selectively available to others.

These studies of articulation work and awareness practices have providedinvaluable insights. However, we will argue that it is fruitful to complement theconcepts of articulation work and awareness with a supplementary means ofdescribing how cooperative work is coordinated and integrated. As indicatedabove, we shall attempt to do this by introducing the concept of stigmergy. On aconceptual level, we will rest the distinction between the concepts of articulationwork and stigmergy on a distinction between coordination done through supra-type activities or second order activities (articulation work) and integrationachieved by virtue of individuals acting on the material evidence of workpreviously accomplished by others (stigmergy).

Furthermore, we shall argue that the difference between stigmergy andawareness is the difference between heeding the material evidence of workpreviously accomplished by others (stigmergy) and rendering activities selec-tively available to co-located others through bodily conduct that these others mayin turn take heed of (awareness).

This formulation is probably too compressed to be really illuminating. Itpresupposes a specific view of the concepts of stigmergy, articulation work and

3Practices of Stigmergy in the Building Process

awareness that must be brought to light. Below, we will attempt to do so by firstintroducing the concept of stigmergy and subsequently comparing it to theconcepts of articulation work and awareness.

2.1. The concept of stigmergy

As mentioned above, we shall argue that we can use the concept of stigmergy tocomplement our descriptions of the coordination of cooperative work. The conceptof stigmergy was not developed in order to describe human practice (Grassé 1959).Rather, it was developed to tackle problems pertaining to the field of entomology.Grassé coined the concept during his study of termite behaviour (Theraulaz andBonabeau 1999, p.97). When looking at a group of termites, they all seem tocooperate in building nests etc., but when looking at single individuals, they seem tobe working, as if they were alone and not involved in any collective behaviour. Thisappeared to be a paradox until Grassé introduced the concept of stigmergy3. Grasséshowed that the regulation of building activities among social insects does notdepend so much on the workers themselves as on the nest structure. A stimulatingconfiguration triggers a building action of a termite worker, transforming theconfiguration into another configuration that may in turn trigger another (possiblydifferent) action performed by the same termite or any other worker in the colony(Grassé 1959). Thus, work in the termite colony is partly coordinated by virtue of theindividual termites acting on evidence of work accomplished previously bythemselves or others.

Grassé and others (e.g. Smith 1978; Wilson 1975) used a stimuli-responsemodel of action characteristic of the field of entomology in their work. We mustbe careful not to transpose this model of action to the context of human practice.CSCW is obviously dealing with humans rather than insects, so we have to leavethe stimuli-response model of action behind.

2.2. Stigmergy in a human context

What we are implying, then, is that when actors coordinate their cooperativeefforts by acting directly on the evidence of work previously accomplished byothers we may describe them as engaged in practices of stigmergy. However, it isimportant to note that before we are in a position to fully embrace this assertion,before we may be comfortable with it, we must exorcise or disassociate theconcept of stigmergy from any stimuli-response model of action that may lingerfrom the concept’s origin in the field of entomology. In this section we willsuggest that the stimuli-response mode of action is not characteristic of human

3 Grassé developed the concept of stigmergy during his study of termites. However, it has since been usedand applied to other groups of social insects (e.g. Wilson 1975), not least in the study of ants (e.g.Hölldobler and Wilson 1994). Over the last decade or so the concept of stigmergy has been introduced tothe field of telecommunication especially in connection with the development of algorithms for networktraffic, these algorithms are sometimes termed ‘ant algorithms’ (e.g. White and Pagurek 1998), and to thefield of robotics (e.g. Dorigo et al. 2000).


practice in complex work domains such as the building process, includingpractices of stigmergy.

The ability to participate in practice in the building process and work with forexample representational artifacts is grounded in skills and techniques that maybe conceived of as acquired by the individual actor not least throughapprenticeship as a ‘feel for the tasks’. The habitus, using Bourdieu's (1977)concept, of for example an accomplished building services engineer acts as a setof dispositions towards certain ways of doing and being, acting and interactingthat are in accord with or reflects the nature of the field of building design. Thesedispositions are in play in practices of stigmergy as the actor acts directly on theevidence of design work previously accomplished by others. As such it is notsimply a matter of stimuli-response.

In addition, stigmergy may be described as performed mostly in the naturalattitude. Schutz (1990, p.227) writes that, ‘this world is to our natural attitude inthe first place not an object of our thought but a field of domination’. Accordingto Schutz (1976, p.73) we normally have to act and not reflect in order to satisfythe demands of the moment. We could suggest that the accomplished actorengaged in practices of stigmergy may mostly have as a basis for their actionssomething quite different from stimuli-response, namely acquired dispositions toperceive, comprehend and act that could be interpreted as oriented towards onetask or another and performed most often in the natural attitude of the actor.

Of course this does not contradict the assertion that termites for their partengage in what may be described in terms of a sequence of stimuli-responseaction.4 We are merely suggesting that the stimuli-response mode of action is notcharacteristic of human practice, including practices of stigmergy.

With these considerations of stigmergy in mind, we may ask what kind ofconcept is stigmergy when used in a human context?

2.3. Stigmergy as a heed concept

In this section we will, based on the work of Ryle (1955), describe stigmergy as a‘heed concept’.

According to Ryle (1955, p.135), the category of ‘heed concepts’ includes:noticing, taking care, attending, minding, applying one’s mind, concentrating,putting one’s heart into something, thinking what one is doing, alertness, interest,intentness, studying, trying. Perhaps stigmergy could also be considered a heedconcept. Let us elaborate.

When a person hums as he walks, he is doing two things at once, either ofwhich he might interrupt without interrupting the other. But when we speak of aperson minding what he is doing e.g. when he is reading (or for exampledesigning) we are not saying that he is doing two things at once. He could not

4 This must be a discussion for the field of entomology, rather than the field of CSCW, and consequentlywe shall refrain from forming an opinion on the subject.


stop his reading while continuing his attention to it (Ryle 1955, p.138). In asimilar vein, we may add that he could not stop his designing while continuing tobe engaged in stigmergy (i.e. acting on physical traces of the design of others).He could of course continue to read but cease to attend (Ryle 1955, p.138), orcontinue to design but cease to engage in stigmergy. The use of pairs of wordssuch as ‘read’ and ‘attend’ or ‘design’ and ‘stigmergy’ suggests that there are twosynchronous or perhaps coupled processes going on whenever both words areproperly used, but that is not the case. This is a feature in the use of heedconcepts (Ryle 1955, p.138).

If we accept at least preliminarily the notion of stigmergy as a heed concept,we may say that performing a task engaged in stigmergy is one, rather than twocoupled activities. The point is not least that stigmergy cannot take place prior tothe performance of a task or afterwards for that matter. It is part of the task, ormore precisely, it is a characteristic way of performing the task. In this mannerstigmergy shares the quality of other heed concepts.

These considerations aside, the central question is this: does the concept ofstigmergy add anything to our ability to account for the coordination ofcooperative work? We will explore this issue by explicitly comparing the conceptof stigmergy to well-established concepts within CSCW, namely articulationwork (e.g. Gerson and Star 1986; Schmidt 1994; Schmidt and Bannon 1992;Strauss 1985; Strauss 1988; Strauss et al. 1985), and awareness (e.g. Heath andLuff 1992; Heath et al. 2002).

2.4. Stigmergy compared to articulation work

In this section we shall compare the concept of stigmergy to the concept ofarticulation work in order to determine if they are interchangeable concepts or not.

According to Strauss (1985, p.8) articulation work is a kind of supra-type workin any division of labour, done by the various actors concerning the meshing andintegration of the interdependent activities inherent to cooperative work . In asimilar vein Schmidt describes articulation work as reflexive second orderactivities (Schmidt 2002b, p.464). It is safe to say that the distinction between thearticulation work and the cooperative work articulated is an inherent feature ofthe concept of articulation work.

In comparison, using the concept of stigmergy does not entail making adistinction between the work and extra activities aimed solely at coordinating thework. That would be a contradiction in terms considering that stigmergy as wedescribed it above is a heed concept. Recall that when we speak of someoneperforming a cooperative work task engaged in stigmergy we are not saying thathe or she is doing two things at once. He or she could not stop the performance ofthe task and continue to be engaged in stigmergy. The usage of a heed conceptsuch as stigmergy, and this is meant to reiterate a point made above, has themerits of suggesting that what is described is one activity with a special character,


rather than two activities that are somehow interrelated in their execution (seeRyle 1955). In comparison, we may say that actors engaged in articulation workin relation to a set of cooperative work tasks, may stop performing the tasks andcontinue any articulation work in relation to their coordination. For example, twocarpenters engaged in distributed cooperative work tasks on a rooftop may stopworking on the roof and continue their conversation concerning how tocoordinate their interdependent efforts—in fact this may often be the case. Inthis manner articulation work may be said to stand in a supra type relationship tothe work tasks articulated, whereas stigmergy may not. That is, articulation workmay be an activity separate from the performance of the cooperative workarticulated, and in comparison stigmergy may not.

The point we are trying to make is that as stigmergy is a heed concept i.e. acharacteristic way of performing a cooperative work task (to a coordinativeeffect) it does not qualify to be described as an effort that may be said to stand ina supra-type relationship to the tasks performed. In this manner the concept ofstigmergy is not interchangeable with the concept of articulation work.

In sum, we have argued that the concept of stigmergy is notinterchangeable with the concept of articulation work, although it may of coursecomplement it.

2.5. Stigmergy compared to awareness

In this section we will compare the concept of stigmergy to the concept ofawareness in order to determine if they are interchangeable concepts or not.

The idea of awareness, at least in CSCW, originally emerged in a numberof work place studies by not least Heath and Luff (1992; 1996) of LineControl Rooms on the London Underground as well as the studies of airtraffic control work by the Lancaster group (Harper and Hughes 1993;Harper et al. 1989a; b). In these studies it was noted how collaborativeactivity in complex organizational environments rests on the individuals’abilities to create awareness through bodily conduct whilst engaged in theirrespective activities. That is, it was described how actors produce awareness byrendering a feature of their conduct or a feature in the environment selectivelyavailable to others. We shall elaborate.

According to Heath and associates (Heath et al. 2002, p.318), the ways inwhich actors produce awareness is inextricably embedded in the activities inwhich they are engaged, and the ways in which those activities unfold. Simplyput, what individuals are aware of depends upon the activities they and others areengaged in. Awareness, then, is a practical accomplishment that arises in andthrough action and activity. This feature of awareness is shared by stigmergy inthe sense that both awareness and stigmergy are inextricably part of performingthe work. However, there are also important differences between awareness andstigmergy, as we shall see.


In the course of their work performance actors may find that the activity inwhich they are engaged becomes potentially relevant for others within thedomain and yet their colleagues are seemingly involved in something else. Insuch circumstances, an actor may modulate an activity (e.g. speak louder, stare inan obvious manner, or overtly move an object about), to enable others to gainawareness of some matter at hand, without demanding that anybody shouldrespond. Heath and Luff (1992) gives a fine example of this as they describe howthe operators in a control room coordinate train traffic and movement ofpassengers on a particular line. In this setting awareness is produced through verydelicate bodily practices. That is, as Heath and associates (Heath et al. 2002,p.321) express it ‘actors may render activities selectively available’ to theircolleagues. How does this feature of awareness compare to stigmergy? We couldsuggest that stigmergy does not involve individuals rendering activitiesdeliberately or selectively available to others through bodily conduct (e.g.modulations in voice, gesture, pose, stance, gaze, glance, etc.). Stigmergy doesnot rely on this sort of selective rendering of activities in the sense that stigmergymerely refers to actors in cooperative work acting on the physical evidence ofwork previously accomplished by others to a coordinative effect.

Furthermore, unlike much production of awareness through bodily conduct,stigmergy does not rely on co-location, as we shall see now. Within CSCW,awareness is commonly associated with a particular type of workplace. In part,this association derives from the fieldwork settings of the studies that contributedto the recognition and understanding of the phenomenon in the first place. Thesesettings have certain characteristics that make awareness pertinent and have beendescribed by Suchman (1997) as ‘centres of coordination’. These include suchsettings as subway control rooms, air traffic control rooms, newsrooms, tradingrooms, and the like. According to Heath and associates (Heath et al. 2002,p.320), one of the important characteristics of such work places is thatpersonnel is co-located in the ‘same’ physical domain (through continuallyinteract with others outside that domain). As indicated, co-location enablesnot least the production of awareness through bodily conduct such asmodulations in voice, gesture, pose, stance, gaze and glance whereby actorsrender a feature in their actions or in the environment selectively available toothers (Heath et al. 2002). How does the notion of co-location relate tostigmergy? We could suggest that in contrast to awareness, co-location isirrelevant for stigmergy in the sense that for an individual acting of thephysical evidence of work previously accomplished by others the co-presence of these ‘others’ is irrelevant or unnecessary. That is, in respectto the notion of co-location awareness and stigmergy differ.

Compared to awareness, then, stigmergy does not involve rendering activitiesselectively available to co-located colleagues through bodily conduct orotherwise. That is, co-location is irrelevant in stigmergy just as there is no placeor need in stigmergy for bodily gestures. Furthermore, stigmergy is in no way


confined to specific domains such as centres of coordination in the sense thatstigmergy may transgress several settings (we shall return to this below).

We could suggest that the difference between stigmergy and awareness is (partly)the difference between heeding the material evidence of work previouslyaccomplished by others (stigmergy) and rendering activities selectively availableto co-located others through bodily conduct that these others in turn may take heed of(awareness). Note that one of the differences is related to the object that is paid heedto. Acting stigmertively involves paying heed to the physical traces of workpreviously accomplished by others, whereas producing awareness involves bodilyconduct that co-present others may take heed of subsequently. That is, in stigmergy itis the state of the material field of work that is heeded, and in awareness the heededobject is mainly bodily conduct.

Having introduced the concept of stigmergy and having established that it isnot interchangeable with the concept of awareness or the concept of articulationwork, we shall now turn to describe practices of stigmergy in the buildingprocess. However, before we do so we shall consider the methods of the studyand provide an introduction to the building process.

3. Methods

This study involves ethnographic fieldwork carried out in the course of 14 monthsin an architectural office and on building sites.

In this period, work within the domains of design and construction in relationto several building projects was studied. One of the building projects, thedevelopment of the new domicile for a publishing house, was a multi-storeybuilding in glass, steel and concrete constructed at the edge of the city ofCopenhagen’s harbour (see Figure 1). A combination of observation andinterviews was used. The fieldwork also included collecting (scanning, takingscreenshots or photographs of) artifacts used and produced by the actors engagedin the building projects.

4. The building process

As indicated above, designing and constructing a large building is a highlycomplex cooperative work endeavour constituted by numerous distributed andinterdependent tasks carried out by a diverse network of actors over a period ofseveral years. For each unique major building project a network of actors iscreated or configured. The network is a diverse ensemble from many differentprofessions, working for many different companies (see Figure 2).

Some actors such as the client and the main architects are with the project fromstart to finish, while others such as the general contractor and the varioussubcontractors, including building engineers, concrete specialists, carpenters,plumbers, electricians, painters, roofing specialists, ventilation specialists etc., are


associated with the project only for the duration of their allotted tasks.Specifically, the network of actors are involved in interdependent taskspertaining to design and construction such as for example ensuring that thedesign of the concrete structure of the building is aligned with the design ofthe building services or in the construction phase that the construction of thesteel skeleton of walls is aligned with the wiring of electricity—one largebuilding project has tens of thousands of tasks and sub-tasks that must becoordinated. The question, then, is how is this coordination performed?

At the onset of the fieldwork that this paper is based on, the focus was on howarticulation work (Schmidt and Bannon 1992; Strauss 1985) was performed bythe different actors involved in the design and construction of the building, howkey coordinative artefacts (e.g. time schedules and Gantt charts) whereinstrumental in the performance of articulation work. Indeed, a large number oftasks and subtasks in the building process are coordinated in meetings througharticulation work with coordinative artifacts (see e.g. ref). However, during thefieldwork another mode of coordination became apparent, namely, practices ofstigmergy. As stigmergy is less well described is the CSCW literature thatarticulation work (or awareness practices) we shall now proceed with adescription and analysis of practices of stigmergy in the building process.5

Figure 1. One of the building projects studied, a new domicile for a publishing house,nearing completion April 2009.

5 Please keep in mind that although we are now highlighting practices of stigmergy, articulation work ine.g. meetings, emails and phone calls is just as important for the coordination of the building process.


5. Stigmergy in the building process

In this section we shall focus on practices of stigmergy in the building process.First, we shall consider stigmergy in the design process. Secondly, we shallconsider stigmergy in construction work.

5.1. Stigmergy in the design process

Contemporary building design is for a large part done by the use of ComputerAided Design (CAD).6 That is to say, in a modern architectural office, the centralrepresentational artifact is the system of CAD models. They incorporate, as anensemble, a project’s trajectory from draft to implementation; they absorb andreflect all decisions taken and changes made, as models are gradually modifiedand rendered more detailed (Schmidt and Wagner 2004).

Initially the CAD models constitute the field of work for the ensemble involvedin the design process, including architects, engineers and other specialists.Subsequently the CAD models are used as an indispensable resource for thecraftsmen and engineers actually erecting the building. It is the former designprocess that we are concerned with here.

The division of labour within the design project is facilitated by the subdivisionof the system of CAD models into partial models or submodels. Someoneresponsible for a particular task such as for example ‘ventilation system design’may work on the submodel for this building part, while others concurrently work

Client Architects General Contractor Authorities

Initiates the process Responsible for overall design Moderates the design of the working plans

Building permit and regulations

Contracts architect and general contractor

Coordinates the design process Plans the construction process Environmental assessments

Formulates building program with architect

Formulates building program with client

Hires subcontractors and retains architect

Users Consultants Subcontractors Vendors

Contributes to requirements Specialists for statics, lightning, building services and more.

Retain the craftsmen that actually construct the building.

Provides building material and fabricate components

Design building components

Figure 2. The network of actors involved in a large building project.

6 In the last 30 years or so, computers have become a prominent tool in the design process, and we speakof Computer Aided Design (CAD) in the context of architectural work and construction. Architects,engineers and other specialists create, share and exchange CAD files in order to be able to coordinate theircooperative efforts in regard to designing a building. The files are shared by the use of a repository on aserver that is accessed with a web browser such as Firefox or Internet Explorer. Here the latest versions ofthe CAD files pertaining to the design of the building are stored and accessed by all the members of thework ensemble. The sharing of CAD files is a crucial resource for coordination of the design process, aswe shall see below.


on other submodels representing other parts of the building. There are forexample submodels for static design, façade elements, roof, ventilation, electricalsystem, sanitation, and etc. The subdivision of the representation of the buildinginto discrete yet interconnected entities (i.e. submodels) enable the actors, forlong and short periods, to proceed concurrently, with only occasional commu-nication, while still acting concertedly. That is, the division into submodelsallows for a distributed work process. The distributed models are joined into acentral model of the building (see Figure 3).

We could interject that the central CAD model is a fine example of a boundaryobject (Star 1989; Star and Griesemer 1989) in as much as it is an integratedsystem of representations that provides an infrastructure that enables distributedactors to make their individual contributions to the overall design in a distributed,incremental, and yet concerted manner. The concept of ‘boundary objects’highlights practices in which activities in distinct local settings are partiallyconcerted by ‘objects’ on the ‘boundary’ between the settings. However, thequestion remains how exactly is cohesion or coordination of cooperative workobtained through such ‘boundary objects’. We shall continue.

The work ensemble including architects, engineers, specialists etc., all makedistinct contributions in the form of CAD models covering their respective areasof expertise. The architect creates the outline of the building. On that basis theconstruction engineer creates the geometry of the concrete structure in a separateconstruction model. Subsequently, the sanitation specialist, for example, will takenotice of the model for the concrete structure and seek to align the sanitation with

Figure 3. The principle of joining of a number of specialists CAD models into a joint modelof a building.


it. In a similar manner, the electricity specialist, for example, will take notice ofthe previously created models and seek to align the wiring of the electricity withit. That is to say, the individual actor creates and changes the form of a CADmodel, not for the purpose of conveying a message, but simply as a part ofconstructing a building; another actor takes notice and acts upon it. In thismanner, components of the building such as concrete structure, sanitation,ventilation and electricity are brought into alignment with the overall design.

Of course actors meet on a regular basis in meetings or over the telephonein order to discuss the ‘big picture’, the overall progress of the project, whodoes what when etc. However, when it comes to the coordination of the‘small things’ the multitude of details involved in designing a building, thisis done not only in meetings or over the phone (there is no time), but alsodirectly with the CAD models exchanged by the use of the repository. Thatis, on the detail level of the CAD models, work is primarily coordinatedthrough the work itself, i.e. through practices of stigmergy, rather than through e.g.meetings about it. Let us look at a concrete instance of this. Take for exampleventilation design.

When creating the model for the ventilation systems the engineer at every turnhas to pay heed not least to the architects’ model of the building in order toensure that the systems is a ‘fit’, that it is in accord with the structural elements ofthe building as well as its layout. Perhaps we could take a closer look at how thisunfolds.

Working with the CAD application the engineer juxtaposes what is to becomethe model of the ventilation system with the architects’ model of the buildingalready made (see Figure 4).

Panning and rotating the architects’ model of the building the engineer is ableto familiarize himself with the space constrains and possibilities that the model

Figure 4. Left, CAD model in three dimensions of ventilation system seen in conjunctionwith parts of the model for the buildings general architectures. Right, CAD model of theventilation system shown in connection with architectural model depicting the spatialdivision of the building.


affords the engineer’s design of ventilation system. The space allotted, forexample, for the air ducts in the building layout is very limited due to the highcost of building construction, so space constraints is a concern. In extreme casesthe engineer has to take such issues up at a meeting with the architects in an effortto have more space allotted for the ventilation system. However, meetings aretime consuming and as a matter of routine the engineer will make every effort tomake the system integrate with the architects’ model of the building withouthaving to meet, talk or otherwise correspond. That is, as a matter of routine theengineer acts on the representation previously made by the architects and in theprocess integrate his own task with theirs. This may involve for examplepositioning ventilation equipment such as air ducts, air dispensers and ductsilencers in conjunction to the space made for building services in the architects’model.

Part of the task of ventilation design, then, is to act on representations createdpreviously by someone else. As mentioned, taking heed of the architects’ modelin his own process of design, the engineer may integrate his own work with thearchitects’ without resorting constantly to meetings, phone calls or emails. That isto say, in addition to relying on meetings, plans and schedules, actors coordinatetheir cooperative efforts through practices of stigmergy.

5.2. Stigmergy in construction work

In construction work, as in design, interdependent tasks may be integratedthrough practices of stigmergy. As a case in point we shall consider theintegration of cooperative work tasks pertaining to the construction of interiorwalls.

During the construction of an office building a large number of partition wallsare constructed. Partition walls are what divide the building into for instance unitsof office space. The construction of these walls is a cooperative work processinvolving a number of different trades such as carpenters, electricians and painters(see Figure 5).

The initial parts of a partition wall is constructed by a carpenter in the form of aframe made of light weight steel grinders fitted with plasterboards on the oneside. At a later point in time, another actor, namely, an electrician will arrive andtake notice of the work carried out and seek to align the wiring of the electricalcircuits with it. That is, the electrician will drill holes in the plasterboard toaccommodate the electrical instillations and he or she will pull electrical cablesthrough little holes in the vertical steel grinders of the frame and connect them tothe electrical system. When the electrician is done and has left the scene, thecarpenter returns to close the wall i.e. clad the second side of the wall inplasterboards in accordance with the previous work done. That is, the carpentermust take notice of the work previously performed by himself and the electricianas he seeks to put up the second round of plasterboards. Subsequently, the painter


shows up to paint what the others have erected. At this point the wall in-the-making will have been worked on to consist of a steel frame, plasterboards on thefirst side, electrical instillations inside, and plasterboards on the second side.Finding the wall in this state the painter paints the wall with several coats ofpaint.

In this manner the work ensemble including carpenter, electrician and painterall make distinct contributions towards the construction of the wall in accordancewith their respective areas of expertise. We could say that the individual actorcreates and changes the form of the wall in-the-making, not for the purpose ofconveying a message, but simply as part of performing their individually allottedtasks, in turn another actor pays heed to and acts upon the material evidence ofthe work of others. This is partly how the cooperative work tasks pertaining to theconstruction of partition walls are integrated.

To allow for full appreciation of the importance of stigmergy in buildingconstruction work, it is prudent to note that no formal construct (e.g. architecturalplan or Gantt charts) exhaustively stipulates a concrete practice. Plans areunderspecified with respect to that which is represented (Suchman 1987), andarchitectural plans and Gantt charts for the construction of for example partition

Figure 5. Interior wall in-the-making. The first and second frame shows the result of thecarpenter’s initial efforts. The third frame, including insert, shows the work of the electricianin progress. Finally, the fourth frame depicts the closed wall ready for painting.


walls are no exception. The actors have to ‘fill in the blanks’ for themselves, so tospeak, and acting on the evidence of work previously accomplished by othersmay be said to be one way of doing this.

Furthermore, architectural plans for specific building parts such as walls are notassembly manuals like those that come with for example IKEA furniture, ratherarchitectural plans represent mainly how parts of the building it are supposed tolook in the final state. Consequently, the assembly of for example partition wallsis not covered in architectural plans.

In addition, the pace of contemporary construction work is such that as soon asone actor (e.g. carpenter) has completed a task, time does not allow for muchstanding around and talking to the next actor (e.g. electrician) even though theirtasks are interdependent and there are numerous details that need to be integrated.Of course articulation work through talk on the building site may contribute tothe integration of cooperative construction work tasks, but so may acting on thematerial evidence of work previously accomplished i.e. practices of stigmergy.

The point is that in addition to various kinds of articulation work with andwithout coordinative artifacts, construction work is coordinated through practicesof stigmergy.

In sum, in construction as in design, cooperative work tasks are (partly) integratedby virtue of practices of stigmergy. Above we have indicated that articulation work,stigmergy and awareness may act in concert as distinct yet interconnected modes ofcoordination in cooperative work. We will elaborate on this proposition next.

6. The interconnectedness of articulation work, stigmergy and awareness

Generally speaking, articulation work may be performed prior, in parallel to, andafter the performance of the tasks articulated. Whereas, for example stigmergymay not be performed prior, in parallel to, and after the performance of the tasksin that stigmergy is a characteristic manner in which cooperative work isperformed to a coordinative effect, rather than a separate activity.7 For example,in the building process the cooperative work task are initially constituted anddistributed through articulation work in for example meetings with coordinativeartifacts such as Gantt charts and described in various time schedules andplanning documents. Subsequently, the tasks may be integrated on the concretematerial level through practices of stigmergy. We have seen examples of thisabove in both design and construction work. On par with stigmergy, awarenesspractices may also play their part in regard to the integration of cooperative worktasks in the concrete, that is, as actors rendering activities selectively available toco-located others through bodily conduct that these others in turn may take heedof. Subsequently, to their concrete performance and integration through practicesof stigmergy and awareness, the performance of the tasks may be evaluated, and

7 As mentioned above, stigmergy (and possibly awareness) are heed concepts.


rearranged through articulation work in, for example, site meetings where timeschedules and other planning documents are adjusted to reflect the actual progressof the tasks on the building site. In this manner articulation work, stigmergy andawareness may act in concert as distinct yet interconnected modes of coordinationin the building process.

More broadly speaking, how articulation work, stigmergy and awarenessinterconnect in cooperative work is an empirical question, that is, something thatvaries from case to case, from setting to setting. Every work setting, everydivision of work, every network of actors and artifacts, frame coordinativepractices in different ways. We may say that in every instance of cooperativework there are (1) tasks that are more or less taken for granted and requires littleor no coordination based on the predefined roles of professionalism, and (2)situations of interdependencies that require coordination via for examplearticulation work, stigmergy and awareness practices.

7. Conclusion and perspectives

In this study, the concept of stigmergy has been introduced, compared to otherconcepts of coordination i.e. articulation work and awareness, and used todescribe the coordination of design and construction work in the buildingprocess. The main contribution of the study may be the concept of stigmergyitself as delimited from the concepts of articulation work and awareness.

Potentially, the three concepts of stigmergy, articulation work and awarenesscan amount to a trinity in the CSCW toolbox for the description and analysis ofthe coordination of cooperative work—each concept pertaining to a unique modeof coordination of cooperative work. As these modes of coordination rarely standalone in practice, an interesting question for further empirical research is exactlyhow stigmergy, articulation work and awareness practices complement each otheras distinct yet interconnected modes of coordination in various cooperative workpractices in for example health care, manufacturing, and the service industries?

Future ethnographic studies of coordinative practices in cooperative work mayattempt to account for the interconnectedness of these various modes of coordinationwith a new heightened sensibility towards the importance of coordination throughthe material field of work. A viable fieldwork strategy may be to track the statechanges in the material field of work at the boundaries and intersections of thevarious cooperative actors in accord with the division of work.

Acknowledgements

I am indebted to PLH Arkitekter A/S and to BPLUS Arkitekter & Designere forletting me take up so much of their time. In addition, I am grateful to the staff ofPHIL & Søn A/S and MT Højgaard A/S. Moreover, Kjeld Schmidt’s opinionsand careful reading of the text are greatly appreciated - needless to say, theresponsibility for any errors is mine alone. Furthermore, the comments and


opinions of Kristian Kreiner are also greatly appreciated. Also, thanks to MortenVisby for proof reading. Lastly, the comments of the anonymous reviewers thathelped improve this paper are greatly appreciated.

References

Bardram, J.E., and C.A. Bossen (2005). A web of coordinative artifacts: collaborative work at ahospital ward. In GROUP 2005: Proceedings of the 2005 International Conference onSupporting Group Work, USA, 6–9 November 2005, Sanibel Island, Florida. New York: ACMPress, pp. 168–176.

Bourdieu, P. (1977). Outline of a Theory of Practice. New York: Cambridge University Press.Carstensen, P., and C. Sørensen (1996). From the social to the systematic: Mechanisms supporting

coordination in design. Computer Supported Cooperative Work (CSCW). The Journal ofCollaborative Computing, vol. 5, no. 4, pp. 387–413.

Christensen, L.R. (2007). Practices of stigmergy in architectural work. GROUP 2007: Proceedingsof the International Conference on Supporting group work, 4–7 November 2007, Sanibel Island,Florida. New York: ACM Press, pp. 11–20.

Christensen, L.R. (2008). The logic of practices of stigmergy: representational artifacts inarchitectural design. CSCW 2008: Proceedings of the ACM Conference on Computer SupportedCooperative Work, 8–12 November 2008, San Diego, USA. New York: ACM Press, pp. 559–568.

Christensen, L.R. (2013). Coordinative Practices in the Building Process: An EthnographicPerspective. London: Springer.

Dorigo, M., E. Bonabeau, and G. Theraulaz (2000). Ant algorithms and stigmergy. FutureGeneration Computer Systems, vol. 16, no. 8, pp. 851–871.

Gerson, E.M., and S.L. Star (1986). Analyzing due process in the workplace. ACM Transactions onOffice Information Systems, vol. 4, no. 3, pp. 257–270.

Grassé, P.P. (1959). La reconstruction du nid et les coordinations inter-individuelles chezBellicositermes natalensis et Cubitermes sp. La théorie de la stigmergie: Essai d'interprétationdes termites. constructeurs. Insect Sociaux, vol. 6, pp. 41–83.

Harper, R.H.R., and J.A. Hughes (1993). What a f—ing system! send 'em all to the same place andthen expect us to stop them hitting: Making technology work in air traffic control. In G. Button(ed.): Technology in Working Order: Studies of Work, Interaction, and Technology. Routledge,pp. 127–144.

Harper, R.H.R., J.A. Hughes, and D.Z. Shapiro (1989a). Working in harmony: An examination ofcomputer technology in air traffic control. Proceedings of the First European Conference onComputer Supported Cooperative Work, Gatwick, London, 13–15 September, 1989, London:Computer Sciences Company, pp. 73–86.

Harper, R.R., J.A. Hughes, and D.Z. Shapiro (1989b). The Functionality of Flight Strips in ATCWork: The Report for the Civil Aviation Authority. Lancaster, UK: Lancaster SociotechnicsGroup, Department of Sociology, Lancaster University.

Heath, C., and P. Luff (1992). Collaboration and control: Crisis mangement and multimediatechnology in London Underground control rooms. Computer Supported Cooperative Work(CSCW) An International Journal, vol. 1, no. 1–2, pp. 69–94.

Heath, C., M.S. Svensson, J. Hindmarsh, P. Luff, and D. Lehn (2002). Configuring awarness.Computer Supported Cooperative Work (CSCW): The Journal of Collaborative Computing, vol.11, no. 3–4, pp. 317–347.


Heath, C.C., and P. Luff (1996). Convergent activities: Line control and passenger information onthe London Underground. In Y. Engeström and D. Middleton (eds.): Cognition andCommunication at Work. Cambridge: Cambridge University Press, pp. 96–129.

Hölldobler, B., and E.O.Wilson (1994). Journey to the Ants.Cambridge, MA.: Harvard University Press.Ryle, G. (1955). The Concept of Mind. London: Hutchinson & Co.Schmidt, K. (1994). Modes and Mechanisms of Interaction in Cooperative Work. Roskilde,

Denmark: Risø National Laboratory.Schmidt, K. (2002a). The problem with ‘awareness’: Introductory remarks on ‘Awareness in

CSCW’. Computer Supported Cooperative Work (CSCW): The Journal of CollaborativeComputing, vol. 11, no. 3–4, pp. 285–298.

Schmidt, K. (2002b). Remarks on the complexity of cooperative work. P. Salembier and T.H.Benchekroun (eds.): Cooperation and Complexity in Sociotechnical Systems, [special issue of]Revue des sciences et technologies de l’information, [série] Revue d’intelligence artificielle(RSTI-RAI), vol. 16, no. 4–5, pp. 443–483.

Schmidt, K., and L. Bannon (1992). Taking CSCWSeriously: Supporting ArticulationWork.ComputerSupported Cooperative Work (CSCW). An International Journal. vol. 1, no. 1–2, pp. 7–40.

Schmidt, K., and C. Simone (1996). Coordination mechanisms: Towards a conceptual foundation ofCSCW systems design. Computer Supported Cooperative Work (CSCW): The Journal ofCollaborative Computing, vol. 5, no. 2–3, pp. 155–200.

Schmidt, K., and I. Wagner (2004). Ordering systems: Coordinative practices and artifacts inarchitectural design and planning. Computer Supported Cooperative Work (CSCW): The Journalof Collaborative Computing, vol.13, no. 5–6, pp. 349–408.

Schutz, A. (1976). Collected Papers II: Studies in Social Theory. The Hauge: Springer.Schutz, A. (1990). Collected Papers I: The Problem of Social Reality. Amsterdam: Kluwer

Academic Publishers.Smith, A.P. (1978). An investigation of the mechanisms underlying nest construction in the mud

wasp Paralastor sp. (Hymenoptera: Eumenidae). Animal Behaviour, vol. 26, pp. 232–240.Star, S.L. (1989). The structure of ill-structured solutions: Boundary objects and heterogeneous

distributed problem solving. In L. Gasser and M. Huhns (eds.): Distributed ArtificialIntelligence. London: Pitman, pp. 37–54.

Star, S.L., and J.R. Griesemer (1989). Institutional ecology, “translations” and boundary objects:Amateurs and professionals in Berkeley’s Museum of Vertebrate Zoology, 1907–39. SocialStudies of Science, vol. 19, pp. 387–420.

Strauss, A. (1985). Work and the division of labor. The Sociological Quarterly, vol. 26, no. 1, pp. 1–19.Strauss, A. (1988). The articulation of project work: An Organizational Process. The Sociological

Quarterly, vol. 29, no. 2, pp. 163–178.Strauss, A., S. Fagerhaugh, B. Suczek, and C. Wiener (1985). Social Organization of Medical

Work. Chicago and London: University of Chicago Press.Suchman, L. (1987). Plans and Situated Actions: The Problem of Human-Machine Communication.

Cambridge: Cambridge University Press.Suchman, L. (1997). Centers of coordination: A case and some themes. In L. B. Resnick, et al. (eds.):

Discourse, Tools, and Reasoning: Essays on Situated Cognition. Berlín-Heidelberg: Springer, pp. 41–62.Theraulaz, G., and E. Bonabeau (1999). A brief history of stigmergy. Artificial Life, vol. 5, pp. 97–116.White, T., and B. Pagurek (1998). Towards multi-swarm problem solving in networks. ICMAS'98:

Proceedings of Third International Conference on Multi-Agent Systems, pp. 333–340.Wilson, E.O. (1975). Sociobiology: The New Synthesis. Cambridge, MA: Belknap Press.


Practices of Stigmergy in the Building Process

Documents

Transcript of Practices of Stigmergy in the Building Process