organizational communication final essay
kkkkpcy
w12McPhee_2004_TextAgency_OrganizationintheLightofStructurationTheory.pdf
Text, Agency, and Organization in the Light of Structuration Theory
Robert D. McPhee Arizona State University, USA
Abstract. This article argues for three linked conceptions. First is a conception of ‘a text’ as characterized by three features: a text is symbolic; it is inscribed as symbols on a relatively permanent, ‘readable’ medium such as paper or CD; and it has a coherent enough structure to be dealt with by interpretive processes or technologies such as financial processing. The second conception is that of an organization as a stable relational system grounded in the capacity to ‘ “bracket space–time” . . . via the reflexive monitoring of system reproduction and the articulation of discursive history’ (Giddens, 1987b: 153). I argue that this capacity depends on the specific properties of ‘a text’ mentioned above. Third, ‘agency’ is conceptualized in line with Giddens’ ‘internalist’ model of agentive powers, which are ones necessary to engage with a text as organization members do. The paper argues that other conceptions risk blinding us to the features and relationships shown to be crucial here. Key words. discourse; text; structuration; organizational constitution; agency; institutional reflexivity
As discursive approaches to organizational analysis become more devel- oped and varied, the array of background assumptions and conceptions they carry has grown in range while also growing more ambiguous and equivocal. It now seems clear, and is even a matter of consensus among commentators, that deconstruction does not carry with it an assumption of utter denial of textual meaning. But that stance leaves the import of the deconstructive challenge to textual meaning even more open to question
Volume 11(3): 355–371 ISSN 1350–5084
Copyright © 2004 SAGE (London, Thousand Oaks, CA
and New Delhi)
DOI: 10.1177/1350508404041997 www.sagepublications.com
articles
and doubt, though less radical. This special issue, aiming to gain some clarity about the two central terms of ‘text’ and ‘agency’ and their relevance to the communicative constitution of organizations, may not be able to lead to consensus about these terms, but at least it may increase the clarity of their development in different theoretic traditions.
In developing this article, I have been influenced by structuration theory and by a concern to analyze the constitution of formal organiza- tional structure using structurational assumptions and central concepts. Anthony Giddens’ emphasis on the concept of agency makes my own treatment of that concept very predictable; Giddens has also written enough about organizations that the special relevance of certain proper- ties of text is rather clear. In this paper I typically follow his lines of argument.
I will operate under two assumptions in this paper. First, technical definitions and conceptualizations are arbitrary and under the control of the theorist, but conceptual choices have consequences. I care which definition we use for the term ‘text’ mainly because some definitions eliminate certain facets, distinctions, and implications of the common- sense meaning while emphasizing others. I will argue that some specific facets and distinctions need to be retained to guide our analysis toward some specific phenomena of organizational communication. Second, I begin this analysis by examining the concepts of ‘text’ and ‘organization,’ which I use as a touchstone to lay out important distinctions involved in the concepts of ‘text’ and ‘agency.’
Definitions and Properties of Texts What conceptions and distinctions are important to retain or emphasize when we consider the term ‘text’? In using this term (and most others) semi-technically, we are drawing on the colloquial meaning of the term and relying on its metaphorical force, while at the same time casting it as a reference point for similarities to and differences from other terms in our theoretical stock. If we collapse common distinctions in the framing of central terms, we may be leading ourselves and others to ignore important differences and inconsistencies; we may also be stretching ordinary connotations beyond their actual range of validity. I will defend certain distinctions that I find in common usage—but of course that means that I will ignore or derogate others. My ultimate ground of defense is that the distinctions help us in understanding the role of texts in constituting organizations.
My first attention will be to the concept of ‘text.’ This concept has great currency in the cultural sciences, but has brought confusion as well as enlightenment. For my discussion, it will help to focus instead on the term(s) ‘a text.’ The more generic term ‘text’ is usually either a label for a medium, especially the ‘hard’ print medium but more generally any medium having the first two characteristics mentioned below, or a term
Organization 11(3) Articles
356
referring to an assemblage of texts, as encountered in the stream of reading consciousness or in cruising the web. I will argue that it is important to retain a distinction between ‘a text,’ ‘an assemblage of texts’ having no necessary principle of design or sequence, and ‘retained results of interaction processes.’ Actually, I will try to make and valorize a number of distinctions among these terms.
Text as Composed of Signs and Symbols For me ‘a text,’ or ‘text,’ has certain important properties, which I will discuss in descending importance. First, ‘text’ is composed of signs or symbols, especially words and numbers. That property excludes most embodied performances and everyday social activities from the realm of ‘text.’ Take, for instance, a family meal where people pass out plates, napkins, etc., take bites of food, talk, etc. We can take part in, or experience, or interpret the events of the meal, but it is not ‘text’—we cannot ‘read’ it in the normal sense of ‘read.’ Of course, we can transcribe the talk of the meal and read the transcript—it would be a text. And we can compose a description of the meal—another text. But thinking that the meal is itself a text is confusing map and territory. (I am not assuming that all texts are maps or representations.)
The outcomes of the meal are also not generally symbolic or, thus, textual. We might gain weight from the meal, or even be poisoned by it. We might become angry during conversation, and retain both discursive memory of the episode and a visceral reaction to it. We might have learned, or begun to learn, to taste wine appreciatively, and memory of the wine flavors may linger. We are socialized humans, so symbolic representations are indubitably involved, in multiple ways, in the reten- tion of these results. But no purpose is served by regarding the meal as the writing of a text.
Why is this claim important for organizational analysis? This question will be addressed in more depth in the discussion of features of organiza- tions. But one reason is that a number of symbolic abilities and techno- logies are available for use with streams or arrays of symbols. One is symbolic formulation itself—putting ideas or experiences into words or counting things, for instance. Humans are so fully linguistic creatures that I would call discursive consciousness the result of an ability, not a technology, but the production of counts or lists or the composition of minutes of a meeting are closer to what might be aptly labeled technologies—studied, rule-governed courses of action that fashion material (speaking broadly) products with recurrent valued features. Some of these technologies are interpretive: legal analysis, for instance, deserves the label of technology because of its reflective, regulated generation of legal implications of any written text, based on its explicit content and circumstances of production (Ong, 1982). Accounting and
Text, Agency, and Organization Robert McPhee
357
finance are equally numerically based technologies generating assess- ments of a company’s financial status from numerical or ordered records.
The second reason symbolicity is relevant is that symbolic formula- tions can have reflexive application to other organizational realities—to other texts and communication processes, as well as to material resour- ces, practices, and members’ positions and relations. Symbolic iterability and reference are the primary way in which different settings, and even different texts, interrelate.
The Permanence of Texts A second property of a text is that it has a sort of permanence, for which terms such as ‘written down’ or ‘inscribed’ are frequently used. A child’s ‘Etch-a-Sketch’ is not a good model for text. Of course, the prototype we all think of is text written on some enduring material—a canvas, to borrow one of Taylor and Van Every’s (2000) metaphors, but a computer’s hard drive should also qualify. I think this is the second primary property we have in mind when discussing ‘text.’ It is also the property that primarily motivated my analysis of organizational formal structure as inscribed (McPhee, 1985) and Dorothy Smith’s discussion (1984) of the textual mediation of organizational interaction, as well as the property behind Taylor and Van Every’s conversation–text dialectic.
Permanence is important for several reasons that are basic to organiza- tion. It is necessary for record-keeping. Since Weber’s analysis, the importance of files, stipulations of formal rules, etc., has been noted. Written statements can be copied, distributed, preserved, and given a guarantee of validity that can be useful in legal and legalistic proceed- ings. Their distribution can be limited, as can access to places where they are stored.
Inscription is also important in allowing reflective analysis: if you have a set of rules (or anything) written down, you can look it over, check for omissions or confusions, ponder its implications, etc.—interpret or pro- cess it in a strong sense. Giddens (1987a) notes the arguments by Ricoeur (1981) and others that writing is importantly different from spoken interaction. In conversation, common situational immersion facilitates a shared interpretive background that facilitates mutual understanding and allows misunderstandings to be quickly detected and corrected. On the other hand, as Ong (1982) argues, writing is so different a practice that it deserves to be called a separate technology—it does not spontaneously emerge through children’s social experience. Texts typically include enough context-generating information that they can be interpreted fairly similarly across a wide spectrum of contexts. A set of supportive tech- nical practices emerges where written records have come to serve as a foundation of organization. For instance, Clanchy (1993) records the struggles of clerks in medieval England to develop the habit of dating letters and records to facilitate later analysis. Those of us who study with
Organization 11(3) Articles
358
email files and database set-ups have to sympathize with their pain, but those pains enabled modern organizational forms. In other words, per- manence has a dual sense: both the material inscription and the founda- tions for its interpretation, in knowledge and context, endure far beyond the immediate setting of a communication episode.
Among the technologies that can be applied to a text are, first, hermen- eutics and interpretive methods. These include the kind of narrative analysis carried out by Greimas (1987). Of course, many of the insights from these apply to spoken discourse and narratives as well, but Giddens follows Ricoeur (1981) in arguing that only with the distancing from the author characteristic of the text does the need for and development of such methods come. Equally important are the technologies of text production, ranging from the dating of records as mentioned above to the standard form stipulated for memos (Orlikowski and Yates, 1992). One other type of technology is important to note: the recent development of knowledge management techniques that allow the processing of verbal texts to generate descriptions of texts and text arrays, or to find specific content themes in large text assemblages. The simplest of these use key words and similar descriptors; for a fuller arraying of options, see Corman et al. (2002). The point is that texts are treated in ways sensitive to their symbolic content but not dependent on understanding and interpretation, ways that are increasingly important to organizations because they lever- age the specific power of organizations over information.
It is important to recognize that texts are not the only permanent elements in societies, not even the only enduring symbolic realities. The memories of organizational members can include words and performed discourse. Moreover, all sorts of material contextual features are perman- ent and thus can constrain us, remind us of organizational matters that a text might also capture, etc. But texts are distinctively functional in their accessibility to multiple people, their ability to be preserved in a legit- imate form, and their flexible utility. Minutes are kept of organizational meetings simply because of these advantages of text over the memories of members present at the meeting.
Text as Coherent Structure There is one more property of a text that deserves development; that is, a text has a certain level of coherent structure. I intend this claim to be minimal, not total—‘coherent’ means ‘coherent enough to be useful in everyday or organizational praxis,’ not ‘completely consistent and self- sufficient.’ I certainly would not claim that organizational texts have a single coherent meaning—that would place them above the aporia-ridden modern novel. Indeed, among the examples of ‘an organizational text’ that I would advance are personnel records, financial statements, and job descriptions, none of which convey a meaning as a parable would. But, for any of these examples of texts to be useful in textual practice, interpretation, or processing, they must have a certain level of orderliness
Text, Agency, and Organization Robert McPhee
359
and a beginning and end. Here, structure or order is practice relative. Hypertexts would not be texts if straightforward reading is our practice of choice, since a hypertext’s multiple links insure that reading produces arbitrary, unique, often puzzling results. On the other hand, a structural analysis might give a very clear answer to questions about the nature of the hypertext. As another example, a set of financial records that would be boring and overwhelming to read might give clear profit-and-loss outcomes if submitted to practices of financial analysis.
To a certain extent this property is an extension of the symbolic nature of a text. In using any technology or practice with respect to a symbolic formulation, the formulation must have enough order for the practice to be employed. A text is very different in nature from a set of experiences accessible in individual or collective memories of organization members. Memories, even if reliably persistent, involve multiple senses and per- spectives; unless reconstructed and rationalized, they amount to an incoherent welter. That is far from saying that they are useless; for instance, collective memories include the mutual understandings that Giddens (1976) labels ‘structure,’ the rules and resources drawn on and reproduced in interaction. They are very different in function from the organizational records or texts whose distinctive powers enable organiza- tions. So I contend that using the term ‘text’ as a metaphor for a whole range of communication processes mixes apples and typewriters, when our goal is to type a paper. A conceptual set that can account for the nature of organizations needs distinctions that highlight the permanence and structuredness of a text. To understand the role of texts in constitut- ing organizations, it is important to review definitions and key properties of organizations.
Properties and Features of Organizations Organizations are the supreme institutions of the late modern age, both because of their wealth and power and because they have properties and powers that have decisively shaped that age. Which properties and powers are especially important in explaining the relevance of text to organizations? I will start with a list of commonsense properties and note their dependence on text, then go to specific features that Anthony Giddens (1987b) sets forth as an explanation for the power of the organizational form.
Organizations as Stable Formations First, organizations are composed of people who impose and bring about relations and interaction among themselves as members. If General Motors closed one day and re-opened as a giant machine, with chutes at one end for raw materials and conveyer belts at the other carrying out finished cars, we would now call it a machine, and reserve the term ‘organization’ perhaps for a group of people officially owning and operat-
Organization 11(3) Articles
360
ing it. This is an important claim today because conceptions of ‘virtual organization’ seem to threaten to make human membership incidental; organizational analysts must not be misled by such reifications. But it is important relative to the phenomenon of text, too. People can interpret texts because they are symbolic and because people have mastered textual technologies such as reading. Thus, they can use texts as enduring guides to action across changing or dispersed contexts, geographically or temporally. A forest ranger, for instance, has at his or her disposal manuals reiterating policies and schedules that can facilitate adherence to park service doctrines despite the ranger’s isolation (Kaufman, 1960).
Second, the organization is a system of relations among its members, and these relations have some properties that I can discuss as a bundle. Organizational membership and relations are prototypically (a) stable, (b) artificially constructed, and (c) power laden. All of these properties are exemplified in the hierarchical organization chart (which is a good example of a text). Of course, the organization chart is a prescribed snapshot that cannot prevent or control the emergence of relationships of solidarity and resistance, as well as practical routine and informal power. But the formal structure is a prescribed context that occasions and is constantly relevant to such organizational processes.
The properties of text mentioned above give it special importance as the ‘medium’ (Smith, 1984) of prescribed organizational relations. First, its symbol system can include both general cultural terms and patterns along with local, artificial signs. It can be used to name members in a list that establishes the bounds of organizational membership, and assign them to artificially generated roles and tasks. Compensation, the employ- ment contract, and task assignments and requirements can be fashioned in a way that is legitimate in the eyes of members reading it, yet legally binding and parallel in form for all members. The symbolic terms of texts such as these set up the vocabulary for power negotiations inside and outside the organization. The text’s permanence is also important. A written list of members and their roles is present or accessible for the long term. The act of creating such a list is exactly what is meant by the artificial construction of an organization. The list states roles and thereby creates an initial distribution of power. Moreover, since it is material, the list’s distribution can be controlled, and the knowledge it carries can be made available to some members but not others—a supportive basis for power, as Smith notes (1984). Finally, the list’s coherence allows it to be counted, used as a basis for budgetary planning, and technologically treated in multiple respects.
Third is a property that is implicit but not obvious in the notion of reflexive monitoring and control: organizations are multi-site. In particu- lar, the local accomplishment and coordination of work (on the shopfloor or in similar loci) call on a different logic than the reflexive design and monitoring of work (via planning in the manager’s office). In their different sites, work, management, and executive control, as well as
Text, Agency, and Organization Robert McPhee
361
manufacturing versus sales, involve different mind-sets and practices. Thus, the assumption that, for example, the logic of guard–prisoner relations is the same as the logic of overall prison administration cannot be sustained (McPhee, 1985; McPhee and Poole, 2001; McPhee and Zaug, 2000). Along one dimension especially, fundamentally different per- spectives are involved in sites of decision-making about formal structure, inscribing it, contesting it, and simply conforming to it (McPhee, 1985). Separation among these sites reinforces power divisions in the organ- ization.
Organizations can be multi-site only because texts exist. Texts can be written in some sites to control other sites. Some of these texts even set up the other sites (and constitute themselves) as structural features. The logic of text encoding in one site can be different from the decoding logic in another site, yet still bring about coordination and control across sites (though not exactly in the way Hall, 1980, describes). And of course, sites can differ partially in the texts they are empowered or assigned to create and in their access to texts.
Fourth, an organization is, in Giddens’ words, ‘a social system which is able to “bracket space–time,” and which does so via the reflexive mon- itoring of system reproduction and the articulation of discursive history’ (1987b: 153). The organization, as a social form, is a special locus of institutional reflexivity, and Giddens argues that this property is very important in explaining the success, the impact, and the distinctiveness of organizations. This conceptualization reframes the commonsense properties noted above and is developed more fully in the next section.
Organizations and Institutional Reflexivity The concept of organization has a place in Giddens’ (1990, 1991) more general theory of modernity. To summarize, Giddens centers his attention on three main elements that explain the dynamic and global character of the modern age: the separation of space and time (through mobility and the uniform scaling of time), the disembedding of institutions (through the replacement of traditional routine and personal trust by expert systems and abstract tokens such as money), and institutional reflexivity, the regular use of knowledge about social life as a resource for guiding and even constituting the social order. The latter involves both the design of social settings to allow for knowledge and control, and the use of that knowledge to guide action within the settings. Similar to generating personally reflexive to-do lists and calorie counts (Giddens, 1991) are the control media present in the chief ‘institutional dimensions’ of moder- nity: capitalism, industrialism, surveillance (mainly as a medium of state control), and control of the means of violence (again by the state). Giddens argues that developments on all four dimensions help explain the rise of modernity, thus avoiding monocausal theories such as tradi- tional Marxism. The peculiar powers of organizations are the basis of
Organization 11(3) Articles
362
modernity on all of those dimensions, although Giddens does not develop this point.
Organizations, then, embody the principle of institutional reflexivity to bracket space–time, that is, to separate traditional connections of times and places and to reintegrate them in a reflexively designed way (Giddens, 1991: 17). Giddens emphasizes three organizational features involved in this reflexivity, to which I add a fourth and perhaps more basic one.
One feature is surveillance, a general feature of modernity but one that is especially powerful in organizations. Giddens notes that surveillance has two aspects: (1) the collection, coding, and retrieval of information, and (2) the direct supervision of activities. The first is a crucial feature of organizations, because it is a necessary condition for informed ‘control over the timing and spacing of the activities’ of organizational members (1987b: 154). Giddens makes special reference here to the notion of ‘system integration,’ by which he means the processes and institutions that tie together spatially distant parts of a social system. Surveillance allows organizations to span large distances, while still maintaining enough control to be ‘power containers.’
Giddens mentions accounting as one medium of surveillance, emphas- ized first by Weber; the second is the file or dossier, the medium of personnel management and of the ‘articulation of discursive history’ mentioned above. Surveillance as information integration is one place where Giddens’ (1987b) analysis merges with Foucault’s (1977) analysis of discipline.
Texts, of course, are essential to the whole phenomenon of surveil- lance. Giddens emphasizes that ‘[w]riting did not originate as an iso- morphic representation of speech, but as a mode of administrative notation, used to keep records or tallies’ (1987a: 41). Symbolic formula- tions are used to make enduring records of information gathered in surveillance. Inscribed records can be moved and collected; they can be stored in a single place to help create the organization’s capacity as a ‘power container.’ Written records or statements can also be duplicated and disseminated, widely or narrowly. Multiple technologies of text treatment allow empowered members to use such records. Texts also can be used imperatively, to guide actions or constrain outcomes. They can thus contribute to the process of system integration; they can also reproduce hierarchic or less formal power divisions of the organization. Written traces of formal structure can be found in payroll records, phone and email directories, and office maps. In all these ways, they help perform the bracketing of space–time that Giddens (1987b) cites as a key organizational feature.
A second juncture with Foucault (1977) occurs in Giddens’ (1987b) second key feature of institutional reflexivity in organizations: knowl- edgeably designed locales. Even traditional temples and palaces were more than symbols of power; they were ‘power containers,’ a favorite
Text, Agency, and Organization Robert McPhee
363
concept of Giddens (1987a,b), here referring to sites where an admin- istrative staff and recorded information were gathered to plan and coor- dinate operations and supervision. The design of organizational locales often facilitates direct supervision, as Foucault argues, whether by phys- ical or mediated oversight. But Giddens develops several differences from Foucault’s account. First, he rejects ‘total institutions’ such as the prison as prototypes for modern organizations. For one thing, typical modern organizations presume a separation of private home life from the more public organizational environments, both in work organizations and in custodial centers such as schools. For another, the concentrated design of the organizational locale typically housed the administrative staff itself, and the locale generated authoritative power by bringing top officials into ‘presence availability’ to one another and to stocks of information, even in cases where many regular employees worked ‘in the field,’ outside of direct supervision. Just as ancient empires spread over vast domains, Giddens (1987b) argues that modern organizations achieve system integration over space and time by emphasizing positional duties and abstract reports, not personal relations of loyalty. He notes that the offices of top officials are designed to allow privacy and to rely on trust, not on direct supervision.
Here again, texts are an invaluable if not necessary medium for organization. As demonstrated earlier, the text form is what grounds the development of the modern array of reflexive technologies that make modern institutional reflexivity so much more powerful than the reflex- ivity of prior ages.
The third feature of modern organizations that Giddens (1987b) emphasizes is the ‘time–space organizing device,’ that is, timetables, schedules, and plans. Here his analysis clashes with that of Weick (1979), who emphasizes the retrospective sense made by plans. For Giddens, and for researchers such as Zerubavel (1981) and Beniger (1986), precision timing is essential in modern complex organizations, not just planning. Giddens notes that managers with highly responsible positions tend to have schedules that are not determined by an overall organizational timetable. Even before the days of Microsoft Outlook, managers kept diaries so that other members could track and determine their move- ments and activities. Devices such as these are, precisely, texts.
Another topic that Giddens mentions, organizational hierarchy, is not one of his key features of institutional reflexivity, but it develops the formally structured character of organizations. Surveillance arrange- ments, the design of spatial zones, and timetables are all media that reflect stipulated formal relationships and rules of the organization. Moreover, authorized members typically determine them, and authority is more or less clearly stipulated in formal documents that articulate organizational structure. Formal structure is intrinsically discursive and textual, both because it typically has legal status derived from the corporate charter and the employment contract, and because explicit
Organization 11(3) Articles
364
discourse is the most flexible medium for reflexively prescribing a range of activities. Formal structure itself works in two important ways: first, it is a substitute for more occasional communication, avoiding having to set up new authority arrangements every day; and it is authoritative meta- communication, setting up a frame of reference that serves as a constraint for all other communication (McPhee, 1985; McPhee and Poole, 2001). Job titles, department names (‘sales’ and ‘loans’ in an auto dealer’s concern), and acronyms, for instance, are treated as preformulated, unquestioned realities even in informal or resistant discourse. ‘Oh, they do things funny over in sales’ might refer to an office location, a group of people, or a boss’s peculiarities, but it involves colonization of informal chat by the text-based organizational system.
The structurational conception of organizations as embodying institu- tional reflexivity is complex. The basic argument is that organizations are media of administrative power through control over time–space and information, as manifested in these four features. These features are explicitly oriented toward reflexive control. If they are ineffective, coord- ination or trust will fail, leading to a palpable need for system redesign. But the system and its possible redesign will be accomplished through text-dependent processes. Another fact worthy of note is that Giddens (1987a,b) retains Foucault’s (1977) emphasis on power in relations, while reorienting and broadening it to focus on issues of coordination and power wielded through the organization as a social unit. The concept of text contrasts interestingly with Foucault’s central notion of discourse as a central axis of difference between the positions.
To summarize the argument so far: a text is a relatively permanently inscribed symbolic formulation. It is important to keep this conception distinct because it is a vital precondition for the development and continuing existence of organizations as we commonly conceive of them. An enduring inscribed record of organizational arrangements is a com- mon if not necessary medium for stability of membership, relationships, and roles. It is the typical medium for organizing and organizational inception. It is the medium for legitimate stipulation of the formal structure, which in turn shapes power relationships and generates multi- ple sites and perspectives. Finally, enduring formulation is the medium of information storage and processing, which ground the institutional reflexivity that gives organizations their amazing power in society.
Agency This discussion of organization and the properties of text has implica- tions for the conceptualization of agency, ones that are consistent with the stratification model developed by Giddens (1984). In brief, that model supports conceptually the claim of the duality of structuration in time– space. Giddens (1984) argues that agents must be empowered to ‘make a difference’ in the course of events and are capable of acting by drawing
Text, Agency, and Organization Robert McPhee
365
on structural resources that they simultaneously reproduce. They must be knowledgeable about relevant local conditions of action, but not about all conditions or implications of action. In particular, agents share in mutual knowledge of the understandings of others, a condition that allows for meaningful, coordinated action. Giddens’ stratification model focuses on six other features of agency, comprising three levels of consciousness and three characteristic processes. The levels of con- sciousness include the discursive consciousness of facts that we learn or process linguistically; the practical consciousness of tacit implications and of how to ‘do more than we can say’; and the unconscious, involving ontological insecurities barred from awareness to avoid paralysis of action. The three basic processes are reflexive monitoring, an intermittent steering of action achieving conformity to purposes and rules without the constant presence of purpose in consciousness; rationalization, or ongoing power to render our actions accountable through linguistic explanation or routines and linkage to the mutual knowledge of others; and deep motivation of conformity to mutually known routines and meaningfulness requirements. Structuration theory rests on an argument that these features are sufficient to account for our ability to act while avoiding postmodernist and other attacks on the rational, self-aware, integrated agent. These features also seem sufficient to account for agents’ powers to create and use texts in ways that contribute to the constitution of organizations.
This sense of agency might be characterized as internalist because it articulates characteristic powers of human, socially capable agents, rather than stressing structural relations or articulated networks that link agents to external influences or construct them symbolically—linkages emphasized by an externalist sense of agency. Internalist concepts of agency do not deny the importance of the network of external linkages and discourses (Emirbayer and Mische, 1998). But a purely externalist sense of agency gives a stunted explanatory vocabulary, ignores crucial human powers, and ends in vitiating the notion of communication.
A broader challenge to the structurational model of agency comes in the argument that a model of agency focused on the power to engage in effective social action must recognize the context and relations of other agents. However, having discursive consciousness and the other abilities in Giddens’ model does not translate into rhetorical agency if no one will listen to our words. In such cases, Giddens’ dialectic of control might blind us to the fact that agents, who are fully competent as described by the stratification model, may still be unable to make a difference. Agents for Giddens have powers, but he often fails to note their limitations. I think Giddens’ reply is that the most common alternative, a Foucauldian or constraint model of agency, might work well in a prison but it is less insightful in an organizational situation in which domination implies the goal of getting others to act in the dominator’s interests rather than the goal of merely constraining the dominated. In the usual organizational
Organization 11(3) Articles
366
context, the dialectic of control, with its recognition of how everyday domination depends on the subordinate’s abilities and cooperation, is vital for insight and liberation.
This argument is especially relevant to this special issue on text. Perhaps the biggest change in organizational life in developed countries is our growing powers and practices to work with texts. We must retain a conceptualization of agency that includes interpretive powers to forge and understand texts, but beyond that to act on them and react to them in meaningful ways, to account for the full range of organizational commun- ication processes.
Relations to Other Views of Text and Agency This paper originated with a specific purpose: to differentiate between structuration-based senses of text and agency, and those developed by scholars who work with other theories, such as the conceptions espe- cially of Taylor and his colleagues (e.g. Taylor, 1993; Taylor and Van Every, 2000), Foucault’s (1977) views of text and agency, and the views of critical discourse analysts. Taylor and his colleagues have remained influenced by the work of Greimas (1983, 1987, 1988) and Latour (1986, 1987, 1993). Several comparisons distinguish notions of text and agency in structuration from ones developed in conversation–text relationships and critical discourse analysis (CDA). Drawing from Taylor and Van Every (2000), I will distinguish between their use of text and that derived from models of structuration. As they comment in their book:
We take the organizational conversation to be the total universe of shared interaction-through-languaging of the people who together identify with a given organization. (2000: 35)
We use the word text to capture the quite ordinary idea that speech is built of words and phrases, strung together in more or less systematic ways to produce a coherent, understandable piece of language. (2000: 37)
[We] conceive the basis of organization to be in the generation and regeneration, within conversational contexts, of texts: multiplexed (in the sense of composed of many layers that correspond to the simultaneously ongoing occasions of communication that collectively realize organiza- tion), to be sure, but nevertheless, conceptually, texts (or what could be called a layered, multiply interconnectable hypertext). If conversation is the indispensable stream of experience of collective existence and text is how that stream is mediated to us so that it is made meaningful, then communication must exist at the intersection of conversation and text. (2000: 37)
Although Taylor and Van Every’s position is more complex than is conveyed here, the difference between a structuration view and their stance on text is clear. They use text to signify formulated results of conversation, mediated grammatically so as to have an organizational
Text, Agency, and Organization Robert McPhee
367
effect. Even though they recognize the importance of permanence or at least iterability, they emphasize the reflexive, sense-made quality of text—what I have rephrased above, following Weick, as ‘the retained results of interaction processes.’ A text, for Taylor and Van Every, could be an impromptu oral speech summing up a group meeting. They do not emphasize the durability of the material medium a text is typically written on, or the special technologies of production and processing of texts such as personnel records and financial data as texts that have important organizational consequences. Although I concur that collective memory of meaning and relational implications of linguistic and narrat- ive grammars are important in the constitution of organizations, I think the specific contribution of structured, inscribed symbolic formulations has distinctive importance that Taylor and Van Every elide in developing their broad sense of the term text.
The traditions of structurational theorizing also stand in contrast to scholars who rely on Foucault’s (1977) notions of knowledge and power in organizational discourse studies. Although Giddens is influenced by Foucault’s notions of power, e.g. power containers, Giddens’ (1984, 1987b) main response is that Foucault understates and neglects the powers of agency as explanatory resources even in his later writings and that Foucault’s emphasis on total institutions leads organizational ana- lysis away from a balanced view of organizations as constraining but also enabling. I concur that Foucault’s sense of agency remains rooted in the image of the subjectivity of the dominated being fashioned in relations of power and manifested in texts. Giddens’ model, like Foucault’s, decenters the subject, but without demoting agentive powers completely. I also take issue with the way Foucault has narrowed organizational analysis to a single logic of disciplinary domination with inchoate resistance. The notion of text that I discuss leaves room for exploring multiple organizational logics and numerous diffuse textual practices that are partly autonomous.
This structuration view of text and agency also differs from critical discourse analysis, as developed by Fairclough (1995; Chouliaraki and Fairclough, 1999). The notion of text as articulated by Chouliaraki and Fairclough and the effects of texts in modernity that they emphasize parallel views of text developed in this essay. However, the emphasis on text as a condition of organizations, which is central in my account, tends to be peripheral and conditional in CDA as they describe and, especially, exemplify it. Albeit the medium of text is an important condition for the ideological effects they find in the specific registers, grammatical forms, metaphors, and textual language focal in prototypical critical discourse analyses, critical discourse analysts do not tend to explore or elaborate on the impacts of texts through institutional reflexivity, particularly features of space–time organizing devices, power containers, and systems integration.
Organization 11(3) Articles
368
Implications In general, traditional discourse studies need to give emphasis to ‘medium’ and to the range of specific material forms mediating discourse. At the institutional level, the works of Yates (1989), Beniger (1986), and Orlikowski and Yates (1992) on organizational communication genres are good examples of work that analyzes the constitutive role and sig- nificance of text, consistent with structuration assumptions of agency. At the level of strategic interaction, Anderson (2003) illustrates the sig- nificance of text creation in organizations. He studied the differential impacts of a written record of meeting outcomes as they became a stable reference for discussions and decisions in subsequent meetings. More importantly, the development of knowledge management and related practices reinforces a second route of influence for texts in organizations. They can be read and interpreted, and they can be processed in varied ways at a greater or lesser distance from immediate, traditional readings. In both ways, texts have a relevance to the constitution and character of organizations that deserves continued focal attention and research.
References Anderson, D. L. (2003) ‘The Textualizing Functions of Writing for Organizational
Change’, paper presented at the ICA Convention, May, San Diego. Beniger, J. R. (1986) The Control Revolution: Technological and Economic Origins
of the Information Society. Cambridge, MA: Harvard University Press. Chouliaraki, L. and Fairclough, N. (1999) Discourse in Late Modernity. Edin-
burgh: Edinburgh University Press. Clanchy, M. T. (1993) From Memory to Written Record: England 1066–1307.
Oxford: Blackwell. Corman, S., Kuhn, T., McPhee, R. and Dooley, K. (2002) ‘Studying Complex
Discursive Systems: Centering Resonance Analysis of Communication’, Human Communication Research 28: 157–206.
Emirbayer, M. and Mische, A. (1998) ‘What Is Agency?’, American Journal of Sociology 103: 962–1023.
Fairclough, N. (1995) Critical Discourse Analysis: The Critical Study of Language. London and New York: Longman.
Foucault, M. (1977) Discipline and Punish: The Birth of the Prison, trans. Alan Sheridan. New York: Vintage.
Giddens, A. (1976) New Rules of Sociological Method: A Positive Critique of Interpretative Sociologies. New York: Basic Books.
Giddens, A. (1984) The Constitution of Society. Berkeley: University of California Press.
Giddens, A. (1987a) The Nation-State and Violence: Volume 2 of A Contemporary Critique of Historical Materialism. Berkeley: University of California Press.
Giddens, A. (1987b) ‘Time and Social Organization’, in Social Theory and Modern Sociology, pp. 140–65. Stanford, CA: Stanford University Press.
Giddens, A. (1990) The Consequences of Modernity. Cambridge: Polity. Giddens, A. (1991) Modernity and Self-Identity. Stanford, CA: Stanford
University Press.
Text, Agency, and Organization Robert McPhee
369
Greimas, A. J. (1983) Structural Semantics. An Attempt at a Method. Lincoln, NE: University of Nebraska Press.
Greimas, A. J. (1987) On Meaning. Selected Writings in Semiotic Theory. London: Frances Pinter.
Greimas, A. J. (1988) Maupassant. The Semiotics of Text. Practical Exercises. Amsterdam and Philadelphia: John Benjamins.
Hall, S. (1980) ‘Encoding/Decoding’, in S. Hall, D. Hobson, A. Lowe, and P. Willis (eds) Culture, Language, Media, pp. 128–40. London: Hutchinson.
Kaufman, H. (1960) The Forest Ranger: A Study in Administrative Behavior. Baltimore, MD: Johns Hopkins University Press.
Latour, B. (1986) ‘The Powers of Association’, in J. Law (ed.) Power, Action and Belief. A New Sociology of Knowledge?, pp. 264–80. London: Routledge & Kegan Paul.
Latour, B. (1987) Science in Action: How to Follow Scientists and Engineers through Society. Cambridge, MA: Harvard University Press.
Latour, B. (1993) We Have Never Been Modern. Cambridge, MA: Harvard University Press.
McPhee, R. (1985) ‘Formal Structure and Organizational Communication’, in R. McPhee and P. Tompkins (eds) Organizational Communication: Traditional Themes and New Directions, pp. 149–77. Beverly Hills, CA: Sage.
McPhee, R. D. and Poole, M. S. (2001) ‘Organizational Structures and Configura- tions’, in F. M. Jablin and L. L. Putnam (eds) The New Handbook of Organiza- tional Communication, pp. 503–43. London: Sage.
McPhee, R. D. and Zaug, P. (2000) ‘The Communicative Constitution of Organiza- tions: A Framework for Explanation’, Electronic Journal of Communication 10 (1–2), URL: http://www.cios.org/getfile/McPhee_V10n1200.
Ong, W. J. (1982) Orality & Literacy: The Technologizing of the Word. London and New York: Routledge.
Orlikowski, W. and Yates, J. (1992) ‘Genres of Organizational Communication: A Structurational Approach to Studying Communication and Media’, Academy of Management Review 17: 299–326.
Ricoeur, P. (1981) Hermeneutics and the Human Sciences, trans. J. Thompson. Cambridge: Cambridge University Press.
Smith, D. (1984) ‘Textually Mediated Social Interaction’; reprinted in D. Smith, Texts, Facts, and Femininity: Exploring the Relations of Ruling, pp. 209–24. London: Routledge, 1990.
Taylor, J. R. (1993) Rethinking the Theory of Organizational Communication: How to Read an Organization. Norwood, NJ: Ablex.
Taylor, J. R. and Van Every, E. J. (2000) The Emergent Organization: Communica- tion as Its Site and Surface. Mahwah, NJ: Lawrence Erlbaum Associates.
Weick, K. E. (1979) The Social Psychology of Organizing. New York: Random House.
Yates, J. (1989) Control through Communication: The Rise of System in American Management. Baltimore, MD: Johns Hopkins University Press.
Zerubavel, E. (1981) Hidden Rhythms: Schedules and Calendars in Social Life. Chicago: University of Chicago Press.
Robert D. McPhee is a Professor in the Hugh Downs School of Human Communication at Arizona State University. His research interests include organizational hier-
Organization 11(3) Articles
370
archies, organizational knowledge, the communicative constitution of organiza- tions and structuration theory. Specializing in communication theory and methods and in organizational communication, he has served as chair of the Organizational Communication Division of the National Communication Association, as associate editor of Human Communication Research, and as book review editor of Communication Theory. Address: Hugh Downs School of Human Communication, Arizona State University, PO Box 871205, Tempe, AZ 85287–1205, USA. [email: [email protected]]
Text, Agency, and Organization Robert McPhee
371
