Mod 4 reading response

Foucault, Power/Knowledge, and Its Relevance for Human Resource Management Author(s): Barbara Townley Source: The Academy of Management Review, Vol. 18, No. 3 (Jul., 1993), pp. 518-545 Published by: Academy of Management Stable URL: https://www.jstor.org/stable/258907 Accessed: 25-03-2020 03:30 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide

range of content in a trusted digital archive. We use information technology and tools to increase productivity and

facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at

https://about.jstor.org/terms

Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy of Management Review

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

? Academy of Management Review 1993, Vol. 18, No. 3, 518-545.

FOUCAULT, POWER/KNOWLEDGE, AND ITS RELEVANCE FOR HUMAN RESOURCE MANAGEMENT

BARBARA TOWNLEY

The University of Alberta

Drawing on the work of Michel Foucault, I argue that human resource management (HRM) may be best understood as a discourse and set of

practices that attempt to reduce the indeterminacy involved in the employment contract. Here I reread HRM practices from a Foucauld- ian power-knowledge perspective and suggest that this provides an avenue to reorient contemporary, historical, and comparative analy- ses of the area.

A glimpse at human resource management (HRM) texts would seem to give light to a comment by Beer, Spector, Lawrence, Mills, and Walton (1984) that HRM is a series of seemingly disjointed activities. This notion echoes an earlier view of Baldamus (1961: 347) that what is encompassed by HRM's subject matter is "anything from supervision, incentives and profit sharing to machine-paced production, methods of training and em- ployee selection." Often this heterogeneity is excused in terms of HRM's reflecting the ad hoc and reactive nature of its origins (Jacoby, 1985; Niven, 1967; Ozanne, 1967). What the heterogeneity of HRM highlights, however, is the importance of an organizing principle, or analytical focus, as opposed to common sense description, which gives HRM practices a theoretical coherence. HRM's heterogeneity stresses the importance of an order "that turns a set of bits, which have limited significance on their own, into an intelligible whole" (Turner, 1983: 191).

The ordering of material necessarily makes reference to an underly- ing theoretical model because statements are made about what subject matter is important, if not why. Underlying most studies of HRM, although often remaining implicit, is what may be identified as a systems main- tenance or functionalist perspective. Reflecting concerns with improve- ment in efficiency that derive from classical management theory, HRM is an organizational mechanism through which goal achievement and sur- vival may be promoted. Its aim is to make the organization more orderly and integrated. In HRM, connotations of goal-directed activity, inputs and outputs, stability, adaptability, and systems maintenance predominate. From this perspective HRM is the black box of production, where organi- zational inputs -employees- are selected, appraised, trained, devel- oped, and remunerated to deliver the required output of labor. Within this framework HRM practices are all too frequently technique oriented, pre- sented as the tools or instruments that enable the effective attainment

518

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 519

of goals-an approach that behaviorist psychology has a tendency to reinforce. Research concerns are usually to make HRM practices more efficient, and they reflect the belief that knowledge of them, through the services of its handmaiden, science, will progressively be made more accurate. As such, the study of HRM stands well within a modernist, and largely positivist, tradition.

As a series of categories, selection, appraisal, training, and so forth, have become so familiar that they are not seen as a "way of ordering," but as "an order which is in the phenomena" (Turner, 1983: 192). Here, I wish to set aside these traditional methods of ordering and examine an alter- native and, it is argued, more productive line of analysis. I do so by drawing upon the work of Michel Foucault, Professor in the History of Systems of Thought at the College de France, Paris, until his death in 1984. One of the central concerns of Foucault's work is to dispel self- evidencies, that is, to indicate that although elements are part of a fa- miliar landscape, they are not "natural," or part of a naturally existing order. Through a number of different examples, Foucault has shown how what counts as truth depends on, or is determined by, the conceptual system in operation. This idea is illustrated at the beginning of one of his early works, The Order of Things (1970), where he quotes a passage from a short story by the Argentinian writer Jorge Luis Borges. This passage cites a Chinese encyclopedia that lists the classification of animals in the animal kingdom. They are classified according to the following catego- ries: (a) belonging to the Emperor, (b) embalmed, (c) tame, (d) sucking pigs, (e) sirens, (f) fabulous, (g) stray dogs, (h) included in the present classification, (i) innumerable, (k) drawn with a fine camel hair brush, (1) et cetera, (m) having just broken the water pitcher, and (n) that from a long way off look like flies (Foucault, 1970: xv). As Philp (1985: 70) noted, "When we classify objects we operate within a system of possibility-and this system both enables us to do certain things, and limits us to this system and these things." Foucault's work provides an avenue to illustrate how established ways of ordering limit our analysis, and it also introduces different ways of seeing.

POWER, KNOWLEDGE, AND THE SUBJECT

Ostensibly, Foucault's work covers a wide range of subjects- nominally, psychiatry (1967), the human sciences (1970), medicine (1973), the penal system (1977), and sexuality (1981, 1985, 1986).1 There are, how- ever, a number of underlying themes that those who are familiar with his work might identify as central concerns. Three of these concerns are per- haps most apparent: power, knowledge, and subjectivity. A necessarily brief explication of Foucault's understanding of these concepts is impor- tant to illustrate the relevance of his work for HRM.

1 Dates cited are the first English translations of Foucault, except for essays printed in edited collections.

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

520 Academy of Management Review July

Dreyfus and Rabinow (1983) stated that part of the difficulty with, and the genius of, Foucault is that he did not accept the usual sociological categories, both in the questions he posed and the concepts he intro- duced. This characteristic is apparent in his refusal to begin his analyses with self-evident concepts: individuals (e.g., managers, employees, workgroups) or institutions (e.g., organizations, the State). The reasons for this refusal lie in Foucault's desire to understand more fully power relations, that is, how mechanisms of power affect everyday lives. His work is critical of views of power that depict it as a commodity (something held or possessed; something embodied in a person, institution, or struc- ture; something to be used for organizational or individual purposes). He wrote, "Power is not something that is acquired, seized or shared, some- thing one holds on to or allows to slip away" (Foucault, 1981: 94). Rather, power is relational; it becomes apparent when it is exercised. Because of this relational aspect, power is not associated with a particular institu- tion, but with practices, techniques, and procedures. Power is employed at all levels, and through many dimensions. Denying the concept of power as a commodity has implications for the way it is studied. Thus, questions such as "who has power?" or "where, or in what, does power reside?" are changed to what Foucault termed the "how" of power: those practices, techniques, and procedures that give it effect. He also offered a different understanding of power as, for example, in the political di- mensions of visibility (rendering something or someone visible) "power is exercised by virtue of things being known and people being seen" (Foucault, 1980: 154).

This view of power informs the concept power-knowledge, "an an- choring device for the unity Foucault gave to his work" (Eribon, 1991: 127). Such a view is perhaps most apparent in Foucault's essay on governmen- tality, in which he examined techniques used in the management of pop- ulations (Burchell, Gordon, & Miller, 1991; Foucault, 1991c). Governmen- tality is a neologism derived from a combination of government and ra- tionality. Government is understood to be not simply political institutions but in a broader sense "the conduct of conduct: a form of activity aiming to shape, guide or affect the conduct of some person or persons" (Gordon, 1991: 2). Rationality is the idea that before something can be governed or managed, it must first be known. It is the acknowledgement that govern- ment is intrinsically dependent upon particular ways of knowing. Pro- grams of government, for example, require vocabularies, ways of repre- senting that which is to be governed, ways of ordering populations (i.e., mechanisms for the supervision and administration of individuals and groups). Rationality is dependent upon specific knowledges and tech- niques of rendering something knowable and, as a result, governable. Governmentality, therefore, is a reference to those processes through which objects are rendered amenable to intervention and regulation by being formulated in a particular conceptual way. Governmentality places an emphasis on regulatory systems, processes, and methods of thinking

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 521

about or perceiving a domain, especially those which may be translated into written inscriptions that claim to authentically depict it. These meth- ods, and so forth, include mechanisms for inscription, recording, and calculation: ways of observing; and ways of coding, (e.g., in balance sheets, audits, population tables, censuses). Once an arena is captured or inscribed, knowledge about it may then be translated to other decision- making bodies.

Foucault does not, therefore, acknowledge a neutral concept of knowledge formation, as his play on the word discipline-at once a branch of knowledge and a system of correction and control- exemplifies:

The exercise of power itself creates and causes to emerge new objects of knowledge and accumulates new bodies of infor- mation ... the exercise of power perpetually creates knowl- edge and, conversely, knowledge constantly induces effects of power.... It is not possible for power to be exercised with- out knowledge, it is impossible for knowledge not to engender power. (Foucault, 1980: 52)

Knowledge is not detached and independent; as a source of illumination, it is integral to the operation of power. From this perspective, procedures for the formation and accumulation of knowledge, including the scientific method, are not, therefore, neutral instruments for the presentation of the real (Steffy & Grimes, 1992). Indeed, scientific discourse and the institu- tions that produce it are part of the taken-for-granted assumptions of knowledge that should be questioned (Knights, 1992). Procedures for in- vestigation and research (e.g., the use of a classificatory table), although operating as a procedure of knowledge, can operate equally as a tech- nique of power. Knowledge is the operation of discipline. It delineates an analytical space and in constituting an arena of knowledge, provides the basis for action and intervention-the operation of power.

The concept power-knowledge has two implications. First, by show- ing how mechanisms of disciplinary power are simultaneously instru- ments for the formation and accumulation of knowledge, Foucault chal- lenged positivism's portrayal of them as independent (Knights, 1992). He dissolved the traditional distinction between power and knowledge, whereby knowledge may lead to power, or power may be enhanced by the acquisition of knowledge. The two are not depicted as having an inde- pendent existence. They are coterminous. Second, according to Foucault, rather than being external, or something which operates on something or someone, power is integral or productive in the sense that it creates ob- jects. Power is the desire to know. Power is not negative; on the contrary, it is creative. As Foucault (1977: 194) wrote:

We must cease once and for all to describe the effects of power in negative terms: it "excludes," it "represses," it "censors," it "abstracts," it "masks," it "conceals." In fact, power produces;

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

522 Academy of Management Review July

it produces reality; it produces domains of objects and rituals of truth. The individual and the knowledge that may be gained of him belong to this production.

This creative element of power introduces a third important aspect of Foucault's work: the concept of the individual. Traditional approaches in the social sciences have taken the individual as a self-evident unit with

which to begin analysis-an observable reality, a unit continuous in time, possessing an essential personal identity. The individual is the basic unit of analysis underpinning many HRM practices, that is, an es- sential human subject whose nature is to be discovered or uncovered, and who is to be motivated through the exercise of correct procedures of re- cruitment, selection, appraisal, training, development, and compensa- tion. Thus, research into HRM practices has been designed to build on previous efforts, in order to make such practices more accurate and less subjective, thereby aiding rational decision making.

Rather than perceive the individual as reducible to an internal core of

meaning, from a Foucauldian perspective, the human subject is not "given" but produced historically, that is, constituted through correlative elements of power and knowledge. "Certain bodies, certain gestures, certain discourses, certain desires come to be constituted as individuals. The individual . . . is I believe one of [power's] prime effects" (Foucault,

1980: 98). When the individual loses his or her privileged epistemological status, it becomes possible to see the individual as a product of the social techniques of power, a perspective that highlights the importance of both identity and identity-securing strategies in the reproduction of power re- lations (Knights & Willmott, 1985). The focus of analysis becomes the "knowability" of the individual-the process by which the individual is rendered knowable, or the process by which the individual is constructed or produced. Such an approach to an understanding of identity is explic- itly adopted in poststructuralism and feminism, where individuality is not

seen as being fixed in its expression (Alvesson & Willmott, 1992; Hen- riques, Hollway, Urwin, Venn, & Walkerdine, 1984). Identity is contingent, provisional, achieved, always in process; not a given or an essential component of the subject. "Identities are not absolute but always rela- tional; one can only ever be seen to be something in relation to something else" (Clegg, 1989: 159). This emphasis on the relational and the consti- tuted nature of self, although strong in the feminist literature (Gilligan, 1982; Keller, 1986), has only recently been incorporated into psychology (Rose, 1990) and occupational psychology (Henriques et al., 1984; Hollway, 1991) and still remains relatively rare as an approach that informs more general studies of management (Calas & Smircich, 1990; Roberts, 1984, 1991).

Described as analyses of regimes du savoir, Foucault's work traces the formation of knowledge and the power structures that result. By an- alyzing the processes involved in the construction of knowledge- processes of classification, codification, categorization, precise calibra- tion, providing tables and taxonomies, in short providing nomenclatures

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 523

or the process of naming-he has shown how discourses on sanity, health, knowledge, and punishment have been developed and the impli- cations that this has had for the individual (the criminal, the madman) who becomes an object of knowledge. He has shown how madness, for example, is not a pregiven entity, but something constituted historically through discourses as both an object of knowledge and a target of insti- tutional practices. In other words, madness is not an object whose history must be discovered, but it is the product of a discourse. Psychiatric knowl- edge invents, molds, and carves out its object-mental illness.

The question that is prompted by a body of knowledge from a Foucauldian perspective, therefore, does not concern the truth or falsity of such processes, or whether the knowledge that is generated is objective or subjective. Issues are not posed in such terms. Rather, the emphasis is on what is involved in rendering an arena or an individual knowable: What are the processes by which they become known? How do these processes become established and used? What are their effects? The em- phasis is on the techniques through which human beings understand themselves and others. It emphasizes the importance of studying, in de- tail, the actual practices that introduce domains and individuals to enun- ciation and visibility-the mechanisms of inscription, recording, and cal- culation that constitute the discursive practices that make knowledge of both arenas and the individual possible. The focus, therefore, is how disciplinary practices operate to create order, knowledge, and ultimately, power effects. Although Foucault never directly addressed production, the view that power operates and is materialized at the level of knowl- edge illustrates how the logic, or rationality, of power is transferable across different domains. This idea provides the basis for a rereading of HRM.

FOUCAULT'S IMPLICATIONS FOR HRM

To illustrate the relevance of Foucault's work for HRM, we must return to the basic building block on which HRM practices are premised-the employment relationship. Analysis of the employment relationship has had a long history, with debate principally colonized by two disciplines: labor law and economics. Within the neoinstitutionalist field of econom- ics, debate has been stimulated by Williamson's (1975) work on transac- tions which, following Commons (1934), he identified as "the ultimate unit of economic investigation" (Williamson, 1975: 254). Williamson's work has centered on the nature of the mediating mechanisms available to the parties to transactions to ensure agreement, given the operation of hu- man factors such as specifically bounded rationality, uncertainty, and idiosyncratic knowledge, and this work has largely been conducted in terms of the relative advantages and disadvantages of the price mecha- nism or administrative rule.

The employment relationship is a paradigm case of a transaction, as Williamson (1975: 59) noted, "Supplying a satisfying exchange relation is

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

524 Academy of Management Review July

part of the economic problem, broadly construed, [which] has special relevance where the employment relation is involved." The employment relationship describes only in general terms services to be provided, al- lowing details to be elaborated later. Baldamus portrayed it in the fol- lowing way:

Though it [the employment contract] stipulates precise wage payments for the employer, nothing definite is ever said about effort or efficiency; nor anything about the components of ef- fort, the acceptable intensity of impairment, the intolerable degree of tedium or weariness. Instead it merely mentions hours of work, type of job, occupational status and similar external conditions. At the most there are vague and con- cealed references to an implied level of effort.... Thus the formal contract between employer and employee is incom- plete in a very fundamental sense. (Baldamus, 1961: 2)

Transactional cost analysis has been incorporated into organization- al studies (see Hesterly, Liebeskind, & Zenger, 1990), most notably in Ouchi's (1980) distinction among markets, hierarchy, and clan and his suggestion that market and administration may be interpreted as differ- ent manifestations, in terms of detail and emphasis, of systems of orga- nizing, rather than as two contrasting or conflicting mechanisms (Ouchi, 1980: 132). Debate, however, has centered on Williamson's problematic, that is, the issue of efficiency. There has been criticism of depicting the employment relationship in neoclassical terms as a transaction between free and equal parties. The labor process literature, which is drawn from Marx's Capital, dismisses this view as fictitious and depicts the exchange in terms of the powerful or powerless, thus emphasizing issues of man- agerial control or domination, rather than efficiency (Burawoy, 1979; Ed- wards, 1979).

The dominance of an economic paradigm and reifications that have gained the status of institutions (the market and administration) have blurred the basic unit of analysis that Williamson's framework high- lights, namely, the gap, or space, that inevitably exists in a transaction between the parties, in terms of what is promised and what is realized, the indeterminacy of a contract. In the employment relationship this gap is the contrast between promise and performance or the capacity to labor and its realization. In this article I wish to refocus attention on the infor- mation gap that arises from the indeterminacy of a contract and, in par- ticular, the requirement to articulate the space that exists between ex- pectation and deliverance of performance.

An underlying presupposition of the dominant neoinstitutionalist ap- proach is that information to determine transaction costs is discovered or uncovered. Thus, for example, relevant prices have to be discovered (Williamson, 1975: 4), and the criteria for efficiency is whether there is sufficient knowledge. Williamson recognized language and computa-

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 525

tional limits. The problem, however, is one of communication. Individu- als may be "language limited" (Williamson, 1975: 255); there may be "in- formation processing limitations of both originators . . . and receivers" (Williamson, 1975: 63). He stated, "The costs of writing and executing complex contracts across a market vary with the characteristics of the human decision makers who are involved with the transaction on the one hand, and the objective properties of the market on the other" (William- son, 1975: 8). Relevant job details may not be identified or accurately described. Information sometimes may be affected. "Sometimes, indeed, the requisite language will not even exist" (Williamson, 1975: 63). Equally, the skills, performance, ability, behavior, and so forth, which knowledge is made of, are preexistent entities. Skills are external and given, fungible or idiosyncratic. Throughout all this there is a view of knowledge as objective. There is the assumption that concepts represent an observable reality. Knowledge is interpreted as giving unmediated access to the world, reflective of an external facticity. As such it assumes the form of being neutral, operating to uncover a naturally existing order.

Williamson's analysis is posited on the positivist distinction between power and knowledge. Attempts to introduce an awareness of power into his framework, although starting from a radically different paradigm, also assume a dissociation between the two and are informed by a con- cept of power, which is portrayed as a commodity. Adopting a Foucaul- dian perspective introduces an alternative analysis of the employment exchange. Following Foucault's eschewal of self-evident categories, the focus is neither institutions (the market, administration) nor individuals (agents, principals). Again, following Foucault, the analysis is not driven by considerations of what (the market, administration) or why (efficiency, shirking, problems of trust, etc.), but how. Also if we are mindful of Mor- gan's (1980) recommendation for a focus in organization theory on the action of organizing, rather than on organizations, the focus of analysis becomes how a relational activity is organized: How is the relational nature of exchange, transaction, or contract and the inevitable indeter- minacy of social relations ordered?

I contend that the provision of knowledge is central to organizing- not in a neutral sense of reflecting what is out there, but a Foucauldian sense of power-knowledge.2 The indeterminacy of a contract provides the analytical space that needs to be rendered governable. The employment relationship is an analytical, conceptual space, which has geographic (at work) and temporal (time at work) dimensions. It also involves a subject the worker. All these dimensions or spaces must be rendered known and articulated before they can be managed. From the employer's point of

2 Ironically, this interpretation is not that far removed from the work of Commons, who identified that "cooperation at work arose from the invention of institutions which produced order out of conflict, defining order as 'working rules of collective action' " (Commons cited in Williamson, 1975: 3, emphasis added).

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

526 Academy of Management Review July

view, there are two principal unknowns in the employment relationship: (a) the detail of activity required and (b) the nature of labor, in embodied form, which is required to carry out the activity. Thus, there is the need for knowledge of two dimensions: the nature of work and the nature of its operator.

HRM is presented here as the construction and production of knowl- edge. It attempts to reduce the space resulting from the unspecified na- ture of a contract. It constitutes a discipline and a discourse, which or- ganizes analytical space-the indeterminacy between promise and per- formance. HRM serves to render organizations and their participants calculable arenas, offering, through a variety of technologies, the means by which activities and individuals become knowable and governable. HRM disciplines the interior of the organization, organizing time, space, and movement within it. Through various techniques, tasks, behavior, and interactions are categorized and measured. HRM provides measure- ment of both physical and subjective dimensions of labor offering a tech- nology that renders individuals and their behavior predictable and cal- culable. In so doing, HRM helps to bridge the gap between promise and performance, between labor power and labor, and it organizes labor into a productive force.

Dividing Practices

The administration of personnel is one of the major requirements of an employing organization. The coordination of large numbers of people, and the ability to differentiate between them-the rational and efficient deployment of a population-requires the development of techniques that enable people to be managed en masse. Managing employees re- quires a vocabulary, that is, a means of knowing and ways of represent- ing and ordering populations. Disciplines provide "tableaux vivants," which transform "confused, useless or dangerous multitudes into ordered multiplicities" (Foucault, 1977: 148). Foucault (1977: 143) wrote:

One must eliminate the effects of imprecise distributions, the uncontrolled disappearance of individuals, their diffuse cir- culation.... [The discipline's] aim was to establish pres- ences and absences, to know where and how to locate indi- viduals, . . . to be able at each moment to supervise the con- duct of each individual, to assess it, to judge it, to calculate its qualities or merits. It was a procedure, therefore, aimed at knowing, mastering and using.

Disciplines begin with the distribution of individuals in space, locat- ing or fixing them conceptually. The art of distribution may use a series of techniques. Foucault (1977) identified three primary methods through which this is effected: (a) enclosure (the creation of a space closed in upon itself), (b) partitioning (each individual has his or her own place and each place an individual), and (c) ranking (the hierarchical ordering of indi-

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 527

viduals). Essentially, what Foucault identified was the process involved in knowing the population and the individual. Discipline organizes techniques of classification and tabulation and introduces individuals into the latter.

Enclosure involves the geographical or spatial separation of a place, and it is reflected in Foucault's studies of the asylum (1967), the hospital (1973), and the prison (1977). In production also, early work organizations tended to be physically enclosed spaces, often surrounded by high walls and fences; they frequently operated under a policy of physically locking the doors once work had commenced (to omit latecomers). Although the physical enclosure of work is now highly modified, some of its conceptual effects remain. The disciplines continue to police the boundaries between work and nonwork. The effects of the discourse of enclosure can be seen most dramatically in the separation between paid and unpaid labor, that is, the public and private divisions that feminist discourse is eager to counter. Work, for example, is conceived of as "employment in the pro- duction of goods and services for remuneration" (Rubin, cited in Pahl, 1988: 13). People are classified according to their location on either side of the enclosure: at home, retired, unemployed, working part-time, and so on. The enclosure of work has had particular implications for definitions of skill and is reflected in gendered divisions of labor. Certain types of labor, particularly emotional labor, nurturing, supporting, and caring are omitted from job descriptions and job analyses, and they receive low levels of remuneration (Hochschild, 1983; Pringle, 1988). Attempts at sus- taining the rigid division between work and nonwork are often breached, however, as is illustrated in Kanter's (1977) work on the "office wife," Pringle's (1988) work on secretaries, and Finch's (1983) illustration of how male managerial careers are enhanced by their wives' social and enter- taining skills.

Within the enclosed sphere of work, individuals are further subject to the art of distribution through partitioning. Partitioning operates on space in a much more flexible and detailed way, and it refers to the divisions that are created internal to the work organization. Partitioning produces both horizontal and vertical divisions between individuals. It may occur geographically or spatially, for example, in the division between head- quarters and other sites; it also can occur through technology, but, essen- tially, partitioning imposes an order by constructing a rational classifi- cation of living beings. This classification (e.g., manual/nonmanual, blue collar/white collar, productive/nonproductive, core/periphery) involves not only the spatial and the analytic, but also the political ordering of people, and it may be enforced through the operation of internal labor markets, the provisions of closed or union shops, and the enforcement of union jurisdictions.

Further partitioning involves ranking, a process of creating a serial, or hierarchical, ordering among employees. This essentially evaluative procedure raises the problem of how relations between beings or things will be conceived: How will an established and ordered succession be-

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

528 Academy of Management Review July

tween units be brought about? Such an order requires a basis for com- parison, that is, a common denominator must be established. Generally, there are two systems of comparing (a) the development of an order through a taxonomy (taxinomia) or (b) the establishment of an order through measurement (mathesis).3 Both procedures define a relation, stat- ing what is to be associated with what. They allow things to be placed in relation to one another, establishing relations of equality and difference. Both also imply a continuum between things. They facilitate the ordering of a multiple, but also allow for the ordering of individuals. These pro- cedures constitute systems of recording, classifying, and measuring-the operation of governmentality.

Several HRM techniques operate to ensure that individuals become classified and hierarchically ordered along a scale. Job classifications and job ladders are examples of ranking systems creating a hierarchy nominally based on skill, responsibility, or experience. Seniority systems and "bumping" rights are other mechanisms through which ranking is materialized. So too are salary administration and job evaluation schemes. The latter determine the relative value of a job in terms of education, skill, experience, and responsibility, and such schemes ex- plicitly include ranking systems. Nonanalytical job evaluation schemes, or the ranking method, compares whole jobs against other jobs on some assessment of value or job content, to provide a hierarchy. Analytic eval- uation schemes are used to subdivide, or further partition, jobs. In this case, jobs are ranked based on independent, measurable job compo- nents, or compensable factors, through factor comparison and the points method. In the pursuit of more objective ranking, some job evaluation schemes are based on matrix of ranking within compensable factors, for example, know-how becomes classified in terms of managerial skills, human resource skills, and practical procedures, each with further gra- dations (Hollway, 1984). The ranking of jobs essentially ensures the or- dering of a population. There is the presumption that different work roles have a certain minimum similarity that enables them to be represented on a hierarchical continuum. Through the reduction of activities to a tax- onomy of job factors, and their subsequent translation into numerical representation or mathesis, the population of jobs becomes ordered to be filled by suitable personnel.

Other familiar tools of personnel management-skills inventories, performance appraisal systems, assessments and evaluation methods, attitude measurements-are all arrangements for ranking, which facili-

3 These terms taken from The Order of Things (Foucault 1970) are approaches to knowl- edge that are identified with the classical episteme (very roughly the Enlightenment or Age of Reason), in which the foundations of science and the scientific method were laid. Essen- tially, users of the scientific method were concerned with the detailed enumeration of ob- jects under investigation, creating a science of order, and empirical forms of knowledge, which involved establishing divisions and classifications.

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 529

tate a serial ordering of individuals. For example, selection testing, as a systematic procedure for observing an individual's behavior and describ- ing it with the aid of a numerical scale of category system, provides an example of partitioning. Most testing takes place through the means of a taxinomia or grid which allows similarities and differences between peo- ple to be recognized. Such similarities and differences render individuals observable, measurable, and quantifiable. The very act of enumerating attributes in a programmatic or codified manner allows for the amount of a particular attribute or quality to be measured and thereafter compared with others. Selection tests provide "a grid of codability of personal at- tributes" (Rose, 1988: 181), and they enhance the calculability of individ- uals by placing them on a comparative scalar measure. Likert scales, for example, render attitudes inscribable and quantifiable, and in doing so enable a population and the individual's place within that population to be known.

Classification schemes are often presented as techniques to analyze labor, reflective of naturally occurring divisions or ordering of ability, skill, aptitude, and so forth. These schemes are, however, very much disciplinary techniques. They proceed by operating primarily through enhancing the "calculability" of individuals, as each classificatory or ranking system designates each individual to his or her own space, and in doing so makes it possible to establish his or her presence and ab- sence. Such classification schemes locate individuals in reference to the whole, and in doing so they operate to reduce individual singularities. Performance appraisal systems, for example, generally attempt to anchor the individual to some type of behaviorally or numerically anchored mea- suring system. Thus, for example, a performance dimension might be "leadership" ability, which might be graded on a five-point scale, rang- ing from "well above" to "well below average." Verbal anchors may at- tempt to make this more explicit; for example, "job knowledge" may range from "extremely well informed about all aspects of the job" to "misin- formed or lacks knowledge on important job dimensions." The act of scal- ing, however, produces gradations, a disciplinary mathesis that deter- mines who is to be seen in relation to whom. It facilitates the hierarchical seriation of a population because overall performance, based on generic dimensions of global rating or the culmination of various individual di- mensions, aids comparisons with others. Indeed, the distribution of the individual within the population may be explicitly achieved through ranking, if it takes the form of paired comparisons or forced distributions.

The questions raised from a Foucauldian perspective relate to how and why, and with what effects, boundaries become imposed, main- tained, and breached, not whether they are accurate or efficient, or whether they reflect reality. For example, judging individuals according to comparative, scalar models not only acts as a disciplinary process, but also as a normalizing one. Ranking, for example, organizes individuals around two poles-one negative, the other positive. "The distribution

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

530 Academy of Management Review July

according to ranks or grades has a double role; it marks the gaps, hier- archizes qualities, skills and aptitudes but it also punishes and rewards" (Foucault, 1977: 181). This can be seen, for example, in pass rates and cut-off scores. By referring individual actions to a whole in a field of comparison, ranking enables individuals to be known through being dif- ferentiated from one another. It measures and hierarchizes according to the value, the abilities, the level, and the nature of individuals. As Foucault (1977: 223) put it: "Disciplines characterize, classify, specialize; they distribute along a scale, around a norm, hierarchize individuals in relation to one another and, if necessary, disqualify and invalidate."

Articulating the Contract

The above are examples of how disciplinary practices may be ap- plied to distribute individuals in space. A continuation of these is the division and articulation of the temporal and physical dimensions of the labor process. "The labor process was articulated, on the one hand, ac- cording to its stages and elementary operations, and on the other hand, according to the individuals, the particular bodies, that carried it out" (Foucault, 1977: 145). From classification and tabulation of a population, the disciplines become more focused, acting on the body, time, and ev- eryday gestures and activities. Disciplines attempt to codify and enumer- ate as closely as possible time, space, and movement, and they attempt to articulate the analytical space of work.

The timetable provides a general framework for activity. An inven- tion of religious orders, "specialists in time" (Foucault, 1977: 151), disci- plines refine temporal regulation. With the onset of industrialization, time becomes measured and paid. It is divided into quarter hours and minutes. The working day itself is segmented and work assigned within discrete time periods -day shift, night shift, split shifts, overtime, and so on-and regulated through the time card and the time clock. There is bankable time in flexi-hour programs and billable time in the accounting and legal professions, where the hour, divided into quarter hours, or tenths, becomes a mechanism of control.

The actions of the body also must be adjusted to these temporal im- peratives, in order to produce an efficiency of movement. This is the temporal elaboration of the act: "The act is broken down into its elements; the position of the body, limbs, articulations is defined; to each movement are assigned a direction, an aptitude, a duration; their succession is pre- scribed" (Foucault, 1977: 152). This temporal elaboration involves the pre- scription of how bodies should act, the nature of gestures and coordina- tion that should be used, and a detailed specification as to how the body should engage with physical objects. Its effects can be seen in the fol- lowing manual skills analysis on fish filleting, which details some of the actions involved: "Hold knife against first and third joints of the fingers. Place upper part of thumb, first joint, against lower blunt edge of knife and the lower part of thumb against upper edge of handle. Do not grasp

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 531

knife tightly. Do not curl tip of figures into palm of hand" (Kenney, Don- nelly, & Reid, 1981: 80). Through minute and detailed regulation, disci- plines make possible the meticulous control of the body. The individual becomes subject to habits, rules, and orders; he or she operates as "one wishes, and with the techniques, speed and efficiency one determines" (Foucault, 1977: 138).

In this respect, the disciplines function in a manner prescribed by Taylor's scientific management, and, indeed, a lot of scientific manage- ment's recommendations represent the extension to production of disci- plines already in operation in other organizations, for example, in schools, in hospitals, and in prisons (Clegg, 1989). To equate disciplines solely with scientific management, however, would be to mistake the ubiquity of their operation and to ignore their continuation in an era in which, we are told, the principles of scientific management no longer apply. They may, for example, be identified in the routine activities of the personnel specialist.

The articulation of activity can be seen, for example, in job analyses and job descriptions, task and skill specifications, appraisal systems us- ing behavioral observation scales, and some training specifications. In all, the nature of performance is detailed with greater precision. For ex- ample, job analysis, the nonindividualized impersonal definition of the nature of work, articulates activity. It begins through the creation of a taxonomy. According to functional job analysis, for example, all jobs involve a relationship with data, people, and things, and these relation- ships can be arranged according to degrees of complexity. Thus, activi- ties become defined and hierarchically arranged on a scale. For example, activities involving relations with people include (from higher to lower skills) mentoring; negotiating; supervising (consulting, instructing, and treating); coaching and persuading; exchanging information; and taking instructions (helping and serving). The relationship with data is codified and ranked in terms of synthesizing, coordinating, innovating, analyzing, computing, compiling, copying, and comparing. These techniques elab- orate the construct "job" through the concept of job skills. Activities are created and serially ordered. Other systems, for example, the Position Analysis Questionnaire, are more elaborate and include measures of time spent on each activity, relative importance of each duty, and learn- ing time in its divisions and hierarchies. The concern with managerial competencies provides the extension of job analysis to managerial jobs- the attempt to articulate and make known managerial activity under the guise of introducing more scientific methods to its study. There is, in all this, the view that detailing managerial competencies is an act of reveal- ing the natural, rather than the product of decisions of an "already en- coded eye" (Foucault, 1970: xx).

The codification and enumeration of activity and movement may be identified in training and performance appraisal systems. For example, interpersonal skills training offers recommendations as to how the indi-

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

532 Academy of Management Review July

vidual should engage with the human "object." The following, for exam- ple, is taken from a skills training document in handling customer com- plaints. The trainee is recommended to "smile to make initial welcome contact . . . smile warmly with direct eye contact . . . watch customer reactions ... note customer's embarrassment by lack of eye contact . . . keep eye contact with customer, nod when customer makes valid point . . . give customer full attention ... smile . . [and after the complaint has been attended to] avoid too much eye contact to allow customer to relax" (Kenney et al., 1981: 82). This articulation of activity may be further reinforced in performance appraisal systems that include behavioral ob- servation scales or behaviorally anchored appraisal scales. Performance assessment with its construction of a system of job factors, evaluation standards, and performance descriptors most obviously involves the pro- duction of knowledge. Aspects of activity become more specifically artic- ulated. For example, behavior dimensions of waiter or waitress "service" may be expanded to include whether the individual "knows the menu and can inform customers about each item, how it is served, and how it is prepared" or "asks customers how they would like their meat and eggs done and brings rolls to the table promptly."

To ensure that both time and activity are rendered productive there is the organization of activity into a series or in a temporal sequence, with each stage successfully graded from the other and leading to a seemingly logical progression. Foucault referred to this as the capitalization of time. An example of this may be seen in management by objectives (MBO) systems. The emphasis of the MBO system is that the action plan speci- fying what is to be achieved should include time limits for particular activities. A person's success with these particular activities is evaluated at the end of each period. In some cases, a person may be reviewed in the middle of the period to see what progress he or she is making. There is a correlation of activity with time, especially if the system is supplemented by regular progress reports. Modifications of the MBO system include a productivity rating index, which combines a time-based index based on the achievement of time-related goals and a quality index that is used to assess how well goals are achieved (Bordman & Melnick, 1990). Again, the MBO system is a process that operates to inscribe activities of work- ers, thereby creating a visibility, which ultimately becomes the basis for constructing norms and trends. This capitalization of time perhaps reached its zenith in human asset accounting, which as Flamholtz (1985: 244) noted, "represents a type of balance sheet of the potential services that can be rendered by people at a specified time." Thus, measuring positional replacement costs, for example, involves not only costs relat- ing to expenses, but also detailed analyses of the time taken and the salaries of those involved in all stages of the activity: requesting a posi- tion, choosing a selection method, setting up a place to interview, review- ing applicants, deciding to offer a position (or not offer a position) to an applicant, and so on. Therefore, capitalization of time is the detail of

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 533

activity through time, related to cost-the partitioning of time, space, and movement allied to a financial equivalent.

Creating the Industrial Subject

Individuals at work must also be rendered visible. In work organiza- tions there must be systems to inspect workers, to observe workers' pres- ence and application, to inspect the quality of the work, to compare work- ers to one another, and to classify workers according to skill and speed. As Foucault (1980: 125) recognized, in order to obtain productive service from individuals "power had to be able to gain access to the bodies of individuals, to their acts, attitudes and modes of every day behavior."

First it must be recognized that making aspects of the labor process more visible has a direct implication for constituting the individual. The precisely adopted codification of action or activities engenders almost automatically the codification of the individual -the one functions to pro- duce knowledge of the other. Undertaking a complete job analysis lays down the dimensions of the individual in terms of intelligence or apti- tudes necessary to do the job. Making activities more visible necessarily renders the individual "known" in a particular manner. As Foucault (1977: 138) noted, "Disciplines create force, they turn into an aptitude or capacity that which they seek to increase," with the result that behavior gradually replaces the "simple physics of movement." Equally, systems of classifi- cation, partitioning, and ranking contribute to the detailed enumeration of the capabilities of organizational members.

Although the parameters of individuality are constituted in this way, there is still the requirement that the individual be "known." As Garland (1987: 853) identified, "The successful control of an object . . . requires a degree of understanding of its forces, its reactions, its strengths and weaknesses. The more it is known the more controllable it becomes." The components of the individual, whatever they may be, personality, atti- tudes, skill, and so on, must be calculated, assessed, and judged. There-

fore, how is an individual rendered an object of knowledge? For Foucault this may be effected in two ways: (a) individuals may become objects of enquiry through being made the subjects of scientific study or (b) individ- uals, through practices labeled technologies of the self, may situate and define themselves by becoming tied to an identity by a conscience or self-knowledge both coming to see themselves and being seen in a par- ticular way (Foucault, 1983). He distinguished two principal practices or technologies that provide knowledge of the individual: the examination, which constitutes the individual as an object of knowledge, and the con- fession, which ties the individual to self-knowledge and establishes con- cepts of subjectivity. Through both activities the individual is rendered more amenable to intervention or management.

The examination, a familiar practice in schools and hospitals, is es- sentially a method of observing. It is a disciplinary process in which several distinct operations are put into play: It is used to measure in

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

534 Academy of Management Review July

quantitative terms, and to hierarchize in terms of value the abilities, the level, and the nature of individuals. It is a technique that makes individ- ual differences and capacities visible, and it allows for these differences to be inscribed or notated. "It [the examination] establishes over individ- uals a visibility through which one differentiates them and judges them" (Foucault, 1977: 184). Through such processes the individual is rendered more easily calculable and manageable. It is the process whereby indi- viduals become compartmentalized, measured, and reported, for the pur- pose of administrative decision making. As such, the examination oper- ates as a technology of governance acting simultaneously to individual- ize and standardize. In doing so, its use has become so widespread that it may be found in "psychiatry, pedagogy, for the diagnosis of diseases to the hiring of labor" (Foucault, 1977: 185).

Where the work force was largely undifferentiated in early factory organization, there was little knowledge of, or interest in, the individual. As classification systems were put into place, greater differentiation arose generally based on observable factors such as skill, age, perfor- mance, behavior, and so forth. Later, the mind, or psyche, became iden- tified as the key to gaining knowledge of performance. "Individuals were moved by attitudes or sentiments . . . internal states that shaped the ways in which an individual apprehended and evaluated events" (Rose, 1990: 27). New dimensions of subjectivity were introduced that helped bridge the external world of conduct and the internal world of the indi- vidual (Hollway, 1991; Rose, 1990). These new dimensions of subjectivity became incorporated into selection procedures in which aspects of the individual are identified, measured, and acted upon. Personality testing (e.g., Cattell's 16PF text, the Minnesota Multiphasic Personality Inven- tory, and the Eysenck Personality Inventory) introduced a standardization to personality, and the techniques of scaling further introduced the sub- jective dimensions of individuals to the sphere of knowledge and regula- tion. This effect is extended in trait and function psychometric testing (Rust & Golombok, 1989). As Rose (1990: 85) noted, "The internal world of the factory was becoming mapped in psychological terms, and the inner feelings of workers were being transmitted into measurements about which calculations could be made." The self directions of individuals the need to achieve, the need for power, the need to self actualize-were incorporated into the production process. Thus, the individual was elab- orated on in further classificatory systems and more compartmentaliza- tion. Human nature is increasingly objectified, and action is reified. Hu- man types are created. Actions are performed because an actor is an X type person, or X type persons perform such actions. Certain work ar- rangements suit individuals with a high "need to achieve" and "type A" individuals. Personnel discourse, allied with specializations and subdis- ciplines of behavioral science, occupational psychology, and industrial psychology, further introduces the individual to more progressive objec- tification and ever more subtle partitioning of human behavior, even to

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 535

the extent of individuals being located on scales of "morningness" (Smith, Reilly, & Midkiff, 1989). These labels may become incorporated into an individual's self-assessment, the means through which individuals iden- tify their feelings and behavior to themselves and others. It is perhaps in this way that Foucault's (1977: 170) statement that "disciplines 'make' in- dividuals" makes most sense.

The assumptions of the examination are legion: that there is an es- sential personality that neither changes according to situation nor is de- finable only in a relational context reflective of interaction; that attributes are distinguishable, isolable, and directly related to a job; that jobs exist rather than being organizational and individual constructs. Essentially, these tests function as a means of measuring and evaluating individuals, rendering them calculable and manageable through "representing in standard forms human mental capacities and behavioral characteristics which previously had to be described in complex and idiosyncratic lan- guage" (Rose, 1988: 195). As Taylor (1986: 76) commented, "To try to bring it (the individual) under the control of reason is to divide what should be a living unity."

The examination has two effects: individualization and individua- tion. The first effect denotes the process of making the individual more identifiable vis-a-vis other individuals or workers, that is, the process of identifying or differentiating individuals. The second effect refers to di- viding practices that are internal to the individual, that is, those pro- cesses that attempt to identify components of individuality. Through the examination, the individual is rendered more and more an object, which is exemplified by the employee's reduction to a final score as, for exam- ple, in graphic rating scales, or overall scores at an assessment center (Bray, Campbell, & Grant, 1974). The translation of individuals into nu- merical equivalents is perhaps the inevitable extension of commodity production. The epitome of such production was described by De Michiel (1983) in her study of occupational health and safety policies, which she believed acted to objectify and commodity the body, denying its totality and "supervising its sale." According to De Michiel, through the operation of these schemes and following a medical examination, the left little finger is the equivalent of $6,723.78, and a leg equals $22,000 (in Austra- lian currency).

As presented here, the rationale of HRM is to create the individual "as an analyzable, describable subject" (Burrell, 1988: 202), to be assessed, judged, measured, and compared with others. There is, however, a diffi- culty for management when faced with this problem, as McGregor (1972: 136) has noted:

The individual knows . .. more than anyone else about his own capabilities, needs, strengths and weaknesses, and goals.... No available methods can provide the superior with the knowledge he needs to make such decisions ... rat- ings, aptitude and personality tests, and the superior's nec-

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

536 Academy of Management Review July

essarily limited knowledge of the man's performance yield at best an imperfect picture.

The principal technology that Foucault identified for accessing the self- knowledge of the subject is the confession (Foucault, 1981, 1985, 1986, 1988a, 1988b). Although the confession is most readily associated with religious practices, Foucault recognized that the basis of its functioning- avowal, the individual's acknowledgment of his or her actions or thoughts-operates in a range of activities, in education, in medicine, also at work. It may be identified, for example, in application blanks that request individuals to acknowledge their main weaknesses or strengths, their ways of coping with success or disappointment, and their pleasures or regrets over past decisions. Other selection devices, for example, pre- screening inventories, ask individuals to be as open and as accurate as possible in acknowledging what words or phrases characterize them "most of the time," "quite often," or are "essentially unlike you." Individ- uals are asked to admit their responses to, for example, being proud, self-sufficient, considerate, encouraging to others, concerned with status, honest and direct in feelings, concerned with what others think, and so on. Selection interviews also operate on the process of avowal, as for example, behavior-descriptive interview questions designed to tap dem- onstrated leadership, which ask individuals to recount times when they influenced or countered people on certain issues and how they handled these obstacles. Team players may be required to give examples of how they resolved differences with others when working on a group project. In all this, however, there is a view of the individual as harboring a secret truth; hence, the techniques recommended to interviewers to get the can- didate to talk openly.

It is, perhaps, in developmental appraisals that the confessional may be most readily identifiable; in these, individuals may be required to comment on, for example, job satisfaction. (Is this area of your work sat- isfying? Are your skills overstretched or underused? Have any recent changes affected your job satisfaction?) Part of the value of the confession is that it produces information that becomes part of the individual's self- understanding. It is also important to notice that these practices shade into other practices based not merely on accessing individuals, but al- lowing or training individuals to access themselves. Training enables individuals to identify what is happening within themselves in order to become more effective. As Rose (1990: 240) noted, "In compelling, persuad- ing and inciting subjects to disclose themselves, finer and more intimate regions of personal and interpersonal life come under surveillance and are opened up for expert judgement, normative evaluation, classification and correction."

Confessional procedures operate in two ways. Not only do they re- quire the individual to break the bounds of discretion or forgetfulness, but they also act as processes that confirm identity. These procedures are ex-

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 537

amples of how "knowing" individuals operate to tie these people to self- knowledge, constituting them as individuals. Other HRM practices func- tion to constitute the individual in a particular manner through tying him or her to a changed sense of self or identity. This may be seen, for exam- ple, in mentoring, as the following recommendation illustrates: "manag- ers can use mentoring to help cultivate desired norms and values in their organization ... through serving as mentors, senior managers guarantee that role models embody core values that best promote desired organiza- tional culture" (White, 1990: 46, emphasis added). The process of consti- tuting the subject may be identified in orientation, socialization, and in- duction programs. Training and development in interpersonal and social skills, communication, and listening skills also constitute individuals in particular ways, tying them to "appropriate" identities. Appraisal sys- tems may operate so as to inculcate the correct behavioral norms (Town- ley, 1989). Health and safety and employee-assistance programs are other means whereby individuals may be encouraged to adopt certain types of behavior or postures that accommodate them to the work environment (De Michiel, 1983). Certain equal opportunity programs in which women are treated "as if they were men" constitute individuals in the image of the dominant model of rational, European/North American male, homo ratio- nalis and homo economicus (Henriques et al., 1984: 130).

Equally, the individual may be constituted through proscribing cer- tain aspects of identity. Disciplinary procedures most obviously define the parameters of acceptable and nonacceptable behavior, and as such, they contribute to the process of individuals' being able to identify the valid from the invalid. Other mechanisms are more subtle. Hochschild (1983), for example, in her analysis of training programs for flight atten- dants illustrated the emphasis on managing negative emotions and sum- moning the ubiquitous smile. The role of these individuals is to become managers of emotion. They, themselves, must suppress anger, any sense of effrontery, that is, their own sense of self, no matter how justified.

All of the above are examples of procedures that constitute the sub- ject with varying degrees of individual engagement and participation. There is the inculcation of required habits, rules, and behavior and so- cially constructed definitions of the norm. However, the status of the in- dividual, that is, the individual's right to be different and everything that makes the individual truly individual tends to get lost in these processes.

IMPLICATIONS FOR RESEARCH

In its rereading of HRM, a Foucauldian perspective reorients research in the human resources area and has implications both for contemporary studies and historical and comparative analyses. By presenting HRM as a process of power-knowledge, I am redirecting interest from the truth or falsity of discourse toward its functioning. The type of questions that are

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

538 Academy of Management Review July

prompted in research relate to the production of knowledge and its effects. In particular, research moves away from the notion of practices as a technicist construct.

A Foucauldian perspective finds some support in research. For ex- ample, in their examination of labor market arrangements Baron, Davis- Blake, and Beilby (1986) found only weak support for hypotheses linking internal labor markets (ILMs) to organizational or sectoral imperatives. They concluded that ILMs are heavily influenced by political and institu- tional forces inside and outside organizations and suggested that re- search should be directed toward how ILMs become institutionalized, and if so, how they become modified. Again Baron and colleagues' (1986) anal- ysis of ranking through job ladders indicates that these rankings could not be easily ascribed to the imperatives of firm-specific skills. Neither could Baron and Beilby (1986) identify efficiency factors as the cause of one organization's structuring a given set of tasks differently from another organization. They suggested that future research should examine the politics of how jobs come to be defined and how those definitions become institutionalized over time. This suggestion is not so far removed from a Foucauldian project.

There have been a few explicit attempts to apply a Foucauldian ap- proach to specific areas traditionally associated with HRM: management education and development (Fox, 1989), performance appraisal (Townley, 1993), health and safety (De Michiel, 1983), and management competen- cies (Pye, 1988). Fox (1989) portrayed how management education and development programs operate to both constitute managers in a specific way and inculcate aspiring managers into this way of seeing themselves. Townley (1993) showed how the act of inscribing the individual through performance appraisal schemes serves to articulate and develop a man- agerial role that becomes institutionalized in structure. De Michiel (1983) illustrated how the body has been fragmented, objectified, and commod- ified through the power-knowledge effects of occupational health and safety programs. Pye's (1988) work on management competencies con- trasts traditional approaches that attempt to isolate, enumerate, and measure these dimensions, with a constitutive and relational concept of competence, in which the latter is constructed in social contexts and de- veloped through social relationships. In doing this, she rejected tradi- tional interpretations that have considered the individual as the self- evident unit with which to begin analysis and the use of scientific enquiry to objectify the individual.

A Foucauldian approach also has implications for conventional his- torical interpretations of HRM and personnel. Traditionally, the concept of personnel has been viewed as stressing the rights of labor and the im- portance of the human side of the organization. But the discourse of wel- fare and the human relations school clouds HRM's role in providing a nexus of disciplinary practices aimed at making employees' behavior and performance predictable and calculable-in a word, manageable. In

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 539

some respects personnel may be a "great web of bureaucracy" that evolves "endless ways to count and classify people" (Hacking, 1986: 34). Its history reflects the gradual accretion of areas of knowledge (Jacoby, 1985; Niven, 1967; Ozanne, 1967), for example, its conjunction with the behavioral sciences in its attempts to gain legitimacy as organizational diagnostician (Lupton, 1971, 1974; Rose, 1975). Military support for behav- ioral research has been particularly influential in extending areas of power-knowledge as, for example, in job analysis, selection, and person- nel classification (Gael, 1988). Supported by government involvement, particularly during the two world wars, there has been an increasing attempt to articulate the dimensions of labor and the labor process and an increased bureaucratization of personnel (Baron, Dobbin, & Devereaux Jennings, 1986; Jacoby, 1985; Niven, 1967). Nor should the role of organized labor in extending practices of power-knowledge be overlooked. Equally, the discourses of professional expertise also have played a role as sys- tems of measurement, classification, recording, and calculation become defined as legitimate knowledge and are disseminated and reinforced through publications, professional journals, meetings, and the operation of professional associations.

Inevitably, however, meanings and discursive practices are constant sites of struggle. Given the asymptotic nature of trying to make the in- tangible tangible, the seemingly insatiable drive toward greater clarifi- cation, through definitions, clauses, codes, and so on, rather than replac- ing texts, adds to them. As Clegg (1989: 193) noted, "Relations of meaning are as resistant to total control as are relations of production." Ultimately, the attempt to articulate the analytical space in the employment relation- ship by trying to impose definitive and unambiguous meaning may result in chaos, for example, the fragmentation of bargaining units (Brown, 1973), backlogs in grievance systems (Hyman, 1972), the proliferation of job classifications with demands for job regrading, the bureaucratization of payment systems through appeals on merit payments, and so forth, that is, the general rules of managing (Kochan, Katz, & McKersie, 1986).

The dysfunction in the degree of rationalization that the disciplines introduce, in particular, in North American HRM and industrial relations (IR), has stimulated the search for alternative systems abroad, particu- larly in Japan. Comparisons, however, tend to come heavily laden with reference to rather nebulous terms such as culture. A detailed, compar- ative analysis, based on Foucault's work, would trace systems of pattern- ing within organizations, or dominant patterns in different countries, for example, how populations are partitioned, how strategies for enclosure are maintained, how activity and time are articulated, and how the indi- vidual becomes known. For example, the Japanese work organization and individual work roles lack the specificity and definition typical of Western organizations. Work by Lincoln, Hanada, and McBride (1986) on a survey of American and Japanese manufacturing plants found significant differ- ences in patterns of partitioning. Although there is a relative lack of

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

540 Academy of Management Review July

occupational or functional differentiation in Japanese organizations, there is a high degree of vertical status differentiation. Groups (i.e., net- works of collective units) are the elementary functional units in Japan. Individual employee positions and roles are rarely specified because em- ployees are expected to perform a range of job functions. The relative lack of visibility of the individual is reflected in selection procedures in Japan, which are more heavily dependent on school grades and the honors sys- tem than on detailed selection testing.

The value of a Foucauldian analysis as a heuristic device is por- trayed here. This analysis may be beneficial in introducing a different way of seeing, one that could provide a framework to reorient research questions; this analysis also may provide the basis for a more systematic comparative and historical analysis. Within HRM generally, and the ba- sis for a long-term research strategy, is a need to conduct a genealogy (Knights, 1992), or a "form of history which can account for the constitution of knowledges, discourses and domains of objects" (Foucault, 1980: 117). Historically what is required is an analysis of the rules of formation of an HRM discourse, where enquiries into the past make intelligible the "nat- ural" or "objective" nature of the present, including its unquestioned ra- tionales (see, e.g., Hollway, 1984, 1991). A discursive analysis would in- volve an analysis of the situations that provoked the discourse of HRM, the consequences to which it gives rise, the practical field in which it is deployed, who is accorded the right to speak, the institutional sites from which discourse derives its legitimation, the position in which it places its subjects, what is recognized as valid, and who has access to the dis- course (Foucault, 1972, 1991a, 1991b). There is no assumption, however, that discourse is an expression of an economic situation, that is, that it may be "read off" from economic criteria or institutional developments. The latter have implications for the conditions of emergence of a dis- course and its functioning but not the discourse per se. Other factors interact to affect the discourse of HRM, for example, educational and scientific discourses as well as accounting and legal discourses.

This latter aspect also introduces the questions of HRM's relationship with other disciplines, an area of relative neglect. For example, judicial involvement through legal cases has implications for the way in which HRM discourse has developed. In North America, for example, this in- volvement has led to the juridification of the employment relationship (Simitis, 1986). Legal decisions have influenced the type of selection tests and appraisal systems used (Rust & Golombok, 1989). In IR, union certi- fication procedures exclude managerial employees, thus requiring judi- cial clarification of the definition of management (an articulation of its activity); the right of a union to exclusive representation of an appropriate bargaining unit reinforces vertical divisions within the work force. Ac- counting practices also may be interpreted as disciplines articulating the analytical space of the organization (Roberts & Scapens, 1985). Using Miller and O'Leary's (1987: 239) broad definition of accounting as "the

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 541

development of a range of calculative programmes and techniques which come to regulate the lives of individuals at work," and understanding HRM as providing the basis of matheses illustrates possible interconnec- tions between accounting and HRM (Townley, In press).

CONCLUSION

To recap the argument, the basic unit of analysis in understanding HRM was identified as the nature of exchange embodied in the employ- ment relationship. Given the essentially indeterminate nature of this re- lationship, the problematic then becomes how this relational, exchange activity is organized. Foucault's concept of power-knowledge was intro- duced to illustrate how HRM acts to impose order on the inherently un- decidable. Attempting to clarify the indeterminacy of contract requires "effective instruments for the formation and accumulation of knowl-

edge-methods of observation, techniques of registration, procedures for investigation and research, apparatuses of control" (Foucault, 1980: 102). The construction of knowledge in HRM operates through rules of classi- fication, ordering, and distribution; definitions of activities; fixing of scales; and rules of procedure, which lead to the gradual emergence of a distinct HRM discourse. Associated with these practices are concepts of rationality, scientificity, measurement, grading-the language and the knowledge of the HRM specialist. Through mechanisms of registration, assessment, and classification-areas of study often neglected or dis- missed as technical or administrative procedures-it becomes possible to illustrate how a body of knowledge operates to objectify those on whomr it is applied. It is also in this discourse that the individual becomes lo- cated as an object of knowledge. Classification schemes, offered as tech- niques of simplification and clarification for the analysis of labor, both as effort and object, become inextricably tied to its disciplinary operation.

A Foucauldian perspective presents an alternative way for perceiv- ing and ordering material. Rather than thinking in functional terms of recruitment, appraisal, remuneration, and so on, in this perspective an emphasis is placed on how HRM employs disciplinary practices to create knowledge and power. These practices fix individuals in conceptual and geographical space, and they order or articulate the labor process. Pro- cesses of individualization and individuation create an industrial subject who is analyzable and describable. As an approach, it allows HRM to be analyzed as the "will to knowledge," that is, as a system of knowledge and modality of power. It is sufficiently detailed an approach to allow for the "micropolitics" of power to be addressed and, by providing examples of how the concern with "knowing" labor as a "population" can percolate down to its effects on the individual, allows for highly individualized practices to be related to an intelligible whole. In doing so, it also pro- vides the basis for reorienting contemporary, historical, and comparative analyses of HRM.

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

542 Academy of Management Review July

REFERENCES

Alvesson, M., & Willmott, H. 1992. On the idea of emancipation in management and orga- nization studies. Academy of Management Review, 17: 432-465.

Baldamus, W. 1961. Efficiency and effort. London: Tavistock.

Baron, J., & Beilby, W. 1986. The proliferation of job titles in organizations. Administrative Science Quarterly, 31: 561-586.

Baron, J., Davis-Blake, A., & Beilby, W. 1986. The structure of opportunity: How promotion ladders vary within and among organizations. Administrative Science Quarterly, 31: 248-273.

Baron, J., Dobbin, F., & Devereaux Jennings, P. 1986. War and peace: The evolution of modern personnel administration in U.S. industry. American Journal of Sociology, 92: 350-383.

Beer, M., Spector, B., Lawrence, P. R., Mills, D. Q., & Walton, R. E. 1984. Managing human assets. New York: Free Press.

Bordman, S., & Melnick, G. 1990. Keep productivity ratings timely. Personnel Journal, 69: 50-51.

Bray, D. W., Campbell, R. J., & Grant, D. L. 1974. Formative years in business: A long-term AT&T study of managerial lives. New York: Wiley.

Brown, W. 1973. Piecework bargaining. London: Heinemann.

Burawoy, M. 1979. Manufacturing consent. Chicago: University of Chicago Press.

Burchell, G., Gordon, C., & Miller, P. 1991. The Foucault effect: Studies in governmentality. London: Harvester Wheatsheaf.

Burrell, G. 1988. Modernism, Post-modernism and organizational analysis 2: The contribu- tion of Michel Foucault. Organization Studies, 9: 221-235.

Calds, M., & Smircich, L. 1990. Thrusting toward more of the same with the Porter-McKibben report. Academy of Management Review, 15: 698-705.

Clegg, S. 1989. Frameworks of power. London: Sage.

Commons, J. R. 1934. Institutional economics: Its place in political economy. New York: Macmillan.

De Michiel, Y. 1983. The subjected body. Australian Journal of Law and Society, 1: 5-44.

Dreyfus, H., & Rabinow, P. 1983. Michel Foucault, beyond structuralism and hermeneutics. Brighton, England: Harvester Press.

Edwards, R. 1979. Contested terrain: The transformation of the workplace in the twentieth century. New York: Basic Books.

Eribon, D. 1991. Michel Foucault. (B. Wing, Trans.). Cambridge, MA: Harvard University Press.

Finch, J. 1983. Married to the job: Wives incorporation in men's work. London: Allen & Unwin.

Flamholtz, E. 1985. Human resource accounting (2nd ed.). San Francisco: Jossey Bass.

Foucault, M. 1967. Madness and civilization: A history of insanity in the age of reason. London: Tavistock.

Foucault, M. 1970. The order of things: An archeology of the human sciences. London: Tavis- tock.

Foucault, M. 1972. The archaeology of knowledge. London: Routledge.

Foucault, M. 1973. The birth of the clinic. London: Tavistock.

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 543

Foucault, M. 1977. Discipline and punish: The birth of the prison. London: Penguin.

Foucault, M. 1980. Power/knowledge: Selected interviews and other writings by Michel Foucault, 1972-77. (C. Gordon, Ed.). Brighton, England: Harvester.

Foucault, M. 1981. The history of sexuality: Vol. 1. The will to knowledge. London: Penguin.

Foucault, M. 1983. The subject and power. In H. L. Dreyfus & P. Rabinow (Eds.), Michel Foucault: Beyond structuralism and hermeneutics: 208-226. Chicago: University of Chi- cago Press.

Foucault, M. 1985. The history of sexuality: Vol. 2. The use of pleasure. New York: Pantheon.

Foucault, M. 1986. The history of sexuality: Vol. 3. The care of the self. New York: Pantheon.

Foucault, M. 1988a. Technologies of the self. In L. Martin, H. Gutman, & P. H. Hutton (Eds.), Technologies of the self: 16-49. London: Tavistock.

Foucault, M. 1988b. The political technology of individuals. In L. Martin, H. Gutman, & P. H. Hutton (Eds.), Technologies of the self: 145-162. London: Tavistock.

Foucault, M. 1991a. Politics and the study of discourse. In G. Burchell, C. Gordon, & P. Miller (Eds.), The Foucault effect: Studies in governmentality: 53-72. London: Harvester Wheat- sheaf.

Foucault, M. 1991b. Questions of method. In G. Burchell, C. Gordon, & P. Miller (Eds.), The Foucault effect: Studies in governmentality: 73-86. London: Harvester Wheatsheaf.

Foucault, M. 1991c. Governmentality. In G. Burchell, C. Gordon, & P. Miller (Eds.), The Foucault effect: Studies in governmentality: 87-104. London: Harvester Wheatsheaf.

Fox, S. 1989. The panopticon: From Bentham's Obsession to the revolution in management learning. Human Relations, 42: 717-739.

Gael, S. 1988. The jobs analysis handbook for business, industry and government. New York: Wiley.

Garland, D. 1987. Foucault's discipline and punish, an exposition and critique. American Bar Foundation Research Journal, 4: 847-880.

Gilligan, C. 1982. In a different voice. Cambridge, MA: Harvard University Press.

Gordon, C. 1991. Governmental rationality: An introduction. In G. Burchell, C. Gordon, & P. Miller (Eds.), The Foucault effect: Studies in governmentality: 1-52. London: Harvester Wheatsheaf.

Hacking, I. 1986. The archaeology of Foucault. In D. C. Hoy (Ed.), Foucault: A critical reader: 27-40. Oxford: Basil Blackwell.

Henriques, J., Hollway, W., Urwin, C., Venn, C., & Walkerdine, V. (Eds.). 1984. Changing the subject. London: Methuen.

Hesterly, W. S., Liebeskind, J., & Zenger, T. R. 1990. Organizational economics: An impend- ing revolution in organization theory? Academy of Management Review, 15: 402-420.

Hochschild, A. 1983. The managed heart: Commercialization of human feeling. Berkeley: University of California Press.

Hollway W. 1984. Fitting work: Psychological assessment in organizations. In J. Henriques, W. Hollway, C. Urwin, C. Venn, & V. Walkerdine (Eds), Changing the subject: 26-59. London: Methuen.

Hollway, W. 1991. Work psychology and organizational behaviour. London: Sage.

Hyman, R. 1972. Disputes procedure in action. London: Heinemann.

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

544 Academy of Management Review July

Jacoby, S. 1985. Employing bureaucracy. New York: Columbia University Press.

Kanter, R. 1977. Men and women of the corporation. New York: Basic Books.

Keller, C. 1986. From a broken web: Separation, sexism, and self. Boston: Beacon Press.

Kenney, J., Donnelly, E., & Reid, M. 1981. Manpower training and development (2nd ed.). London: Institute of Personnel Management.

Knights, D. 1992. Changing spaces: The disruptive impact of a new epistemological location for the study of management. Academy of Management Review, 17: 514-536.

Knights, D., & Willmott, H. 1985. Power and identity in theory and practice. Sociological Review, 33: 22-46.

Kochan, T., Katz, H. C., & McKersie, R. B. 1986. The transformation of American industrial relations. New York: Basic Books.

Lincoln, J., Hanada, M., & McBride, K. 1986. Organizational studies in Japanese and U.S. manufacturing. Administrative Science Quarterly, 31: 338-364.

Lupton, T. 1971. Management and the social sciences. London: Penguin.

Lupton, T. 1974. Industrial behaviour and personnel management. London: IPM.

Marsden, R. 1993. The politics of organizational analysis. Organization Studies, 14: 93-124.

McGregor, D. 1972. An uneasy look at performance appraisal. Harvard Business Review, 50 (5): 133-138.

Miller, P., & O'Leary, T. 1987. Accounting and the construction of the governable person. Accounting, Organisations and Society: 235-265.

Morgan, G. 1980. Paradigms, metaphors and puzzle solving in organization theory. Admin- istrative Science Quarterly: 605-622.

Niven, M. 1967. Personnel management, 1913-1963. London: IPM.

Ouchi, W. 1980. Markets, bureaucracies and clans. Administrative Science Quarterly, 25: 129- 141.

Ozanne, R. 1967. A century of labor-management relations at McCormick and International Harvester. Madison: University of Wisconsin Press.

Pahl, R. E. 1988. Editor's introduction: Historical aspects of work, employment, unemploy- ment and the sexual division of labour. In R. E. Pahl (Ed.), On work: Historical, com- parative and theoretical approaches: 11-15. Oxford: Basil Blackwell.

Philp, M. 1985. Michel Foucault. In Q. Skinner (Ed.), The return of grand theory in the human sciences: 65-82. Cambridge, England: Cambridge University Press.

Pringle, R. 1988. Secretaries' talk: Sexuality, power and work. London: Verso.

Pye, A. 1988. Management competence in the public sector. Public Money & Management, 8 (4): 62-67.

Roberts, J. 1984. The moral character of management practice. Journal of Management Stud- ies, 21: 287-302.

Roberts, J. 1991. The possibilities of accountability. Accounting, Organisations and Society: 355-368.

Roberts, J., & Scapens, R. W. 1985. Accounting systems and systems of accountability- Understanding accounting practices in their organizational contexts. Accounting, Or- ganisations and Society: 443-456.

Rose, M. 1975. Industrial behaviour. London: Penguin.

Rose, N. 1988. Calculable minds and manageable individuals. History of the Human Sci- ences, 1(2): 179-200.

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms

1993 Townley 545

Rose, N. 1990. Governing the soul: The shaping of the private self. London: Routledge.

Rust, J., & Golombok, S. 1989. Modern psychometrics: The science of psychological assess- ment. London: Routledge.

Simitis, S. 1986. The juridification of labour relations. Comparative Labour Law, 7: 93-142.

Smith, C., Reilly, C., & Midkiff, K. 1989. Evaluation of three circadian rhythm questionnaires with suggestions for an improved measure of morningness. Journal of Applied Psychol- ogy, 74: 728-738.

Steffy, B., & Grimes, A. 1992. Personnel/organizational psychology: A critique of the disci- pline. In M. Alvesson & H. Willmott (Eds.), Critical management studies: 181-201. Lon- don: Sage.

Taylor, C. 1986. Foucault on freedom and truth. In D. C. Hoy (Ed.), Foucault: A critical reader: 69- 102. Oxford: Basil Blackwell.

Townley, B. 1989. Selection and appraisal: Reconstituting "social relations"? In J. Storey (Ed.), New perspectives on human resource management: 92-108. London: Routledge.

Townley, B. 1993. Performance appraisal and the emergence of management. Journal of Management Studies, 30(2): 27-44.

Townley, B. In press. Managing by numbers: Personnel management and the creation of a mathesis. Critical Perspectives on Accounting.

Turner, S. 1983. Studying organization through Levi Strauss's structuralism. In G. Morgan (Ed), Beyond method: 189-201. Beverly Hills, CA: Sage.

White, H. 1990. The self method of mentoring. Bureaucrat, 19(1): 45-48.

Williamson, 0. 1975. Markets and hierarchies. New York: Free Press.

Barbara Townley received her Ph.D. from the London School of Economics. She is an associate professor in the Department of Organizational Analysis at the University of Alberta, Edmonton, Canada. Before coming to Canada she taught in the United Kingdom in the Department of Behaviour in Organisations, Lancaster University, and the Department of Industrial Relations and Organisational Behaviour, Warwick University. Her current research interests include a Foucauldian critique of person- nel practices and the relationship between personnel and accounting.

This content downloaded from 129.219.247.33 on Wed, 25 Mar 2020 03:30:16 UTC All use subject to https://about.jstor.org/terms