TY - JOUR
AU - Béguin,, Pascal
AB - Abstract In the instrument-mediated activity approach, it is argued that artifacts are far from being finished when the final technical specifications leave the research and design office. It is up to the user, in and through its use, to turn the artifact into an instrument. If the design process continues as the artifact is being used in real situations, then how can we conceptualize the design process? This article proposes an understanding of project management as a mutual learning process that takes place during exchanges of activity. After discussing how such activity exchanges can be extended to mutual learning among users and designers, a concrete case is presented to illustrate the approach: designing an alarm system to guard against chemical runaways in chemical plants. 1 Introduction For some time now, French-speaking ergonomists have been devoting considerable effort to the issue of inventiveness as it is manifested in the activity of users confronted with a technique. As early as 1955, Ombredane and Faverge argued that “while certain significant aspects of tasks are planned and incorporated into the instruction process, there are an indefinite number of others that are unforeseen and are subject to discovery by the worker” (p. 15, our translation). The inventiveness of subjects as users in real situations has been stressed ever since, so much so that one of the premises of this trend in ergonomics holds that the inventiveness of operators is the key feature of work activity (Montmollin, 1992). This position was summarized quite clearly by Weill-Fassina et al. (1993) who wrote, “Actions can never be reduced to the effecting of responses to received stimuli in a more or less passive way, to motor acts, to executed procedures. They are a manifestation, in the facts, of the processes through which operators explore, interpret, use, and transform their technical, social, and cultural environment” (p. 21, our translation). The ‘instrument-mediated activity’ approach (Béguin and Rabardel, 2000; Rabardel and Bourmaud, 2003) is part of this tradition and contributes to its ongoing renewal. The design of an instrument, which cannot be confused with an artifact, is far from being finished when the final technical specifications leave the research and design office. It is up to the user, in and through its use, to turn the artifact into an instrument. In the instrument-mediated activity approach, an instrument is seen as a composite entity made up of ‘the artifact’, in its structural and formal aspects, and the subject's social and private schemes. Instrumental genesis accounts for a rarely-modelled but nonetheless strategic dimension of the user's activity: the constructive dimension: instrumental genesis (its psychological and material components), but also competencies and conceptualization, dependent upon collective work and work groups. As a general rule, the constructive dimension includes the development, by the subjects, of the conditions and resources of their productive activity. Productive activity is rooted in the subject's intentional commitment to pursuing an objective and attaining goals. A dialectic relationship links the productive and constructive dimensions of the activity: failure or resistance at the productive level will lead to new constructions (Rabardel and Béguin, 2003). How can designer–user interaction be organized in order to articulate the inventiveness of both parties? The fact that users do not utilize the system as might be expected, modifying it momentarily or durably, has of course been pointed out by many authors (Greenbaum and Kyng, 1991; Henderson, 1991; Bannon, 1991; Bannon and Bødker, 1991; Robinson, 1993; Béguin, 1993). However, these processes are the object of various interpretations. One possible interpretation is that an insufficiently elaborated design produces artifacts unsuited to users' needs. This results in many ‘gaps’ (Thomas and Kellogg, 1989): ecological gaps, by omission of factors in the real world, ‘problem formulation gaps’, which have to do with how the user realizes that a tool is appropriate for a task, or ‘work context gaps’ pertaining to the social setting, the culture of the work place, and so on. Design continues in usage because designers did not sufficiently consider and anticipate the users' requirements or practices. A second interpretation, well-established and supported by the French-speaking school of ergonomics, is that in work settings, users encounter oppositions linked to fluctuations of their own state and to ‘industrial variability’ (e.g. systematic deregulation of tools, instability of the matter to be transformed, etc.) (Daniellou, 1992). Suchman (1987) used the term ‘situated action’ to generalize this aspect. Whatever the effort put into planning, performance of the action cannot be the mere execution of a plan that fully anticipates the action. One must adjust to the circumstances and address situation-related contingencies, for instance, by acting at the right time and by seizing favorable opportunities. A third interpretation sees the situated nature of activity as not only coming from the dynamic variability of circumstances, as postulated by the proponents of situated action, but also as the result of the constructive activity of subjects. Driving a vehicle as a job, for example, is different from driving for personal reasons, regardless of the circumstances and the contingencies of the situation. This stems from the fact that constructive activity involves both instruments and the subject's resources as a whole (namely his/her scheme and the associated conceptualizations, objects, and activity criteria; for a discussion on this point, Rabardel and Béguin, 2003). In this context, design-in-use is not rooted in the extrinsic dimensions of the subject's activity (an insufficiently elaborated design process or the dynamic variability of circumstances), but in its intrinsic dimensions, particularly its development. We believe that instrumental geneses has both an extrinsic origin (the design process itself, or ontological situation characteristics) and an intrinsic origin (the subject's activity, which must be sought in activity development and constructive activities). However, the different interpretations of design-in-use presented above lead to very different understandings of the relationships between use and design. In the first approach, designers try to fully anticipate action. By contrast, a situated approach to actions leads to ‘design for unanticipated use’ (Robinson, 1993), or ‘specifying boundaries on action’, allowing workers to finish the design in a ‘space of functional action possibilities’ (Vicente, 1999). These two approaches develop along diverging pathways. In the first one, the user's activity is made to be a resource for the designer's activity. In the situated approach, it is the outcome of the designer's activity that is a resource for the user. But they both omit the constructive activity. No matter how the subject's activity is anticipated before the design process, user will develop the resources of his/her own productive activity. Similarly, with continuing design-in-use, subjects encounter new needs, which leads to further design activity (sometimes through instrumental genesis; Béguin,1993; Galinier, 1997; Duvenci-Langa, 1997). In a developmental perspective on activity, design seems to be without any real beginning or end: it is more a cyclical and never-ending process (Henderson, 1991) design and use mutually shape one another. The result of one person's activity (the designer or the user) constitutes a resource for the activity of another. This is very different from the traditional engineering approach, where design is perceived as a change of state during which a problem is solved. The organization of mutual activities between user and designer are as strategic as the problem to be solved. Then, how can one conceive of the design process? And how can one interrelate and organize the alternation between instrumental genesis and constructive activity on the one hand and formal acts of design on the other? In the first part of this article we will propose to understand design as a mutual learning process, mediated by ‘intermediary productions’ and framed by power relations. In the second part, we will examine how this understanding of design can be applied to designer–user interaction. We will propose an approach called the ‘artifact-based learning process’ in which a key role is given to instrumental genesis. In the third part, this approach will be illustrated by means of a concrete case: designing an alarm to signal chemical runaways in plants. 2 Design as a mutual learning process between designers Taking action in the design process presupposes having a model to guide one's action: What is design? From a phenomenal viewpoint, design is both an individual and a collective process (Falzon et al., 1996). Two principles lay the groundwork for this individual and collective dimension (Béguin, 1994; Hatchuel, 1994). The first is an actor differentiation principle,1 1 It is a ‘horizontal differentiation’, opposed to ‘vertical differentiation’ which is based on a hierarchy. Horizontal differentiation is more a question of knowledge classes. made necessary by the fact that the tasks to accomplish are complex. No matter what object is being designed (a factory, a work of art, a vehicle, etc.), it is too complex for a single person to be able to represent all of its inherent problems and possess all of the skills to solve them. Differentiation serves to reduce this complexity by distributing tasks among the members of a team. The second principle is interdependence; no specialist alone holds the key to a successful product or a high-performance industrial system. These two principles are potentially contradictory. Differentiation reduces complexity at the task level; but interdependence increases it at the collective level: different individuals perceive the object being designed from different or even divergent ‘points of view’, which must be articulated. What, then, are the different forms the design process can take on? 2.1 Integration: establishing coherence and mutual learning Interdependence is manifested directly in the object being designed. Any modification or improvement made to one of an artifact's components may have an impact on the other components of the technical system. For instance, increasing the performance of an airplane engine may cause the builders to change and redesign the shape of the wings and the fuselage structures, even though the latter had been considered fixed (Vincenti, 1990). From the standpoint of the object, the integration process is the process that links its constituents into a coherent system. How can the integration process be characterized in terms of activity? The first characteristic is that each designer, by his/her own activity, learns. No action can be the pure and simple implementation of prior knowledge. Every action reconstructs the knowledge it needs. This is a vital point in the genetic epistemology proposed by Piaget (1970), who stated that knowing consists in acting upon the real and transforming it (in appearance or in reality) in order to understand it. In design, this process is even more noticeable. The famous metaphor of the ‘reflexive conversation with the situation’ proposed by Schön (1983) is relevant here. According to this author, the design process can be described as an open-ended heuristic during which the designer, striving to reach a goal, projects ideas and knowledge. But then the situation ‘replies’, ‘surprises’ the designer by presenting unexpected resistances. These serve as a learning basis for the designer. However, due to the principle of interdependence and the need for integration at the object level, the result of this learning process is only temporary. For the other designers in the project, the new knowledge will be incorporated into the definition of the problem to solve. Learning has taken place, so now the initial knowledge can be validated, or perhaps refuted, in which case a new exploration process begins (Béguin, 1994; Hatchuel, 1994). In other words, the ‘reflexive conversation with the situation’ often takes place in a ‘dialogue’ with the object of design. But the object is not alone in talking back; the other actors ‘reply’ and ‘surprise’ too (Granath, 1991). The result of one designer's activity is at best a ‘working hypothesis’ that will be validated, refuted, or set in motion based on actions and learning performed by another actor involved in the design process. The integration process can be described as a mutual learning process, rooted in ‘activity exchanges’. During this type of exchange, and through the ‘working hypothesis’, the activity of some designers sets the activity of others in motion. We will come back later to the terms ‘activity exchange’ and ‘working hypothesis’. 2.2 Learning mediated by ‘intermediary productions’ Activity exchanges are not direct. Many kinds of mediation punctuate mutual learning. Semiotic mediation occurs when a symbolic language is used to generate graphic descriptions such as maps and diagrams. But mediation also comes in other forms, such as scale models, mock-up, prototypes, etc. (Vinck, 2001). Let us call these productions ‘intermediaries’ insofar as they link the individual and collective dimensions of design. At the individual level, intermediaries are artifacts, that can act as instruments for the designer's activity in the sense of the instrument-mediated activity approach. When associated with the designer's scheme, an ‘intermediary production’ acts as a mediator between the designer and the object being designed. Béguin and Rabardel (2000) provide an example on the design of an electrical device. Their analysis showed that the graphic artifact was associated with a cyclically-organized ‘design scheme’. The cycle contained periods of graphic production, during which the designer project a hypothesis, in alternation with periods during which the designer used a graphic space to simulate the operation of the device just designed, learning something new, and then made corrections in order to bring the specifications closer to the goals pursued. Others authors have noted that designers proceed by representing a design idea in some medium, reflecting on the representation, and then modifying it (Herbert, 1993; Schön, 1983). Such productions are also intermediaries because they play a role in the context of exchange between actors (Erickson, 1995). The exchanges may be diachronic or synchronic. In diachronic exchanges, a graphic production, for example, can appear as a ‘graphic message’ (Waern and Waern, 1993) or as a mediator between different types of participants (Bødker, 1998). In synchronic interactions, intermediary productions supply the basis for joint reflection involving several people, at the same time as they influence them (Hammond et al., 2001). In both cases, and despite their differences, intermediary productions capture working hypothesis at the same time as they are the vector for mutual learning. 2.3 Mutual learning and conflict During mutual learning, disagreements or dilemmas may appear between designers. They play an important role and constitute a driving force of the design process. But in fact, among designers, viewpoints often diverge considerably, as do the goals they pursue and their criteria for success. Highly interdependent situations, as in design, exacerbate opposing goals, and reaching the goals of some may run counter to reaching those of others (Easterbrook et al., 1993). In this context, it is not surprising that the design process appears conflict-ridden (Moran and Carroll, 1996). It is necessary to make a distinction between ‘disagreements’ and ‘conflicts’. In accordance with the differentiation principle (see below), designers can legitimately disagree. By modifying the characteristics of the object currently being designed, or by changing the criteria for attaining the goal, mutual learning is one way of dealing with disagreement. But other routes so may also be taken, such as conflict (e.g. the exercise of authority or even the exclusion of certain actors whose goals appear too contradictory). In this case, the characteristics of the object under design become less central, leaving the actors in a situation of face-to-face contention where the difficulties are ascribed to others. Design is achieved within a community, where divergence legitimately surfaces. In certain cases, the divergence will be handled in the mutual-learning mode, while in others, conflict will ensue. 3 Design as a mutual learning process between users and designers A key reason why an approach in terms of mutual learning is essential, in our mind, is that it allows the inventiveness of both designers and users to be grasped within the same framework. However, achieving mutual learning between users and designers is a difficult task (Bratteteig, 1997) that must be treated in all of its specificity, especially since designers may not draw any benefits from the constructive activities of users (the same holds true, or so it seems, in the general case). How and on what basis should the designers learn from the users? Mutual learning between users and designers is an important part of participatory design, where the emphasis among researchers practicing this type of design “has shifted from issues of power and industrial democracy to making the system design process more cooperative” (Kraft and Bansler, 1992). It leads to ‘cooperative design’ or ‘user participation’ (Bjerknes and Bratteteig, 1995). In this context, mutual learning between users and designers paves the road for more participatory design (Bjerknes and Bratteteig, 1987, 1995; Bratteteig, 1997; Mogensen and Trigg, 1992; Trigg and Bødker, 1991; Bødker and Grønboek, 1996; Bødker et al., 1987). In a recent discussion of the ‘Florence project’, Bratteteig (1997) clearly demonstrated the heuristic value of an approach in terms of learning. This author argues that “the objective of mutual learning is to enable the future users to participate in design, i.e. to do parts of the design” (p. 11). The computer programmers taught the users about computers in such a way that the mutual learning provided them with sufficient knowledge about computers to suggest a design. Another idea put forward by Bratteteig is that the participatory process is a learning process in itself. Accordingly, the users who did participate in the mutual learning were in charge of instructing their colleagues who did not. However, we can see that clearly distinct processes exist here. In their ‘cooperative prototyping’ approach, Bødker and Grønboek (1996) are more explicit about mutual learning, and closer to our view. Bødker et al. (1987) stress that joint action between users and designers is often seen as creating a new shared activity that is different from that of the designers and that of the users. The ‘cooperative design’ approach (Kyng, 1995) aims to establish a design process wherein both users and designers are participating actively and creatively, based on their differing qualifications. In ‘cooperative prototyping’, the multitude of activities (instead of a shared one) implemented around a prototype constitute a place where the future artifact and its use will be developed. We share this viewpoint: design is achieved by separate actors, engaged in an interdependent process, during which mutual learning is achieved on the basis of the differing qualifications and expertise of the actors. However, our approach, which we call ‘artifact-based learning’, differs in several respects from cooperative prototyping. 3.1 The artifact-based learning approach To present the artifact-based learning approach, let us first come back to the term ‘activity exchange’ used above: exchanges of activity constitutes the backdrop for mutual learning between designers, which should be extended to exchanges between users and designers. 3.1.1 Activity exchanges: a dialogical process Many authors have insisted on the merits of viewing design processes as communication processes. References to this idea are sometimes metaphoric, as in Schön's idea of a ‘dialogue with the situation’, and sometimes more formal (Winograd and Flores, 1987; Granath, 1991; Erickson, 1995; Brown and Duguid, 1994). Day (1995) proposes a model that can be applied to describe how developers of tools communicate process preferences to tool users via the human–computer interface. We share this viewpoint: exchanges between actors are carried out through the mediation of intermediary productions. However, the language plays a role that cannot be undermined, we think it is more accurate to describe design in terms of dialogical processes. Introducing Bakhtin's ideas about ‘dialogicality’ as a way to expand the Vygotskian approach, Wertsch (1998) (who considers language to be a cultural tool, and speech a form of mediator) stresses the importance of appropriation of utterances during production. The term ‘appropriation’, borrowed from Bakhtin (prisvoenie) refers to a process where someone takes something that belongs to others, and makes it his/her own. As Wertsch outlined, producing an utterance inherently involves appropriating the words of others, and making them, at least in part, one's own. “Because words are half-ours and half-someone else's […], one is invited to take the internal word as a ‘thinking device’, or as a starting point for a response that may incorporate and change the form or meaning of what was originally said” (Wertsch, 1998, p. 67). It is contended here that the dialogical process described by Wertsch can be generalized to all situations where intermediary productions are used. Our goal is not to argue that there is no difference between artifacts and words. But it seems that language should be regarded as one of the possible dialogical forms. In the design process, exchanges between actors are carried out through the mediation of temporary outcome, not just through words, whose use in action reshapes, enriches, or shifts the characteristics of the object currently being designed. It is in this sense that the term ‘activity exchange’ is employed: to emphasize this dialogical process, during which the result of a designer's activity is brought back into play in the activity of another person, through intermediary productions. Activity exchanges constitute the basic foundation upon which ‘mutual learning’ between users and designers can be built. The consideration that it is too restrictive to limit activity exchanges to language exchanges allows one to grant a status to a critical point in the instrument-mediated approach: instrumental genesis. Instrumental genesis is a way of conceptualizing appropriation processes (Béguin and Rabardel, 2000; Rabardel and Béguin, 2003). The ‘someone else's half’ (the artifact) is associated with ‘one's own half’ (the scheme) to bring about instrumental genesis.2 2 By postulating a dialogical structure in the Bakhtinian sense, as we have just done, we are letting it be understood that instrumental genesis can have exchanges with others as its driving force. Stated in other terms, it is the multivoicedness of one activity that leads to instrumental genesis during a process of appropriation. This aspect of the development of the instrument-mediated activity approach cannot be considered here. How does instrumental genesis fit into activity exchanges? 3.1.2 Working hypothesis and instrumental genesis We underlined above that an intermediary production affords a working hypothesis.3 3 We are only interested here in the role of artifact, even though we think that an object-based approach is necessary. As objects, intermediary productions are extremely diverse: videos of users, snapshots of usage contexts, transcripts of user interviews, scenarios, prototypes, mock-ups, and so on. These different objects offer different advantages. Briefly, in our approach, a prototype is indispensable if appropriation is to occur. Extended to a dialogical approach, the term ‘working hypothesis’ not only suggests that the appropriation of the designers' output in the users' activity is such that it reveals implicit hypotheses, but what is more, that it is itself a producer of novelty. In what way is an artifact a hypothesis, and what is significant for the designer in this novelty? Leont'ev (1978) stressed that operator, through his/her activity, must appropriate the ‘motor operations’ and the ‘word operations’ that are ‘crystallized’ in the artifact (Bannon and Bødker, 1991). In a study analyzing the designer's activity, Nicolas found supporting arguments for this position (Nicolas, 2000). By means of functional analysis sessions, designers of computer-assisted aids for automobile driving attempt to foresee users' schemes, and the predictions they make have an impact on the characteristics of the object they are designing.4 4 The anticipation process is itself problematic: they imagine they are at the wheel of a vehicle, using the device they designed. However, although the artifact conserves the trace of these hypotheses, the artifacts do not cover all of the designer's hypotheses. Many aspects are not ‘crystallized’, as clearly shown by the work on the ‘reuse of solutions’ in design. Visser (1992) found that designers need to anticipate future reuse by creating ‘reusable structures’, and Voss and Schmidt-Belz (1993) showed that whenever such structures do not exist, designers look for or need their original ‘context’ or purpose. In other words, utilization of the artifact in the users' activity only constitutes an opportunity, a ‘catalyst’ (Erickson, 1995), for revealing the hypotheses of designers or their consequences, and possibly enriching or shifting them. We think that the instrument-mediated activity approach furnishes conceptual resources for reexamining the hypotheses of designers and revealing novelty. The instrument, composed of an artifact associated with a scheme, is aimed at acting on an object in order to reach a goal and realize a motive. It is this system as a whole (scheme associated with artifact, goal, and motive)—set in motion by the users' activity through the processes of instrumentalization and instrumentation—that has to be captured during activity exchanges. 3.1.3 Work analysis as objectification We have just argued that the instrumental approach supplies a conceptual framework for activity exchanges, during the working hypothesis. But framing instrumental genesis in the mutual learning process has an impact on the concrete management of a given process. An approach like “rapid system prototyping in collaboration with users” (Shapiro, 1994, p. 425) probably only grants a minor role to the users' activity. A detour, consisting in analyzing work in which instrumental genesis manifests itself, is necessary. A work analysis enables an ‘objectification’, i.e. making the constructive activity of users available and legitimatized within the mutual learning community. We will take a look at that these two aspects of objectification and at the role work analysis plays therein. We have seen that during a project, mutual learning is achieved within a community, where the diverging views that surface may lead to mutual learning, but also to the exclusion or compliance of certain actors. The possibility of excluding certain actors is even more likely during joint action among users and designers, as shown by Gärtner and Wagner (1996), for whom “an important part of agenda setting is to create legitimacy” (p. 211). These authors consider that one must “redesign [the actor network] in a way that helps establish and maintain participatory structures” (p. 212). How can we ‘redesign’ the actor network? Above, we introduced the distinction between ‘divergence’ and ‘conflict’, stressing that divergence among actors can be legitimate, but that conflict is one of the ways of treating disagreement that results in a face-a-face confrontation between actors. This means that the fact of accompanying exchanges among actors with knowledge production—which makes it possible to ascribe the divergences not to relationships between human beings but to the properties of the things—probably facilitates mutual learning. Cicourel (1994) showed how a shift in sociological thinking can be achieved by placing knowledge production at the core of the analysis of the social organization. The latter is less related to an established order than to a negotiated order that sets in step by step as new transactions take place. We think that work analysis can play an essential role in this context. It is capable of eliciting knowledge that highlights the nature of the problems operators encounter on the job (Wisner, 1972) and of showing that their actions, skills, and know-how are objectively adapted to the situation. In other words, work analysis can help operators to be treated as worthy contributors.5 5 Note that this technique works less well in branches of industry with little added value, where operators are unskilled, in short, in the less envied occupational milieus. In these sectors, workers' knowledge is in fact worthless. In this respect, work analysis is a prerequisite to setting up a mutual learning process. But there is more. The appropriation process, during which constructive activities emerge, accounts for an ‘internalization’ process, to borrow Wertsch (1998) term. Yet such a process is not entirely observable. Instrumentalization, for example, which is the process by which a user attributes a new function to an artifact during instrumental genesis (Béguin and Rabardel, 2000), does not necessarily have immediate consequences that can be observed from the outside. Blombreg et al. (1996) also stressed this aspect, arguing for the need to ‘look for invisible work’ (p. 260). But in addition to being difficult to see, not all constituents of activities with instruments are directly verbalizable. This is particularly true of schemes in our approach, which have to be grasped. Action is known to mobilize ‘incorporated knowledge’ which is difficult to verbalize (Leplat, 2000). This dimension also applies to schemes, which are supported by ‘operating concepts’ (Vergnaux, 1996) or ‘pragmatic concepts’ (Samurçay and Pastré, 1995) that cannot be put into words by their users. A large part of the activities at work, even in very modern settings, may fall into this category. Work analysis in this case has a twofold objective. It is aimed first of all at reexamining designers' hypotheses, since it can inform us of users' schemes, goals, and motives during the working hypothesis phase. But it is also a condition for user participation in the transformation of their work situations. This is because the method equips users; it supplies them with the conceptualization they need to talk about their own activity. To analyze the work of users, we use an extension of the ‘ergonomic work analysis’ which we apply to the instrumental approach (Béguin and Pastré, 2002). Ergonomic work analysis is a kind of ethnographic way of studying work, which French-speaking ergonomists have been developing for more than thirty years (Ombredane and Faverge, 1955) but which differs from ethnological approaches by the fact that it is aimed at solving concrete problems (see Wisner, 1972; De Keyser, 1991; Wisner, 1995).6 6 The key point consists in relating observations of actions in real situations to dialogue about work with users, during self- or cross-confrontation in order to conceptualize action. During this process, it is the work activity that is taken to be the object of activity. It is not possible here to present a comprehensive description of the entire ergonomic work analysis, but let us say that it is a long and difficult process. It seems to be underestimated in approach like cooperative prototyping, where the artifact is considered to be a ‘trigger’ that ‘provokes’ discussions in the work setting (Mogensen and Trigg, 1992; Trigg and Bødker, 1991). In order for these discussions, and more broadly activity exchanges, to be fruitful, the positions of operators must first be considered as legitimate and valid within the community, and the task of conceptualizing the action must be carried out in concrete situations where constructive activities emerge. It is not until this conceptualization task has been completed that ‘discussions’ are possible. The latter are aimed at reactivating the designers' hypotheses in order to arrive at a ‘record’ (Meyerson, 1949), i.e. a collective interpretation of events (“We observed something. What should we do, what lesson can be drawn from it, what decisions should we make accordingly?”). The record supplies the basis for defining the essential points that need to be retained; it promotes generalization, capitalization, and memorization. Record supports and facilitates learning (Bruner, 1996). 4 Mutual learning with an alarm Based on a model of the activities of design, we have just presented the dialogical approach we call ‘artifact-based learning’, where instrumental genesis can be a constituent of activity exchanges between users and designers if an objectification process is carried out. We attempted to apply this method to the design of an instrumented safety system. 4.1 The Project The project7 7 The project was conducted in close collaboration with INERIS (National Risk Institute), a branch of the French Ministry of the Environment and Industry. I thank G. Beaumont (IPSN), P. Vicot and M. Kaszmierczak (INERIS). was launched following an inquiry conducted in a chemical plant where an operator had been killed due to an explosion caused by a ‘chemical runaway’. Five operators were on-site at the time of the accident. They all said they had realized that something was wrong two hours before the explosion, but did not attempt to leave the premises until a few seconds before it occurred. There are several reasons why the operators stayed on site too long, including the desire to ‘recover’ production and avoid destruction of the installation, etc. But the main problem was their difficulty identifying the time remaining before the explosion. This point was confirmed by other analyses (accidents like this are not rare in chemical plants, especially in small or middle-size companies). Chemical runaways are one of the major causes of human death in chemical plant. The decision was thus made to develop an instrumented safety system8 8 Three types of process control devices are defined by norm NE 31, depending on the extent of the risk: control systems, which assist operations; monitoring systems, used to detect events; and instrumented safety systems, which are specifically devoted to deteriorated situations. (Namur, 1995) in order to prevent the occurrence of this serious type of accident. An algorithm to detect the critical moment of chemical runaway (the explosion) was developed, and then tested in an experimental situation (large-scale test). Once the algorithm was developed, the engineers contacted us to ask questions about ‘the appropriation of the device’ (as they call it). They wondered, for example whether the device would lead operators to push the reaction process to its limits for the purposes of production. After having explained our position, we proposed to install a prototype of the device (beta version) on a pilot site interested in having such an alarm at their disposal. The rest of the study was run in two phases: – The first phase was a preparatory phase for activity exchanges; it was aimed at objectifying the operators' activity based on an analysis of their work. This phase was conducted while the engineers were developing the alarm prototype. – It was during the second phase that mutual learning was set up. The artifact was put into operation one day a week, in our presence. A total of 11 working-hypothesis sessions of about hours each9 9 This period corresponds to the duration of the product's production cycle. were held. This phase lasted eight months, during which two other artifacts were produced. The operators participated throughout the study. The engineers attended all of the working-hypothesis sessions. Follow-up group meetings were held to present the main results of the study, during both the preparatory phase and the mutual learning phase. They were attended by management (the head of technical operations), foremen, and the head of safety operations, in addition to the concerned operators, engineers, and ourselves. Fig. 1 summarizes the entire design process. Fig. 1 Open in new tabDownload slide Flow chart showing the steps in the design of the instrumented safety system. Fig. 1 Open in new tabDownload slide Flow chart showing the steps in the design of the instrumented safety system. 4.2 User work activity at the pilot site Let us briefly describe the pilot site and the activity of the operators. Without some basic knowledge of the operators' activity, instrumental genesis cannot be understood. 4.2.1 Characteristics of the pilot site and the product The project took place in a catalyst production unit that makes synthetic optical lenses. Five operators put out two ‘loads’ of the product daily. A ‘load’ is several tens of kilograms of product from the catalyzer, where it is synthesized in two glass reactors. The operators work right beside the reactors, with no ‘control room’ between them and this highly explosive product. The product can take on three states: – A gaseous state. The synthesis of the catalyzer is exothermic: the temperature of the product can rise on its own and release gaseous by-products. If the exothermic process ‘takes off’, the build-up of gas can be so great that there is an explosion. This is ‘chemical runaway’. – A solid state. To avoid chemical runaway at the pilot site, the catalyzer is cooled during synthesis by glycolated water circulating in coils placed in the reactors. Thermal homogeneity is ensured by agitators. But over-cooling of the product can lead to an unstable ‘super-undercooling’ phase. The product then solidifies: this is what the operators call ‘crystallization’. – A liquid state. Between the solid ‘super-undercooling’ state and the gaseous state in chemical runaway, the product is liquid. 4.2.2 Users' work activity 4.2.2.1 Cold-based process control The activity analysis conducted during the preparatory phase of the study showed that the operators ran the process ‘as close at possible to crystallization’, based on the low temperature threshold rather than on the high threshold (the chemical runaway point). This strategy is rooted in two facts: – Obviously, operators are well aware of the high risk of chemical runaway in the upper temperature ranges of the product. Keeping the process running close the crystallization state lowers this risk. – Production experience has shown that the colder the product during synthesis, the higher the quality of the resulting catalyzer (from the physico-chemical standpoint). In controlling the process, then, operators maintain the product as close as possible to the lower temperature threshold for production purposes. 4.2.2.2 Ongoing risk However, with this low-temperature strategy, there is a risk of ‘crystallization’. Although less dangerous than chemical runaway, crystallization is a serious incident that must be avoided. It leads to solidification of the product, which may cause equipment breakage (especially of the agitators, which help prevent solidification and promote product cooling). In addition, when crystallization does occur, the product must be allowed to warm up in one way or another, with all the risks the instability of this polymer entails. 4.2.2.3 Conceptualizations for action Controlling the process close to the lower temperature threshold is a difficult task which has corresponding schemes and requires certain operator skills and conceptualizations. This aspect of the problem is illustrated here by the ‘crystallization theories’ uncovered during the activity analysis. Two process control strategies were identified, despite the small number of operators. – In one strategy, it is deemed important to avoid propagation of the ‘beginnings of crystallization’, i.e. the formation of crystals in the vicinity of the cooling system. In this strategy, the agitators are stopped as soon as these beginnings of crystallization appear. This strategy rests on a ‘propagation theory’ regarding the onset of crystallization. – In another strategy, it is deemed, on the contrary, that the agitators must be kept running, especially after the last bout of cold. In this approach, it is believed that agitation optimizes thermal exchange between the cooling system and the product. This strategy can be said to rest on a sort of ‘thermal equilibrium theory’. The ‘propagation theory’ and the ‘thermal equilibrium theory’ elicited during the activity analysis, were not based on scientifically validated knowledge, but were conceptualizations about what action should be taken to avoid the ‘ongoing risk’ of crystallization. Although these theories were relied upon at all times, they were only poorly conceptualized by the operators, were unknown to the company, and were considered a priori to be insignificant by engineers. It is in this sense that the work analysis allowed the voice of operators to be considered legitimate and valid. 4.3 Designers learn from users' activity Let us now present the initial results of alarm installation. First, the initial prototype will be described. Then we will see how instrumental genesis ensued. Some paradoxes were encountered after the genesis phase, especially on the designer side. Resolving the paradoxes led to the reinterpretation of the purpose of the project from a new angle. 4.3.1 The initial prototype The artifact consisted of an anti-deflagration box, two alarms (visual and auditory), and two digital displays. One display gave the temperature and the other, the ‘time-to-maximum-rate’ (TMR). TMR predicts the time remaining before a chemical runaway, on the basis of the product's dynamic. From the designer's viewpoint, the purpose of the device is to structure the operators' actions: an initial visual alarm would indicate the need to take preventive action, and then a sound alarm would be a warning to leave the premises.10 10 Although simple in form and function, this safety device is nonetheless based on complex principles and is highly innovative. No comparable device was available on the market at the onset of the study. 4.3.2 Instrumental genesis The results obtained in the weeks following the introduction of the artifact showed that the operators looked at the alarm increasingly often, which was positive a priori. During the first session, they consulted the interface only 1.7% of the total process control time, whereas they did so 31.5% of the time by the end of the sixth session (Table 1: the comparative evolution of operators' gazes—in percentage of total session observation time—at the prototype and at previously available thermometers during the first, third, and sixth sessions). However, a finer analysis of information intake (gaze direction; for additional methodological details see Guérin et al., 1997) pointed out the paradoxical nature of this ‘appropriation’. The device merely replaced previously available thermometers (Table 1). In other words, operators read the temperature information off the artifact, but not the TMR, even though this was the main advantage offered by the new system. Table 1 Comparative evolution of operators' gazes . First session (%) . Third session (%) . Sixth session (%) . Gazes at prototype 1.7 8.1 31.5 Gazes at previously available thermometers 31.3 24.5 4.2 Other gazes (at product, co-worker, etc.) 67 67.4 64.3 . First session (%) . Third session (%) . Sixth session (%) . Gazes at prototype 1.7 8.1 31.5 Gazes at previously available thermometers 31.3 24.5 4.2 Other gazes (at product, co-worker, etc.) 67 67.4 64.3 Open in new tab Table 1 Comparative evolution of operators' gazes . First session (%) . Third session (%) . Sixth session (%) . Gazes at prototype 1.7 8.1 31.5 Gazes at previously available thermometers 31.3 24.5 4.2 Other gazes (at product, co-worker, etc.) 67 67.4 64.3 . First session (%) . Third session (%) . Sixth session (%) . Gazes at prototype 1.7 8.1 31.5 Gazes at previously available thermometers 31.3 24.5 4.2 Other gazes (at product, co-worker, etc.) 67 67.4 64.3 Open in new tab This is a case of instrumental genesis, and more specifically, of instrumentalization: the users assimilated the artifact into their existing operating schemes by assigning it a function that differed from the one initially planned by the designers (Béguin and Rabardel, 2000). In the present case, the operators gave the artifact the function of a thermometer, and not the function of an alarm. Two considerations account for this instrumentalization. As we have seen, these operators were accustomed to running the process close to the crystallization point, so they assessed the utility of the artifact in terms of its relevance to the prevention of crystallization. Now it just so happens that the devices available to the operators before the prototype was introduced were inadequate in this respect. They were sensors that indicated the temperature to the nearest half-degree, whereas controlling the process close to crystallization requires a precision level of about one tenth of a degree. As a result, the operators had to make estimations. Such estimates were no longer needed with the alarm, which displayed the temperature in tenths of a degree, thereby allowing for greater process-control precision. However, instrumentalization was not immediate: the operators first compared the behavior of their thermometers to the thermometers on the prototype.11 11 This dimension was not as easy as it might seem: the numbers displayed were highly dependent upon the location of the sensors in the reactors. This instrumentalization process taught the designers something, and with this in hand, they developed a second artifact. An analogical temperature display with memory was added to the artifact along with the digital display. An analogical temperature curve would facilitate the interpretation of a trend in the thermal kinetics of the product, and would act as a preventive variable of crystallization and probably chemical runaways. 4.3.3 Redefining the project The engineers were nevertheless in an uncomfortable position. From their point of view, the instrumentalization was a sign of failure, for two reasons. The use the operators made of the alarm violated norm NE 31, which stipulates that ‘regular control systems’, ‘monitoring systems’, and ‘instrumented safety systems’ should be separate. The way the operators used the new artifact, initially designed to be an instrumented safety system, made it into a monitoring system. Furthermore, while the device enabled the operators to prevent crystallization, what role did it play relative to the main risk? The artifact, whether anyone liked it or not, had turned into a mere thermometer. Given the research and development costs, it was a very expensive thermometer indeed! In order get around this obstacle, we suggested ‘reflective backtracking’ sessions where the plant foreman, the operators, the engineers, and the ergonomists would go back over and think about the study results. The goal was to establish a new record (in the sense of the term defined here). This reflective backtracking led to a redefinition of the purpose of the project, which can be described in three steps. The first step consisted in weighing the different ‘points of view’. The introduction of the alarm had put two forms of expertise ‘face to face’. The first pertained to crystallization, which was a lesser but still ongoing risk. The operators had developed their skills in this realm. They were crystallization experts. The second type of expertise pertained to the risk of explosion. The artifact, or more exactly, the theoretical knowledge upon which it was based and embodied, was the outcome of chemical runaway expertise. The second step involved defining the interdependency of these two ‘points of view’. The point of view based on the major risk of chemical runaway, which served as the frame of reference for the engineers, did not contribute to controlling this particular process, with all its various characteristics (size of reactors, load variability which depends on the outside temperature, etc.). Nor did it help cope with the daily risk. Reciprocally, crystallization, which was the frame of reference for the users, was also insufficient. Chemical runaway could appear following poorly controlled super-undercooling, or simply following equipment breakage. If these events were to happen, the operators would have to ‘cope with the unknown’. The opportunity for building knowledge of chemical runaways on the job were nonexistent: one must produce, and produce without risk. They had to function under uncertainty, as in all risky situations: the more efficient the operators were at producing and at protecting themselves, the less opportunity they had to learn! The third step involved establishing a new orientation for the project. While there was agreement upon the goal to attain (lower the risk of death), the range of possible solutions was only very partially dependent upon the design of the alarm. Not only was crystallization expertise a necessity that had to be instrumented—the second version of the artifact fulfilled that function—but at the same time, the designers' chemical runaway expertise had to be incorporated into the operators' activity. From our standpoint, the record redefined the direction of the exchanges between users and designers. So far we had confronted the outcome of the operators' work with the activity of the designers. To complete the activity exchange process, the next step was to incorporate the conceptual results of the designers' work into the workers' activity in such a way that they could experience, in action, the concrete conditions of a chemical runaway. 4.4 Users learn from designers' activity The characteristics of the artifact made it possible for the users to learn. Indeed, the artifact operated on the basis of an algorithm that modeled the thermal kinetics of the product until chemical runaway occurred. The artifact thus enabled one to simulate a runaway. This involved developing a new version of the alarm and devising scenarios so that the operators could experiment in action with how they would have to act in case of a runaway. – Of course, the artifact had to be operated without the product, so it was necessary to develop a new version of the prototype. During the working-hypothesis sessions, the polymer was replaced by an inert liquid. – Scenarios were defined on the basis of the company's safety procedures. These procedures were seen as methods geared to answer to well-defined situations which are all the more important as the situation is rare as is often the case here. Three scenarios were defined, corresponding to the different ways of inhibiting the process (destruction of the product or chemical inhibition of the reaction). – A simulation was run with the operators and other persons in charge (as stipulated in the procedures), along with the engineers. For each scenario, a ‘record’ was produced by ourselves and the engineers. The records provided the opportunity to transmit knowledge to the operators about chemical runaways. However, from our point of view, the most important fact was not the enhancement of the operators' conceptualizations and skills. In experimenting with the concrete conditions of product destruction, it became apparent in two of the scenarios that it was impossible for the operators to prevent chemical runaway! The first scenario pointed out the need to change the organization's means of action. Indeed, there were not enough operators on site to handle product destruction. The second scenario pointed out the need to modify the working conditions rather than the means of action. In this second scenario, it was discovered that the architectural characteristics of the production room (in particular, the state of the floor) hindered proper action. 5 Discussion In the presentation of our method, the main question we raised concerned how to set up a mutual learning process, and even more specifically, under what conditions can designers learn from users' activity. The results just presented pointed out the importance of ‘activity exchanges’—first from users to designers and then from designers to users—during which intermediary versions of the artifact being designed serve as a vector of learning. The introduction of the first version of the artifact led to its instrumentalization by the operators. This instrumental genesis, ‘objectified’ through a work analysis, led to the development of a second version of the prototype by the designers. After this first phase, the project was reoriented in order to allow the users to experience the concrete conditions of a chemical runaway. Once implemented, this phase showed that the users would fail under these circumstances, due to the organizational and architectural conditions of the site. This launched the final cycle of design in which we did not participate: organizational and architectural modifications were made a few weeks after our project had ended. An additional operator was hired, and the floor of the production room was modified. When sufficiently long and sufficiently objectified, activity exchanges like these substantially modify the object being designed. Initially pragmatic in nature (given the remaining amount of time, inhibit the reaction), the device took on its final status of a subject-oriented instrument with a quasi-didactical purpose. The results just presented thus confirm the merits of incorporating instrumental genesis into the design process, but at the same time, they raise some new questions. Even though one must be careful not to make generalizations on the basis of a single case, we will briefly bring up two issues where a better understanding would be useful in the future. The first deals with the nature of the learning achieved by the designers, the second, with the learning object. It is impossible to fully account for the activity exchanges that took place here without postulating different levels of learning. We have seen that to begin with, the designers ‘learned’ from user appropriation (Section 4.3.1). Based on this, they designed a second version of the alarm. But this was only the first ‘learning’ level. It was a prerequisite to the second level, a much more important one since it provoked a reorientation of the project and the development of a third prototype (Sections 4.3.3 and 4.4). In other words, in the first step the designers draw a new version of the alarm out of the user' appropriation of artifact; in the second step, the designers gain a new understanding of their own activity from the users' activity. This suggests two levels during activity exchanges: learning and development. This is an important distinction. It has been the subject of many debates, but they cannot be reported in this conclusion. Let us simply stress the importance of what Bateson calls a ‘double bind’. Bateson worked out a well-known, complex hierarchy of learning processes based on a “hierarchic classification of the types of errors which are to be corrected in the various learning processes” (Bateson, 1972, p. 287). A central aspect of Bateson's theory is that it insists on the role of the subject's inner contradictions. In a double-bind situation, the individual receives two messages, or commands, which negate each other. It is these inner contradictions at one level that generate learning at a higher level. During the activity exchanges, the designers themselves expressed this double bind, “If there was appropriation it didn't work, and if there wasn't, it was useless.” Resolving this dilemma is what triggered the reorientation of the project, which was then attributed a new meaning. Engeström (2001) also emphasized the crucial role of contradictions as a source of development at the activity-system level. The appearance of novelty can be regarded as a collective invention in the face of felt dilemmas and contradictions, which impede ongoing activity and impel movement and change. Such contradictions “generate disturbances and conflicts, but also innovative attempts to change the activity” (Engeström, 2001, p. 137). Our case confirms this position. Indeed, the double bind of the designers appeared as a highly critical phase, one of discouragement in the designers and the desire to end the study. From a methodological standpoint, this leads us to pay careful attention to what we have called ‘record’, in line with Meyerson and Bruner. The idea is to seek a way out of the contradictions, and even more broadly, a way of producing something new, that does not exist. The outcome of this process is uncertain. It also encourages us to be attentive to conditions where users can be true actors, i.e. where they will be able to have an impact on the choices made during the search for a new way. The second point concerns the object of the learning process. This question is just as important as the preceding one. It seems that a third item in activity exchange should be added to our schema: the situation in which the action takes place. The properties of the situation in fact structure the learning process. Phenomena such as ‘chemical runaway’ and ‘crystallization’ cannot be regarded as second-order variables. Dewey (1938/1990) insisted on the importance of ‘situations of action’. This author contended that when a subject's usual way of acting enters into a state of crisis, it is the ‘situation of action’ that must be determined at a new level. One is tempted here to define the ‘situation of action' as a ‘world’, i.e. a way of grasping a situation (that is neither entirely objective nor entirely subjective) whose function is to conceptually organize reality and orient action (Cassirer, 1942/1991). The notion of world has been widely examined in the literature, sometimes in diverging ways. In design, it is most often employed to account for the social and cognitive ‘features’ of a given specialist (e.g. a methods engineer; Bucciarelli, 1994). This approach is interesting insofar as it shows that the conceptual, axiological, and praxeological backgrounds of designers form a system with the object they are specifying or developing. But above all, it shows that there exist different possible and acceptable worlds for understanding the same situation of action (Goodman, 1978). In our case, the users mobilized a ‘world of cold’, composed of cooling systems, the ‘beginnings of crystallization’, and ‘crystals’, a world constructed for taking action during the process and whose frame of reference was crystallization. This ‘world of cold’ was totally different from the engineers' ‘world of hot’, a gaseous and explosive world understood through an initial in vitro study followed by minute observations loaded with calculations and formulas, a world also constructed for taking action, but whose frame of reference was explosion. The simulation, during which the users learn from the designers' activity, was aimed at confronting the operators' instruments and conceptualizations with the designers' ‘world of hot’. Here, the prototype no longer fulfilled the function of artifact for the users, but served instead as a means of representing the designers' world, in a manner that enabled them to construct their own ways of seeing and acting with respect to that world. The same could be said about the working hypotheses with the first prototype: the alarm had to be implemented in the users' world of cold in a manner that enabled the designer to design a new device with respect to that world. One could even be tempted here to speak not only of ‘exchanges of activity’, but of ‘exchanges of activity within a world’. However, granting such an important role to these ‘worlds’ requires a conceptual framework for grasping them. Acknowledgements An earlier version of this article was presented in February 2002 at the ‘Center for Activity Theory and Developmental Work Research’ (Helsinki). I thank R. Engeström, Y. Engeström, R. Miettinen and J. Virkkunen for their comments on this article, and all the people at the CATDWR, with whom I have enjoyed spending time. Thank you also to the anonymous reviewers of the article and to the guest editors of this special issue P. Rabardel and Y. Waern who provided useful comments on prior drafts of this article. References Bannon, 1991 Bannon L. , From human factors to human actors Greenbaum J. Kyng M. Design at Work, Cooperative Design of Computer Systems 1991 From human factors to human actors , Hillsdale, New Jersey 25 – 45 OpenURL Placeholder Text WorldCat Bannon and Bødker, 1991 Bannon L.J. Bødker S. , Beyond the interface: encountering artifacts in use. Designing interaction Caroll J.M. Psychology at the Human–Computer Interface 1991 Cambridge University press , Cambridge 227 – 253 OpenURL Placeholder Text WorldCat Bateson and 1972 Bateson G. , Steps to an Ecology of Mind 1972 Ballantine Books , New York Béguin, 1994 Béguin, P., 1994. De l'individuel au collectif dans les activités avec instruments (from individual to collective it the instrument-mediated activity approach). Doctoral Dissertation. CNAM, Paris, 189 pp. Béguin, 1993 Béguin, P.,1993. Field evaluation and collective work design in practice with CAD. Position paper for the Workshop on evaluation studies in CSCW, ECSCW'93, Milan, September 1993. Béguin and Pastré, 2002 Béguin P. Pastré P. , Working, learning and designing through simulation Bagnara S. Pozzi S. Rizzo A. Wright P. Cognition, Culture and Design Proceedings of 11th European Conference on Cognitive Ergonomics. Pomel, Roma 2002 5 – 12 [ISBN 88-850 59-13-9] OpenURL Placeholder Text WorldCat Béguin and Rabardel, 2000 Béguin P. Rabardel P. , Designing for instrument-mediated activity , Scandinavian Journal of Information Systems 12 ( 2000 ) 173 – 191 OpenURL Placeholder Text WorldCat Bjerknes and Bratteteig, 1987 Bjerknes G. Bratteteig T. , Florence in wonderland Bjerknes G. Ehn P. Kyng M. Computer and Democracy—A Scandinavian Challenge, Aveburry, Aldershot 1987 279 – 295 OpenURL Placeholder Text WorldCat Bjerknes and Bratteteig, 1995 Bjerknes G. Bratteteig T. , User participation and democracy. A discussion of Scandinavian research on system development , Scandinavian Journal of Information Systems 7 ( 1 ) 1995 ) 25 – 45 OpenURL Placeholder Text WorldCat Blombreg et al., 1996 Blombreg J. Suchman L. Trigg R.H. , Reflection on a work oriented design project , Human–Computer Interaction 11 ( 3 ) 1996 ) 237 – 266 Google Scholar Crossref Search ADS WorldCat Bødker, 1998 Bødker S. , Understanding representation in design , Human–Computer Interaction 13 ( 2 ) 1998 ) 107 – 127 Google Scholar Crossref Search ADS WorldCat Bødker and Grønboek, 1996 Bødker S. Grønboek K. , Users and designers in mutual activity: an analysis of cooperative activities in systems design Engeström Y. Middleton D. Cognition and Communication at Work 1996 Cambridge University Press , Cambridge 130 – 158 OpenURL Placeholder Text WorldCat Bødker et al., 1987 Bødker S. Ehn P. Kammersgaard J. Kyng M. Sundblad Y. , A utopian experience Bjerknes G. Ehn P. Kyng M. Computer and Democracy—a Scandinavian Challenge, Aveburry Aldershot 1987 251 – 278 OpenURL Placeholder Text WorldCat Bratteteig, 1997 Bratteteig T. , Mutual learning. Enabling cooperation on systems design Bras K. Monteiro E. Proceedings of IRIS 20 Conf. Proc. N°1, June 97 1997 Department of informatics, University of Oslo 1 – 19 OpenURL Placeholder Text WorldCat Brown and Duguid, 1994 Brown J.S. Duguid P. , Borderline issues: social and material aspects of design , Human–Computer Interaction 9 ( 1 ) 1994 ) 3 – 36 Google Scholar Crossref Search ADS WorldCat Bruner, 1996 Bruner J. , The Culture of Education 1996 Harvard College University Press , Cambridge, MA Bucciarelli, 1994 Bucciarelli L.L. , Designing Engineers 1994 The MIT Press , Cambridge Cassirer, 1942/1991 Cassirer E. , Logique des Sciences de la Culture 1942/1991 Ed. du Cerf , Paris Cicourel, 1994 Cicourel A.V. , La connaissance distribuée dans le diagnostic médical , Sociologie du travail 36 ( 4 ) 1994 ) 427 – 450 OpenURL Placeholder Text WorldCat Daniellou, 1992 Daniellou F. , Le statut de la pratique et des connaissances dans l'intervention ergonomique de conception Thèse d'habilitation à diriger des recherches 1992 Université de Toulouse-Le Mirail , Toulouse pp. 230 OpenURL Placeholder Text WorldCat Day, 1995 Day D. , Adaptive discovery and least commitment: An extension of cognitive fit Hasan H. Nicastri C. HCI: A Light into the Future. Proceedings of the 4th conference of the Computer Human Interaction, Special Interest Group of the Ergonomics Society of Australia (OZCHI95), Wollongong, Australia, 27–30 November 1995 pp. 256–261 OpenURL Placeholder Text WorldCat De Keyser, 1991 De Keyser V. , Work analysis in French language ergonomics. Origin and current research trends , Ergonomics 34 ( 6 ) 1991 ) 653 – 669 Google Scholar Crossref Search ADS PubMed WorldCat Dewey, 1938/1990 Dewey J. , Logique, la théorie de l'enquête 1938/1990 PUF , Paris Duvenci-Langa, 1997 Duvenci-Langa, S., 1997. Evolution de l'activité et des compétences en situation d'automatisation: le cas des machines-outils. Doctoral thesis in ergonomics. CNAM, Paris, 198 pp. Easterbrook et al., 1993 Easterbrook S.M. Beck E.E. Goodlet J.S. Plowman L. , C.S.C.W. cooperative or conflict Easterbrook S.M. A Survey of Empirical Studies of Conflict 1993 Springer , Berlin 1 – 68 OpenURL Placeholder Text WorldCat Engeström, 2001 Engeström Y. , Expansive learning at work: toward an activity theorical reconceptualization , Journal of Education and Work 14 ( 1 ) 2001 ) 133 – 156 Google Scholar Crossref Search ADS WorldCat Erickson, 1995 Erickson T. , Notes on design practices: stories and prototypes as catalysts for communication Caroll J.M. Scenario-Based Design Envisioning Work and Technology in System Development 1995 Wiley , New York 37 – 58 OpenURL Placeholder Text WorldCat Falzon et al., 1996 Falzon, P., Darses, F., Béguin, P., 1996. Collective Design Process. In: Proceedings of the Second International Conference on the Design of Cooperative Systems, COOP'96, Juan-les-Pins, France, 12–14 Juin 1996. pp 43–59. Galinier, 1997 Galinier V. , Concevoir autour des schèmes d'utilisation: l'exemple d'une boite de vitesses semi-automatique , International Journal of Design and Innovation Research 10 ( 1997 ) 41 – 58 OpenURL Placeholder Text WorldCat Gärtner and Wagner, 1996 Gärtner J. Wagner I. , Mapping actors and agendas: political frameworks of systems design and participation , Human–Computer Interaction 11 ( 3 ) 1996 ) 187 – 214 Google Scholar Crossref Search ADS WorldCat Goodman, 1978 Goodman N. , Ways of Worldmaking 1978 Hackette Publishing Company , Indianapolis Granath, 1991 Granath J.A. , Technology and Human Factors Design in a Socio-Technical Context 1991 Chalmers University of Technology, School of architecture , Göteborg, Sweden OpenURL Placeholder Text WorldCat Greenbaum and Kyng, 1991 Greenbaum J. Kyng M. Design at Work: Cooperative Design of Computer Systems 1991 Lawrence Erlbaum associates Publishers , Hillsdale, New Jersey Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Guérin et al., 1997 Guérin, F., Laville, T., Daniellou, F., Durrafourg, J., Kerguelen, A., 1997. Comprendre le travail pour le transformer, la pratique de l'Ergonomie, Second ed. ANACT, Lyon. Hammond et al., 2001 Hammond J. Richard J.K. Craig M.H. , Distributed collaboration for engineering design: a review and reappraisal , Human Factors and Ergonomics Design in Manufacturing 11 ( 1 ) 2001 ) 35 – 52 Google Scholar Crossref Search ADS WorldCat Hatchuel, 1994 Hatchuel A. , Apprentissage collectif, et activités de conception (Collective learning and design activity) , Revue Française de Gestion, Juin, Juillet, Août 94 ( 1994 ) 109 – 120 OpenURL Placeholder Text WorldCat Henderson, 1991 Henderson A. , A development perspective on interface, design and theory Carroll J.M. Designing interaction: Psychology at the Human Computer Interface 1991 Cambridge University press , Cambridge 254 – 268 OpenURL Placeholder Text WorldCat Herbert, 1993 Herbert D.M. , Architectural Study Drawings 1993 Van Nostrand Reinhold , New York Kraft and Bansler, 1992 Kraft P. Bansler J. , The collective resource approach: The Scandinavian experience In: Proceedings of the PDC92 Participatory Design Conference ‘Computer Professionals for Social Responsability’, Cambridge, MA 1992 pp. 127–135 OpenURL Placeholder Text WorldCat Kyng, 1995 Kyng M. , Creatring contexts for design Carrol J.M. Scenario-Based Design. Envisioning Work and Technology in System Development 1995 Wiley , New York 85 – 108 OpenURL Placeholder Text WorldCat Leont'ev, 1978 Leont'ev A.N. , Activity, Consciousness, and Personality 1978 Englewood Cliffs, NJ , Prentic-Hall Leplat, 2000 Leplat J. , L'analyse psychologique du travail en ergonomie (Psychological analysis of work in ergonomics) 2000 Octarès , Toulouse Meyerson, 1949 Meyerson Y. , Les fonctions psychologiques et les œuvres 1949 J. Vrin , Paris Mogensen and Trigg, 1992 Mogensen P. Trigg R.H. , Artifacts as Trigger for Participatory Analysis In: Proceedings of the PDC92 Participatory Design Conference ‘Computer Professionals for Social Responsability’ Cambridge, MA 1992 pp. 55–62 OpenURL Placeholder Text WorldCat Montmollin, 1992 Montmollin M. , de Macroergonomics: hodge podge or new foundation , Le travail humain 55 ( 2 ) 1992 ) 112 – 125 OpenURL Placeholder Text WorldCat Moran and Caroll, 1996 Moran T.P. Caroll J.M. , Design Rationale 1996 Lawrence Erlbaum Associates Publishers , Hillsdale, New Jersey Namur, 1995 Namur, 1995. Safety of Process Plants Using Process Control Systems. NE31, Standardisation Association for measurement and control in process industries, 11 Jan. 1995. Nicolas, 2000 Nicolas, L., 2000. L'activité de conception durant l'analyse fonctionelle: vers des outils anthropocentrés pour la conception de produits automobiles. Doctorat d'Ergonomie. CNAM, Paris, 178 pp. Ombredane and Faverge, 1955 Ombredane A. Faverge J.M. , L'analyse du travail 1955 PUF , Paris Piaget, 1970 Piaget J. , Biologie et connaissance 1970 PUF , Paris Rabardel and Béguin, 2003 Rabardel, P., Béguin, P., 2003. The instrument-mediated activity approach. Theoricals Issues In Ergonomics Sciences (in press). Rabardel and Bourmaud, 2003 Rabardel, P., Bourmaud, G., (2003). From computer to instrument system: a developmental perspective. Interacting with computers (in press). Robinson, 1993 Robinson M. , Design for unanticipated use Michelis D. de, Simone C. Schmidt K. Proceedings of the Third European Conference on C.S.C.W 1993 Kluwer , Dordrecht 187 – 202 OpenURL Placeholder Text WorldCat Samurçay and Pastré, 1995 Samurçay R. Pastré P. , La conceptualisation des situations de travail dans la formation des compétences , Education Permanente 123 ( 1995 ) 13 – 32 OpenURL Placeholder Text WorldCat Schön, 1983 Schön D. , The reflective practitioner How Professionals Think in Action. Basic Books 1983 Harper Collins Publisher , USA OpenURL Placeholder Text WorldCat Shapiro, 1994 Shapiro D. , The limits of ethnography: combining social sciences for CSCW Proceedings of the CSCW94 Conference on Computer Supported Cooperative Work 1994 ACM , Chapell Hill OpenURL Placeholder Text WorldCat Suchman, 1987 Suchman L. , Plans and Situated action 1987 Cambridge University press , London Thomas and Kellogg, 1989 Thomas J. Kellogg W. , Minimizing ecological gaps in user interface design , IEEE Software 1989 ) 78 – 86 OpenURL Placeholder Text WorldCat Trigg and Bødker, 1991 Trigg R.H. Bødker S. Grønbaek A. , A video-based analysis of the cooperative prototyping process , Scandinavian Journal of Information Systems 3 ( 1991 ) 63 – 86 OpenURL Placeholder Text WorldCat Vergnaux, 1996 Vergnaux G. , Au fond de l'action la conceptualisation Barbier J.M. Savoirs théoriques et savoirs d'action 1996 PUF , Paris OpenURL Placeholder Text WorldCat Vicente, 1999 Vicente K.J. , Cognitive Work Analysis: Toward Safe Productive and Healthy Computer-Based Works 1999 Lawrence Erlbaum Associates Publishers , Hillsdale, New Jersey Vincenti, 1990 Vincenti W.G. , What Engineers Know and How They Know It? 1990 Johns Hopkins , Baltimore Vinck, 2001 Vinck D. Ethnography of Design and Innovation 2001 MIT Press , Cambridge, MA Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Visser, 1992 Visser W. , Use of analogical relationship between design problem–solution representations , Studia psychologica 34 ( 4–5 ) 1992 ) 351 – 357 OpenURL Placeholder Text WorldCat Voss and Schmidt-Belz, 1993 Voss, A., Schmidt-Belz, B., 1993. Case-oriented knowledge acquisition for architectural design. In: Visser, W. (Ed.), Proceedings of the Workshop of the Thirteenth International Joint Conference on AI ‘Reuse of designs: an interdisciplinary cognitive approach’, Chambery (France), August 29, 1993, pp. 34–46. Waern and Waern, 1993 Waern Y. Waern K.G. , Computer assistance in design engineering , Behavior and Information Technology 12 ( 3 ) 1993 ) 165 – 173 Google Scholar Crossref Search ADS WorldCat Weill-Fassina et al., 1993 Weill-Fassina A. Rabardel P. Dubois D. Représentation pour l'action 1993 Octarés , Toulouse Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Wertsch, 1998 Wertsch J.V. , Mind as Action 1998 Oxford University press , New York Winograd and Flores, 1987 Winograd T. Flores F. , Understanding Computer and Cognition: A New Foundation for Design second ed. 1987 Addison-Wesley , Reading, MA Wisner, 1972 Wisner A. , Diagnosis in ergonomics or the choice of operating models in field research , Ergonomics 15 ( 6 ) 1972 ) 601 – 620 Google Scholar Crossref Search ADS PubMed WorldCat Wisner, 1995 Wisner A. , Understanding problem building: ergonomic work analysis , Ergonomics 38 ( 8 ) 1995 ) 1542 – 1583 Google Scholar Crossref Search ADS WorldCat © 2003 Elsevier B.V. All rights reserved.
TI - Design as a mutual learning process between users and designers
JF - Interacting with Computers
DO - 10.1016/S0953-5438(03)00060-2
DA - 2003-10-01
UR - https://www.deepdyve.com/lp/oxford-university-press/design-as-a-mutual-learning-process-between-users-and-designers-cgZ22rsIgE
SP - 709
EP - 730
VL - 15
IS - 5
DP - DeepDyve
ER -