9
An Orientation-Based Approach to Expertise
Technical Research Center of Finland
The expertise of operators of complex dynamic processes expresses itself in the operators’ ability to interpret the process situation on the basis of mediated and often uncertain information of the process. An orientation-based approach was developed to study mastery of this demand from the point of view of operators’ diagnostic and prognostic decision making. According to this approach, in problem solving situations, operators adopt qualitatively different orientations towards the target process. It is, further, supposed that differences in orientations correlate with differences in action strategies and efficiency of actual process control performance and, perhaps, even with differences in the acquisition of the mastery of the target process in the long run. The results of the studies reported in the paper give tentative partial support to these assumptions. As relatively stable regulatory structures orientations would effect operators’ decisions both in regard with normal and disturbance situations. Thus, this framework opens up a coherent way to investigate activities under these, seemingly polarized conditions, promoting understanding of the development of operators’ expertise in daily work.
INTRODUCTION
There are two mutually related tendencies in the development of work associated with the increasing utilization of information technology. These are automation and “informatization” of work (Zuboff, 1988; PAQ, 1987). Automation refers to the materialization of human cognitive skills in machines, which increases the subject’s physical distance from the object of work. One of the most important aspects of this distancing is that knowledge of the process events is not immediate. On the contrary, it is mediated through measurements and transformed into information that appears in symbolic form. In this sense, information technology not only automates, but also simultaneously informates work. Utilization of this information prerequires its conceptual interpretation. As a result, representations of the process events may be deepened as unperceivable events become conceivable and as relationships and dependencies between events are established. The realization of the informing potential of information technology is thus dependent on the acquisition of the relevant conceptual skills among the personnel and on the creation of organizational requirements for learning and utilization of such skills (Zuboff, 1988).
Increasing reliability in the implementation of automation seems to imply the existence of two discrete operating states in production, the normal operation and the rare disturbance or accident situation. This fact has commonly been considered to create a polarized cognitive demand structure in the process control activity, characterized by the daily routines and the highly demanding novel problem solving, sometimes in an actual crisis. The question of how operators in the daily routine are able to learn those skills which are necessary in an accident situation, challenges the idea of polarized demands and calls for an elaboration of the connections between them.
In our research, we have attempted to define the general common demands of expertise that must underlie the specific demands in routine and crisis situations. We see that interpretation of information in a complex situation is the key problem through which expertise in process control activities could be studied.
In order to investigate the work demands as a unity we need to focus on the processes that control actions in different situations. Because the object of activity, its motive, is the controlling instance of actions, we become interested in how the persons define the object of their activity (Leontyev, 1981). This process is called orientation which becomes manifest in a person’s way of framing a problem situation (Galperin, 1979). In our study, the concept of orientation is used to define the operators’ way of coping with the problems of interpreting information.
In the following sections, an orientation-based approach to expertise is described. The presentation consists of three major parts. First, the role of disturbances in work as a challenge for the personnel’s expertise is discussed. Second, the personnel’s optional approaches to disturbances in work are identified, and two ways to utilize orientations as indications of expertise are explained. Third, the development of expertise is discussed from the point of view of the orientation approach. This approach has been developed through the analysis of the mastery of different kinds of industrial work, both from the point of view of daily work, and from the point of view of control of difficult disturbance situations. Some results of these studies are brought up in order to elaborate the steps in the development of the approach, and references to publications that contain more detailed results are indicated.
DISTURBANCES IN THE SYSTEM AS A CHALLENGE TO EXPERTISE
Due to the principal difficulty, even impossibility, of precisely anticipating the functional and economical constraints of the system in its future operation, the system deviates in design from the system in operation. Future operational demands of a complex system, and direct faults in design create uncertainty in the process. This becomes apparent in different kinds of everyday problems, deviations and disturbances in the normal planned operation. It is the operators who directly face these problems, and, as a result of their handling of them identify the features that complete the design. The more complex the system is, the more difficult it is to anticipate such problems. At the same time, due to safety and economical reasons, the more important it becomes to identify them. Thus, the operators who handle difficulties are not restricted to identifying problems that are predictable, but they are also required to define the world of the possible. The users’ interventions can be interpreted as an opportunity for them to complete the design and continue the construction of the system. Through the handling of problems, a potential for learning is opened up for the user.
The intersection of the top-down design activities and the bottom-up user interventions are the disturbances caused by the unanticipated behavior of the system. Disturbances have a double nature. On the one hand, they are threats to the proper functioning of the system, and on the other hand, they include the possibility of developing the system.
It is a typical theoretical system idea to study a system’s functions from the point of view of its deviations. Thus, the approach is well known in the field of industrial safety where it is applied to controlling occupational accidents (Kjellen, 1987). In their book Individual Behaviour in the Control of Danger, Hale and Glendon (1987) referred to a number of accident sequence models based on this concept. These models define different sequential states that a system may take when a disturbance is imminent. Not only the functions of the technical system but also the system operators’ activity has been analyzed from this point of view. There is a wide range of models of human error mechanisms from which taxonomies of errors are derived (Rasmussen, 1982; Reason, 1987).
These “orthodox disturbance models” refer to the first aspect of disturbance, that is disturbance as a threat to the functionality of the system. The deviation is typically interpreted as a state that falls out of the set norm which can be, for example, the planned functioning of the system. When analyzing disturbances, interest is concentrated on the deviation process itself, for example, failure causes and mechanisms are studied and disturbance classifications constructed with the aim to prevent the threat caused by the deviation. Corrections of the system are carried out by design engineers. As a consequence of the assumptions of the orthodox disturbance models, a high disturbance rate during implementation, a steady low rate at normal operation due to the feedback control of disturbances, and an acceleration of disturbances due to normal wear during later years of the system’s operation can be predicted. Possible design failures as signs of “teething problems” are supposed to be identified and eliminated during the implementation.
In a research project on the implementation of a new flexible manufacturing system (FMS), different research interventions and a follow-up after 1 1/2 years of normal operation were carried out. The disturbances in the system were investigated in this study (Toikka, Norros, Hyotylainen, & Kuivanen, 1991). The results of the follow-up indicate that, against predictions of the orthodox disturbance model, the failure rate at normal operation was very high, the disturbance time accounting for 25% of the total time. An analysis of the causes of the failures further indicated that the proportion of design failures and failures of undefined cause were higher than could be expected on the basis of the disturbance model.
Thus, the results indicate that neither the rate of the disturbances nor the distribution of the disturbance types supports the prediction based on the aforementioned orthodox disturbance model. The results became more comprehensible when the unpredictability of complex systems was taken into account. When analyzing the design-based failures in detail, it was found that only one part of the cases could be considered design failures in the strict sense; that it could have been possible to make a better design decision. Instead, some of these disturbances were caused not by deficient, but rather by limited design. In other words, unpredictability of some functional requirements and unanticipated interactions of the complex system had made complete design unattainable. The high rate of design-based disturbances during the normal operating period would thus indicate that new questions and knowledge of the system had been created during the operating period.
Given that the data supports the assumption of the unpredictability of a complex system, and the existence of disturbances as the more or less “normal” state of the system, it would further imply that there exists pressure on the users, as well as opportunities for them to develop the system during the operation. Thus, disturbances should also be taken into account as a basis for innovation and change. Support for this implication can be found in the studies made by Hutchins (1988), who identified such innovations in his study of a disturbance situation on the bridge of a large ship. Hutchins conceived the construction of practices in an unexpected situation as an evolutionary transformation of the socio-technical system and discussed it as a complement to direct design. Also, Friedrich (1992) studied the operators’ role in the implementation of modern production systems in several case studies and came to similar conclusions. Of course, this hypothesis contradicts the orthodox disturbance model and typical assumptions derived from the psychology of work, that after implementation (learning phase), the developmental demands of a new system decrease, and the operators’ activity becomes routine.
It is an obvious conclusion from the above that the actual handling of disturbances by operators should become the focus of investigation. It might also be concluded that the afore mentioned data afford a simple extension of the orthodox model of system disturbances. However, we claim that such an extension is inadequate. We argue that it is necessary to reconceptualize system disturbances as an essential feature rather than merely an eliminable feature of system functioning. This and other studies described in this chapter aim to do this.
As was mentioned, the rate of disturbances in normal operation of the FMS studied was high. In our analysis, we found that during the 24-hour observation period, each operator in the shift used on average 1.5 hours (i.e., 21% of his working time) for disturbance handling. However, considerable differences among individual operators were found. The proportion varied from 0% to 47% of the total working time, depending on each shift’s interpretation of the official division of labor.
These facts demonstrate, of course, that disturbances threaten the functionality of the system. They also suggest individual differences in reacting to them. It is now claimed that, depending on the users’ choices in reacting to the disturbances, the developmental challenges of the present system can be exploited to a greater or lesser extent. In a disturbance situation, the system is, in a sense, restored back to its design phase. The problem situation tells the users that design is not yet completed and that something can and must still be done. To catch the decision space and the learning potentials of the users’ problem-solving activities, we have defined optional problem orientations (Norros, 1989a) which can act as a heuristic for exploiting the developmental potential of system disturbances. Orientation is discussed in detail in the next section.
Considering disturbances as an essential feature of system functioning not only means an extension of the orthodox disturbance model but it also affects the conception of expertise. Disturbances are faced by the operators, whose reactions to them indicate the expertise that is available (or that operators are willing to make available in the situation). This indication is based on the afore mentioned double nature of disturbances; that is., disturbances as threats to the functionality of the system, and as possibilities to develop the system. It is the expertise of the operators that defines the extent to which the possibilities for development are exploited. At the same time, the utilization of these possibilities, that is, developing the system, is also an opportunity for the users to construct their expertise.
The way of defining expertise just described has two major advantages. First, through relating expertise with disturbances, the human skills that make up expertise are described in connection with the functional demands of the system. This means that a contextual definition of expertise can be achieved. Second, essential in the conception of expertise is the idea that the operators construct their own expertise through identification of development needs of the system and through participation in the implementation of changes to the system. In this point, we can agree with Engeström (1989), who emphasized that the development of expertise means the ability to reconceptualize the skills of existing experts, and the ability to create new forms of practice. Thus, instead of concentrating on the processes of assimilation and application of expertise, attention should be directed to an analysis of the construction of expertise.
ORIENTATION AS INDICATION OF EXPERTISE
The concept of orientation was originally used by Galperin (1979) to refer to a subject’s cognitive, motivational and volitive reactions to the task he or she is supposed to solve, or to a situation to be mastered. “Most important in life is to orientate adequately in a situation that demands activity, and to direct one’s actions properly” (Galperin, 1979, p. 93). We have adopted this concept to indicate the subject’s way of framing the problem in a situation that requires activity. Orientations are, on the one hand, thought to control the subject’s actions and efforts in the utilization of the development potentials. On the other hand, problems in actions may require reorientations. It is assumed that, through analysis of different orientations, it is possible to make inferences of operators’ expertise because the orientations are some kind of construction principles of expertise.
Orientations represent different ways of framing the problems. Thus, they also define qualitative differences in expertise. Engestrom (1989) pointed out the existence of qualitatively different types of expertise. He noted that this fact questions the common tradition of conceiving expertise in universal terms, which, further, has led to a more or less quantitative handling of expertise in research. This is expressed, for example, in the prevailing novice-expert continuum and in its wide adoption as a research paradigm in studies on expertise. There are, however, examples of studies that challenge the implicit notion of universality. Thus, Lawrence (1988), in her analysis of judicial decision making introduced the concept of frames of reference to point out the qualitative differences in judges’ definitions of their problem spaces. Schön (1988) also draw attention to differences in people’s framing of problems, and he considered them as important determinants of expertise.
The development of our orientation approach to expertise has, until to now, comprised three temporally partly parallel and conceptually interrelated research phases. In the first phase, a typology for problem orientations was created in a research on implementation of flexible manufacturing systems (FMS). In the second phase, a method for description of everyday work orientations was developed in a study within the maintenance organization of a nuclear power plant (NPP). The third phase comprised the analysis of the role of orientation in the control of problem solving in an acute disturbance situation. This study was carried out during the NPP operators’ simulator training.
Development of Problem Orientation Typology
In our attempt to conceptualize the optional orientations that operators might adopt when handling a problem or disturbance, two aspects were found to be important. First, it should be asked what is conceived as the object of disturbance handling activity. The decisive distinction is whether the disturbance is conceived and handled locally as a temporary deviation from normal situation, or in relation to the system’s global functions, including diagnosis of the problem and feed back to the system in a normal situation. Second, it can be asked who actually handles the disturbance and how widely distributed the activity in disturbance handling is; that is, who is the subject of the disturbance handling activity. The typology that includes five basic options was originally based on these two criteria (Norros, 1989a). A schematic representation of the model is depicted in Figure 1.
The typology was developed during the analysis of the data of the aforementioned follow-up study of the implementation of FMS (Norros et al., 1988; Norros 1991; Toikka et al., 1991). A description of the options, including the major functional restrictions of each orientation from the point of view of the management (top-down), and from the operators (bottom-up), are presented in the following.
Withdrawal From Disturbance Handling.
This option refers to the possibility that an operator does not handle the disturbance when it occurs. The basis of this orientation is thought to lie in the acquisition of a deterministic representation of the system through direct tacit learning of normal operation routines. Restricting activity to mere normal execution of routines is a rather extreme orientation. It might express workers’ attempts to achieve short term efficiency in work, or it might also be an indication of workers’ active resistance. Besides being a possible individual choice, this orientation can also be due to Tayloristic organizational principles.
The top-down limits of the withdrawal orientation are met when the restricted capability and motivation of the operators get economic weight that is big enough. The bottom-up need to overcome the withdrawal orientation is a result of the meaninglessness of this orientation from the users’ standpoint. This is the case when the personal costs of this type of activity (monotony and frustration from underutilization, stress caused by the contradiction of responsibility for expensive technology, or the feeling of inability to adequately respond to unexpected situations) exceed the benefits.
In our FMS case study, this type of problem orientation was principally ruled out by the management through its choice for a flexible organizational strategy that emphasizes homogeneous and high expertise. However, we did observe withdrawal during the implementation period.
FIGURE 1. Typology of problem orientations.
Routine Disturbance Handling.
The normal work of the user is interpreted as carrying out prescribed tasks. Handling disturbances in a routine way without making an attempt to analyze the causes of the disturbance or change the system or situation is a part of this.
The user is identifying problems and troubleshooting, reacting to some familiar disturbances one by one, according to known procedures, and locally, without searching for any general reason or systematic connection behind them. Restrictions and inadequacies of the operators’ representation of the system prevent them from anticipating more complex and latent interactions and possible failures of the system.
Our data from intensive observation of the FMS operation include examples that demonstrate routine problem orientation. This orientation first prevailed in the handling of a disturbance that occurred during the observation. However, repetition of the problem increased the pressure to change the orientation. The attempt to diagnose the problem, an indicator of a more advanced orientation, was facilitated by the interaction with the fellow operator.
Private Development Activity.
The transition to more developed management of disturbances takes place when this reactive disturbance orientation loses its meaning for the user, and anticipation of disturbances, and activities to change the system, replace it. This is the case when the routine becomes too arduous or challenges his or her safety, and the user invents a better way to solve the problem by making improvements to the system. The users privately take the authority to define their tasks or make a change to the system aimed at making their jobs smoother. Such an activity prerequires relational models of the system for the systematic collection of experiences in order to consider the competing functional goals of the system. The activity is still, in most cases, triggered by repeated problems, or in some cases, acute demands of the situation.
In transition from the routine disturbance handling to this level of orientation, the operator is making a diagnosis of the disturbance. This involves reflection and redefining the constraints, and possibly the goals of activity in a problem situation. In a way, the user unofficially adopts the authority of the supervisor, and even that of the designer. Paradoxically, this unofficial developmental work maintains the traditional division of labor between operation and design by preventing the revelation of the design deficiencies. This also means that the operator takes personal responsibility for the improvements, which in many cases includes the risk of making errors. This is why, in complex risk processes, the users are advised to rely on known practices, rather than very eagerly invent their own improvements. This conflict creates pressure for public and cooperative development activity.
Public System Optimization.
As unofficial development efforts do not lead to sufficient improvements in the system, and, at the same time as it is clear the user’s knowledge is not efficiently exploited, a tendency to establish the user participation might officially appear. At this stage, the design efforts are directed to optimizing the system functions within the given boundary conditions.
From the operators’ point of view, this orientation requires maintenance of the disturbance as a problem and object of development activity, and explicit knowledge of the system as the condition of its optimization. Learning can remain mainly experiential and on-the-job learning, but some conceptual elements might be needed (e.g., adoption of common planning and working routines, independent use of process-oriented quality control systems, or use of documentation systems for collecting operation experiences)
Explicit rules are made for group work and for the relationship between work groups and the rest of the organization. The critical decision is the transition from the traditional resistance of mass production or craft work culture into participation in official development activities of the organization and cooperation with management.
Expansive System Development.
In this problem orientation, development activity is not limited to optimizing system functions within the given constraints but, instead, directed to redefining the boundary conditions of the system. Operative experiences are systematically collected and fed back to the development of the structure, functional principles and organization of system. Thus, for the first time, the users adopt systematic conceptual design methods. The design tasks include the socioeconomic and technical basis of production, and new work methods are created to organize the cooperation between designers and users.
As the continuous development of the system through its lifetime becomes an economical and functional necessity (product quality, safety, ecological effects, etc.), the sporadic cooperation between design engineers and users becomes insufficient. The integration of the expertise of these two acting groups has to become institutionalized through top-down decisions to form new organizations, and a common system of conceptual tools and working methods must be created. This development includes coping with conflicts of power and social relations as Zuboff (1988) demonstrated in her case studies.
Evaluation of the FMS operators’ disturbance orientation was carried out in normal operation with the help of a check list including the characteristic features of disturbance orientation from the cognitive and cooperative point of view in different sequential phases of disturbance handling. We evaluated those activities that were carried out as a reaction to 36 different disturbances during 24 hours of intensive observation. Further information of the operators’ disturbance orientation was obtained in interviews held 2 weeks after the intensive observation. The results indicate that 67% of the disturbances were handled in a routine way, whereas in 33% of the cases, a more advanced orientation was typical. It was also found that, if a disturbance was a repetitive one, it was more likely to be handled in a nonroutine way.
On the basis of the operators’ and system leaders’ interviews, we were also able to collect information on developmental measures that had been initiated by the operators during the 1 1/2 years of normal operation. According to the operators’ description, it was estimated that 69% of the developmental measures were optimizing activities without direct connection to disturbances. The rest of the cases were immediate reactions to the disturbances.
According to common expectations, it was shown that users participate in system development during implementation. But the data further indicates that the acquired developmental orientation was also maintained during normal operation. Moreover, not only disturbances but also increasingly anticipated problems trigger these activities. This tendency is related to the bottom-up pressure to reach the type 5 disturbance orientation, expansive system development.
Among the developmental measures we were able to register, there were two that indicated the operators had approached the fifth orientation option as they clearly challenged the frames of the given system. These included measures to overcome bottlenecks in the production and developments of the control system, which were not achievable within the present frames, but still close enough to be within reach if new design decisions were considered. However, our interview data also elaborates power issues related to the users’ increased abilities to develop the system. The impression of the users was that their proposals and ideas for the future development of the manufacturing were not especially welcomed by the management. This management attitude became even more apparent through direct omission of engineering resources that were needed to carry out the development suggestions.
The significance of the FMS study from the point of view of the development of the orientation-based approach to expertise was the formulation of types of orientation, which could be considered relevant basic options that system operators would choose when facing a problem situation in their work. In the next phase, an attempt was made to utilize the orientation typology to describe work orientation in daily work on the basis of interviews with personnel.
Orientation in Everyday Work
During the FMS study, it became more and more evident that it would be inadequate to consider orientations merely determined as straightforward individual choices. Instead, the fact that optional orientations to the system disturbances are socially constructed became evident through the study. Thus, adopting an orientation option is a contradictory situation of problem-solving and decision-making. The contradictions to be coped with are already structured top-down by techno-organizational design, but there are also social contradictions within the management, between management and the shop-floor, and within the shop-floor organization itself, as well as psychological contradictions of the organization’s individual members. Setting the orientations of different personnel groups in relation with each other (i.e., operators to maintenance personnel, operative personnel to management or to design), a diagnosis for the development potential of the organization can be achieved. This would signify the viability of the organization.
Further study focused on the organizational assessment of a maintenance department at a nuclear power plant, in which the problem orientation typology was utilized (Reiman & Norros, 1994; Wahlstrom, Norros, & Reiman, 1992). For this kind of “high-hazard-low-risk” plant, it is particularly significant to evaluate the personnel’s expertise from the point of view of the control of safety and quality in their daily work activities. The everyday judgments made by the maintenance personnel in the control of safety and quality are affected by problems in interpreting information of a complex phenomenon in which different sources of uncertainty are present. These judgments affect the prediction of future maintenance tasks. Control of safety and quality have gained increasing attention within the nuclear community, as the recent discussion on the “Safety Culture” indicates (International Atomic Energy Agency, 1991).
This study attempts to diagnose the personnel’s conceptions of the problems of maintenance activity, and the conceptions of their own tasks. It was assumed that these conceptualizations of their daily work would reveal the personnel’s general work orientation, which could be interpreted as a contextual indication of expertise. Thus, comprehensive interviews were carried out with three different personnel groups — plant management, maintenance foremen, and maintenance technicians. The interviews focused on the main functions of the subject’s activity, and particular interest was devoted to the means and constraints of the achievement of high quality and safety in maintenance. The questions were aimed at bringing forward practical cases, where problems in the execution of managerial activities and maintenance work had existed.
The problem orientation typology was utilized in the description of orientations. A new adaptation of the basic orientation types was created for the analysis of managerial orientations, whereas the typology just described was utilized for maintenance foremen and technicians. The typology was operationalized by using particular dimensions of work activity. Thus, an attempt was made to predict expressions of the different orientation options with regard to four dimensions, object of work, work motivation, cooperation, and work culture. Each of these four activity dimensions comprised several subcategories. Descriptive matrices, including the orientation types and dimensions, were constructed for the personnel groups to serve as reference tables in the assessment of orientations. The cells of the matrixes included descriptions of work practices that are thought to manifest different orientation types. On the basis, orientation profiles for each person and average profiles for each personnel group could be achieved.
As a result of the analysis of the interviews, a diagnosis of the major problems of maintenance activity as a whole could be achieved. These problems found specific expressions in the work of each group. Further, it turned out that expertise and development potential, as assessed with the help of orientations, only partially met with the demands expressed in the problems of work. The inadequacy to tackle major demands of work became particularly clear with regard to the utilization of information technology and conceptual tools of activity. This applied to all personnel groups. Underestimation of the role of new tools can essentially weaken the efficiency of the personnel in their attempts to predict future demands and formulate adaptive maintenance strategies. Management was efficient in the control of technical development activities, but less adequate in the identification of the role of communication and of personnel’s participation in goal-setting as resources of management functions. Foremen and technicians seemed to make use of cooperation in task performance, but they did not consider cooperation as a resource to be developed consciously. The quality and safety conceptions of the latter groups were adequate, the functional demands of the whole production process being their reference. In summary, the evaluation indicated that the present expertise within the personnel was internally contradictory, and partly inadequate to tackle the development demands of the maintenance.
Some important conclusions concerning the orientation approach as a contextual assessment method of expertise can be made. The methodical lessons learned through this study concern the criteria for defining the optional orientation types. As was mentioned earlier, there were two aspects of significance in defining orientation: The conceptions of the object of the disturbance-handling activity and the cooperation in disturbance handling.
The use of the typology in the evaluation of everyday work orientations supported the assumption that the way of conceptualizing the object of activity is a significant aspect of orientation. This aspect can be called conceptual orientation.
The basic question is how the persons conceptualize the functional principles of the object they control, which is reflected in the persons’ conceptions of the disturbances of the system. This can be inferred on the basis of how persons handle actual problems or potential disturbances that they identify in the system. In the aforementioned study, this was diagnosed through the personnel’s safety and quality conceptions, through their anticipation of disturbances or problems, and their definitions of their responsibilities in the control of problems. The general distinction that can be used to define the options is between the conception of disturbances as discrete events, and the conception of them in their relation to the global functioning of the whole system. The more functionally a disturbance is conceived, the more general the concept of disturbance is. This also means that disturbance and normal situations are not discretely separated but rather interpreted through a common reference and represented as a continuum. This promotes anticipation of disturbances and their detection on the basis of early signs; this is the central task in daily work.
The conceptual aspect of orientation has been in the foreground of those studies in the literature that identify different frames of reference as determinants of expertise, and also in those works in which the concept of orientation has been used deliberately (e.g., Engeström, 1989).
The second aspect for making a distinction in orientations turned out to be more problematic in the analysis of the data. As we have discussed so far, the crucial problem of expertise is the construction of meaning for the available information. Many authors have pointed out that, due to the complexity of the process, different views of the process coexist. Thus, the construction of meaning has to be carried out cooperatively. For the same reason, Brehmer (1991) emphasized the necessity of communication and distributed decisionmaking in the control of systems. Thus, cooperation should be taken into account when describing activity but, as such, it is not an aspect of orientation.
What makes cooperation and communication interesting with regard to orientation, is a further aspect that is demonstrated in a study by Abercrombie (1989). She studied difficulties in perception and construction of meaning in complex objects during medical training. In her teaching, she noticed, e.g., that students had difficulties in constructing the meaning of the object seen in the microscope, and the tendency of the students to see the expected object described in the textbook. She carried out an experimental training series with medical students, during which she could point out the advantages of multivoiced dialog in the context of making diagnostic judgments. It was shown that dialogue contributed to awareness of the factors that affect judgments, and, thus, it enhanced judgments.
The connection between communication and self-monitoring, or critical evaluation, is important for our reasoning. Communication was, in this Abercrombie’s experimental training, deliberately used as a resource for enhancing task performance. Focusing on communication as a resource resulted in critical evaluation of one’s own activity. Through such changes in the focusing, which can be called reflective relation to problem situation, a distancing effect is achieved, and a redefinition of the activity and the means of activity may result. This seems to promote task performance. This effect has been identified in the studies on expertise. One of the main results of the research on expert skills is that experts typically have strong self-monitoring skills (Glaser and Chi, 1988). Olssen and Rasmussen (1989) also emphasized reflectivity as a characteristic of expert activities.
Thus, we conclude that the critical relation to the problem situation, just described should be considered the second aspect of orientation, reflective orientation. Both aspects of orientation, conceptual orientation and reflective orientation serve as criteria for distinguishing orientation types. This conclusion was affected by the methodological results achieved in the study on process operators’ disturbance-handling activity, which will be discussed in the next section. The redefinition of originally the second aspect may cause a need to refine the typology. Instead of utilizing cooperation as criterion, we should see how the subjects conceive resources of activity, for example., communication and cooperation, as a means and as an object of development within the whole activity.
Orientation in the Control of Actual Disturbance-Handling Activity
In the study just described, the inferences on orientations and respective evaluations of expertise were based on the subjects’ verbal reactions to diagnostic questions. In this case, the subjects were “talking about practice” (Lave & Wenger, 1991), but equally important is to study practice itself. When doing so, orientation can be inferred from the persons’ utilization of available information. We now turn to this question.
Process operators of large automated systems ultimately control the system in real time. The demands faced by the operators represent the requirements of informated work, in which the interpretation of actual process information on the basis of previous experiences and conceptual knowledge of the process are essential. Interpretation of the situation becomes critical when the process reaches an unstable or disturbed state. Operating experience and research results indicate that in such a situation the operators do not always adequately succeed in their interpretations. Typical problems are either nonoptimal utilization of process information or inability to operationalize former knowledge of the system. These problems also could be identified earlier in our own study on nuclear power plant operators’ errors in simulated disturbance situations (Norros & Sammatti, 1986).
In order to understand the origin of these problems, the question of the role of orientation in the control of the crew’s diagnostic and prognostic interpretations of the problem situation was raised in the new study (Hukki & Norros, 1993a, 1993b).
Before starting the new study, the methodology of the earlier one (Norros & Sammatti, 1986) seemed to require consideration and a critical evaluation. In the latter study, the operators’ task was described as an ideal sequence of actions in the predefined disturbance situation. This ideal sequence was then used to define errors and evaluate the adequacy of the operators’ performance. This implies a traditional interpretation of expertise as an ability to reproduce actions that have been defined to represent excellency. This definition of expertise is not explicit but is embedded in the experimental technique used and the model of human cognition behind it. As has been pointed out by Collins (1990), this widely accepted model implicitly assumes that human actions are governed by plans because, retrospectively, they can be described as following plans. This planning metaphor of action was criticized by Suchman (1987) who saw action as situated. “The fact that we can always perform a post hoc analysis of situated action that will make it appear to have followed a rational plan says more about the nature of our analyses than it does about the situated actions” (Suchman 198, p.53).
The notion of situated action draws attention to the constructive aspect of activity, which we had found essential to our conception of expertise, as noted in the third section. Instead of analyzing the fulfillment of the predefined sequence of acts, it is necessary to scrutinize, from moment to moment, how the operating crew navigates in the genuinely uncertain situation and gradually succeeds in structuring the situation.
Based on the critique of the earlier study and the research on the concept of orientation just described, the model of the interpretation of a disturbance situation was created. The focus in the model was the role of orientation in the control of actual disturbance handling, and the purpose of it was to guide the analysis of complicated data of the plant events and the crew’s performance. The model that represents the assumed basic relationships is depicted in figure 2. It was constructed during the evaluation of six process operator crews’ activity during a simulated difficult disturbance situation (Hukki & Norros, 1993a, 1993b).
FIGURE 2. Model of interpretation of a disturbance situation.
As indicated in the model (Figure 2), task performance is supposed to consist of intertwined diagnostic activities (i.e., selection of information and definition of the problem situation) and prognostic activities (anticipation of the development of the situation in relation to possible action alternatives) leading to operations. Carrying out these activities requires an interpretation of the meaning of the whole situation, both from the diagnostic and prognostic point of view. The interpretation is thought to be based on the utilization of available information (dynamic process information and operating instructions, especially emergency procedures). These interactions between interpretations and information (indicated by the horizontal arrows in figure 2) are directed by the crew’s orientation to task performance.
Orientation is naturally dependent on the operators’ knowledge of process dynamics. The coherence of the representation manifests the conceptual aspect of orientation. The other aspect of orientation, the actors’ reflective attitude towards their activity, i.e. their critical evaluation of their own activity, is also important for the interpretation of the situation. These two aspects of orientation contribute to the construction of interpretation of the meaning of the situation.
Because of the collaborative nature of the process control activity, the operators’ activity has to be considered as group activity, as collective decision-making. The construction of the interpretation is collective, and therefore the operators’ orientation to the task performance affects their interpretation of the disturbance situation through collaboration.
By using this model, the construction of the interpretation of the disturbance situation and, inferred from the basis of it, the orientation of the crews can be evaluated by analyzing the way the operators use available information. The possibility of making inferences from conceptual orientation is based on the idea that the informating power, that is the informativeness of information available, may differ. The better information reflects the dynamically essential functions of the process, the better it can be used as evidence of the functional state of the process on a global level, and the higher its informativeness is. The orientation is supposed to be reflected in the degree of utilization of this informativeness. The functionality of conceptual orientation can be evaluated on this basis.
The results of the analysis indicated that there were differences between the crews’ utilization of informativeness of available process information. Those crews whose conceptual orientation was evaluated to be functional seemed to be more efficient in their diagnostic and operative activities. This was interpreted as resulting from a more functional orientation by these crews, which had created a coherent reference for diagnostic and prognostic interpretation of the situation. The results concerning the reflective orientation were not equally confirmative, but clear indications of lack of reflectiveness by the less functionally oriented crews could be found. This was associated with signs of subjective uncertainty, which could be seen as a reflection of incoherence between diagnostic and operative interpretations of the situation among these less functionally oriented crews.
The results of the application of information interpretation model were interpreted as indicating the efficiency of the model in the analysis of the crews’ performance in a dynamic situation. With this model it seemed to be possible to conceive the crews’ activity as it was constructed in the situation. In relation to that, instead of primarily emphasizing the situatedness of activity (Suchman, 1987), we would, on the basis of this analysis, stress the contextuality of activity. Orientation can be seen to serve this contextualization both through conceptual and reflective control processes.
DEVELOPMENT OF EXPERTISE
Two different ways of analyzing orientations as indications of expertise have been constructed in our studies. In the first, orientation was inferred on the basis of verbal descriptions of work practices, whereas, in the second, specific features of the utilization of information was the basis of our inferences. The basic aspects of orientation, the conceptual and reflective orientation are utilized both for constructing the typology of problem approaches and for explaining the control processes that affect interpretation of information in an actual situation. Coherency of the two analytical approaches is important for the possible combination of the two approaches in a study.
A further question that can be raised is whether a functional and reflective orientation in a problem situation would predict effectivity of learning from actual problem experiences and, thus, the development of expertise.
One of the most influential theories about the development of expertise is the one proposed by Hubert and Stuart Dreyfus (1985). In their model of development of skills or expertise, Dreyfus and Dreyfus make the distinction between analytical and intuitive thinking. According to them, it is the level of expertise that defines the prevalence of one or the other type of thinking. Expertise is gradually developed from an initial novice stage through a less committed use of simple and context independent rules. As a result of repeated confrontation with practical situations, the cognitive skills are supposed to transform into expert thinking. This is characterized as contextually adequate, comprehensive and intuitively “seen” decisions that show high intellectually mediated commitment to the object of activity.
Although we find this description of an expert intelligible, we would like to understand better how such a novice could develop into such an expert. Thus, there are a number of questions we would like to have clarified. The first questions are related to the formation principle of intuitive thinking. In the description of the stages of expertise, it seems that the basic mechanism is repetition and identification of similarities on the basis of experience of the repetitive but variable practical situations. This leads to the conclusion that, in a novel situation, an expert acts analytically and becomes a novice again. This seems to be a contradiction to the everyday notion of expertise, and also to, for example, Hammond’s cognitive continuum theory, which assumes that, in a novel situation, an expert acts intuitively (Hammond, 1981). In addition, if intuition is exclusively thought to rely on nonsymbolic forms of representation, it becomes difficult to explain the ability of an expert simultaneously to be a good teacher, which many experts are known to be.
A further question is: Would it be necessary to assume something that is common to or similar between a novice and an expert for explaining the development from a novice to expert, and to understand why every novice does not become an expert?
Could there be an alternative formation principle for intuitive thinking other than repetition that would allow intuition also to be usable in novel situations, and also by inexperienced persons? The Russian author Dawydow (1977) would offer his conception of theoretical thinking as a candidate to that. Dawydow makes a distinction between empirical and theoretical thinking derived from the Hegelian dialectics and principles of dialectical logic (see, e.g., Ilyenkov, 1977). He claims that empirical, or formal, thinking is directed to visible overt features of the object and comparison between them. Consequently, it produces a descriptive and classificatory concept, the operativity of which is questionable. Theoretical or substantial thinking, to his mind, could be characterized as an attempt to reveal internal relations and contradictions of the object that could explain its development and essential nature. A theoretical concept is operative because it includes the construction principle of the object, and, thus, it can also be useful in novel situations. To speculate even further, we have found interesting the notion of intuitive thinking that Spinoza used (1677/1975) in his Ethics to signify the highest form of thinking. Instead of stressing repetition or similarities, he focused on conceiving the essential structure of the object, which can principally be possible straight away.
A further question regarding the theory of Hubert and Stuart Dreyfus relates to the assumption of the theory that learning starts with procedural rules. This implies the notion that the context of expertise is given principally — it exists but it is not yet known by the novice. The development of expertise could thus primarily be conceived as internalizing given expertise through a large amount of practice. This strikes a contrast to some of the results in literature to which we referred earlier in this chapter and also to our own results. On the basis of this, essential to expertise is its constructive character.
Our ongoing work is to study the development of expertise on the basis of the orientation approach. Research work has to be directed to operationalization of the conceptual and reflective aspects of orientation from the point of view of development of expertise. The previously mentioned theoretical thinking, typical to which is the conception of internal dynamical principles of the object, seems to be closely related with what we have called functionality of conceptual orientation. Our results support the conclusion that the functionality of orientation explains the expert’s capability to interpret a concrete situation in a coherent way, in its relations. This capability to contextualize the particular situation seems to be very much the same as what is often meant by intuitive problem solving, typical to which is operative efficiency based on immediate insight of the problem situation. Emphasis on the contextuality rather than intuitiveness of problem solving would have the benefit that the former not only needs to be restricted to characterize experts’ thinking, but could as well appear at least in some novices. Our further assumption is that these novices would learn more effectively, due to the fact that the coherent frame, which allows an effective interpretation of a particular situation, could also serve as a frame for the accumulation of experiences. A bridge between disturbance situation and routine could also be created, as the functional conception of disturbance of the system would offer a frame to anticipate the evolution of disturbances, and to interpret the role of small changes in the system from this point of view. Likewise, a basis for development activities could be found.
The aforementioned hypothesis, that functional orientation would promote learning and development activities, is incomplete without additional assumptions of the role of reflective orientation. Larkin, McDermot, Simon, and Simon (1980) pointed out an important paradox when concluding that teaching physics problems to students does not result in good performance in practice, because teaching is too practical, it is oriented directly to learning operations.
This conclusion demonstrates the potentials of theoretical instruction that for some reasons are often not realizable. Needs to reconsider one’s own tools or to invent new ones are, however, actually present themselves in any real-life work process. In our Nuclear Power Plant maintenance study for example, it was evident that lack of conscious development of the tools of one’s own activity becomes an obstacle to overcome the restrictions of one’s own expertise. Munley and Patrick (1990) demonstrated that, in simulated fault diagnosis, those persons who were reminded of the possibility of further types of failures than that included in a typology of faults, could act more adequately in a novel situation. When the operators were made conscious of the possible restrictions of their disturbance model, they became able to complete the model and act more effectively.
Thus, it can be hypothesized that promoting the reflectiveness of orientation is an important condition for the development of expertise. Deliberate learning of conceptual tools that enhance the functionality of orientation is one way toward this aim. However, reconsideration and refinement of other resources of activity is as important. Multivoiced dialogue and participation in common activity are resources of learning too often neglected (Lave & Wenger, 1991).
CONCLUSIONS
The way expertise is conceived has practical relevance because this conception influences the way expertise is developed through training. The concept adopted influences at least three activity areas. First, one would expect that it effects research on expertise. This chapter has brought up a number of conceptual distinctions that can be considered as suggestions to revise prevailing conceptions of expertise. Thus, our approach has brought up the issue that expertise should not be considered as individual mastery of tasks but should be defined in the social context of activity. The formation of orientations as a contradictory process within the social organization of labor is an indication of this.
Through the adoption of the idea that disturbances reveal both the developmental needs of the system and the possibilities to develop the system, we have adopted a view of expertise that emphasizes its constructive nature. In our approach, expertise manifests in the ability to utilize the potentials for development that are overt in the disturbance. Thus, this view demonstrates that expertise is constructed through the actual problem-solving or development activity.
Further, it follows that expertise is analyzed concretely in connection with the object of control. In this sense, our conception of expertise is context-dependent or contextual. Furthermore, contextuality also refers to relating the particular situation with the functions of the system and with the constraints of activity on a more global level. Contextuality in this sense was seen to relate with the functionality and reflectivity of orientation. This point of view was seen to open possibilities of tackling important theoretical questions of the relationships between the intuitiveness of thinking and theoretical thinking, which are central issues in the development of expertise.
On the basis of the redefinitions concerning the nature of expertise, transformation in research paradigms may be expected. This question was not deliberately discussed here but was brought up as our own reorientation in choosing the research methodology in our studies on Nuclear Power Plant operators’ activity in disturbance situations.
The concept of expertise adopted should have implications on the deliberate development of expertise. It seems that the approach to expertise presented here could lead to changes in the learning of expertise and, thus, new training methods should be derived. Essential reorientation would be needed towards the theoretical elements in training. Instead of understanding theory as a set of declarative knowledge, knowledge as activity should be emphasized. One way to develop substantial knowledge of the process could be a training process that uses extensive modeling activities (Norros, 1989b). Tools for the analysis of the basis of the operators’ decision-making in problem situations would also promote the critical evaluation of their own activity and enhance learning from experience. But, even more profound reorientation would be needed in conceiving the relationship between the development of expertise and development of production technology and tools. There is increasing evidence of incremental technological innovations that are made by operating personnel. Simultaneously, this demonstrates that the expertise of the operators exceeds the level expected. The need to utilize this eventual potential is reflected in the growing interest in the participatory design. Much more emphasis should, however, be put on the conceptual tools of the participating personnel.
Third, the concept of expertise should affect our conceptions of operators’ daily activities. Expertise in automated work includes conceptual comprehension of the uncertainty inherent in the process. Insight into these facts might cause a revision of the attitude toward daily routines. If work is conceived as mere routine, being prepared for problems and learning are hindered. As a result, overconfidence of control and regression of actual expertise may appear simultaneously. This is a combination that gives little possibility to master a novel situation. On the other hand, an understanding of the complexity of the system and its inherent uncertainty could create an attitude of respect and curiosity toward routine and small problems. A consciousness of the need of change embedded in the routine, together with a comprehension of oneself as a subject of this change, adds to the routine the possibility of learning and development of expertise. Our concept of expertise would suggest qualification measures and restructuring of daily routines in such a way that the polarization of demands between routine and crisis would tend to decrease.
REFERENCES
Abercrombie, M.L.J. (1989). The Anatomy of Judgement. An investigation into the processes of perception and reasoning. London : Free Association Press.
Brehmer, B. (1991). Distributed decision making: Some notes on the literature. In J. Rasmussen, B. Brehmer, & J. Leplat (Eds.), Distributed decision making. Cognitive models for cooperative work (pp. 3–18). Chichester, UK: Wiley.
Collins, H.M. (1990). Artificial Experts. Social Knowledge and Intelligent Machines. Cambridge, MA: The MIT Press.
Dawydow, W.W. (1977). Arten der Verallgemeinerung im Unterricht [Types of generalization in teaching]. Berlin: Volk & Wissen.
Dreyfus, H.L., & Dreyfus, S.E. (1985). Mind over machine. New York: Macmillan/The Free Press.
Engeström, Y. (1989). Developing thinking at the changing work place: Toward a redefinition of expertise. San Diego, CA: University of California, Laboratory of Comparative Human Cognition.
Friedrich, P. (1992). Kompetensutveckling vid lokal teknikförändring: Operatörsarbete och datorstödd automatisering i verkstadsindustrin [Development of expertise on incremental changes of technology: Operator’s work and computerization in manufacturing industry]. Unpublished doctoral dissertation, The Royal Institute of Technology, Stockholm.
Galperin, P.J. (1979). Introduction to psychology (in Finnish). Helsinki: Kansankulttuuri..
Glaser, R., & Chi, M.T.H. (1988). Overview. In M.T.H. Chi, R. Glaser, & M.J. Farr (Eds.), The Nature of Expertise (pp xi-xxvi). Hillsdale, NJ: Lawrence Erlbaum Associates.
Hale, A.R., & Glendon, A. I. (1987). Individual behaviour in the control of danger. Amsterdam: Elsevier.
Hammond, K.R. (1981). Principles of organization in intuitive and analytical cognition (Tech. Rep. No. 231). Boulder: University of Colorado, Center for Research on Judgment and Policy.
Hukki, K., & Norros, L. (1993a, September). An orientation based approach to evaluation of process operators’ diagnostic activity. Paper presented at the Fourth European Meeting on Cognitive Science Approaches to Process Control. Hillerød, Denmark.
Hukki, K., & Norros, L. (1993b). Diagnostic orientation in control of disturbance situations. Ergonomics, 36, 1317–1328.
Hutchins, E. (1988, July). Organizing Work by Evolution. Paper presented at the conference Work and Communication. La Jolla, CA.
IAEA Safety Culture. (Safety Series No 75-INSAG-4). (1991). Vienna: IAEA.
Ilyenkov, E. (1977). Dialectical logic: Essays on its history and theory. Moscow: Progress.
Kjellen, U. (1987). Deviations and the feedback control of accidents. In J. Rasmussen, K. Duncan, & J. Leplat (Eds.), New technology and human error (pp. 143 - 156). Chichester, UK: Wiley.
Larkin, J., McDermot, J., Simon, D.P., & Simon, H.A. (1980). Expert and novice performance in solving physics problems. Science, 208, 1335–1342.
Lave, J., & Wenger, E. (1991). Situated learning. Legitimate peripheral participation. Cambridge: Cambridge University Press.
Lawrence, J.A. (1988). Expertise on the bench: Modelling magistrates’ judicial decision-making. In M.T.H. Chi, R., Glaser & M.J., Farr (Eds.), The nature of expertise (pp. 229–260). Hillsdale, NJ: Lawrence Erlbaum Associates.
Leontyev, A.N. (1981). Problems of the development of the mind. Moscow: Progress.
Munley, G.A., & Patrick, J. (1990, September). A strategy based approach to training transferable fault-finding skill. Operator training and aqcuisition of cognitive skills. Paper presented at the first meeting of the international network on Cognitive Approaches to Dynamic Environment Supervision. Cardiff, UK.
Norros, L. (1989a, October). Responsibility for system development as an element of process operators’ professional expertise. Paper presented at the Second European Meeting on Cognitive Science Approaches to Process Control. Siena, Italy.
Norros, L. (1989b). Simulation in industrial work training. In L. Bainbridge & A. Ruitz-Quintanilla (Eds.), Developing skills with information technology (pp. 315–328). Chichester, UK: Wiley.
Norros, L. (1991, July). Development of operators’ expertise in implementing new technologies. Work and Welfare. Paper presented at the Second Karlstad Symposium on Work. Karlstad, Sweden.
Norros, L., & Sammatti, P. (1986). Nuclear power plant operator errors during simulator training (Tech. Rep. No. 446). Espoo, Finland: Technical Research Centre of Finland.
Norros, L., Toikka, K., & Hyotylainen, R. (1988). Implementation of a FMS — Results of a case study. In J. Ranta & P. Huuhtanen (Eds.), Information technology and work environment part III. The Finish Work Environment Fund Publications A3 (pp. 139–189). Espoo, Finland: Technical Research Centre of Finland.
Olssen, S.E., & Rasmussen, J. (1989). The reflective expert and the prenovice: Notes on skill, rule and knowledge-based performance in the setting of instruction and training. In L. Bainbridge & A. Ruitz-Quintanilla (Eds.), Developing skills with information technology (pp. 9–34). Chichester, UK: Wiley.
Projektgruppe Automation und Qualifikation (PAQ) [Project group automation and expertise] (1987). Wiederspruche der Automationsarbeit. (Contradiction of automation work). Berlin: Argument Verlag.
Rasmussen, J. (1982). Human errors. A taxonomy for describing human malfunctions in industrial installations. Journal of Occupational Accidents, 4, 311–333.
Reason, J. (1987). A framework for classifying errors. In: Duncan, K., Leplat, J., and Rasmussen, J. (Eds), New technology and human error (pp. 5–14). Chichester, UK: Wiley.
Reiman, L., & Norros, L. (1994, March). Organizational assessment of a maintenance department at a nuclear power plant. Paper presented at PSAM-II:An international Conference Devoted to the Achievement of System-based methods for the Design and Operation of Technological Systems and Processes. San Diego, CA.
Schön, D. A. (1988). Educating the reflective practitioner. San Fransisco: Jossey Bass Publishers.
Spinoza, B. (1975) Ethik (J. Stem, Trans.). Leipzig: Verlag Philipp Reclam jun. (original work published 1677).
Suchman, L. (1987). Plans and situated action. The problem of human-machine communication. Cambridge: Cambridge University Press.
Toikka, K., Norros, L., Hyotylainen, R., & Kuivanen, R. (1991). Control of disturbances in Flexible Manufacturing (Tech. Rep. No. A14) (in Finnish). Tampere: Work Environment Fund Publications.
Wahlström, B., Norros, L., & Reiman, L. (1992, August). Human Factors Research in the Nuclear Power Field in Finland. Paper presented at the Fifth IEEE Conference on Human Factors in Power Plants. Monterrey, CA.
Zuboff, S. (1988). In the age of the smart machine. New York: Basic Books.
ACKNOWLEDGMENTS
Development of this approach has taken place at the Technical Research Centre of Finland in diverse research projects carried out in collaboration with Kristiina Hukki, Raimo Hyötyläinen and Kari Toikka. Also Lasse Reiman from The Finnish Centre for Radiation and Nuclear Safety, and Ari Kautto from Imatran Voima Power Company have contributed to the development of the approach.