Surveillance Mitigation: Identify a form of surveillance that you encounter in your everyday life

Warning Concerning Copyright Restrictions

The Copyright law of the United States (Title 17, United States Code) governs the making of photocopies or other reproductions of copyright material. Under certain conditions specified in the law, libraries and archives are authorized to furnish a photocopy or other reproduction. One of these specified conditions is that the photocopy or reproduction not be "used for any purposes other than private study, scholarship, or research." If a user makes a request for, or later uses, a photocopy or reproduction for purposes in excess of "fair use," that user may be liable for copyright infringement.

314 TECHNIQUE: THE MATERIAL DIMENSION OF POWER

CHAPTER NINE

THE INFORMAnON PANOPTICON

authority; he made abrupt and arbitrary decisions in order to instill fear and to test for obedience; he insisted that authority could be delegated for only a short period. The ruler might exclude officials from responsi- bility over areas in which they had personal ties, or even take their sons or'other relatives as courtiers, retainers, or hostages so that the official would fear for their safety.1 Managers, too, have used a wide variety of techniques to safeguard

their authority and increase the certainty of their control. Part 3 ex- plores how information technology malees itself available as such a means. First, chapter 9 illustrates how the informating capacity of the new technologies can heighten the transparency of workers' behavior as both the substance ofworle and the way it is performed become part of the new electronic text. In chapter lOwe see computer technology used as a communications medium by professionals and managers who unwittingly expose their daily dialogues to an unprecedented degree ofhierarchical scrutiny. Both chapters examine how managers discover and decide to exploit the new opportunities for control that are laid open by information technology and how the consequences of their choices tend to be manifested.

Home Returning to earth after his life ofweightlessness, the astronaut cannot lift the small bouquet of flowers the child gives him. He cannot raise his head off the pillow, pulled down by the gravityofa dream.

He remembers nothing, no sound. In the absolute zero ofdeep space, the wild poundingofhis baffied hean. He lifted abuilding in hand and apencil in the other. This was what he wanted: the world like a worn stone cast into the water. He wanted to break the promise of the body to the earth, to stop the long descent ofeveryone he loved under the ground. He wanted to rise, an angel in II paradise ofexact data.

He 'pills his milk on his shirt. The earth I has darkness, and then light. The earth has bird; bickering over the last seeds. His fork slips clattering on the plate. The road is shining. The magnolia is shameless in the rain.

,

-}OHNWrrrE

THE TEXTUALIZATION OF WORK BEHAVIOR

It was 3:00 A.M. when the phone rang in the home of Cedar Bluff's plant manager. The production coordinator was calling to infoI'IIJ. him that a tanlc had exploded-there was one minor injury. The night crew

As another manager put it:

The system fulfills an important need because our plant is so interde- pendent. I can't imagine managing without it. It captures history. If a screwup happens. you can go to the precise moment in time to see what happened. It helps you troubleshoot and coordinate data to figure out how to solve a problem.

Mostof the keypromotional criteria, including leadership in problem solving, use of discretionary time, training others, being a resource to peers,·correcting problems, understanding the interdependencies of the process, project work, and hard number results are all deriv- able, at least to some extent, from information in the Overview Sys- tem. The system gives you a larger and more objective data base for making technically related assessments ofperformance.

The Overview System had been designed for these purely technical purposes-to accumulate, monltor, and analyze comprehensive data reflecting each moment of the production process. Once the system was in place, managers discovered that another parallel history was also being captured. This was the history of human decisions at the data interface, the history of operators' attentional behavior, the history of operator probJem solving. Managers had learned to use the objective record as.a means of deducing this behavioral history. One manager explained:

" ,

When the tankblewup, managers turned to the Overview System, pour- ing through each minute of the last several days in order to see exactly h al had changed up until the moment of the exploSlon.w at process v ues

This info<mation in turn would enable them to deduce the cause: had some piece ofequipmentor instrumentation failed, had anoperator made a poor decision, or had an operator simply failed to take notice? . Some managers appreciated the system because it prOVIded a power-

ful tool for accelerating individualleaming and improving the plant's performance. It could help a manager determine who needed more training and in what areas; it could be a source of instructive feedback with which operators could refine their understanding.

The manager should use the data to say. "I've got some data for you, am 1off calibration or did you do something incorrectly?" The person would then receive the data and make the right move. The manager is motivated by the need to know what's going on in his orher area- that's their responsibility. The system is a way managers can learn what operators need without taking their confidence away.

That vision of a cool, unruilled, nonthreatening partnership in the data was not shared by all managers, however. Others recognized that although using the system as a learning tool sbunded like a idea, the path of least resistance tended point in another dIrectIOn.

With the Overview System, I can find out everything you did for what went on in your head. This can be good or bad. It becomes easier to catch people in judgment situations if you have to use as a basis. It makes feedback easier and more effective. someone makes an obvious mistake, the tendency is to feel that It 15 easier to fire them than to help train them.

As managers tried to unravel the tank catastrophe, operators waited uneasily, wondering if the llhard facts," the "straight scoop," the "real- time reality" available from the data captured in the Overview System would mean that one of them would soon be without a job.

The Inlormation ranoptlcon

the roaches scurry out. If the employee blows it, it's clear right You can know who didn't do what. If a person stops concentratlng, you know it from the data.

J.t.t:HNH.JUE: tHE MATERIAL VIMENSION OF POWER10

had begun cleaning and repairs, and production was certain to be affected for several days. The next day, managers began to unravel the causes of the explosion. To do this, they summoned up data from the Overview System, which had been built to read and record key instru- ment values throughout the plant (more than 2,500 pieces of data) every five seconds. These values could be organized by unit or summa- rized to provide a bird's-eye view of the total operation. The data were available on-line for three days and then were stored indefinitely. This meant that it Was not necessary to know in advance what data might be important or why; the data could be retrieved and analyzed at a later date, giving rise to ever new interpretative possibilities. People learned moreabout what they had been doing as they found new ways to engage with this voluminous history ofminute events.

This computer is like X-ray vision. It makes the people who operate the process very unfriendly. It's like lifting up the rock and letting

It is 8:00 A.M., and a craftsworker for Metro Tel reports to work on the outskirts of town in a fortress-like building that houses the

jiM 1ECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 319

telephone-switching circuits for a portion of the city. Not many years ago, entering the building meant reporting in to the foreman and ex- changinggreetingswith several othercraftsworkers. The foreman di.trib- uted assignments, pertinent information was discussed, and the crafts- wor:kers set to their tasks. Sometimes they would discover problem cir- cuits, and they would put aside their assignments in order to address the more urgent problem. At times, they sought out one another or the foreman to confer over a task, which they referred to as a "trouble." On this morning, the building is empty. The only sounds the craft.-

worker hears as he enters the work area are the low hum of the elec- tronic equipment and his own footsteps moving across the yellowed linoleum floor. He moves towar-d a computer terminal and enters his password, time, and location. Within seconds, the screen is filled with his assignments for the eightChour workday. The assigned tasks are listed in the order in which they are to be undertaken, and each task is accompanied by a "price"-the amount of time in which it is to be completed. As he finishes each task, the craftsworker checks in with the computer terminal to note the completed assignment and the amount of time it actually took to do the work. He also checks to see if his work load for the rest of the day has been altered in any way. If he was unable to solve a problem, he may indicate that the task was not completed, and it will be passed along automatically to a worker on the next shift. Sometimes a task will take longer to complete than its assigned time. Thismay have beenbecause the craftsworkerdawdled or because he simply wasn't able to figure out the trouble. Usually it is because a very.complicated assignment was underpriced. Once a price is assigned. by the system, it cannot be altered, except by a foreman who is identified by a special password. Managers in the central office want to know howworkers perform against the prices they are assigned. The system uses these ratios to compute "efficiency ratings" for each worker, ratings that are later used to evaluate performance and that become part of the managerial data base used for determining prices, assigning work loads, and judging organizational efficiency. Today this worker is concerned. Three of his assigned tasks were

complex and required more than the allotted time. Rather than finish the day with a poor efficiency rating, he decides to change the original prices designated for each of these jobs. He feels fortunate to know the foreman's password, which will allow him to enter the system and alter the prices; he wonders, briefly, if anyone will notice.

• • • An informating technology textualizes the objects, events and pro-

cesses that constitute an organization's work. The factories and offices described in the previous chapters have illustrated how this textualiza- tion process engenders new possibilities for the production and distri- bution of knowledge, and finally challenges a system of managerial authority that has depended on exclusive control of the organization's knowledge base. Conversely, managers can choose to design and deploy the technology in ways that reproduce old experiences of limited ac- cess to information. Because knowledge is becoming ever more depen- dent upon intellective mastery, this approach both preserves and ex- tends the distance that has separated workers from managers. In each of the organizations I studied, information technology had

textualized not only the content of work but also the task-related be- haviors of the men and women who engaged with the data interface. This parallel text presented managers with a new set of divergent choices. To what extent was it the occasion for organizational members to more fully their own reflections, learn from what they could now see, and increase their opportunities for autonomous action? To what extent would the text be treated as a technical convenience en- abling more assiduous behavioral control? Was the text to be the key- stone of a new learning environment, or would the increased, visibility of behavior be exploited in the service of managerial control as an antidote to the pressures ofuncertainty? I

PANOPTIC POWER AND THE DISCIPLINARY SOCIETY

Techniques ofcontrol in the workplace became increasingly important as the body became the central problemofproduction. The early indus- trial employers needed to regulate, direct, constrain, anchor, and chan- nel bodily energies for the purposes of sustained, often repetitive, pro- ductive activity. Still struggling to establish their legitimate authority, they invented techniques designed to control the laboring body. The French historian Michel Foucault has argued that these new techniques of industrial management laid the groundwork for a new kind of soci- ety, a "disciplinary society," one in which bodily discipline, regulation, and surveillance are taken for granted.

320 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 321

This type ofpower is in every aspect the antithesis of the mechanism of power which the theory of sovereignty described.. .. The latter is linked to a form of power that is exercised over the Earth and its products, much more than over human bodies and their operation. .The theory of sovereignty is something which refers to the displace- ment and appropriation on the part ofpower, not of time and labour, but ofgoods and wealth. It allows discontinuous obligations distrib- uted over time to be given legal expression but it does not allow for the codification of a continuous surveillance. It enables power to be founded in the physical existence of the sovereign, but not in continuous and permanent systemsof surveillance. ...This new type ofpower . .. which lies outside the form ofsovereignty, is disciplin- ary power. I

The Panopticon was an architectural innovation developed by the moral philosopher Jeremy Bentham. In Foucault's view, it is both a sign of and metaphor for this new disciplinary society. Bentham extolled the Panopticon as a technological triumph capable of exploiting the productive potential of those who had flouted, thwarted, or otherwise escaped social authority. His initial targets were convicts and paupers, though he did not fail to see the potential usefulness of thePanopticon for students, asylum inmates, and workers. The architectural plan conceived by Bentham was first constructed

by his brother Samuel, an engineer, in 1787 for a factory at Critchef in Russia. It was later adapted to the problems of housing convicts and paupers. The structure consisted of a twelve-sided polygon formed in iron and sheathed in glass in order to create the effect ofwhat Bentham called "universal transparency." A central tower, pierced with wide windows, opened onto the inner wall of the surrounding polygonal structure, which itself was divided into narrow cells extending across the width of the building. Each cell had a window on both the inner and outer walls, allowing light to cross it and illuminate all the inhabi- tants to an observer in the central tower, while that observer could not be seen from anyone of the cells. Mirrors were fixed around the tower to direct extra light into these apartments.

The external skin of small glass panes was held in a mesh of leadings, in a web of mullions and transoms, in a ladder of iron columns and lintels, without one inch of walling. ... The function of inspection was greatly enhanced by the use of these materials. With masonry. nothing like the same panorama could have been achieved.2

,

l

The Panopticon's genius was that illumination and visibility provided the possibility of total control.

Hence the major effect of the Panopticon: to induce on the inmate a state ofconscious and permanentvisibility that assures the automatic functioning of power. So to arrange things that the surveillance is permanent in its effects, even if it is discontinuous in its action; that the perfection of power should tend to render its actual exercise unnecessary; that this architectural apparatus should be a machine for creating and sustaining a power relation independent of the per- son who exercises it; in short, that the inmates should be caught up in a power situation of which they are themselves the bearers. ... Bentham laid down the principle that power should be visible and unverifiable. Visible: the inmate will constantly have before his eyes the tall outline of the central tower from which he is spied upon. Unverifiable: the inmate must never know whether he is being looked atat anyone moment; but he must be sure that he may always be so. ... In the peripheric ring, one is totally seen. without ever seeing; in the central tower, one sees everything without ever being seen.3

The Panopticon represents a form of power that displays itself auto- matically and continuously. Panoptic power lies in its material structure and presence; the Panopticon produces the twin possibilities of obser- vation and control. It functions independently of the observer in the central tower and is meant to function effectiveJ.y without an observing presence. It can achieve results independent of the motives, intentions, or ideologies of the observer. The allure of the panoptic world view is, above all, the promise of

certain knowledge based upon the totality ofobservation it affords. The psychological effects ofvisibility alone are enough to ensure appropri- ate conduct. Certainty can be achieved even without observational effort. "He who is subjected to a field of visibility,. and who knows it, assumes responsibility for the constraints ofpower; he makes them play spontaneously upon himself; he inscribes in himself the power relation in which he simultaneously plays both roles; he becomes the principle of his own subjection. By this very fact the external power may throw off its physical weight.... It is a perpetual victory that avoids any physical confrontation and which is always decided in advance. ,,4- Bentham's Panopticon relied upon the materials and techniques of

his day to create a structure that could autonomously reproduce an individualized sodal control, providing a central authority with certain

322 TECHNIQUE: THE MATERIAL DIMENSION OF POWEll The Information Panopticon JLJ

knowledge of an institutional population through the architectural in- vention of Huniversal Bentham's extensive plans for the reform of prison management created both controversy and interest within the British Parliament.5 Though his management proposals were not implemented, the central principle ofcontinuous observation made possible by technical arrangements was to influence the administrative and architectural orientation of bureaucratic organizations from schools, to hospitals, to workplaces in which individuals are taken up as unique problems to be managed and measured against appropriate norms:6 llPanopticisrn is the general principle of a new 'political anat.. amy' whose object and end are not the relations of sovereignty but the relations of discipline. .. . What are required are mechanisms that analyze distributions, gaps, series, combinations, and which use instru- ments that render visible, record, differentiate and compare.... It is polyvalent in its application.... Whenever one is dealing with a multi- plicity of individuals on whom a task or a particular fonn of behavior must be imposed, the panoptic schema may be used."7

THE PANOPTIC POWER OF INFORMATION TECHNOLOGY

InformationJptems that translate, record, and display human behavior can provjde the computer age version of universal transparency with a degree of illumination that would have exceeded even Bentham's most outlandish fantasies. Such systems can become information panopticons that, freed from the constraints of space and time, do not depend upon the physical arrangement of buildings or the laborious record keeping of industrial administration. They do not require the mutual presence of objects of observation. They do not even require the presence of an observer. Information systems can automatically and continuously record almost anything their designers want to capture, regardless of the specific intentions brought to the design process or the motives that guide data interpretation and utilization. The counterpart of the central tower is a video screen. The web of windows is replaced by procedures for data entry such as microprocessors built into operating equipment, or the control interfaces that record operator inputs at

Cedar Bluff, or the daily system updates provided by the craftsworkers in their remote field sites at Metro Tel. What is the allure of panoptic power? Why do some managers look

toward information technology as a way ofachieving control, when the same results might be accomplished by simply managing the authority relationship? The manager's world is an interpersonal vortex ofrelent- less demands upon personality, ego strength, and empathy. Managers face the continual problem of getting others to do what the organiza- tion requires ofthem, and the manager's action·centered talents in the service of actina-with are a crucial factor in determining the likelihood of success. With the professionalization of management, a vast body of literature has generated amountain of sometimes contradictory advice on howmanagers can cope most effectively with this central and often terrifying requirement of face-to-face engagement with an "other." Managers are exhorted to revamp their methods of communication, invite feedback, give feedback, listen, coach, facilitate, manage by ob- jectives, encourage autonomy, provide vision, et cetera. These demands are further complicated by the mutual dependency

of the authority relationship, which is based upon reciprocity and so requires mutually advantageous conduct. Mutual dependency derives from the fact that both the manager and the managed have the means with which to counter the behavior of the other. Reciprocal relations are notmechanical-they must be continually nurtured and cultivated if reciprOcity is to be adequately they are inherently uncertain because reciprocity also c,an be withheld. Managers can wish to avoid reciprocal relationships because of the psychological effort they require and the necessary ambiguity they entail. Under these con- ditions, any alternative that promises to achieve managerial objectives while sidestepping these uncertainties may be irresistibly attractive.8

Techniques of control and the panoptic power they convey offer one such alternative. Infonnation systems can alter many of the classic contingencies of the superior-subordinate relationship, providing cer- tain information about subordinates' behavior while eliminating the ne- cessity of face-to-face engagement. They can transmit the presence of the omniscient observer and so induce compliance without the messy conflict-prone exertions of reciprocal relations. As an informating technology textualizes a wide range ofworkplace

behaviors, new patterns of conduct and sensibility emerge from the heart of the panoptic vision-the secret comfort of the one-way

324 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 325

mirror, the yearning for omniscience in the face of uncertainty, the conformity-inducing power of involuntary display. The panoptic power of information systems was evident at both Cedar Bluffand Metro Tel, and the social-psychological dynamics unleashed by these information panopticons unfolded with clockwork consistency across each organization.

PanapricPawerat Cedar Blujf Though managers at Cedar Bluffwere frequently drawn to the advan-

tages of the information panopticon. the systems they relied upon to achieve universal transparency had not been developed for that purpose. For example, the Overview System received real-time data from the plantwide control system. It was referred to as a "snapshot of what's happening"-one glance at the summary screen and the operations of the entire plant werevisible. The system hadbeen developedas a solution to problems that were pragmatic, immediate, and technical: to achieve better process control, to achieve higher levels of reliability. and to have better infonnation for thepurposesofprocess optimizationand improve- ment. Once in place, these technical systems offered opportunities to better cope with some of the chronic exigencies of the managerial pre- dicament. particularly the felt need for certainty as a form of release from the personal risk, exertion, and dependency associated with the authority relationship. The techniques of panoptic power were typically developed after the fact, by coincidence and by accident, as managers discovered ways to colonize an already functioning portion of the tech- nological and use it to satisfy their felt needs for additional certainty and·control. The plant manager admitted:

Our information capability did not evolve this way conSciously. The evolution of the systemwas not seenas an opportunity to gather total information;- rather, it was seen as an opportunity to create better process control. We are a giant laboratoryi everything is new and up for grabs. We don't have the answers.9

To be sure, some managers saw the enhanced certaintyoftheir infor- mation about subordinates as an opportunity to accelerate learning and improve performance. They found that the data from the Overview System made it possible to survey a wide range ofbehavior for evidence of irregularities. To the extent that the sources of such irregularities could be discovered, the data could be used to "coach" subordinates

and to help them realign their behavior within acceptable standards. One manager recounted:

We have tremendous amounts of information about people in our computer, and there are positive reasons for it. It can be used for coaching. When a person goes to the storeroom, he will give his employee number when he takes something out. 1£ a person consis- tently seems to get the wrong work order number or charge to the wrong department, we can give them coaching as to why it's impor- tant to give the right data, how we are going to use that data, et cetera. Right now, we only know that certain individuals aren't giving the right coding inFormation, but we don't know the reason. ... Is it a problem in knowledge or'in execution? We don't know if they are trying to pervert the system or sabotage it, or just don't understand how the data can be used later. By finding out who the individuals are and who we need to coach, we stand abetter chance ofresolving the situation.

On the other hand, many ofCedar Bluff's managers were more inter- ested in the possibility ofobtaining foolproof data than in the opportu-, nity of using such data for coaching. In the context of a conventional technology. managers had to rely on their subordinates for valid infor- mation. The quality of the data they received was likely to be a function of the quality of the reciprocal relationship they had established. In contrast, an informating technology could pr.pvide valid information under any conditions.

When I managed in the older plant, ··we depended on our hourly people keeping logs. We had to rely on the operators' putting true numbers in the logs. The computer takes away this need to rely on someone's integrity. People were very apt to fudge the numbers to cover their own ass. Now we no longer have that problem because the computer doesn't lie, at least not intentionally.

This capability is more than just a matter of convenience. It alters familiar contingencies in the process of management and creates new ones. The computer can be less an aid to supervision than a substitute for the supervisory function. For example, in most organizations, termi- nation decisions are considered grave enough to require exhaustive documentation. It has become increasingly important for managers to document carefully the examples ofan employee's behavior that might warrant separation from the company, because supervisors can be ac-

326 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 327

cused of bias in their assessment. Such evidence might take a year or longer to compile. During that period, the employee would be likely to learn of the e/fort and perhaps even receive feedback that may help to im_prove his or her behavior. In the Cedar Bluff"laboratory," manag- ers discovered that the impartiality and detail of data from the Over- view System provided an ironclad basis for some difficult personnel decisions that in a conventional context would have taken months or years to execute. As the plant manager described it:

We have disciplined and terminated people based on information from the Overview System. It provided information on incidents which .showed that the individual was not performing to basic knowl- edge requirements. By recording what happens to all of the instru- mentation in a given part of the process on a five-second basis, we can see exactly what was done, what should have been done, and what was not done. Since you know the people that were there, you know what they did or did not perform. This becomes independently verifiable because the system knows it. You can take a piece of paper right out of the Overview System

and put it into someone's performance file, and this contributes to their promotion. It can be used for counseling. for guidance, for promotion, and for discipline. Without the information coming from the system, it would be difficult to have absolute proof ofpoor per- formance. If the poor performance had occurred on shift, maybe it would have been picked up. In general, it would have been deduced at some point through greater accumulation of incidents. Therefore, the termination might have occurred a year later because a case would have had to be built, and it would have cost the company a lot more. ..

Some managers who observed the tendency to use the Overview System as a substitute for personal supervision believed that it allowed managers to distance themselves from the kind of qualitative knowl- edge about one's subordinates that was so vital in maintaining good reciprocal relations.

If I didn't have the Overview System, 1would walk around and talk to people more. 1 would make more phone calls and digress, like asking someone about their family. I would be more interested in what people were thinking about and what stresses they were under. When I managed in another plant without it, I had a better feeling ofthe human dynamics. Now we have all the data, but we don't know why. The system can't give you the heartbeat of the plantj it puts you out oftouch.

They realized that the computer system could not tell why an opera- tor made an error or what failures of commitment or comprehension contributed to poor decisions and missed opportunities for process im- provement. Such insights could be gathered only in direct interaction. Whether managers experienced panoptic power as an unfortunate crutch or as an invaluable solution seemed to depend on their tolerance for uncertainty and the degree of confidence they felt in the interper- sonal skills that could assure reasonable levels of control through healthy reciprocal relations. Cedar Bluffwas an organization committed to participative manage-

ment "and thus was charaet,erized by a better-than-average ability to tolerate ambiguity. Still, the interdependency of the production pro- cess, the systemic and costly consequences ofhuman error, the respon- sibilities felt by managers to have the answers, and the psychological exertion of the face-to-face relationship conspired to make universal transparency irresistible for many of its managers. As a result, panoptic power was slowly transfonning the administrative assumptions and practices the managerial hierarchy.

Panaptic Pawer at Metra Tel Metro Tel was part ofa large and highly centralized telecommunica-

tions company. Both technology and work organization were standard- ized across all the firm's operating companie,. While many of the men and women at Metro Tel had worked together for as long as fifteen years, the lines between workers managers were firmly drawn. Workers were unionized, members of the Communications Workers of America. Though there was little overt hostility, subtle adversarial relations and mutual suspicion pervaded the organization. Some manag- ers admitted that evaluationsofworkers were biased according to "who we want to become part of our group." In their turn, many workers seemed intent on limiting their e/fort to fulfilling only the minimum standards ofadequacy. In Metro Tel, the drift toward increased reliance on panoptic power

was given a boost with the computer-based centralization of telephone- switching functions in the 1960s. Just as operators at Piney Wood had closely interacted with discrete pieces of equipment before the advent of centralized computer control, telephone-switching craftsworkers had worked in buildings that housed switches (switching stations) for a partic- ular sector of the city, and an individual had primary responsibility for a

328 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 329

small group of switches. Each craftsworker was a member of a crew as- signed to a single building and headed by a foreman. like the operators at Piney Wood, the telephone craftsworker's eyes, ears, and hands were essential to accomplishing good work. Some said that hearing was the mOst important ability of the craftsworker, that they could troubleshoot the switches based solely on what they heard. "The worst sound ofall in the switching office would be the sound of total silence. If you didn't hear a lot of c1ackitty clack, you knew something was gravely wrong." Other craftsworkers were known for their ability to "run their fingers over a bank ofswitches and feel the one switch that was off." The computerization of SWitching brought with it centralized con-

trol in switching control centers (SCCs). The SCCs were established to integrate monitoring and control of the switching functions from a number of electronic switching stations (ESSs) over an extended geo- graphical area. Information from the computerized switches was sent directly by computer to a central office to be analyzed by a second level of computers and monitored by a new breed of craftsworkers known as "analyzers." Centralization meant that instead of having crews as- signed to an ESS, it was now possible to dispatch individual craftswork- ers to a given ESS with instructions as to which switches were out of order and what remedies they required. Foremen, who constituted the first level of management, could now rotate between field and SCC assignments. Field foremen might supervise as many as seven ESSs, while center foremen supervised the analyzers and aSSigned work loads to craftsworkers in the field. The centra!izat10n of operations in the SCC created a series of per-

pleXing administrative problems. The SCCs were "drowning in paper- work" as they nowwere reqUired to administer a range ofcraftsworkers dispersed across a wide area with complicated assignments and follow- up procedures. With centralization, the methods of work administra- tion had far more specific and complex. Before, a foreman decided what had to be done in his or her building and assigned the tasks to acraftsworker. Now, analyzers used computerized information to determine the work load for each switching station. These tasks were then prioritized, and clerks, under the supervision of SCC foremen, were responsible for assigning them to a particular craftsworker. An elaborate pricing gUide, based on thousands of cases, specified how much time each job would require. Thus, computerization had meant an absolute increase in paperwork, as well as the centralization of these administrative functions in the SCC.

A second problem involved the difficulties of remote supervision. Craftsworkers were dispersed among a variety of locations, and they usually worked in isolation. A foreman could drive around to each of the buildings under his or her supervision, but this was the only means of determining what each worker was actually dOing. Did the worker report to the building? How efficiently was the worker accomplishing his or her work? What problems might require further effort? The only data on work load completion available in the SCC were the paper tickets submitted by each craftsworker at the close of the day. Bins corresponding to each ESS were crammed with these paper tickets, and in order to get an overvie:w of work in progress, it was necessary to spread out these bits of paper and "eyeball" them for a sense of task completion, outstanding work, priorities, overtime hours worked, scheduling, potential roadblocks, quality of troubleshooting, et cetera. Collapsing under the new paperwork demands and the logistics of

remote management, the SCCs clamored for a new kind of information system, one that would allow them to effectively centralize the admin- istration of,the work force, just as the SCC had centralized the adminis- tration of switching eqUipment. The system developers' response to this demand was dubbed the Work Force Supervisory System (WFSS). The WFSS integrated two major data bases: one that contained informa- tion on the work to be done in each ESS and a second that contained information on the work force available fO{ these assignments. The pricing guide was automated, so the system determined the hours and minutes necessary for each task. the integration of the two data bases, the computer system could respond to a work identification number with calculations on the amount of time the job would require and its priority status. Each job would then automatically be assigned to a worker according to an algorithm that created up to an eight-and- one-half-hour workday, while minimizing travel time. Each crafts- worker who reported to an ESS began the day logging On to the WFSS computer, which printed out the day's assignments, their sequence, and their price. Workers would then update the system as they completed their tasks. Any work not completed would be reloaded automatically to the next shift, with an increment in its priority rating. Foremen in the SCC could now review workers' progress at any time of day or night. This meant real-time monitoring of work behavior: "what is pending and what is in jeopardy. The computer is the foreman's eyes and ears." Managers were now able to "monitor things that they could not even dream of before." As one foreman put it:

We direct him on what to do and when to do it at all times. He doesn't do what he wants to do when he wants to do it.

The systemhadalso absorbed a number ofvital supervisory decisions. For example, the WFSS's designers described the days ofheated discus- sion that went into finding the algorithm that would automatically form a work load for the craftsworker. The data base was built to include infonnation on site location, travel time, and even the number of floors in each building. This made it possible to create work loads that mini- mized the amount of travel by car and by elevator during the workday. Decisions were made as to how jobs would be reloaded automatically if someone called in sick. Another important decision involved how to load the worker for an eight-hour day. Suppose that while loading a worker with the highest priority items, the computer had a choice of forming a seven-and-one-half or an eight-and-one-half-hour work load? Should it create a load for eight-and-one-half hours or scan the priority list until it found a less important half-hour job in order to total exactly eight hours? One of the system's designers described the decision-making process this way:

jj1330 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon

gencies meant that a work load would have to be adjusted during the shift, but such adjustments could be made only by the foremen in the see whose special password allowed him or her to enter the WFSS and alter the work assignment or its price. Computerization meant that the work itself-the condition of the

switches-had become transparent. The WFSS meant that workers' be- havior was nowalmost as visible as their work. Gone were the bins filled with paper "trouble tickets"; gone were the ledger books and files. All the information was "at your fingertips" in a moment. The foremen were full ofpraise:

Ican mow if the work is getting loaded properly and if the craft has done his job Tight. You can see roadblocks instantly. I can look each day and see everything that I have loaded, if it's done, if not, why not. Ican know everything in an instant.

It is beautiful now. I can track my people's work. All Ihave to do is type the craft's initials in and see how he is progressing and see what his total work load was. What is his productivity? Before, we had to judge pe,ople more on hearsay. Now we have it in black and white.

There were hundreds of gutsy human issues that had never been faced down before which were necessary to tackle in order to auto-' mate these decisions. You can't believe how many hours we argued about this. Finally, we reached a consensus. The idea was to drive the craftsperson as hard as possible. Instead of going down to the next lowest priority, take that one-hour job. This way, we end up with an eight-and-sme-half-hour load, but ifwe push them, they will do the work and probably do many of the jobs in less time than we have

Unfortunately, this process was completely undocumented. We have no record of what these issues were or why we made these decisions. When the people who made them leave, no one will know why we designed it this way.

When analyzers in the see sat down in front of their computer ter- minals, the data they viewed was received directly from the computer- based switching equipment in the field stations. When they noticed lIt.roubles" (problem switches requiring craft attention), these were re- corded as work requests and downloaded into the WFSS computer. The WFSS in tum transformed these work requests into work loads assigned to each craftsworker in the field. Work loads could be altered as often as fifteen minutes before the start of a new shift. Sometimes, emer-

The system automatically computed efficiency ratings for worker by comparing the amount of time it took to complete a. Job with the amount of time for which it had been priced. Such calculatIOns could have been generated only by acompute{ system, since "we could never have enough clerks to compute these kinds ofratios." The ratings were considered management information and were not accessible to the craftworkers. Many foremen against taking the ratings at face value. The most experienced craftsperson frequently was loaded with the most difficult assignments. That individual might show the worst efficiency rating Simply because his or her assignments were the most complex and required more than the assigned time. Some fore- men argued that the only effective way to evaluate their workers was to compare data from the WFSS to the output from the see computers.

By what he told me on the WFSS about how long it took him. what he did, I can look back at the machine and see what he did m order to best evaluate him. With these two pictures, one of the ma- chine and one of the man, I can know what he did, the exact time, how long it took. Everytime he does something to the machine, no matter what he does, I can see it and usually the reasons why. Even if Iwas out there visiting the buildings, Icould never see all of this.

333 TECHNIQUE: THE MATERIAL DIMENSION OF POWER

Other foremen were more cynical about the way in which WFSS was u'ed. There seemed to be something compelling about the "black- and-white" infonnation from the system. Because uncompleted work would be reassigned automatically, these foremen were less motivated to dig into background data in order to verify or understand the causes of the workers' behavior.

People see that someone did not do the job, but they would never go bade and see why; whereas if you were there in the building with the guy, you would know. No one seems to backtrack to find out why or when, because it will get done eventually and it is all in the computer. You tend to get away from questioning.

In fact, many of the foremen believed that the WFSS had usurped their role as the primary supervisors of the craftsworkers. There was less room for individually negotiated accommodations between foreman and craftsworker. Earlier reciprocities-you help me out now, and I'll take care ofyou later-had no place in a world ofunilateral computer- ized instructions.

Before, craft and management cared about each other if they could be useful to each other. Now the computer could care less about who you are. A person can be dealt with, but not a computer. 10

Just as the architecture of the Panopticon was deSigned to accom- plish supervision without the direct engagement of the observer and the observed, the architecture of the WFSS absorbed supervisory func- tions and diminished the necessity of, and the opportunities for, face- to-face engagement. Most foremen acknowledged that they sometimes found themselves "supervising" workers whom they had never seen. In these cases, their evaluations were based solely on data from the WFSS. The consequence, they felt, was a new impersonality and fonnaHty. Even foremen who had known many of their crew members for several years found that they now rarely interacted with their crew and knew far less about their personal problems and needs-precisely the sort of information upon which reciprocity depends.

In the old environment, the supervisors would arrive at the same building as the craft and greet you in the morning. He would be able to walk around and talk to you, and he would see how you were dOing. You had that camaraderie of communication through the vi.. sual. Now you are basically just dealing with computers and checking

I !

I I

I i I

The Information Panopticon

on people. I can't see my men in the field now. because I look through the computer. Ifsomeone has a problem, I can'twork it out with him like I used to. Does he need some time off? The computer doesn't know his problems. It doesn't want to know if his kid just passed away or if his wife has problems. Maybe I ne.ed to layoff him for a while and then later I'll know I can count on hIm.

The foreman's job had become one of monitoring the new text of worker behavior and responding to major gaps, inconsistencies, prob- lem" or evidence of lagging skill,.

The manager's job has had several components. There is the face part-giving orders,'direction, organizing. using the s authority. Then there is administration, and then developmg and evaluating people. A lot of the face-to-face stuff has been built the WFSS, which increases the administrationaspect. Even though It.' a face-to-face job, I don't have to do it face-to-face. The computer IS being my face. It's telling them what to do.

The computer is programmed to do all the thinking for me. : .. It figures out vacation time and scheduling. It calculates the a person gets paid for. The WFSS i.s not a tool. It has become my Job. Now I track their work to see what they are doing. Now I can see where there are problems and spend more time training people.

The computer has become an extension of the supervisor. It is the law now black and white. There is no question there about what you. , have to do.

Some foremen argued that laking that extra step to engage with the craftsworker in a more direct fashion would result in better perfor- mance. They spoke of the value of periodic site visits. These believed that it was important to maintain the personal relationshIps that would provide leverage for valuable reciprocities. Others stressed the importance of personal contact so that the would not feel like an anonymous number in an indifferent orgamzation.

I d n't have to see the people anymore, but I think if you disregard pe:ple, they will feel like just a number. Most people will rebel if they are made to feel that way. So I think. it's important to call them up once in a while and say, HHey, I've over all reports. and I see that you are doing a good Job. ThIS way, he won t think. all of his interaction with the computer is just wasted. At least you mak.e the person feel a little more connected to something. They know that someone is paying attention.

334

If you confront a guy, even if he's a poor worker, you become the bad guy. You are alone. They corne back at you with the union.

own peers or from other workers. The tendency, then, was to take advantage of the technical buffer as a means of exercising formal con- trol without risking the threat of ostracism associated with personal confrontation.

Foremen could afford to avoid such confrontations because they knew the system would automatically bump any uncompleted tasks along to the next worker. This meant that difficult assignments would repeatedly drift to the best workers, and poorer perfonners would be able to float through the system, at least for a while. The system could make low perfonners visible, but it could not ensure that they learned from that visibility or that organizational perfonnance benefitted from their transparency. For many foremen, transparency instead had be- come one way to avoid managing, even as it tightened the reins of technical control.

335The Information PanopticonTECHNIQUE: THE MATERIAL DIMENSION OF POWER

Despite these insights into the authority relationship thatpunctuated the foremen's reflections, most agreed that there was a distinctive shift

less interaction, less engagement, and more impersonally ad- mlnIstered relationships. Once, knowledge of workers' behavior de-

the interpersonal effort associated with maintaining the recIprOcItIes 0,£ the relationship through face-ta-face engage- ment.. admInIStration had increased the difficulty of direct supervIsIon, and .the of behavior accomplished by the WFSS now made It pOSSIble to circumvent interpersonal effort entirel

had become an option rather than a necessity. Under condIt.lons, many foremen seemed irresistibly drawn to the WFSS for a from the uncertainties and personal terrors associated with managing reciprocities. They felt more comfortable knowing that they could rely on the computer to give orders.

Back in the old days, if a guy says he didn't want to do somethin you had to start trading favors with him. Now it doesn't work w.ay. The computer says, "Bingo-you got this assignment Tuesday night. Here is the information, do itl"

There were still cases when people failed to fulfill th' .e InstructIons passed on to them through WFSS. Managers then had to decide whether to confront these individuals or avoid the confrontation letting th slide by until so much objective documentation had that the .union have to respond to management's complain:s. One to aVOld confrontation was that a subordinate with an a _ gresslVe pe.rsorlality could make life unpleasant. g

There are a lot of people we do not want to confront. If it is the kind :fperson may give you a hard time, you can just ignore it. The omputer IS such a buffer to me as a manager that I can call up the

and deal with them. If the others do not do their work, I Will know it from the computer. I can go around their back and write things about them and turn them in and never reall have to confront them. y

Others wanted to avoid potential conflicts that might lead to union flare-ups or grievances. They complained that it was not worth the aggravation to go after the shirkers, because they Were Sure to accuse management of harassment and invite union involvement. When that happened, a foreman could not be certain of support from his or her

If I don't confront a man because he didn't do the job, who is going to know? The men lcnow among themselves, and they become dis- gruntled. Yet it's a big game with the computer between us. You can keep pushing the work to the people who will do it. The other group, they can just keepgoing through the computer, not being confronted by anyone. No questions are asked. The top people stay at the top of the rating, and the bottom never moves. People are protected by the system because nobodywants to tonfront them. You lcnowwho really benefits from it? Nobody. That man is allowed to do what he wants, and I am allowed to not do my job.

System designers were aware that the WFSS made it possible for foremen to avoid face-to-face conflict with their craftsworkers. They were proud of having designed the system to accomplish just that. For example, the union limited the degree to which foremen could specify which workerwould receive which assignment. The goal was to protect the top workers from always beingassigned the most complicated tasks. Designers had programmed the WFSS so that each task was categorized according to its level ofcomplexity and each worker was given a corre- sponding designation based on assessments of his or her ability. The system then automatically matched difficult tasks with excellent work- ers. Workers receiving these assignments had no way of knowing that

1ECHNIQUE: THE MATERIAL DIMENSION OF POWER

the program was biased in its method of distributing the work. As one designer explained:

We have been able to use the computer this way with a high degree of refinement. We can have any number of designations we please which really works Out to being able to have a designation for person ifwe want to. This allows Us to get around our problems with the union.

Another designer recognized that programming choices were made that would have increased the potential for conflict in a face-ta-face situation. The designers felt comfortable maJcing choices that "pushed the craftsworker harder than usual," knowing that the technical buffer would muffle conflicts that might have erupted in a more personal at- mosphere. For example, one designer explained how he programmed the computer to load a worker for up to eight-and-one-half hours.

There might be a different set of algorithms that we would have developedon the basis ofpeople-to-people interaction than the ones we developed on the basis of envisioning computer-ta-people inter- action. Without the computer, we might not have deliberately de- cided to load eight-and-one-half-hours' worth of work, because if you did that, the craft might caB you up and yell at you. With the

you can avoid friction. It doesn't go away, even if you Ignore it-it will keep banging away, day in and day oUt. We knew they could not yell at the computer.

In the inforlnation panopticon, many managers are tempted to cir- Cumvent the exertionsofthe authorit)' relationship-the responsibility for decisions that help shape the work life of subordinates, the perpet- ual effort required to nurture reciprocity. the emotional demands of face-to-face engagementwith its attendant riskofconflict and confron- tation. They seek other meansof increasing the certainty that subordi- nates will produce desired behaviors. In Metro Tel, many managers relied on the technology to make behavior transparent in the belief that people are more likely to do what they are told when they know their actions will be translated instantly and displayed as electronic text. That behaVioral text then becomes incontrovertible proofofobe- dience or neglect; this kind of proof can overpower reciprocities and bolster the manager's claim to control.

Under these circumstances, managers confront the electronic text

The Information Panopticon

not only in the service of their substantive tasks but also as a medium of administration. Intellective skill demands appear to crowd out an earlier reliance on action-centered skill, as managers must interpret behavior from the text rather than observe it directly. Managers at Cedar Bluff and at Metro Tel had become adept at using the data from the behavioral text to make inferences and informed evaluations. Their oral culture, highly charged with attractions and antagonisms, seemed to be giving way to a different kind of workplace. As one manager at Metro Tel described it:

Thereare fewerargumentsnow, hardly any at all. There is less friend- ship, too. We were once atight little group. Now we aremechanized. It is impersonal now. You are no longer close to the people you work closely with.

Still, the textualization ofbehavior has raised some intriguing ques- tions that have not yet been answered sufficiently. There were at least some managers who saw the potential of the behavioral text as an in- strument of learning and improvement. What are the organizational

under which that potential might be exploited? The answer requires a closer examination of the information panopticon. What is the extent of the transparency it promises? Does it in truth accomplish its intended effect-to induce conformity with the foreknowledge of certain and involuntary self-display? ,

HIERARCHY IN THE INFORMATION PANOPTICON

One way that the information panopticon departs from Bentham's prin- ciples is that it is hierarchically organized. At every level of the organi- zation, the observer is as likely to be a target of technical control as its vehicle. Observation is double-edged: observer and observed are enmeshed in one body but are of two minds. 1n the American Paper Company, networks of information systems

linking plants to divisional and corporate headquarters were in the early stages of implementation. Many corporate executives had yet to be- come involved in using information systems, but there were already

JJU J. J:LHNfllUI:i: I HE MATERIAL DIMENSION OF POWER The Information Panopticon 339

significant numbers of executives who enthusiastically anticipated ac- cess to a real-time data base that would reflect the organization's entire business. One of the top leaders in APe's corporate managementvoiced his expectations for the development of the corporate information sys- tem"capability:

My vision is that one wall of my office wiH be a screen. I can hit buttons and see my Teports or any other data I want. The data base will integrate the entire organization, and all the data will be in agreement. We'll all have the same numbers. I'll be able to manipu- late the data, and much of it will be of a quality that doesn't yet exist. We must have the whole company looking at data the same way. I want the president of the company to have a screen on his wall. We should be able to look at the data on a basis, and the screen should be continuously updated. That way, we can see how different businesses are doing and who needs help.

Work. waS under way to integrate divisional offices with plant opera- tions through real-time information systems. One divisional manager explained how he envisioned utilizing information systems to improve management results:

With the new integrated systems we are developing, I can push the button and in one minute see which of the plant managers has the situation in control. I can manage by exception. I anticipate a Screen on the plant manager's desk that flashes when something goes out of control or flags exceptions at the end ofeach day. The systems allow us to supplement all the bullshit with the real thing. I can call up some screens.end see where something is wrong and walk right to that place. Then I can talk from real data, not just things that they are dishing up. I can use that to coach them in how they are managing their business. It gives you that check at the bottom level-are they doing what they should bel

Throughout 'the division, plant managers discussed their concerns about becoming transparent to their divisional hierarchy, even as divi- sional executives seemed insulated from the mounting anxiety. A se- nior manager from corporate headquarters who had been tracking the controversy described it this way:

The plant managers are worried. They know the divisional executives want to have total on-line real-time access to the raw data from the mills, and they don't want this to happen. They have jokingJy tried

to communicate this to the divisional vice president, but he hasn't heard it. There are others trying to head it off, but the push for it is very strong.

Early in the implementation process, Cedar Bluff's plant manager, the same man who coolly admired the Overview System's capacity to increase the objectivity of performance evaluation, spoke with disdain for and suspicion of the plans for a divisional information system:

They want to have access to real-time data from the mills. But as plant !"anager, I do not want to be I don't ,;ant them breathing down my' neck or wasting their tIme on the mtty- gritty from the shop floor. Right now, I have flexibility.in. terms of deciding what data they see, when they see It, and how It IS pre- sented. If I lose control over that, it is an important piece of my job over which I will have less control.

In the course of a long discussion, he began to appreciate the irony of his case. He discovered the extent to which he was animated by the same fears'lirritations, and mistrust that characterized many members of the work force. With a wide grin and nervous laughter, he let me know that he was aware of the incongruity and apparent hypocrisy of his position. In a tone of self-mockery, he reflected:

Well, of course, when my people say it, it matter, but when I say it, it really is a problem.

, As the first stages of implementation were completed, the new divi-

sional capability began to alter the plant manager's daily life:

When the vice president calls me at 9 A.M. from his office over five hundred miles away to ask what happened last night because he has spotted something irregular in the data, I have to have the answers for him. This means that I must get in and ask the questions earlier. He would not be calling if it were not for the information power we have-instant access. The mere presence of the information makes you ask the questions.

Some months later, he discussed Cedar Bluff's new Data Access Sys- tem. He spoke with enthusiasm of the access he would have to an integrated data base reflecting the totality of plant operations and busi- ness functions. He looked forward to programming the system to flag deviations from the strategic plan, to check budget expenditures

340 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 341

against the forecast and to portray unit productivity, reliability, absen- teeism, overtime, et cetera.

I expect the information will let me keep closer track of the units and will mean that Iwill do more following up with my subordinates. It will mean additional pressure on them. I plan to use the data to track my managers. My belief is that this will push them to become more computer literate, which will become increasingly crucial to our success.

Panoptic power had also found its way through every level of the Metro Tel organization. While foremen observed their craftsworkers, the second level ofmanagement used the WFSS asa means ofmeasuring and evaluating foremen, particularly those foremen assigned to the field. During the earlier phase of centralization, which had culminated in the see organization, a rivalry had developed between the foremen who rotated among field sites and those who were assigned to the sec. Work load assignments were supposed to originate in the see and could be altered only by ansee foreman. However, "the field foreman didn't like not being the one who told the guys what to do, because that meant he was no longer thought of as 'The Boss.' " Without that authority, field foremen had lost the basis for reciprocal relations with their craftsworkers. They responded to this situation by frequently "breaking the load" that had been assigned from the central office. Upper levels of see management considered this one of the biggest obstacles to efficient operations. The WFSS, which pro- vided a recqrd ofhow many times a load was broken, was seen as a way to combat the problem. Because changing a load required entering a password, it was easy to determine whether the load had been changed from the central office or by a field foreman. These data were now considered an.important indicatorof the quality ofwork force manage- ment in a switching control center and were monitored by upper levels of management. Metro Tel's second-level managers relied on these numbers, noting that "now we can see what is going on." They also insisted that their own managers should have no use for the WFSS and the varieties of data it provided:

I don't think my boss should get involved with it. ... If anything is wrong, he can come and a.sk me. I don't think he should be involved with the nitty-gritty of running the job on my level.

Despite their concerns, the third-level manager was an avid user of the WFSS. He claimed to want a terminal on his desk in order to be able to sit in his office and punch up data that would indicate how the work in the see was being performed. The principal system designer had agreed to collaborate with him. As he put it, "We took care of each other. We had a secret pact, and we worked together to under- stand what those measures meant." The systems designer described the various ways that the third-level manager had learned to use

the system:

He was after the Playboy magazines and the sleeping bags-all the guys who take advantage of the night shift in our remote locations. He would compare the night- and day-shlft statistics and try to use those to get the night people to do more work. He would use the data to look for the broken load and go right to that field foreman and ask for an explanation. He used the measures to look at the analyzer's job-that is, the person who spots troubles on the SCC screen and creates work requests for the craft. Ifyou find a group of workers who say they can't find the trouble they are supposed to fix, and they are all getting their requests from the same analyzer, then we have to ask, does this person know his job?

Part of the collaborative agreement between this manager and the systems designer included protecting the manager from scrutiny at higher levels: "We agreed that the work b\'ing tracked by the WFSS would not go any higher than the third level. I didn'twant the computer to show that he was doing a crappy.. job. He promised never to bad- mouth WFSS, because it was my system." Still, it seemed only a matter of time before higher levels of Metro Tel management would have access to similar forms of data. Their interest was not in measuring craftsworkers or foremen but in developing aggregate indicators of the effectiveness of third- and fourth-level management. Another system designer described the growing pressure for such data:

The vice presidents of Metro Tel see they have an opportunity to increase levels ofmeasurement of the SCC. They want us to add new dimensions to the WFSS so that they can measure on a district or regional basis and assess the effectiveness of managerial decisions at the third- and fourth-level and potentiaUy right on up the hierarchy. This introduces a whole new set of terrors.

IHE MATERIAL DIMENSION OF POWER Ihe Inlormation Panopticon 343

THE PSYCHOLOGY OF VISIBILITY

At each level of these organizations, people searched for ways to adapt to the intense illumination of the information panopticon. There was less room for interpersonal ploys or negotiated solutions and fewer hidden comers in which to enjoy anonymous repose or indulge in overtly negative behavior. To some, it seemed that only the human heart retained its privacy, out of reach and recalcitrant. Cedar Bluff's operators recognized the Overview System's supervi-

sory power and tried to adapt themselves to the fact of involuntary display.

Managers use the system to see -how we are dOing. It's a good thing, unless you are trying to hide something. You can't hide mess-ups from the There are no fibs with this system.

When I do something, it shows up on the Overview System so I have think .about what I tel1 the boss. The system can give manager

infOrmatIOn of anything that happened. Usually there is no punish- ment. We have a lot of pride, too, because they can also see if you straightened something out.

There are only a few standard operating procedures we can cheat on because it can't be traced. When they can't trace it, we feel more freedom.

Some operators had discovered that the truth value ofthe data could also work in their.... favor. There were examples of operators using the data from the. Overview System to prove to managers that they were not to blame for a particular problem. There were other cases where managers used the data to protect their own subordinates from being blamed for an error. Finally, operators wondered how far managers would go in their reliance upon the Overview System as the basis for decisions regarding how rewards and punishments should be distrib- uted. Would the data be used as a guide to termination, promotion, or pay? Would this lead to an unbearable work environment or a more just organization?

Maintaining reciprocal relations also required psychological exertion on the worker's part. Some operators hoped that the new data-rich environment would remove the burden of interpersonal effort, substi- tuting objective facts for political skills.

There must be a reason for them to save all of this data, but I am not sure what it is. As the plant begins to run more smoothly, I think managers will start to use the data more. They will get to managing at agreater level ofdetail because ofthe Overview System. Will they start to manage against the system? How long will it be before they start to use it for performance verification? If they do, some people will probably do better as judged by objective data instead ofhaving to excel at the political stuffofgetting along and getting over.

As the operators pondered the implications of universal transpar- ency, Cedar Bluff's head of systems development was grappling with the prospect of the plant's sudden visibility to divisional management.

We don't want them to second-guess our minute-to-minute deci- sions. We want them to evaluate us based on our ability to meet our monthly goals. So our concern is that they will be on our backs and we will all end up with ulcers.

His search for a solution ricocheted back and forth between the two dimensions qf power-authority and technique. First, he appealed to the inherent rationality of authority and the covenant implied in the structure of imperative control.

Those at higher levels in the hierarchy know their jobs, and they know that we are doing our job. They shoulc¥t't be spending their time on these data, and I think they know that. I am sure these issues canbe worked through with objectivjty.

According to this view, managers would be expected to rely on the binding power of imperative control rather than on their capacity to view subordinates in an ever-present electronic text. It assumes that decency and rationality will prevail among those who are joined by a common faith in the values that rank them and that ensure their mutual dependency. Yet, as the discussion unfolded, this manager revealed his own misgivings about the civility with which the issue would be han- dled. "You have to deal with the politics of it too," he added confes- sionally. "We will have to push back all the way up the hierarchy." What is the method for "pushing back"? His second solution was a technological one: the plant's systems engineers would make sure that the plant retained ultimate control over divisional managers' access to data.

VlAJhKIAL UIMENSION OF YOWER.

res OK for them to have an on-line capability if we control what they see. All they will have is a terminal designed for the purpose of receiving what we release to them. They will not have a keyboard. If they ask for more, then we tell them we don't have the technical capability to do what you asked for. Our technological solution in- cludes the idea that we will transmit to them, and they will never request data from us. This way, we can control what they look at, and they could not just come in and snoop.

This example illustrates how a lack ofconfidence in the sharedvalues of the authority relationship can lead to a new reliance on technical means of promoting group interests. Once superiors reveal their lack of trust in the power of legitimate authority to guide behavior, then subordinates are alerted to the danger of a broken covenant. When superiors betray a tenuous faith in the hierarchy, then subordinates will develop means ofself-protection. When superiors doubt their own legitimacy and so tum to technique as a means of enforcement, subor- dinates, too, cast about for "extralegitimate" techniques ofdefense. The adversarial vocabulary of"us" and "them" invaded the language

of operators toward their managers, and plant managers toward divi- sional executives. This mistrust was not rooted in a perception of evil or malicious intent. It was, instead, the feeling. evoked in the silent dance of the observer and the observed. To be visible in this way evokes a sense of vulnerability and powerlessness. The person observed begins to wonder, "Am I exposed in some way that Iwould not choose to be? How can I be certain about precisely what I have exposed? What is it that they might see?", The resist:mce to such exposure reflects in part an effort to retain a

sense of self-control and to avoid feelings of shame. What is shame and how is it activated in this context? The developmental psychologist Erik Erikson has described shame in the following way:

Shame supposes that one is completely exposed and conscious of being looked at-in a word, self-conscious. One is visible and not ready to be visible; that is why in dreams of shame we are stared at in acondition of incomplete dress, in night attire, "with one's pants down." Shame is early expressed in an impulse to bury one's face out of sight, or to sink, right then and there, into the ground. I I

One way to minimize the risk of shame would be to look for ways to circumvent the observer, to thwart the power of the panopticon.

The Information Panopticon

This motivates managers who seek a technological solution that will ensure some measure of shadow, of privacy, and thus of self-control, in the increasing glare of the information panopticon. These feelings rarely surface directly. Instead, they must be read

from tactics, innuendo, and humor. For example, a small group ofengi- neers at Cedar Bluffmet to discuss a new automation project. The con- versation proceeded in an entirely "rational" mode until the subject of real-time divisional access to the new automated data base was raised. The mood in the group shifted abruptly. The men began to giggle and to send one another sidelong glances, like schoolchildren who had hid- den a frog under the teacher> desk. They attempted a few statements that paralleled the head of systems development's earlier comments on the rationality of authority. Suddenly there was a crack in the facade. One member of the group looked around and, with a sly smile, an- nounced, "Look, if it gets to the point that they really want access, we'll just make sure it doesn't work. Right, fellas?" The group's response was immediate and unequivocal-"Rightl" At that moment an awareness of the feelings they had just revealed flashed across their faces, and as if choreographed, they returned just as abruptly to an emotionless discussion of their engineering plans.

Too much shaming does not result in a sense of propriety but in a secret determination to try to get away with things when unseen.

12

, There is another tactic for avoiding,the experience of shame. The

behavioral expectationsof the observer can be so keenly anticipated by the observed that the foreknowledge of visibility is enough to induce confonnity to those nonnative standards. Anticipatory confonnity is a tactic for avoiding the dread associated with the possibility of shame. It accepts visibility and adapts to it by producing behavior that mini- mizes the risk of unwanted discovery. As the immediate environment is saturated with measurement, the pressure ofvisibility begins to reor- ganize behavior at its source, shaping it in confonnity with the nonna- tive standards of the observer. On one pulp-drying machine, a micro- processor recorded each time there was a break in the sheet of pulp. The operators discovered that if the machine could be brought up in less than thirty seconds, the break would not register. As a group of managers recalled:

,.,." 1ECHNIQUE: IHE MATERIAL DIMENSION OF POWER The Information Panopticon 347 The operators began to really race to beat the clocK. They started to beat it regularly, so we changed the programming to give them, only 6fteen seconds. We found out that when they tried to beat the thirty seconds, they were getting too sloppy and the overall number of brealsincreased.

An operator on the pulp-drying machine explained:

With this infonnation on the computer, there has been a psychologi- cal effect. We know there is something that will tell on us exactly. We can't fudge it now, so we hustle more.

Though Cedar Bluff was struggling with the challenge to managerial authority provoked by the informating process, its commitment to team-based participativemanagement within the hourly work force and its highly informated task environment had generated a policy of open access to a wide range of operating data. Managers used the Overview System to monitor their workers, but the workers had access to these data as well. The plant manager believed that universal access would create joint ownership of the data and thus joint responsibility for the behavior it reflected.

As long as operators get more information, too, why should they care that I have the data? They should expect we will question them for purposes of improvement. It is not my data. It is our data, all of it. It is OK for you in the pulping unit to question someone in woodyard. We want to build in a healthy contesting nature. If it is a legitimate business question in an area that you have some control over, you have to respond to questions, no matter who puts it to you.... Lead- ership-should be functional, not hierarchical.

In such a view, anticipatory conformity is not only a response to the threat of being shamed by the boss but also a method for managing one's appearance to peers and minimizing the pressure they are likely to exert. That pressure derives in part from internaliZing management's goals and values, and it is fueled by the functional interdependence of production units. No single group can fulfill the normative standard independent of the others; mutual visibility becomes an additional mechanism for ensuring the viability of that interdependence. Univer- sal access, as a correlate of universal transparency, diminishes the feel- ing of oppressive surveillance. Many of Cedar Bluff's operators voiced feelings that did indeed reflect a sense of collective responsibility:

Everything that anyone does is documented. Anything you do that interfaces with the equipment is recorded. It is our data, even if we don't have a need for most of it or wouldn't even know how to use it, Because we all have access, it is like we are all watching each other, not just like management watching us.

With these systems, there is no doubt. The results are the truth. They bring the truth to management. This means managers c:an really see what is happening, and they have to buckle down and focus on prob- lems. It creates a joint awareness. We end up working together more than fighting over what really happened.

considerable evidence that operators used their access as a way ofasserting control within their own spheres of influence. When they observed peers making mistakes, they felt the right and the obliga- tion to respond.

One time, a guy from another area looked at this screen and caught a problem and told me about it. Ifanother group acts crazy, we can see what's happening right on our screen. You can tell what your buddies are doing right quick. Having the system like this makes for better communication. We use the screens and that eliminates word of mouth. You don't have to have a conversation; you just pull up the screen. It's a way we provide each otherwith checks and balances.

Though many of Cedar Bluff's managers were determined to limit divisional access to their data, they also reaated differently when data was treated as a communal resource. They believed that the truth value of the data tended to reduce the arr/ount ofbickering involved in estab- lishing the facts and to eliminate the feelings of resentment or frustra- tion that typically followed such disagreements. The information sys- tem was experienced as putting the facts on a higher plane. Cedar Bluff's plant manager claimed:

In those areas of the plant that are most highly penetrated with auto- mated information, there are fewer interpersonal problems. Those areas that don't have it complain about management judgment. Where I have less objective data, Thave to deal with emotions. With data, I am able to get around the emotions. In those areas with less data, there is only individual subjective judgment, and they have a more difficult time resolving their differences.

As their traditional forms of influence began to languish, subjected to increasingly closer scrutiny and measurement, and faced with a

.lHb MATERIAL lJlMENSION OF POWER The Intormation Panopticon 349

growing sense of the ineffectiveness of their functional roles, many managers found the new data-rich environment to be a humbling expe- rience. Humility was a word repeated countless times to express the sense of muted subjectivity engendered by the presence of so much

information. It comes from the Latin humus, meaning soil, and it expresses the sense of a new groundedness and realism, a fall from a grace that had drawn upon faith in the rightness and necessity of managerial control. The information panopticon tended to expose the managers' territory, preempting opportunities to fashion his or her own methods of illumination and thus construct a portrait of "what is." The panopticon also enabled managers to see more of the processes and behaviors that affected their areas, without necessarily making it any easier to influence or control those events. The heightened aware- ness of these contingencies increased the complexity of managerial data, while the skills they had developed, the constraints imposed by the definition of their roles, and the criteria by which they continued to be evaluated limited their capacity to respond. These parallel experi- ences-a felt loss of control and a heightened awareness of complex- ity-often left a manager feeling more diminished than augmented.

If you assume you don't know all the right answers, you can handle it. But if you like to control, then it is frustrating. Data opens up the organization. Everyone has to be more humble, modest, open.

The role requirements associated with managerial authority have de- manded that managers have the answers and be in control. The spiraling complexity and insense illumination afforded by the information pres- ence conspire ...to create new contingencies for influence and success. In the midst of transition, these contingencies are only dimly under- stood, and the skills, roles, and structures that would define the new middle-management function are not yet apparent. Like the operators they are supposed to direct, many of these managers felt robbed of the familiar ways in which they have experienced mastery, with little to embrace as a substitute. This feeling ofdiminished importance may be a necessary prelude to

a new kind ofaugmentation. The yearning for omniscience that exploits technology to create the information panopticon represents a craving for truth as much as a need for more predictability and control. The all-seeing eye is not satisfied with a welter of detail; rather, those who observe usually believe that out of the particularity of real-time reality

the contours ofsome essential truth will emerge. Given the right meth- ods ofobservation, the right way to bow, the data can be made to open up their center, the still point where the one best way, the summary of all that is rational, will be revealed. Humility is the price one pays for revelation. The individual is weaker than the force of truth that shines through the facts. If one is willing to endure temporary humiliation, then the prize may be a truth more vast and comprehensive than one individual's efforts could ever obtain. Many of Cedar Bluff's managers expressed this reverence for the truth value of their data.

With the kind of data base we are developing, the problems should jump out at you from the data base, not from someone jumping on you after the problem has occurred. When everybody looks at the data base, it should be obvious-data becomes the focus, not the people. People hear words differently, but with the right data, we have something objective beyond what each person hears.

With all of this data, you can be more objective and less subjective. You don't have to debate; you can see what is true.

This perception of the truth value conveyed in the information sys- tem brings organization members together as brethren in the data. The world that is disclosed comes closer to Frederick Taylor's dream of mutuality born of a perfect science than anything the time-study ex- perts ever conceived. The information creates the fantasy of a world that is not only transparent but also shorn of the conflict associated with subjective opinion. Many,who now dwell in that trans- parency suppose that reasonable individuals confronting the same data will find the same truth revealed, and that in this way the system will eliminate disagreement about "what is." Their attitudes betray an ur- gent desire for light without heat. Thus rationalized, human energy can be turned toward the implementation of solutions whose appropriate- ness is amply demonstrated by the "facts." The powerful, conformity-inducing effects of shared information

had not been lost on those divisional executives most actively involved in supporting new information systems to link the plant with divisional headquarters. As they developed their plans for the future design and utilization of such systems, several executives recognized the technol- ogyas an opportunity to increase conformity to management's expecta- tions. Individuals would learn to respond to the silent, calm, steady, and unalterable "information presence" once they understood the in-

350 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 351

evitability of their own transparency. In order to achieve this effect. the same data would have to be equally accessible to each level of the organization. In this vision, the data base would become a shared reality-the grounds upon which to establish consensus as to "what is re'ally going on. II The system would communicate senior management's expectations clear down to the shop /Ioor. This could be accomplished by establishing parameters, to define the desired range for various out- comes, and by designing the information system to flag deviations from these parameters. Everyone with access to the system would be able to identify the sources of deviance, whether human or mechanical. Individuals would be able to see their own behavior reflected in the system, while knowing that others (peers, subordinates, and superiors) could see it as well. The system would also include industry data in order that everyone might share the frame of reference from which management's expectations were derived. In this way, the real-time system could both communicate the normative standard and measure its fulfillment. As one divisional executive put it:

The plants will be able to see industry data, and this will make me believable. I will have put it at their fingertips and said, "Here is the basis for my expectations." They will know what we know. They will be able to judge themselves in real time, see their record compared to industry and the real-time impact of their actions. You set a norm for what your expectations are, and then they know what is acceptable.

This tacti<; relies upon individual internalization of the evaluation process. People would be able to see where they fell shorr of the norm and would know that everyone else could see it as well. The potential for experiencing shame under these conditions would fuel individual energies toward anticipatory conformity. Executives discussed this dy- namic in the language of "exception management." The system could be used to flag exceptions, activities that fell too far outside the norm, and to focus the organization's efforts on those problem areas in order to bring them into aHgnment with corporate expectations.

The system will make different levels ofexception management pOSe sible. As an organization, we will be able to target the exceptions and work on where the pain is. These exceptions might be at the indlvidual, group, or plant level. The information system will force awareness ofexceptions down into the organization.

The knowledge that one could show up as an "exception" was consid- ered by this executive to be an important motivator in ensuring antici- patory conformity. The beauty of the arrangement lay in its bloodless automaticity.

From my desk I can look at any plant. They all know what I am looking at. They can see itl too, and how they stack up. They will compare themselves internally, and this creates a certain level of competition. It can e>l:tend all the way down to the individual. The hourly worker will know that if this exception goes up the line, it will come right back to him. Once he learns that he is going to be questioned on it, he will his own responses more effectively. A worker under these conditions does not need to be controlled- you simply e>l:pect him to respond to the information the same way you do.

This rendering ofpanoptic power reflects an important evolution of the original concept. It rests on a new collectivism in which "the many" view themselves and each views "the other." Horizontal visibility is created even as vertical visibility is intensified. The model is less one of Big Brother than of a workplace in which each member is explicitly empowered as his or her fellow worker's keeper. Instead of a Single omniscient overseer, this panopticon relies upon shared custodianship of data that reflect mutually enacted behavior. This new collectivism is, . an important antidote to the unilateral use of panoptic power, but it is not a trouble-free ideal. Horizontal tr<'Jnsparency breeds new human dilemmas as well, many of which will be· discussed in more depth in the conclusion.

Metro Tel did not share this collective orientation; thus, there was little to mitigate the oppressive sense of unilateral surveillance. There were a few workers who welcomed computer-mediated remote man- agement because it increased their sense ofautonomy.

I prefer feeling self-sufficient. ObViously, I am not-there are many invisible reins on me-but I like the idea of being given tasks and accomplishing them as qUickly as I can. I like the invisible presence of a foreman. I know he's there, but I don't have to deal with him unless I need him. At least there is some imagined independence involved.

352 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 353

Most of the craftsworkers reacted differently. One man recently had been promoted to foreman after many years as a craftsworker, and his reaction to the WF55 represented that ofmany workers in the field.

Ihated it. It was too dose. I could no longer hide anything. Manage- ment could monitor me hourby hour, and that was kind of scary.

It seemed that the system generally had succeeded in inducing a considerable degree of conformity to work standards. Foremen noted that the majority of workers could be counted on to finish the work load they were assigned. For the high achievers, this might mean doing less work than usual, but for the mediocre workers, it meant doing more: "People do what is expected of them now, so there is an equal- ization between those who used to do more, and those who did less." On the surface, the situation seemed orderly enough. The WFSS

computed daily efficiency ratings, which were accessed by all of the management hierarchy for a variety of evaluative purposes. Below the surface, however, things were not as tidy. The culture of adversarialism and the emphasis on the WFSS as a means of unilateral control also stimulated barely latent antagonisms and evoked a generous display of sly, subversive, and frequently imaginative adaptations. For many, surveillance represented an exciting challenge: "Go ahead, try and catch me." The goal was to thwart management's efforts at omniscience. Some workers called this approach "passive resistance." They chose

to ignore one or more work requests, later insisting that it simply had not been· included in their work load assignment. "The computer missed it," they would argue. "The computer isn't perfect yet." Work- ers also reported that they had worked on a trouble without being able to fix it. This meant their time was accounted for, but they might have been sleeping or reading a book during the hour or two the system recorded them as working on the problem. If a worker took this ap- proach only once every five or six days, he or she was unlikely to be detected. A third approach involved learning how to "wheel and deal" with the WFSS. This meant figuring out how to accomplish an eight- hour work load in less time in order to garner the extra hours for relaxation. As one man put it, "Now Iam starting to realize this system can work for me, not only me working for the system." The techniques of panoptic power applied at Metro Tel differed in

some important respects from those used at Cedar Bluff. One significant difference involved the dependence on worker's data entry in order to maintain the WFSS. In Cedar Bluff, the control interface and computer- ized equipment registered data regardless of the operators' intentions, but the WFSS at Metro Tel required workers to access the system for their daily work load and to update it throughout the day as those work loads were completed. It was this dependence on worker input that created the opportunity for the most cunning adaptations. Some crafts- workers had learned how to "snow the computer" in order to "make yourself look like a star." Instead of conforming to the standards it conveyed, theyhad figured out how to make the data look good regard- less of their actual behavior. How? "All you need is a terminal and a password, and if you are smart enough, you can change things." These workers had come into possession of their foreman's password, and they used it to change the prices they were aSSigned. For example, if a job priced for one hour actually took two, the worker would receive a 50 percent efficiency rating-not very flattering. By changing the price to two hours, the rating would be 100 percent. How did these workers come to have the passwords that would allow them to alter the original price? In many cases, their foremen had given them these passwords just for this purpose. Why? Some foremen did it outofa sense ofjustice. If a worker had put in the time to do a job well, why should he or she be penalized with a poor efficiency rating? One fbreman agonized:

If a man puts the time in and the computer makes him look bad, where is the justice? We have to manipulate the computer to show how good or bad people are trying.

Typically, however, when foremen shared their passwords with workers or when they themselves changed the prices in order to "clean up" the ratings, it was because they feared the conclusions their own bosses would draw when they noticed poor efficiency ratings on the WFSS reports. These foremen believed that if the boss questioned their ratings, it wouldmean a "blackmark against my name" or that the boss would conclude that the foreman did not have things "under control" in his or her area and, therefore, "must be doing the job wrong." Be- cause they did not trust their managers to interpret the data correctly, and because they did not trust their own ability to explain the data adequately, foremen preferred to manipulate the data so that it re- flected something even more meaningful than behavior-the still un-

354 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon ClSS

measurable intentions, commitments, and aspirations of the foremen and their workers. As one foreman explained:

If we let the computer run us, we look bad, so we manipulate the computer. We are not trying to cheat anybody or steal. We are trying to deal with the human element involved.

A small fish is eaten by a bigger one, which is in tum swallowed by an even larger fish: there was some evidence that second-level manag- ers knew of, and quietly condoned, the price changes because they, too, were being evaluated by third-level managers on the efficiency of their areas. As one foreman recollected:

It is a vicious cycle. If my boss sees that he did not meet his boss's productivity, can't he change the data too? He wants everybody's printout to lookgood. How much cheating is going on? Who knows?

The WFSS had been designed as a technical means by which to control workers' behavior and increase managers' certainty about what was ac- tually occurring in remote field locations. It was a classic example of what organizational theorists James March and Herbert Simon called an "uncertainty absorber"-a vehicle for transmitting and legitimating lIfacts" to levels of the organization that no longer have contact with the concrete realities this abstracted information is meant to repre- sent.13 What is even more striking is that the WFSS was able to fulfill its role as a of uncertainty, even when many individuals had cause to. either reject or suspect the validity of these "facts." These managers colluded in treating the behavioral text as a legitimate re- flection of the workplace, because it could continue to fulfill their emotional and pragmatic requirements for certainty. 14 Each of the hierarchy at Metro Tel chose to emphasize the

opportunities for increased unilateral surveillance provided by the new behavioral text. The message communicated throughout the organiza- tion was one of tenuous faith in the binding strength of the authority relationship. Nurturing reciprocities had become less important than these more certain means ofcontrol. Each level responded to its man- agers in a similar fashion, as they developed extralegitimate techniques of defense to protect themselves from involuntary display. Even as technique battled technique and faith gave way to cynicism, the emerg- ing conditions of work bore the stubborn mark of universal transpar-

ency. How much cheating was really going on? No one could be sure, but it seems that relative to the earlier work environment, there were indeed fewer shadowy comers. According to Metro Tel foremen, "cheating" had been a persistent characteristic of their adversarial, measurement-prone organization, regardless of the technology in- volved. The crucial point is not just that people continued to find ways to subvert measurement, but that almost everyone agreed with the new information systems it was becoming increasingly more difficult to do so. Several foremen recognized this trend:

Pe<;tple- were cheating before on paper-it was called a pencil whip. That meant you take a monthly report and lie with the numbers. That went to the boss, he signed it, and everyone believed that was "reality." Now once people get skilled with the computer, it's harder. Cheating can't even be done on certain levels because they have tenninals and can monitor you.

They were right. The systems designers already had wind of the effi- ciency ratings scam and were preparing their own response.,

There are ways that people in the field will find to get around mea- surements. But what you want to do with the system is to continually look for ways to trap people by recording their actions. If the craft are trying to get around their efficiency ratings by changing the price, all I have to do is start measuring how often,.they change the price and use that as a performance measure.

MANAGING COLLECTIVE RESPONSIBILITY IN THE INFORMATION PANOPTICON

The information panopticon adds another dimension to the infonnating process. It enables organizational processes, in addition to production or business-related processes, to be explicated by information technol- ogy. Who does what and how well they perform can be translated into "objective" data, along with the hundreds of variables that describe the production process or other business functions. In many cases, this means that managers' subjective judgments and personal sources of in-

356 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 357

fluence lose some of their force, while the intellective skills required to competently address the electronic text become more important. The system of imperative control permitted the middle manager a

fair degree of latitude to exercise personal judgment in operating his or-her own area of responsibility. The development of the techniques represented by the information panopticon were intended to protect andpreserve that authority. However, as Cedar Bluff's middle managers began to discover, the objective information presence enshrines ratio- nality over authority. Reality is codified independent of those who would benefit from it orbe ill-served by it. Now anautonomous domain of llfacts" can be used to check the command structure. Techniques of universal transparency, even when their goal is to

elicit anticipatory conformity, do not benignly serve the interests of authority. Because rationality has no gUile, it can also work to limit authority. Workers and managers together can find themselves "fo- cused on the data rather than the people," as one manager put it. The ambigUity of reciprocal relations and the subjective quality of behav- ioral data are displaced by a new sense of"black-and-white" certainty. AmbigUity has worked for and against both managers and workers. It has allowed managers to be more arbitrary and powerful; it has allowed workers to find ways to exert some control over their immediate envi- ronment. The price of that ambiguity has been that the manager's word counted for more than the worker's. The informating process has not only provided workers with the language to confront their managers but also equalized their respective realities, since the objective record stands as finaI...rbiter of what has happened. For the textualization of organizational processes represented in

panoptic power to have this effect, two elements are absolutely neces- sary, though in the longer term not sufficient, to engender a new frame- work of relations. One element involves egalitarian access to the emerging electronic text as it increasingly represents the full spectrum of organizational functioning. The second element involves the pres- ence of a sufficient depth of intellective skill so that those who have access to data also have access to their meaning. The very distinct dynamics engendered by an emphaSiS on collective,

as opposed to hierarchical, panoptic power indicate the central impor- tance of the rules that govern access to the data base. As the electronic text becomes the symbolic surrogate for the organization's vital activi- ties, access rules become the functional equivalent of organization

structures. As in the case of Metro Tel, these rules can mirror the hierarchy as they enable unilateral observation. Alternatively, universal access provides a structure in which a redistribution of authority be- comes pOSSible. One operator at Cedar Bluff summarized this neatly when he said, "If I can control my own access to data, I can control my own learning." It is this learning, more than anything else, that ex- plodes the confines of the worker's conventional role. Not one of the organizations I studied had yet confronted this crucial

issue. Rules ofaccess were beingconstructedad hoc, as individualmanag- ers were motivated to dip into data normally reserved for their subordi- nates and as subordinates found it in their interests to restrict or impede their superiors' access. Most 'people said that it would "takea lot ofguts" for their organizations to develop a clear policy for data access. Cedar Bluff's plant manager reflected on his and others' use of the

Overview System. He recognized the growing tendency to use the sys- tem in ways that contradicted the collaborative spirit of the plant.

Now that total information is possible, it requiresmanagement disci- pline ndt to ask for information. The effort at limiting information wUI be greater than the effort to generate information. It ls very hard for a top manager to have information and not want to do something about it. If you give me the data from that system, Iwill make more decisions. What is the right level of problem for each level of the organization? '"

Managers throughout these returned repeatedly to this theme of management discipline in the face of the vast new capability afforded by information systems. They seemed to be searching for some code ofdiscipline that managers would use to restrain their enthusiasm for information, but most wondered how this discipline could be mus- tered. Why should a fourteenth-century author ofa primer on etiquette compose a lengthy discussion on the evils of blowing one's nose into the tablecloth, if it were not for the fact that many people routinely did just that? Similarly, the volume of commentary on the need for managerial discipline is likely to point to the frequency with which managers are in fact tempted by panoptic power. The elusiveness of a managerial code of information discipline and

the difficulty of explicitly confronting the policy issues related to ac- cess rules may stem in part from the fact that there is no longer an obvious rationale to justify the choices one makes regarding the appro-

358 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon 359

priateness of information at each level of the organization, and there- fore the manager's optimum span of control. Historically, the criteria used to establish such norms were shaped in part by a recognition of what was actually feasible in the daily press of organizational affairs. It is clearly inappropriate for a divisional vice president to have real-time operating data from the manufacturing site when acquiring such data would engage extreme amounts of time and resources. The process of such acquisition would be ridiculously awkward. Would the vice president spend days in the plant? Would he or she have someone con- tinually available by telephone? Would the vice president ask to read operator logs?15

Technology radically alters the context of what is possible. How different it is for a high-level executive to sit in his or her office and use a desktop terminal to view summariesofoperating data or to review exceptions to optimal ranges of performance. It is still reasonable to argue that such an executive should not spend time responding to data that others are better positioned to manage. Once the awkwardness of the data acquisition process is eliminated, however, the immediate face validity of such arguments is diminished.16

As the material circumstances of the enterprise evolve, so do the de- mands on the manager. In a highly informated organization such as Cedar Bluff, the autonomy of the data base and the extent ofboth vertical and horizontal channels of access to the data added considerable weight to the already severe pressures that middle managers felt. Traditionally, the middle manager's role has been largely defined by the collection, analy- sis, and dissemination ofinformation. The most skilled aspects of the job involved nOrIToutine information, usually culled in face-to-face interac- tions and through personal networks. The less skilled aspects were those that concerned routine information associated with a particular bureau- cratic position or function. Both Metro Tel's and Cedar Bluff's managers discovered that the power of their information systems had undercut both dimensions of their job. Personal sources of information once lodged in informal networks were now deemed less important. Their control over the information related to their own functions had been severely curtailed. At Cedar Bluff, managers said:

Our systems now require a difference in how people interact with each otherand what role they play. Things no longer must be filtered from the top down. Our informationand control systemsmake infor- mation automatically available. Unless someone destroys the system

I

I, l

or pulls the plug, it all exists independently now. The operators can go to the data and bypass the manager. It's even data that managers have never had.

Managers don't collect and assimilate data themselves. It is a lost luxury. This means more communication at night via the telephone to managers while things are happening. I don't want to be surprised in the morning when my manager needs answers because his manager is trying to respond to the divisional VP who already has data on how we ran last night. I don't want to be caught cold. The subordinate used to control the size of the supervisor's window. Now the infor- mation system does it.

Both the skilled and the semiskilled aspects of the middle managers' job had been invaded by the information presence. Remember Chester Barnard's definition of a communications system in chapter 3? He viewed it as managers placed around the organization who came to- gether to discuss what they knew. How different it is to consider com- munication in the context of a comprehensive real-time system, such as the systems at Metro Tel or Cedar Bluff, where information consti- tuted an domain independent of those who once gathered, recorded, analyzed, discussed, and stored it. The familiar forms of in- fluence associated with these activities were losing much of their force; yet, for most managers caught in this transformation, new sources of influence remain undefined. ... Much of the manager's sense of artistry is derived from the manner

in which he or she garners and comnuUl,icates information. One conse- quence of the new electronic text is that it diminishes the sense of personal style, initiative, and art that many managers had felt. A man- ager at Cedar Bluff expressed it this way:

The more automated the information becomes, the more managers perceive themselves to have less choice, Jess opportunity for artistry or initiative. The information will be more integrated, and it will be more obvious what has to be done, and different people will know data from different parts of the organization, so the data won't be owned by asingle manager.

At Metro Tel, a manager explained:

I had ways of doing things that other people never knew about. My own approach to administration-getting and storing information. Other people had their own, but] always felt mine was better. That

360 TECHNIQUE: THE MATERIAL DIMENSION OF POWER The Information Panopticon JU'

was my skill. Now my administrative skill is not required, because everything comes out of the computer.

The integrative nature of the data base, together with its increased scope and complexity, also created new questions about the utility of the categories that defined managerial functions. There was a wide- spread feeling in Cedar Bluff that the amount of data now widely avail- able to the management group demanded a more team-centered, problem-solving orientation. Like the employee at the data interface, these managers faced a broad array of data that integrated disparate production, business, and organizational processes. Their capacity to deal adequately with such data seemed to require interactive discussion and analysis as observed in the control rooms as well as a new interde- pendence among the various business and operations functions. Some managers believed that the conventional logic of middle-

management functions not only was outmoded but also had become a dangerous barrier to constructive use of the data. Just as it was no longer useful to occupy an operator with data pertaining to only one piece of equipment, or for workers at Metro Tel to have responsibility for only one set of telephone switches, so there was also the risk that managers who regarded the data base from a narrow functional per- spective would misinterpret or overlook important information.

There is a great opportunity for misinterpretation of data when ev- eryone can see what is happening but their narrow perspective means that they can't tell why it is happening. Most people have a one- sided, functibnally oriented sense when they look at the data. It gets worse in 'that the technology lets you look down that data tunnel at lightening speed-then the tunnel turns into a dot. You end up with one number, one reason, and you react to it.

Maintaining the faith that undergirds imperative control is hard work-psychologically demanding, time-consuming, and inevitably prone to ambiguity. It requires action-centered skills oriented toward nurturing reciprocity among those bound together in the mutual de- pendency of direction and execution. Sometimes technologies appear to offer an easier way to secure managerial objectives, and alternatives that promise more certain results are difficult to resist. Managers are frequently persuaded to employ these techniques of control to satisfy their conceptions of their own roles ("I must have the answers") or as a means of circumventing the emotionally risky and sometimes fright-

I I t

l

ening exertions of face-to-face engagement. It is in this context that managers can choose to "automate" their own action-centered skills and to build managerial functions into machines that convey instruc- tions and tirelessly display even minute behavioral events. When these techniques are accompanied by an emphasis on unilateral surveillance, the triumphs ofenhanced certainty and control come at some cost. Unilateral techniques ofcontrol tend to evoke techniques ofdefense

from subordinates who resent their own involuntary display. While these defensive measures can be thwarted, they also can contaminate the validity of data. Even more important this battle of techniques of control versus techniques of defense signals the erosion of reciprocal relations as information becomes the field upon which latent antago- nisms are set loose. When managers choose to automate the action-centered skills in-

volved in managing reciprocities and in gathering and disseminating information, they increase their dependence upon the electronic text, not only as a means of accomplishing substantive tasks but also as a medium ofadministration. The electronic text can so insulate managers from the f';it realities of their workplaces that they will no longer have available the means with which to rekindle reciprocities if they should want to. Paradoxically, that very insularity increases the vulnerability of the text to contamination while it simultaneously heightens the re- quirements for valid objective data. Thus insu!ated, managers often col- lude in ignoring the ever more slender relationship between their data and the organizational realities that,they are meant to reflect. The horizontal dimension of Cedar Bluff's information system sug-

gests another view of panoptic power. Shared universal transparency can create a sense ofmutual participation in and responsibility for oper- ational and behavioral events. Joint access to the behavioral text can mean opportunities for joint learning. While key dimensions ofmanag- ers' action-centered skill may be absorbed by the information system, new skill demands emerge. Prominent among these is the capacity to understand the business more broadly: in Cedar Bluff, functional desig- nations seemed as outmoded as job classifications that once assigned a man to watch a single furnace for his entire work life. Managers also are likely to face new communicative demands-for sharing interpre- tations, problem-solving, and making sense ofcomplex and interdepen- dent forms ofdata about people and events.

•