THEMES IN DISCUSSIONS OF ORGANIZATIONAL STRUCTURE

Many times, when researchers look at organizational Structure, they fail to attend to the variety of its meanings and implications. It is easy to depict "span of control" in a textbook or measure it in an empirical study, but rarely is the multifaceted significance of such Structural dimensions fully developed in comments about the picture or explanations of the study's findings. To keep us alert to the breadth of implications of Structure, a survey of some of the themes most commonly developed about Structure in the social scientific literature is useful.

STRUCTURE AS AN EMPIRICAL OBJECT

Social scientists have long been fascinated with organizational Structures as simply raw objects of study, almost like the skeletons of biological individuals. Within this perspective, specific traits or dimensions describing organizational Structure are identified and their theoretical and empirical interrelations are assessed.

Indeed, some of the most distinquished founders of organizational theory fall, in part, within this tradition. Max Weber (1968) listed a number of features of bureaucracy-a hierarchy of authority, formal rules governing work, selection and promotion of employees on the basis of work qualifications, and the like-which he found in many real organizations. Although this list is a derivative part of his general position (see McPhee, 1981), the list has been highly influential in later organizational theory. For instance, Gouldner (1954) describes three alternative forms of bureaucratic structure that differ in source and nature from that identified by Weber. Again, perhaps resting on the keystone work of Burns and Stalker (1966), a debate developed in the 1960s and early 1970s over whether organizational structures varied unidimensionally between Weber's "mechanistic" bureaucracy and an opposite pole, the minimally formal "organic organization," or whether such Structures are multi-dimensional, with relatively independent dimensions such as formalization of work processes and centralization of authority (see Pugh et al., 1968).

More recent work has attempted to isolate multiple, distinctive types of organizational Structures. For instance, on empirical and theoretical grounds Mintzberg (1979) differentiates five (or six; 1983) prototypical "structural configurations" of organizations, each of which is internally coherent and consistent with particular contextual patterns. McKelvey (1982) goes even further by adopting biological taxonomic methods, likening particular organizations to individual organisms, and arguing for an "organizational species" concept. Species would be sets of organizations with similar competences and practices; McKelvey holds that general environmental forces condition the evolution of species and that they can be identified using quantitative methods of numerical taxonomy.

This theme, though rather abstract and macroscopic, is important to students of organizational communication for two main reasons. First, many of the traits or dimensions analyzed by these theorists specifically describe communication processes-for instance, Weber's stipulation of written messages and vertical consultation about exceptional problems. Second, underlying many of these discussions is an understanding of an organization as an information-processing entity. For instance; Mintzberg's five structural configurations rest, in part, on his distinction of five coordination mechanisms that the configurations embody and depend on. Each coordination mechanism is either a form of communication or a Structural substitute for communication. This perception of Structure as related to information processing leads to our next theme.

STRUCTURE AS AN INFORMATION PROCESSING/

COORDINATION MECHANISM OR TOOL

The utility of organizational Structure as a means of organizing the efforts of vast numbers of individuals has been recognized in writings about organizations from the very beginning. This idea is fully consistent with functionalism and the confluence of ideas in social theory described by Parsons (1937). But a very modem and influential formulation of the idea appears in James March and Herbert Simon's Organizations (1958).

March and Simon noted that human beings are characterized by "bounded rationality"-their capacities to gather and process information, to notice problems and solve them, are very limited. In particular, they are completely incapable of effectively adjusting their activities to those of hundreds of other unorganized individuals. So, March and Simon argued, organizational Structure, as well as a variety of other less formalized practices in organizations, are created to ease the decision-making tasks of individuals. Structural information tells the employee roughly what to do, whose commands to follow, and whom to inform about activities and results. It is designed to ensure that there is someone to do every task the organization requires, and that they are surrounded by an appropriate information environment to make proper decisions.

The clearest recent celebrants of this theme have been the "contingency theorists" (Perrow, 1967; Thompson, 1967; Lawrence and Lorsch, 1967; Galbraith, 1973), whose views I have considered in more detail elsewhere (McPhee, 1983). Their central argument is that organizational Structures vary because, as mechanisms for coordination or information processing, they should be and generally are adapted to the organization's needs for coordination/information processing. For instance, an organization is likely to be highly decentralized when matters requiring decisions are so turbulent or complex that high-level managers are too far removed from the action to make informed decisions. Generally, these theorists have gone further in analyzing organizational information processing needs, rooted in task complexity or environmental dynamism, than in describing the exact capacities of Structural elements or dimensions for information processing.

I have argued that contingency theory is one pole defining the domain of organizational communication studies (McPhee, 1983). For one thing, their conception of Structure makes communication central to the nature of organization. But beyond that, their theories have inspired an impressive number of studies of communication patterns in organizations (Van de Ven et al., 1976; Hage, 1974; the review in Tushman and Nadler, 1980).

STRUCTURE AS SYSTEM FORM

As we already noted, the theme of Structure as information processing or control mechanism involves a functional logic-Structure is purposefully and appropriately organized for its ends. This sort of logic is very tempting in the organizational arena since Structures are apparently purposeful human creations, belonging to social subsystems-organizations. So we should not be surprised at the pervasiveness of "systems" interpretations of structure, drawn from a long intellectual tradition of holism and vitalism (Phillips, 1976).

Systems theories of organizations have their immediate sources in the social systems views of Parsons (1937) and the general systems approach of Bertalanffy (1968), but the locus classicus of the theme is Katz and Kahn's Social Psychology of Organizations (1965, 1978). Katz and Kahn begin by theoretically describing general social systems, then account for Structure (and organizational processes) in terms of the components and requisities of the organization as social system. Organizational Structure separates and delimits the bounds of subunits; it also creates means for their integrative coordination. In short, it determines, to a large extent, the form or arrangement of subcomponents of the system. Systems theory serves both as a basis for integration of a wealth of other literature, and as an analytical instrument that stimulates attention to new organizational features.

Of course there are other and earlier systems-theoretic treatments of organizations-Simon's theory and sociotechnical systems theory (Emery and Trist,1960) come immediately to mind. But in the field of communication, Katz and Kahn's chapter, "Communication: The Flow of Information," was required and inspiring reading: "Communication-the exchange of information and the transmission of meaning-is the very essence of a social system or an organization" (1966: 223). Much of the study of networks (see Rogers and Kincaid, 1982) and vertical communication (see Jablin, 1979) is inspired by the sense that important information flow processes are guided by Structure, or by a desire to discover the nonstructured, supplementary contributions of communication.

STRUCTURE AS NEGOTIATED

The other pole (in addition to contingency theory) for orienting communication studies in organizations is the set of theories that describes

organizational regularity as negotiated. The "negotiated order theorists" (Strauss et al., 1964; Strauss, 1978; Maines, 1977; Day and Day, 1977), with strong affinity to symbolic interactionism, argue that order in at least some organizations has little to do with Structure. In looking at professional organizations, especially psychiatric hospitals, they note that professionals from a variety of disciplines must achieve a working consensus on duties, relationships, and influence despite their variant interests and world views. This "working consensus" is usually based on compromises achieved during work on specific concrete cases, where direct, intense involvement of low-status employees gives them disproportionate influence. It generally involves departures, both from the dictates of Structure, and from the disciplinary world views of the professions involved (for example, psychiatry, medicine, and nursing). And, in an important finding by theorists along these lines, the working consensus achieved in particular cases is often a basis for long-term, general Structural change. Negotiated order theorists have achieved influence in occupational sociology, where this theme has been extended to nonprofessionalized organizations (see Goldner, 1970).

Another line of research that develops this same theme is the recent work of James March (March and Olsen, 1976; Cohen and March, 1974). In particular, March and Olsen have demonstrated that organizational decision making often departs from the dictates of rationality. Organization members have limited resources of attention, and, in complex, ambiguous contexts, can pay attention to only part of the decisions and other matters that, by formal criteria, should concern them. In organizations where complexity and ambiguity are typical, problems, solutions, participants, and choice opportunities are thrown together in a random way that Cohen, March, and Olsen call a "garbage can model." Such organizations are called "organized anarchies."Although these often have Structures that look rational, the Structures are poor-quality, outdated pictures of the organizational processes.

STRUCTURE AS POWER OBJECT/RESOURCE

In viewing Structure as subject to negotiation, the writers just discussed have sometimes been accused of paying too little attention to power as a resource distributed in its own right. In a relatively recent development, organizational theorists have advanced power as an important topic of discussion in connection with Structure. Generally, even in traditionally Marxist literature mentioned below (and in Conrad and Ryan, this volume), intra-organizational power has been seen as derivative of Structure, which always confers power with authority and responsibility. But in this theme, the opposite occurs: Power is separated out as a discrete variable, something that organization members strive for, and a matter of popular concern.

A good example of this theme is Jeffrey Pfeffer's Organizational Design (1978). Pfeffer analyzes the various implications of Structure for power. He begins by noting bases of power-formal position, to be sure, but also such matters as assigned responsibility for critical and uncertain tasks, a structural location that results in information and access to decision makers, and control over important resources-all of which are at least partly determined by Structure. He also describes the power implications of Structural specialization, differentiation, formalization, and information systems. Correlatively, he describes major strategies employed by individuals and coalitions in the organization to control Structure so as to maximize their own power or minimize Structural effects on it.

Of course, this theme has been developed at a number of levels in recent popular and scholarly literature, with great appeal. For instance, in Korda's Power (1975), along with the various informal strategies mentioned are such Structural matters as control of the office files and the regular scheduling of meetings. At the more macro-economic end of the scale, John Kenneth Galbraith's The New Industrial State discusses the reallocation of power over large organizations, from their owners to the set of employees who "contribute information to group decisions," a set Galbraith calls the technostructure, the growth of which has been determined, of course, mainly by the growth of Structural complexity (1967: 82).

Like the preceding themes, this is one in which Structure is very intimately related to the communicative functions of information transmission and influence. Power is treated as an unseen force, ordering flows of information and influence attempts, at once their goal and the resource on which they depend.

STRUCTURE AS CARRIER OF

SOCIAL PSYCHOLOGICAL PROCESSES

In the next two themes, Structure is not exactly the central focus: It becomes instead a stimulus or backdrop to other phenomena. Here, our interest is in social-psychological processes that are transformed in organizations due to the presence of formal Structure. Many people would include organizational communication as such a process (I give it a broader status-see below, and McPhee, 1983), but numerous others have been examined.

One example, widely studied in organizational contexts, is role conflict-a person is torn as he or she tries to meet the incompatible expectations of various different role senders. Various researchers have noted that Structural features in organizations make certain persons or groups (for example, one's boss, one's subordinates) especially likely to "send" expectations that will result in role conflict (Kahn et al., 1964). Indeed, certain Structural forms-the matrix organization, in which the typical employee reports to two different superiorsare well-known for their tendency to engender role conflict.

Two more examples are the dual referents of the term "intergroup processes." The naive sense of this term includes what Doise (1977) calls "group differentiation"-any human group forms in contrast to some "foreign" entity, to which it often opposes itself. The formal differentiation of Structure sets groups of people apart, labels them in common as a work unit, relates them to contrasting groups on which they are dependent-differentiation, misunderstanding, and conflict are natural results. But the more common sense of "intergroup" involves racial or other minority groups. Here Kanter (1977) has made clear how formal structures can affect social psychological processes. Even where clear-cut prejudice is not present, majority perceptual processes such as attention, contrast, and assimilation, cued by the "token" status of minority individuals rising in organizations, can lead to a vicious cycle of real and perceived failure.

Other commonly studied social-psychological processes can be listed ad infinitum: alienation, conformity, motivation, status comparison, attribution, social perception of various sorts, and the like. One process emphasized by Simon, identification, has been dealt with in detail by Tompkins and Cheney (this volume). Another process that might belong here is formation of "climate" (but see the more complex interpretation of Poole, this volume).

STRUCTURE AS CARRIER OF SOCIAL PROCESSES

Structure is not merely a backdrop for relatively microscopic processesmore or less vast historic social movements have also found a worthy medium in organizational Structures, according to many social theorists. Indeed, the progressive rationalization of society was the Weberian theme within which the articulation of the bureaucratic model discussed above took place. Weber argued that modern history was distinguished by rationalization in many respects-the growth of modern science and the corresponding "disenchantment of the world" in the evolution of religion, for instance. Bureaucracy can exist only because of the general change in world view that makes "rational-legal authority," the motivational basis of bureaucracy, understandable and convincing. But it also contributes to rationalization, by making the actions of organizations efficient and predictable.

Organizational Structure has also been linked to the division of modern societies into social classes. Beginning at least with Marx, social theorists have argued that class divisions have been mirrored in the distinction between owners of organizational resources and laborers in organizations, who can only sell control over their labor power. The dynamics of capitalist production "reproduce" class distinctions by supplying a capital return to upgrade, replace, and expand factories and machinery, while furnishing workers with enough money to live and raise children-the next generation of workers-but not enough to let them accumulate capital and escape their class. Alternately, it has been argued that organizational Structuring has allowed the emergence of a "middle class" of professionals and managers (Walker, 1979; Giddens, 1971).

Beyond these major social processes, some writers have found relatively transient social trends to be reflected in Structures. For instance, Mintzberg has argued that fashion can be a determinant of Structure, as organizations adopt fads such as long-range planning, management by objectives, organizational development, matrix Structure, and so forth, in response to "pitches" by

Robert D. McPhee 157

business periodicals and business schools (1979: 295; see Meyer and Rowan, 1977).

STRUCTURE AS A CONTROL/DOMINATION MECHANISM

Theorists who see the organization as reproducing societal divisions also often see Structure as a tool for one side in such conflict: generally, as a tool of those in charge of an organization in attempts to manipulate other organizational members aganist their interests. The most notable elaborations of this analysis have come from recent theorists in the Marxian tradition (Braverman, 1965; Edwards, 1979; Friedman, 1977; Clegg and Dunkerley, 1979; Salaman, 1982, 1983; Storey, 1983), who portray the evolution of Structural forms as part of a series of attempts by owners and their managerial stooges to bribe, bludgeon, or brainwash labor into cooperation. Current corporate Structures are seen as forms of organization that maximize, not general output or social utility, but profitability for the few.

This theme focuses consistently on a number of Structural features. For instance, the development of the modern managerial hierarchy and technical staff served to "deskill" or proletarianize formerly skilled work, by separating "conception" (planning, design, and coordination) from "execution" (the actual labor) and assigning them to different ranks of employees, in order to make laborer's work less valuable, less highly paid, and more easily substituted for if the worker quits or disobeys and is fired. The consequent possession by managers of crucial information and functions serves to conceal the inequity of organizational control by giving managers special claims to legitimacy and authority. Structure can also increase control by rewarding workers with autonomy-room to act freely, set up comfortable group cultures (Katz,1968), work on technically interesting projects, or wield organizational power-in exchange for compliance; and, for loyalty and long tenure, with increasing salaries and eventual promotion. This sort of bribery, in conjunction with the specialized division of labor, also can weaken the position of the working class by dividing it against itself, rewarding some occupations and thereby leading them to resist the claims of less favored groups. Indeed, managers might be seen as simply workers who have been recruited by this tactic.

An interesting twist of this theme is the argument that top management groups, via Structural elaborations, have gained power over owners and important environmental elements (Galbraith, 1967; Pfeffer and Salancik, 1978). In a number of respects, this theme is a complement to the view of Structure as an information-processing mechanism: Structure is again viewed as communicatively significant, but here as a form of or substitute for compliancegaining strategies.

STRUCTURE AS COUNTERPRODUCTIVE/DYSFUNCTIONAL

The views quoted in the beginning of this chapter exemplify the line of argument that organizational Structure often has unintended and unfortunate

consequences. This theme is very old, and early attached the connotation of rigidity and red tape to the term "bureaucracy" (Albrow, 1970). It is popular with students of communication, who prefer the flexibility and humaneness of interaction to the formality of Structure.

One variant development of this theme is the argument that these unfortunate consequences are the result of a self-reinforcing cycle started by Structure: Its consequences stimulate further reliance on Structure, which eventually produces extremes of the consequences (March and Simon, 1958; Crozier, 1964). Such consequences include: rigid rule-following due to a sense that behavior must be defensible; goal displacement of organizational goals by subunit goals; and increasing "closeness of supervision" due to conformity to minimum subunit norms. Structure, then, may not be an ill itself, but often is the nutrient for the growth of ills.

Another variation is the perception of a dilemma, and resulting dialectic, between Structure and creative initiative. Blau and Scott (1962) argue that organizations need both the coordination, regularity and discipline, and central planning fostered by Structure, and the general problem-solving communication, professional orientations, and individual initiative that. are stifled by Structure. Blau and Scott argue that effective organizations undergo a sort of dialectic of change, pendulating (as with the negotiated order theory) between innovation or conflict and systematization, but learning and improving through the experience. Here Structure, while dysfunctional, is a necessary and contributory moment of development.

STRUCTURE AS RESISTED, EVADED, OR IGNORED

Naturally, a class dominated or harmed by Structure has reason to oppose the control exercised through organizational Structure. And the working class has done so, according to the Marxian analysis. Braverman, for instance, has been criticized repeatedly for neglecting the active resistance of the working class to the successive forms of capitalist organization, especially through the labor movement (Edwards, 1979). But this argument joins a broader theme that analyzes the almost omnipresent failings of Structure to secure conformity from members of organizations.

This theme is also an old one: Organizational theorists as early as Taylor (1903) decried "systematic soldiering" or restriction of output by workers. But workers also ignore or break a variety of rules-they exchange jobs; attribute more authority to some managers than to those managers' bosses; alter work processes; communicate when, where, and what they should not; circumvent key employees in a bureaucratic chain of work-all against regulations. A number of alternative explanations for this phenomenon exist. Most interestingly, violations of Structure may actually be useful to the organization, accomplishing work more efficiently (Gross, 1953). But most violations seem organizationally dysfunctional, including stealing (Mars, 1982) and other worker self-interested acts, conformity to deviant social norms such as output restrictions, and oppositional union activity.

Deviations from Structure-derived predictions about behavior have fascinated students of communication, who have explored the growth of informal deviations from rules (Roy, 1960) and suggested that some deviations may be the result of ambiguous communication-orders interpreted as suggestions and the like (Burns, 1954).

STRUCTURE AS ENACTED/ACCOMPLISHED

This final theme derives from an important insight: Rules of any sort do not simply apply themselves. They always require some degree of creativity and judgment, at its highest in such formalized arenas as the legal system. But the process of adaptation or accomplishment is not the same as negotiation-it resolves not discrepant member interests, but an uncertainty that is present in all human action. Organizational Structure is a clear case in which this insight itself applies.

One school of thought that has developed this theme is ethnomethodology. Ethnomethodologists argue that our sense of the "everydayness" of life is actively accomplished by people, not a natural state of affairs. In organizations, people can and do "follow a routine" but only by being more creative than standard organizational theories ever imply. For instance, Zimmerman (1970) studied a clinic in which a receptionist was supposed to assign patients to various physicians by writing their names on one or another physician's list. But some physicians were delayed by difficult cases; to prevent inordinately long waits, especially for seriously ill patients, the receptionist sometimes juggled the lists, more or less radically, depending on her judgment. Even that simple rule had to be adapted to a variety of exceptional circumstances, an adaptation best regarded as "common sense," not an underlying, more complex decision rule.

Another influential and related development of this theme relates to the concept of "organizing" initially developed by Weick (1969, 1979). For Weick, our Structure is retained in the multitude of possible organizational memory banks; as such, it is limited by many kinds of phenomena such as assimilation to prior memory organization, memory deterioration or forgetting, and problems and costs of recall. Moreover, Structure is not simply recalled and conformed to. Its application is part of a process whereby the organization's environment is engaged or "enacted," and features from the equivocal whole are "selected" or interpreted in an organizational response that is itself retained as a gloss on Structure. Enactment, retention, and especially selection are all social processes, so once again communication determines the practical implications of Structure.

REPRISE

All these themes allow and deserve much more discussion than I have given them here; all have their weaknesses that have been criticized, and their answers to critiques. I think of them as theoretic "molecules" or "words" that can be connected in many different developments of thought about organizations. I have separated them to allow re-mention in what follows, and to make several

160 Formal Structure

general points. First, in many of these themes, Structure has consequences that are undesirable from many points of view. Indeed, only the theme of Structure as information processing or coordination mechanism describes it as clearly and generally useful. This negative tone recalls the views of Argyris and Kanter, mentioned above, and is in line with a long-term revelatory and critical current in social science that organizational communication theorists, until recently, have tended to ignore. Its opposition to the rational, scientific efforts of organizational designers remind one of the Romantic classic Frankenstein: Are organizations the social-scientific equivalent of the monster? I mean this question (perhaps it should be phrased a bit less emotionally) seriously-Is Structure, in principle, an ill? Even in its identity as coordination mechanism? And what are the implications of this question for the study of organizational communication?

The rest of this chapter will fall into three sections. In the first, I will discuss some peculiarly communicational implications of Structure. Then I will introduce the theory of structuration, a perspective that will, in the third part, facilitate an enriched account of organizational Structure integrating many of the themes just discussed.

THE NATURE OF STRUCTURE

In the introductory section of the chapter, defining qualities of organizational Structure were stipulated. Here I want to emphasize the qualities of Structure that are related to the way it is communicated or established. The very existence of Structure is dependent on its communication in a formal, explicit, authoritative way, often in writing, in a document that has recognized status (though often limited circulation) in the organization. Also, with accompanying changes in Structure there is sometimes a ritual of annoucement, of public proclamation of an employee's promotion, a new department, or a new system. Although Structure has effects on behavior, it is not fundamentally a behavior patternwe would operationalize Structure, not by observing behavior (which is often informal deviation or resistance), but rather finding and reading an authoritative pronouncement (and making sure, by asking an authority, that it still reigns). Structure is fundamentally communicational, then. Of course, in observing Structure we should not stop with document-reading. On the contrary, an advantage of this view of Structure is that it forces us to view, as a system, three aspects in a balanced way: the process of design (that is, writing the documents), the process of communication, and the process of response (interpreting and responding to the documents over time). The design and response aspects have been studied repeatedly, but the central aspect, communication, has often been treated simply as a matter of implementation. I want to give more theoretical attention to this central aspect, actually a distinct subsystem of organizational communication with its own distinctive type of sublanguage. Two features of the communication of Structure (called "Structure-communication" below) will be emphasized in this section.

THEMES IN DISCUSSIONS OF ORGANIZATIONAL STRUCTURE

Many times, when researchers look at organizational Structure, they fail to attend to the variety of its meanings and implications. It is easy to depict "span of control" in a textbook or measure it in an empirical study, but rarely is the multifaceted significance of such Structural dimensions fully developed in comments about the picture or explanations of the study's findings. To keep us alert to the breadth of implications of Structure, a survey of some of the themes most commonly developed about Structure in the social scientific literature is useful.

STRUCTURE AS AN EMPIRICAL OBJECT

Social scientists have long been fascinated with organizational Structures as simply raw objects of study, almost like the skeletons of biological individuals. Within this perspective, specific traits or dimensions describing organizational Structure are identified and their theoretical and empirical interrelations are assessed.

Indeed, some of the most distinquished founders of organizational theory fall, in part, within this tradition. Max Weber (1968) listed a number of features of bureaucracy-a hierarchy of authority, formal rules governing work, selection and promotion of employees on the basis of work qualifications, and the like-which he found in many real organizations. Although this list is a derivative part of his general position (see McPhee, 1981), the list has been highly influential in later organizational theory. For instance, Gouldner (1954) describes three alternative forms of bureaucratic structure that differ in source and nature from that identified by Weber. Again, perhaps resting on the keystone work of Burns and Stalker (1966), a debate developed in the 1960s and early 1970s over whether organizational structures varied unidimensionally between Weber's "mechanistic" bureaucracy and an opposite pole, the minimally formal "organic organization," or whether such Structures are multi-dimensional, with relatively independent dimensions such as formalization of work processes and centralization of authority (see Pugh et al., 1968).

More recent work has attempted to isolate multiple, distinctive types of organizational Structures. For instance, on empirical and theoretical grounds Mintzberg (1979) differentiates five (or six; 1983) prototypical "structural configurations" of organizations, each of which is internally coherent and consistent with particular contextual patterns. McKelvey (1982) goes even further by adopting biological taxonomic methods, likening particular organizations to individual organisms, and arguing for an "organizational species" concept. Species would be sets of organizations with similar competences and practices; McKelvey holds that general environmental forces condition the evolution of species and that they can be identified using quantitative methods of numerical taxonomy.

This theme, though rather abstract and macroscopic, is important to students of organizational communication for two main reasons. First, many of the traits or dimensions analyzed by these theorists specifically describe communication processes-for instance, Weber's stipulation of written messages and vertical consultation about exceptional problems. Second, underlying many of these discussions is an understanding of an organization as an information-processing entity. For instance; Mintzberg's five structural configurations rest, in part, on his distinction of five coordination mechanisms that the configurations embody and depend on. Each coordination mechanism is either a form of communication or a Structural substitute for communication. This perception of Structure as related to information processing leads to our next theme.

STRUCTURE AS AN INFORMATION PROCESSING/

COORDINATION MECHANISM OR TOOL

The utility of organizational Structure as a means of organizing the efforts of vast numbers of individuals has been recognized in writings about organizations from the very beginning. This idea is fully consistent with functionalism and the confluence of ideas in social theory described by Parsons (1937). But a very modem and influential formulation of the idea appears in James March and Herbert Simon's Organizations (1958).

March and Simon noted that human beings are characterized by "bounded rationality"-their capacities to gather and process information, to notice problems and solve them, are very limited. In particular, they are completely incapable of effectively adjusting their activities to those of hundreds of other unorganized individuals. So, March and Simon argued, organizational Structure, as well as a variety of other less formalized practices in organizations, are created to ease the decision-making tasks of individuals. Structural information tells the employee roughly what to do, whose commands to follow, and whom to inform about activities and results. It is designed to ensure that there is someone to do every task the organization requires, and that they are surrounded by an appropriate information environment to make proper decisions.

The clearest recent celebrants of this theme have been the "contingency theorists" (Perrow, 1967; Thompson, 1967; Lawrence and Lorsch, 1967; Galbraith, 1973), whose views I have considered in more detail elsewhere (McPhee, 1983). Their central argument is that organizational Structures vary because, as mechanisms for coordination or information processing, they should be and generally are adapted to the organization's needs for coordination/information processing. For instance, an organization is likely to be highly decentralized when matters requiring decisions are so turbulent or complex that high-level managers are too far removed from the action to make informed decisions. Generally, these theorists have gone further in analyzing organizational information processing needs, rooted in task complexity or environmental dynamism, than in describing the exact capacities of Structural elements or dimensions for information processing.

I have argued that contingency theory is one pole defining the domain of organizational communication studies (McPhee, 1983). For one thing, their conception of Structure makes communication central to the nature of organization. But beyond that, their theories have inspired an impressive number of studies of communication patterns in organizations (Van de Ven et al., 1976; Hage, 1974; the review in Tushman and Nadler, 1980).

STRUCTURE AS SYSTEM FORM

As we already noted, the theme of Structure as information processing or control mechanism involves a functional logic-Structure is purposefully and appropriately organized for its ends. This sort of logic is very tempting in the organizational arena since Structures are apparently purposeful human creations, belonging to social subsystems-organizations. So we should not be surprised at the pervasiveness of "systems" interpretations of structure, drawn from a long intellectual tradition of holism and vitalism (Phillips, 1976).

Systems theories of organizations have their immediate sources in the social systems views of Parsons (1937) and the general systems approach of Bertalanffy (1968), but the locus classicus of the theme is Katz and Kahn's Social Psychology of Organizations (1965, 1978). Katz and Kahn begin by theoretically describing general social systems, then account for Structure (and organizational processes) in terms of the components and requisities of the organization as social system. Organizational Structure separates and delimits the bounds of subunits; it also creates means for their integrative coordination. In short, it determines, to a large extent, the form or arrangement of subcomponents of the system. Systems theory serves both as a basis for integration of a wealth of other literature, and as an analytical instrument that stimulates attention to new organizational features.

Of course there are other and earlier systems-theoretic treatments of organizations-Simon's theory and sociotechnical systems theory (Emery and Trist,1960) come immediately to mind. But in the field of communication, Katz and Kahn's chapter, "Communication: The Flow of Information," was required and inspiring reading: "Communication-the exchange of information and the transmission of meaning-is the very essence of a social system or an organization" (1966: 223). Much of the study of networks (see Rogers and Kincaid, 1982) and vertical communication (see Jablin, 1979) is inspired by the sense that important information flow processes are guided by Structure, or by a desire to discover the nonstructured, supplementary contributions of communication.

STRUCTURE AS NEGOTIATED

The other pole (in addition to contingency theory) for orienting communication studies in organizations is the set of theories that describes

organizational regularity as negotiated. The "negotiated order theorists" (Strauss et al., 1964; Strauss, 1978; Maines, 1977; Day and Day, 1977), with strong affinity to symbolic interactionism, argue that order in at least some organizations has little to do with Structure. In looking at professional organizations, especially psychiatric hospitals, they note that professionals from a variety of disciplines must achieve a working consensus on duties, relationships, and influence despite their variant interests and world views. This "working consensus" is usually based on compromises achieved during work on specific concrete cases, where direct, intense involvement of low-status employees gives them disproportionate influence. It generally involves departures, both from the dictates of Structure, and from the disciplinary world views of the professions involved (for example, psychiatry, medicine, and nursing). And, in an important finding by theorists along these lines, the working consensus achieved in particular cases is often a basis for long-term, general Structural change. Negotiated order theorists have achieved influence in occupational sociology, where this theme has been extended to nonprofessionalized organizations (see Goldner, 1970).

Another line of research that develops this same theme is the recent work of James March (March and Olsen, 1976; Cohen and March, 1974). In particular, March and Olsen have demonstrated that organizational decision making often departs from the dictates of rationality. Organization members have limited resources of attention, and, in complex, ambiguous contexts, can pay attention to only part of the decisions and other matters that, by formal criteria, should concern them. In organizations where complexity and ambiguity are typical, problems, solutions, participants, and choice opportunities are thrown together in a random way that Cohen, March, and Olsen call a "garbage can model." Such organizations are called "organized anarchies."Although these often have Structures that look rational, the Structures are poor-quality, outdated pictures of the organizational processes.

STRUCTURE AS POWER OBJECT/RESOURCE

In viewing Structure as subject to negotiation, the writers just discussed have sometimes been accused of paying too little attention to power as a resource distributed in its own right. In a relatively recent development, organizational theorists have advanced power as an important topic of discussion in connection with Structure. Generally, even in traditionally Marxist literature mentioned below (and in Conrad and Ryan, this volume), intra-organizational power has been seen as derivative of Structure, which always confers power with authority and responsibility. But in this theme, the opposite occurs: Power is separated out as a discrete variable, something that organization members strive for, and a matter of popular concern.

A good example of this theme is Jeffrey Pfeffer's Organizational Design (1978). Pfeffer analyzes the various implications of Structure for power. He begins by noting bases of power-formal position, to be sure, but also such matters as assigned responsibility for critical and uncertain tasks, a structural location that results in information and access to decision makers, and control over important resources-all of which are at least partly determined by Structure. He also describes the power implications of Structural specialization, differentiation, formalization, and information systems. Correlatively, he describes major strategies employed by individuals and coalitions in the organization to control Structure so as to maximize their own power or minimize Structural effects on it.

Of course, this theme has been developed at a number of levels in recent popular and scholarly literature, with great appeal. For instance, in Korda's Power (1975), along with the various informal strategies mentioned are such Structural matters as control of the office files and the regular scheduling of meetings. At the more macro-economic end of the scale, John Kenneth Galbraith's The New Industrial State discusses the reallocation of power over large organizations, from their owners to the set of employees who "contribute information to group decisions," a set Galbraith calls the technostructure, the growth of which has been determined, of course, mainly by the growth of Structural complexity (1967: 82).

Like the preceding themes, this is one in which Structure is very intimately related to the communicative functions of information transmission and influence. Power is treated as an unseen force, ordering flows of information and influence attempts, at once their goal and the resource on which they depend.

STRUCTURE AS CARRIER OF

SOCIAL PSYCHOLOGICAL PROCESSES

In the next two themes, Structure is not exactly the central focus: It becomes instead a stimulus or backdrop to other phenomena. Here, our interest is in social-psychological processes that are transformed in organizations due to the presence of formal Structure. Many people would include organizational communication as such a process (I give it a broader status-see below, and McPhee, 1983), but numerous others have been examined.

One example, widely studied in organizational contexts, is role conflict-a person is torn as he or she tries to meet the incompatible expectations of various different role senders. Various researchers have noted that Structural features in organizations make certain persons or groups (for example, one's boss, one's subordinates) especially likely to "send" expectations that will result in role conflict (Kahn et al., 1964). Indeed, certain Structural forms-the matrix organization, in which the typical employee reports to two different superiorsare well-known for their tendency to engender role conflict.

Two more examples are the dual referents of the term "intergroup processes." The naive sense of this term includes what Doise (1977) calls "group differentiation"-any human group forms in contrast to some "foreign" entity, to which it often opposes itself. The formal differentiation of Structure sets groups of people apart, labels them in common as a work unit, relates them to contrasting groups on which they are dependent-differentiation, misunderstanding, and conflict are natural results. But the more common sense of "intergroup" involves racial or other minority groups. Here Kanter (1977) has made clear how formal structures can affect social psychological processes. Even where clear-cut prejudice is not present, majority perceptual processes such as attention, contrast, and assimilation, cued by the "token" status of minority individuals rising in organizations, can lead to a vicious cycle of real and perceived failure.

Other commonly studied social-psychological processes can be listed ad infinitum: alienation, conformity, motivation, status comparison, attribution, social perception of various sorts, and the like. One process emphasized by Simon, identification, has been dealt with in detail by Tompkins and Cheney (this volume). Another process that might belong here is formation of "climate" (but see the more complex interpretation of Poole, this volume).

STRUCTURE AS CARRIER OF SOCIAL PROCESSES

Structure is not merely a backdrop for relatively microscopic processesmore or less vast historic social movements have also found a worthy medium in organizational Structures, according to many social theorists. Indeed, the progressive rationalization of society was the Weberian theme within which the articulation of the bureaucratic model discussed above took place. Weber argued that modern history was distinguished by rationalization in many respects-the growth of modern science and the corresponding "disenchantment of the world" in the evolution of religion, for instance. Bureaucracy can exist only because of the general change in world view that makes "rational-legal authority," the motivational basis of bureaucracy, understandable and convincing. But it also contributes to rationalization, by making the actions of organizations efficient and predictable.

Organizational Structure has also been linked to the division of modern societies into social classes. Beginning at least with Marx, social theorists have argued that class divisions have been mirrored in the distinction between owners of organizational resources and laborers in organizations, who can only sell control over their labor power. The dynamics of capitalist production "reproduce" class distinctions by supplying a capital return to upgrade, replace, and expand factories and machinery, while furnishing workers with enough money to live and raise children-the next generation of workers-but not enough to let them accumulate capital and escape their class. Alternately, it has been argued that organizational Structuring has allowed the emergence of a "middle class" of professionals and managers (Walker, 1979; Giddens, 1971).

Beyond these major social processes, some writers have found relatively transient social trends to be reflected in Structures. For instance, Mintzberg has argued that fashion can be a determinant of Structure, as organizations adopt fads such as long-range planning, management by objectives, organizational development, matrix Structure, and so forth, in response to "pitches" by

Robert D. McPhee 157

business periodicals and business schools (1979: 295; see Meyer and Rowan, 1977).

STRUCTURE AS A CONTROL/DOMINATION MECHANISM

Theorists who see the organization as reproducing societal divisions also often see Structure as a tool for one side in such conflict: generally, as a tool of those in charge of an organization in attempts to manipulate other organizational members aganist their interests. The most notable elaborations of this analysis have come from recent theorists in the Marxian tradition (Braverman, 1965; Edwards, 1979; Friedman, 1977; Clegg and Dunkerley, 1979; Salaman, 1982, 1983; Storey, 1983), who portray the evolution of Structural forms as part of a series of attempts by owners and their managerial stooges to bribe, bludgeon, or brainwash labor into cooperation. Current corporate Structures are seen as forms of organization that maximize, not general output or social utility, but profitability for the few.

This theme focuses consistently on a number of Structural features. For instance, the development of the modern managerial hierarchy and technical staff served to "deskill" or proletarianize formerly skilled work, by separating "conception" (planning, design, and coordination) from "execution" (the actual labor) and assigning them to different ranks of employees, in order to make laborer's work less valuable, less highly paid, and more easily substituted for if the worker quits or disobeys and is fired. The consequent possession by managers of crucial information and functions serves to conceal the inequity of organizational control by giving managers special claims to legitimacy and authority. Structure can also increase control by rewarding workers with autonomy-room to act freely, set up comfortable group cultures (Katz,1968), work on technically interesting projects, or wield organizational power-in exchange for compliance; and, for loyalty and long tenure, with increasing salaries and eventual promotion. This sort of bribery, in conjunction with the specialized division of labor, also can weaken the position of the working class by dividing it against itself, rewarding some occupations and thereby leading them to resist the claims of less favored groups. Indeed, managers might be seen as simply workers who have been recruited by this tactic.

An interesting twist of this theme is the argument that top management groups, via Structural elaborations, have gained power over owners and important environmental elements (Galbraith, 1967; Pfeffer and Salancik, 1978). In a number of respects, this theme is a complement to the view of Structure as an information-processing mechanism: Structure is again viewed as communicatively significant, but here as a form of or substitute for compliancegaining strategies.

STRUCTURE AS COUNTERPRODUCTIVE/DYSFUNCTIONAL

The views quoted in the beginning of this chapter exemplify the line of argument that organizational Structure often has unintended and unfortunate

consequences. This theme is very old, and early attached the connotation of rigidity and red tape to the term "bureaucracy" (Albrow, 1970). It is popular with students of communication, who prefer the flexibility and humaneness of interaction to the formality of Structure.

One variant development of this theme is the argument that these unfortunate consequences are the result of a self-reinforcing cycle started by Structure: Its consequences stimulate further reliance on Structure, which eventually produces extremes of the consequences (March and Simon, 1958; Crozier, 1964). Such consequences include: rigid rule-following due to a sense that behavior must be defensible; goal displacement of organizational goals by subunit goals; and increasing "closeness of supervision" due to conformity to minimum subunit norms. Structure, then, may not be an ill itself, but often is the nutrient for the growth of ills.

Another variation is the perception of a dilemma, and resulting dialectic, between Structure and creative initiative. Blau and Scott (1962) argue that organizations need both the coordination, regularity and discipline, and central planning fostered by Structure, and the general problem-solving communication, professional orientations, and individual initiative that. are stifled by Structure. Blau and Scott argue that effective organizations undergo a sort of dialectic of change, pendulating (as with the negotiated order theory) between innovation or conflict and systematization, but learning and improving through the experience. Here Structure, while dysfunctional, is a necessary and contributory moment of development.

STRUCTURE AS RESISTED, EVADED, OR IGNORED

Naturally, a class dominated or harmed by Structure has reason to oppose the control exercised through organizational Structure. And the working class has done so, according to the Marxian analysis. Braverman, for instance, has been criticized repeatedly for neglecting the active resistance of the working class to the successive forms of capitalist organization, especially through the labor movement (Edwards, 1979). But this argument joins a broader theme that analyzes the almost omnipresent failings of Structure to secure conformity from members of organizations.

This theme is also an old one: Organizational theorists as early as Taylor (1903) decried "systematic soldiering" or restriction of output by workers. But workers also ignore or break a variety of rules-they exchange jobs; attribute more authority to some managers than to those managers' bosses; alter work processes; communicate when, where, and what they should not; circumvent key employees in a bureaucratic chain of work-all against regulations. A number of alternative explanations for this phenomenon exist. Most interestingly, violations of Structure may actually be useful to the organization, accomplishing work more efficiently (Gross, 1953). But most violations seem organizationally dysfunctional, including stealing (Mars, 1982) and other worker self-interested acts, conformity to deviant social norms such as output restrictions, and oppositional union activity.

Deviations from Structure-derived predictions about behavior have fascinated students of communication, who have explored the growth of informal deviations from rules (Roy, 1960) and suggested that some deviations may be the result of ambiguous communication-orders interpreted as suggestions and the like (Burns, 1954).

STRUCTURE AS ENACTED/ACCOMPLISHED

This final theme derives from an important insight: Rules of any sort do not simply apply themselves. They always require some degree of creativity and judgment, at its highest in such formalized arenas as the legal system. But the process of adaptation or accomplishment is not the same as negotiation-it resolves not discrepant member interests, but an uncertainty that is present in all human action. Organizational Structure is a clear case in which this insight itself applies.

One school of thought that has developed this theme is ethnomethodology. Ethnomethodologists argue that our sense of the "everydayness" of life is actively accomplished by people, not a natural state of affairs. In organizations, people can and do "follow a routine" but only by being more creative than standard organizational theories ever imply. For instance, Zimmerman (1970) studied a clinic in which a receptionist was supposed to assign patients to various physicians by writing their names on one or another physician's list. But some physicians were delayed by difficult cases; to prevent inordinately long waits, especially for seriously ill patients, the receptionist sometimes juggled the lists, more or less radically, depending on her judgment. Even that simple rule had to be adapted to a variety of exceptional circumstances, an adaptation best regarded as "common sense," not an underlying, more complex decision rule.

Another influential and related development of this theme relates to the concept of "organizing" initially developed by Weick (1969, 1979). For Weick, our Structure is retained in the multitude of possible organizational memory banks; as such, it is limited by many kinds of phenomena such as assimilation to prior memory organization, memory deterioration or forgetting, and problems and costs of recall. Moreover, Structure is not simply recalled and conformed to. Its application is part of a process whereby the organization's environment is engaged or "enacted," and features from the equivocal whole are "selected" or interpreted in an organizational response that is itself retained as a gloss on Structure. Enactment, retention, and especially selection are all social processes, so once again communication determines the practical implications of Structure.

REPRISE

All these themes allow and deserve much more discussion than I have given them here; all have their weaknesses that have been criticized, and their answers to critiques. I think of them as theoretic "molecules" or "words" that can be connected in many different developments of thought about organizations. I have separated them to allow re-mention in what follows, and to make several

160 Formal Structure

general points. First, in many of these themes, Structure has consequences that are undesirable from many points of view. Indeed, only the theme of Structure as information processing or coordination mechanism describes it as clearly and generally useful. This negative tone recalls the views of Argyris and Kanter, mentioned above, and is in line with a long-term revelatory and critical current in social science that organizational communication theorists, until recently, have tended to ignore. Its opposition to the rational, scientific efforts of organizational designers remind one of the Romantic classic Frankenstein: Are organizations the social-scientific equivalent of the monster? I mean this question (perhaps it should be phrased a bit less emotionally) seriously-Is Structure, in principle, an ill? Even in its identity as coordination mechanism? And what are the implications of this question for the study of organizational communication?

The rest of this chapter will fall into three sections. In the first, I will discuss some peculiarly communicational implications of Structure. Then I will introduce the theory of structuration, a perspective that will, in the third part, facilitate an enriched account of organizational Structure integrating many of the themes just discussed.

THE NATURE OF STRUCTURE

In the introductory section of the chapter, defining qualities of organizational Structure were stipulated. Here I want to emphasize the qualities of Structure that are related to the way it is communicated or established. The very existence of Structure is dependent on its communication in a formal, explicit, authoritative way, often in writing, in a document that has recognized status (though often limited circulation) in the organization. Also, with accompanying changes in Structure there is sometimes a ritual of annoucement, of public proclamation of an employee's promotion, a new department, or a new system. Although Structure has effects on behavior, it is not fundamentally a behavior patternwe would operationalize Structure, not by observing behavior (which is often informal deviation or resistance), but rather finding and reading an authoritative pronouncement (and making sure, by asking an authority, that it still reigns). Structure is fundamentally communicational, then. Of course, in observing Structure we should not stop with document-reading. On the contrary, an advantage of this view of Structure is that it forces us to view, as a system, three aspects in a balanced way: the process of design (that is, writing the documents), the process of communication, and the process of response (interpreting and responding to the documents over time). The design and response aspects have been studied repeatedly, but the central aspect, communication, has often been treated simply as a matter of implementation. I want to give more theoretical attention to this central aspect, actually a distinct subsystem of organizational communication with its own distinctive type of sublanguage. Two features of the communication of Structure (called "Structure-communication" below) will be emphasized in this section.