In this chapter I review an approach to studying work that I call “work as practice.” (I borrow the term from Orr 1996, the last of the authors I review in this chapter.) In doing so, I aim to construct a foundation for a later discussion of how a technical practice such as software development spreads and stays synchronized across space. The literature that I bring together under the heading of “work as practice” shares important insights, as it looks at work as intertwined with individual identities and membership in social groups. At the same time, the literature brings together several theoretical traditions, with two being of particular importance: the Chicago School of Sociology (exemplified here most clearly by Hughes and Becker) and the Marxist theories of practice (exemplified by Lave & Wenger, 1991). My synthesis relies on the Chicago School tradition as the basis, bringing in elements of the theory of practice to the extent that they become necessary. (One of those elements is the concept of “practice” itself.) I do so in part because I find that the down-to-earth approach that characterizes the Chicago School fits more naturally with an ethnographic project than does the often highly abstract discussion of practice theory. This approach also helps me link my project with a longer tradition of studying modern work ethnographically.

      I present in this chapter a conservative synthesis of the existing literature, aiming to organize it, to discuss some its problems, and look at some of the solutions to those problems, arriving at an approach that I believe is defensible against the criticism to which some of the work in this tradition has been subjected. I do not aim to extend the theory of “work as practice” in this chapter. In the two chapters that follow, I subject this synthesis to my own criticism — for failing to provide a theory of practice in space that would work at a world scale. I do so first by presenting ethnographic observations that challenge the initial theory (chapter 1.2), and then by providing a theoretical critique and proposing a resolution (chapter 1.3), drawing on additional ideas from practice theory (Giddens in particular). Parts 2 and 3 of the dissertation then illustrate and expand the theory presented in chapter 1.3.

      I present my review chronologically, starting with 1950s. The reader may wonder if books written half a century ago can help us understand such modern occupations as software development. I intend to show that they do, in three ways: by reminding us of certain ideas that are too often forgotten, by putting more recent literature in perspective, and by showing the ways in which peculiarities of software development represent not a break from the past, but rather a radicalization of phenomena observed for many decades.33 I also believe that understanding the ideas expressed by the earlier authors, allows us to focus on the specific contributions of their intellectual heirs. Putting each author in their historical context also makes it easier to see how their writings respond to the challenges leveled at their predecessors.

      An important problem that has characterized some of the literature in the “work as practice” tradition is the insufficient attention to the economic side of work. Many kinds of modern work cannot be understood without considering them as a form of cultural practice, that is a system of activities that is linked with the collective of people who engage in such activities and which cannot be understood without considering how individual practitioners become members of this collective, how they establish their claims to membership, and how their membership interacts with their identity. As I argue in chapter 2.1, for example, the work of software developers cannot be understood without considering the developers’ identity as “nerds” and the process through which they become “nerds.” At the same time, however, it is important to avoid looking at work as just a matter of culture. Work is also a form of participation in a system of production and is entangled with economic relationships. Such relationships are nearly always invested with power and the goals of the workers may not always align with those of the organizations that employ them. I refer to this aspect of work as “work as labor.” Authors writing in the “work as practice” tradition have been often accused of politico-economic naiveté by those of the “work as labor” conviction (e.g., Braverman 1974). I take this charge seriously and attempt in my review to show where the literature on practice may in fact be guilty of ignoring the “labor” side of work, point out the places where this side is acknowledged and incorporated, then bring out and expand those elements. As I attempt to show in the later chapters, using software development as my case, work must be understood simultaneously as a matter of both practice and labor. Such simultaneous analysis is especially important for understanding how practice is reproduced in space, since the reproduction of labor relationships may either run ahead or fall behind the reproduction of the cultural practice.

From Knowledge to Practice

      Software development is often described as “knowledge work,” and sometimes presented as the quintessential example of it — something that can help us understand other types of modern work. The immaterial, informational elements of software development are easily apparent and are quite clearly important for understanding globalization of software work. In this chapter and this dissertation, however, I take a different approach to software development, looking at it as practice. Such approach does not reject the importance of knowledge, but avoids reducing software work to manipulation of “knowledge.” I do so out of a belief that such reduction inevitably proceeds to further reduction of “knowledge” to “information” and then to a purely informational understanding of work. The concept of “practice” helps us buttress against such a reduction. The rest of this section briefly reviews some conceptualizations of knowledge and presents an argument for looking at “knowledge work” as practice.

      Modern understanding of knowledge is strongly influenced by what has been called the “codification view” of knowledge.34 This tradition, which goes back to Diderot and other Encyclopedists of 18th century, has stressed the understanding of technical knowledge as a collection of reproducible facts and techniques. Under this “codification view,” knowledge possessed by skilled artisans can — and should — be converted into textual form to then be made available to all who need it. New information technologies developed in the 19th century (such as microfilm photography, developed in 1839), inspired by the end of the century a wave of excitement about the potential for reproduction and free dissemination of knowledge, led by the Documentation movement of Paul Otlet (e.g., 1925/1990) and supported by figured such as H.G. Wells (1938). Some of those ideas were later popularized by Vannevar Bush in his articles on Memex (e.g., 1945), and went into the Rapid Selector (Shaw 1949). With the diffusion of the digital computer after World War II, the dream of freely reproducible knowledge started to take a concrete shape. Work of early information retrieval researchers such as Luhn (e.g., 1961) and early networks researchers (e.g., Davis et al, 1967) laid the foundation of what we now know as “googling.”35 Doug Engelbart’s 1962 report and Ted Nelson’s visions of Xanadu (a hyper-text system later described in Nelson 1974/1987) provided a revised version of the Memex. The spread of Internet access and the development of the World Wide Web technologies in 1990s made this view even more appealing. In my own interviews with software developers, questions about how they learn what need to know to do their work often led to the same short response: “the Internet.”

      The codification view of knowledge has been quite influential in economics and business. Since Drucker’s (1969) realization that of the effect of knowledge and innovation on productivity and long term development, there has been growing interest in understanding how knowledge could be “managed.” While business literature on knowledge management has sometimes acknowledged the fact that knowledge cannot be fully reduced to information in computer files, much of it bears the imprint of the codification view. When the work of Paul Romer (1990, 1993), established knowledge as a legitimate topic in economics, it brought with it a codification perspective on knowledge.36 Romer’s discussion focused on the sharing of “designs,” which he described as “non-rival” — that is, in principle usable by any number of people at the same time.37 Consequently, in Romer’s (1990) view, designs/knowledge “can be accumulated without bounds on a per capita basis” (p. S75) — an assumption that established a new paradigm for economic treatment of knowledge.38 Romer then argued (1993) that the misfortunes that afflict many countries would be best understood by considering the “idea gap” between them and the more successful countries like the United States. Such “idea gaps,” Romer argued, are “relatively easy to solve,” and consequently “idea gap explanation for persistent poverty offers a more optimistic picture of the potential for rapid development” (p. 546).39

      The increasing popularity of the codification vision led to a critique of this view. The question of whether all knowledge can actually be codified in principle (apart from the technical issues of storage, retrieval and transmission explored by engineers) first attracted the attention of philosophers. Ryle (1949) proposed a distinction between “knowing that” and “knowing how,” and Polanyi (1952, 1966) introduced the concept of “tacit dimension” of knowledge, arguing that “we know more than we can tell” (1966, p. 4). The idea of tacit knowledge was then developed further by scholars in the Science and Technology Studies tradition, such as Collins (1974) who looked at the diffusion of knowledge regarding a laser design (and later, in Collins 2001, at the difficulties experienced by British physicists attempting to replicate a Soviet experiment), and MacKenzie & Spinardi (1995) who discussed “the uninvention of nuclear weapons” in the United States. Some of the discussion of ‘tacit knowledge’ focused on the simpler observation that communication always relies heavily on prior common ground and typically references the physical environment.

      The term “tacit knowledge,” however, suffers many of the shortcomings of the codification view of knowledge. The use of the word “knowledge,” a noun, positions tacit knowledge as another type of “knowledge,” somewhat of an evil twin of the “normal” (codifiable) knowledge. (See, for example, Cowan et al. 2000.) The term “tacit knowledge” invites us to continue thinking of knowledge as a “thing” and something an individual (or a group) can “have.”40 Yet when we ask where it is, tacit knowledge proves to be an illusive, mysterious concept. Additionally, while authors who argue for tacit knowledge typically introduce a distinction between “knowledge” and “information,” such distinction is often forgotten or employed inconsistently when the ideas developed by those authors are popularized. I suspect that the term “knowledge” is today so strongly connected to the idea of replicable designs and electronic transmission of files, that I cannot expect my reader to simply put those associations aside.

      Instead, I use a different term, and look for a different theoretical foundation. I find such a foundation in the work of Becker (1953, 1972), Lave & Wenger (1991) and other authors discussed below, and systematize it using Lave & Wenger’s term “practice.” I define “practice” as a system of activities, a collective way of doing certain things. Understood this way, “practice” subsumes both tacit and codified dimensions of knowledge. In this process, “tacit knowledge” disappears as a distinct category. That is, I do not find it useful to draw a distinction between “tacit knowledge” and “practice” and limit myself to saying that “tacit knowledge” is a part of practice. In contrast, knowledge codified in documents becomes highlighted as an element of practice — an important one, yet one of many. I focus on what people do and how they come to change how they do things, and then approach the specific knowledge-processing activities — such as reading, writing, and other forms of communicating — in the context of this doing, considering them side-by-side with the many aspects of practice that cannot be reduced to acquisition of information.

      Looking at practice rather than acquisition of knowledge, however, means more than just looking at what people do. To follow Becker (1953, 1972) and Lave & Wenger (1991) is to recognize that practice is a deeply collective and relational phenomenon which transforms the individual, making her or him a member in a collective of practitioners, enmeshed in a new web of relationships. Looking at engineers, for example, it is not sufficient to ask how someone acquires the knowledge of engineering or starts to engineer things. Instead, we must consider the full implications of becoming an engineer — the individual’s entry into a network of socio-technical relationships, which will transform how they see other people, objects and themselves, and how others will see and interact with them. Learning, according to Lave & Wenger, is “the historical production, transformation and change of persons” (p. 51). Such “relational view of the person and of learning” (p. 53) implies that “activities, tasks, functions, and understandings do not exist in isolation; they are part of a broader systems of relations in which they have meaning” (p. 51). To understand practice and learning, we will therefore have to look at them in the context of such broader systems, considering also how the participants themselves reflect on those broader systems. This view also implies that newcomers do not merely acquire the knowledge that was out there before they arrived, but rather become a part of a system, changing it by their very arrival.

      The argument that learning must be understood as a matter of progressive engagement in a community of practitioners was most famously presented in Lave & Wenger’s (1991) Situated Learning: Legitimate Peripheral Participation.41 This important book has been widely read in a number of different fields, and interpreted in a number of ways. I look at Lave & Wenger in the context of a particular tradition, the “Second Chicago School” of sociology,42 which I believe gives it its fullest meaning, while also helping us see the work’s most important limitations. While certain important elements of Situated Learning do not fit in this tradition, looking at Lave & Wenger in connection with the (second) Chicago school makes it possible to link their theoretical discussion to a wider body of ethnographic work, as well as to very similar theoretical accounts.

      In this chapter I review the work of Hughes, Becker, Van Maanen & Barley, Lave & Wenger, and Orr, providing my synthesis of their respective theories of work and practice. (Later, in chapter 1.3, I also incorporate some insights from the work of Anselm Strauss and other authors who contributed to the Chicago school literature on “social worlds.”) While discussing their respective takes on the relationship between learning, practice and communities, I consider the fact that they use different terms to talk about similar, but not identical, concepts. Hughes writes of “work” or “occupations” (1958), Becker of “activities” (1953) or “groups” (1963) or without using any technical terms at all (1972),43 Van Maanen & Barley (1984) talk of “occupations,” Orr (1996) or “work” and “practice,” and Lave & Wenger (1991) talk of “practice.” “Work,” “occupation” and “practice” are distinct but intertwined concepts. I look at work as a kind of practice, and one of the most important ones to study, since work is in many ways central to the functioning of society (see the discussion of Hughes below). Yet we cannot fully understand modern work without looking at how it relates to other types of practice — those that we would hesitate to call “work” — and understand what it has in common with those practices. In case of software development, we must find a way to account for the fact that some of the world’s most important infrastructure today runs on an operating system much of which was written either “just for fun” or out of moral calling.44

      The research tradition that I trace from Hughes has been subjected to an important critique by Marxist scholars of work, who have come to dominate sociology of work in 1970s (see Simpson 1989 and the discussion below). Similar critiques have been later applied to the more recent authors in this tradition (see Contu & Willmott 2003). I believe that the two traditions can be reconciled and that we must see work as being a matter of participating simultaneously in a cultural community and system of labor relations. Without attempting to work out such reconciliation fully in this chapter, I try to make explicit the disagreements between the traditions, setting up ground for a more balanced discussion in the later chapters.45

Occupations as Named Parts in a System of Division of Labor

      Everett Hughes, a transitional figure between the “first” and “second” Chicago schools of sociology, is often recognized as a founding figure in the modern sociology of work.46 Hughes offers a theory of occupations that highlights the relationship between work and identity (at least for skilled work), looking at occupations as social groups, while at the same time considering their relationship with other occupations and their position in the societal division of labor. This second aspect of Hughes writing is often downplayed in the work of his followers, who often focus more fully on occupations as cultural groups, rather than units in the larger society. Revisiting the work of Hughes will help us establish a balance between “work as practice” and “work as labor” later in the chapter.

      Hughes writes about “work” without defining this term but seemingly using it in a way that does not differ much from the vernacular usage. Unlike Marx, who sees “work” as a fundamental relationship between man and nature (see below), Hughes looks at work mostly in the context of contemporary industrial society. He associates “work” with participation in the “labor force” and “formally organized work activities… named and categorized in payrolls, organization charts and union-management contracts, and in income-tax, licensing, and social-security legislation” (1959/1994, p. 21-22).47 He also contrasts “work” with “leisure” (p. 22-23). Hughes sees “work” as often defining the course of an individual’s life: “A man’s work is as good a clue as any to the course of his life, and to his social being and identity” (1958, p. 7). The type of work the individual does is the most important determinant of social status in the industrial society: “Race, sex, marital status, and other characteristics formerly determined civil estate quite directly; now it is work that counts… and the other characteristics take their importance by virtue of their influence on one’s place in the labor force” (1959/1994, p. 22).

      Hughes focuses on skilled work, which he sees as divided into occupations (or more specifically named occupations). An occupation is defined simultaneously by a particular type of work activities, a group of people who carry them out, and their role in the division of labor. Hughes’s stress on the division of labor between occupations (drawing heavily on Durkheim’s notion of “organic solidarity”) is important to note here, since it is largely left out by much of the later “Chicago School” literature on work and practice. Specifically, Hughes writes that “an occupation, in essence, is not some particular set of activities; it is the part of an individual in any ongoing system of activities” (1959/1994, p. 24). He then re-defines an occupation as “a more-or-less standardized one-man’s part in some operating system” concluding from this that an occupation “cannot be described apart from the whole” (p. 30). In introducing the notion of “a more-or-less standardized one-man’s part” (my emphasis) in the larger division of labor, Hughes draws a link between the division of labor between individuals in a specific enterprise (as in Adam Smith’s famous example of the pin factory), and the division of labor between occupations in the context of a larger industrial society. He suggests, in other ways, that the division of labor between individuals is not organized on an ad hoc basis in every enterprise. Rather, the work tends to be divided in more or the same roles, which allows us to say that people who take the same role in different enterprises comprise the same “occupation.”48

      At the same time, Hughes stresses the personal significance and the cultural aspects of occupations. He sees many occupations as characterized by shared “culture” and “technique.” Hughes ties the technique on the one hand to the culture shared by those who engage in the occupation, and to “a set of collective representations, more or less peculiar to the occupation and more or less incomprehensible to the community.” He stresses that the practitioners of occupations have a distinct way of relating to certain objects and activities: “The technique of the physician is in relation to the human body, which must be for him a different sort of object from what it is for the layman. To the layman it is a sacred thing, and an object of sentiment.” (1958, p. 35.) Members of the occupations see familiar objects in different ways. Even more importantly, they typically have a distinct way of seeing the activity itself.

      Named occupations “are a combination of price tag and calling card” writes Hughes (1951/1958, p. 42). The “calling card” metaphor links two sides of the occupation: an occupation is a crucial source of individual identity, but it is important as a source of identity largely because it identifies the individuals’ role in the industrial society. This metaphor and the term “named occupations” (my emphasis) bring together what I later (chapter 1.3) call the “external” and the “internal” view of “named” categories of practitioners.

Insiders and Outsiders

      Hughes’s notions of “culture” and “technique” are closely tied to those of membership and license. “An occupation consists, in part, of a successful claim of some people to license to carry out certain activities which others may not, and to do so in exchange for money, goods and services,” writes Hughes (1958, p. 78). The “license”49 to perform certain activities is given to a group. The individual who seeks permission to engage in those activities must gain membership in the group that has the license: “the practitioners, by virtue of gaining admission to the charmed circle of colleagues… exercise the license to do things others do not do” (1958, p. 79). This leads to a question of how membership is granted — a question that Hughes, to the best of my knowledge, does not answer. The work of Hughes’s student and co-author Howard Becker, however, helps shed light on this issue.

      Howard Becker, one of the most well-known figures of “the Second Chicago School,” started his career with a study of Chicago jazz musicians. His 1951 paper on the subject (Becker 1951) introduces the study as a case of conflicts experienced by “members of service occupations” (p. 136) and Hughes incorporates Becker’s study as an example of an occupation (1951/1958, one of Hughes’s early publications on the subject of occupations). However, Becker later reprints the study in Outsiders: Studies in the Sociology of Deviance (1963). This re-conceptualization of the practice of Chicago jazz musicians from “an occupation” to membership in “a deviant group” may seem puzzling, but in fact provides us with an opportunity to get a deeper understanding of Hughes’ discussion of occupations as groups.

      While viewing jazz musicians as first an occupation and then as a “deviant group” may seem strange, we should remember that one of the key aspects of an occupation for Hughes is that its practitioners do things that other people do not do. Becker’s notion of “deviance” does not correspond precisely to Hughes’s (and early Becker’s) “occupation,” but those two kinds of groups share enough similarities to warrant a comparison. Becker defines as “deviant” activities that are labeled as such by groups that do not engage in them.50 He explains his inclusion of Chicago jazz musicians among “deviant groups” (in a chapter following one about marijuana users) by the fact that “their culture and the way of life are sufficiently bizarre and unconventional for them to be labeled as outsiders by more conventional members of the community” (1963, p. 79). Both types of groups are thus characterized by “a culture” and engage in activities that other people do not do. The question of whether the musicians form “an occupation” or a “deviant group” would thus presumably rest on whether they have a license to engage in the activities that they engage in. There is a marked difference, however, between Hughes and Becker as to whose licenses is relevant. Hughes talks abstractly about “the society” allowing people to do certain things. Becker is careful to be more specific. The jazz artists do not generally lack society’s license to play: their activities are legal and they have audience that enjoys their music. Rather, they are labeled as outsiders by specific (if powerful) other groups.51

      Becker’s insight has important consequences for the discussion of occupation. Becker urges us to avoid looking at rule-breaking without asking who creates the rules and when they choose to enforce them. Deviance becomes a matter of group boundaries and their enforcement. Applied to studies of work this suggests that we might want to question Hughes distinction between activities for which society does or does not give license to certain groups (and more generally the idea of a monolithic “industrial society”), and instead look at people who engage in certain activities, form a group (recognized by themselves and often by outsiders) by virtue of such engagement, have a distinct “technique” for performing the activity, and a special way of relating to this activity and the relevant objects. Looking at it this way, we quickly realize important similarities between Hughes’s doctors and Becker’s marijuana users.

      Becker’s study of deviant groups (1953, 1963) develops Hughes’s notions of “culture” and “technique” and presents a model of the process by which a newcomer enters the group, presaging some of the ideas of Lave & Wenger (1991). Becker’s theory of progressive engagement in the activity (read with the hindsight assisted by Lave & Wenger) can be summarized as follows. The novice starts off with a limited interest in the activity, perhaps seeing little point in it (or worse, seeing it negatively). If they were to try to engage in the activity by themselves anyway, they would likely fail to obtain the expected results, lacking the necessary technique. (A first-time user of marijuana will likely burn her throat by inhaling incorrectly. A wannabe programmer will likely write code that will not run.) If they do obtain the results, they might fail to recognize or appreciate them. (A novice marijuana user might fail to recognize being “high” and if she does might fail to find it enjoyable. A novice programmer might not recognize the joy of getting the code to compile.) Becker writes that the user must learn to define the effect of the activity as enjoyable, acquiring what Becker calls “a conception of the meaning of the behavior, and perceptions and judgments of objects and situations that make the activity possible and desirable” (p. 235). (I will further use the terms “perceptions and judgments” or simply “perceptions” as short-hand.) Experienced practitioners become important here, as they help the novice learn the basic technique and to develop the necessary “perceptions and judgments.” This is not to suggest that they become the primary source of learning: rather they help the novice make early steps into the activity, and it is from the combination of their advice and own experience that the novice learns the perceptions and the technique.52

      The shared techniques and dispositions are for Becker (1963) the culture of the group united by those activities.53 While the novice acquires the culture of the group by engaging in the activity, this should not be understood as two distinct steps. As Becker shows, engagement in the activity is not a matter of a single decision, but is often a sequence of steps, each step leading the novice to changes in dispositions that make the next step sensible. Becker illustrates this best when he describes (Becker 1998) a deliberately extreme case of gradual changes that might lead a man to a sex change operation. Deepening engagement in the activity thus happens in lockstep with adoption of the culture of the group that practices that activity. We might add to this, that the entry into a subculture of the group defined by an activity may well precede direct contact with the actual activity.54     

      Becker (1963) sees this shared culture as arising from shared problems: “Many people have suggested that culture arises essentially in response to a problem faced by a group of people, insofar as they are able to interact and communicate with one another effectively” (p. 81). He notes, however, that not all activities develop a culture around them, since some activities are carried out in isolation, perhaps because they are stigmatized so much that those who engage in such activities choose to share this fact with nobody. (Becker uses the kleptomaniac as an example.) It is also clear that some activities, on the other hand, are so common that they do not lead to distinct group culture — instead the techniques and dispositions necessary for such activities are already present in the mainstream culture of the place.55

      Becker’s theory, as summarized so far, illuminates the process by which the individual may come to engage in an activity, in a way that can be clearly applied to the process of entry into what Hughes would call occupations. It mostly leaves out, however, Hughes’s stress on the division of labor between occupations. It seems clear, that to understand the continuous use of marijuana in the United States or the continuous practice of jazz, software development, or medicine, we would need to not only look at how an individual comes to engage in those activities, but also at how such activities interact with other activities, considering, for example, the relations between users of marijuana and dealers or police. Becker raises this question in his first paper about the Chicago jazz musicians (1951), looking at their interactions with the clients, but limits his discussion to the musicians conflicted feelings about the clients.56

Work as Labor

      While most sociology of work in 1950s and 1960s shows influence of Hughes and other Chicago School authors, the field was transformed in 1970s under the influence of a different tradition: Marxism.57 Although the Chicago School relied heavily on Durkheim (for instance, the ideas of “organic solidarity” and “industrial society”) and Weber, and while Durkheim and Weber had in turn been influenced by Marx (see Giddens 1971), Chicago School sociologists do not draw on Marx directly and their approaches differ from Marx’s substantially, usually lacking interest in radical criticism, longer-term historical change and relationships of production. Marxist literature similarly developed up to 1970s independently of American sociology.58 Braverman’s (1974) Labor and Monopoly Capital (subtitled “The Degradation of Work in the Twentieth Century”) and other Marxist writings of 1970s, however, had a great impact on American sociology of work.

      Braverman, a Marxist coming from the socialist movement rather than academia, explicitly enters a debate with the literature on occupations, arguing that under capitalism work is increasingly deskilled and alienated. Alienation of labor in the capitalist society is one of core principles of Marx’s thought, which Marx contrasts with what he sees as the natural (unalienated) labor. In Marx’s writing, labor is understood as human efforts to transform their material environment to fit their needs for survival. Such efforts are defining of the human existence and serve as the foundation of any human society. In other words, no human society can be understood without paying attention to how it organizes efforts towards producing things that people need for survival. Unalienated labor involves collective effort to transform the material environment to serve the needs of the group. Under capitalism, however, the workers do not transform their material environment towards their needs. Instead, they offer their labor to the capitalist in exchange for money. Labor then becomes “alienated” (alternatively “estranged” or “externalized”59), that is external to the worker (Marx 1978, Giddens 1971, Braverman 1974). Marx then makes an additional argument that the division of labor inside a capitalist enterprise requires the workers to increasingly focus on small tasks, further alienating them from the product of labor. Further, mechanization makes those tasks increasingly simple, making the worker “a living appendage of the machine” (Marx 1867/1977, p. 614).

      Braverman (1974) extends Marx’s theory, arguing that since the capitalist purchases from the worker abstract labor rather than specific products of labor, organizing this labor in order to maximize productivity becomes the responsibility of the capitalist. (Such organizing is achieved through capitalist’ agents: the management.) Since the workers have no intrinsic interest in the work, the capitalist cannot rely on the workers to apply their labor in the most productive ways. The capitalist’s solution, according to Braverman, is to capture the knowledge that was once in possession of the practitioners of different crafts, and then prescribe to each worker exactly what to do. (Braverman thus relies on a theory of knowledge that is akin to “the codification view” described earlier, though he focuses on the disempowerment of the worker, rather than the empowerment of “the society.” Braverman’s acceptance of the positive but not the normative side of the codification view, points out the political assumptions — and class interests — inherent in the standard articulation of the codification view.)

      Such organization of “the labor process” completes alienation of labor, as it removes the need to rely on workers’ creativity, and allows the capitalist to simply buy workers’ time. This “degradation of labor” depresses wages, since any worker can in theory do any work. Additionally, Braverman argues, the worker’s “working time becomes sharply and antagonistically divided from nonworking time, and the worker places an extraordinary value upon this ‘free’ time, while on-the-job time is regarded as lost or wasted” (p. 278).

      Braverman also attempts to update Marx by presenting an argument against “middle layers” of employment. Much of those workers have been absorbed into the labor market, he argues, and their work is also increasingly deskilled. While Braverman makes his argument most convincingly for clerical workers, he suggests that even “professionals” (Braverman uses this term in quotations) are not exempt from the alienating effects of capitalist labor process.60 Braverman thus presents an image of work that stands in stark contrast with that painted by Hughes (1958). Hughes’s understanding of work is most certainly influenced primarily by his understanding of skilled work. While he considers briefly unskilled work (“jobs,” which require no more than “to present one’s self at the proper time and place when manpower of a certain age, sex, and perhaps a certain grade of intelligence, is wanted,” p. 34), he seems to treat such work as atypical of the modern society, focusing instead on work that does require skill (which he calls “professions,” “arts” or “trades”). Braverman, on the other hand, treats the professions as an exception, suggesting that for most workers work increasingly lacks any meaning.

      Braverman’s book greatly influenced sociology of work (see Simpson 1989, Ritzer 1989), leading to a focus on deskilling and on the importance of differences among the different workers. Much of the ensuing discussion questioned Braverman’s argument.61 Form (1987), for example, argues that Braverman overestimates the importance of skilled work in the 19th century and that Taylorism had less effect on work than Braverman assumes and was only implemented in industries that already relied primarily on unskilled labor.62 After considering the difficulties involved in measuring or even defining “skill,” Form concludes that there is no support for deskilling across the board, and that the effects of technology and mechanization vary greatly by industry. The argument about deskilling continues until today, without any unambiguous conclusions, but with overall stronger support for upskilling than for deskilling. Perhaps the most important lesson that we learn from this literature is that there are very different kinds of work.

      Some authors have used this diversity of skills to argue for the emergence of a “dual labor force” (see Simpson 1989 for a review), divided into skilled and privileged “core” (or “primary”) labor force and deskilled “peripheral” labor force. This argument is later restated by Castells (1996/2000), who identifies the core labor force with “information-based managers” (p. 295). As the rest of this dissertation suggests, the boundaries between the two cannot always be drawn clearly.

      Braverman’s (1974) argument assumes, following Marx, that workers engaged in repetitive manual labor cannot find meaning in such work. Some of the other Marxist authors have disagreed with this. Burawoy (1972) presents an interesting fusion of Marxism and non-Marxist sociology:63 drawing on his ethnography at a plant he argues that workers do identify with their work and put substantial effort into it because they redefine it as a game. In other words, while Burawoy accepts the Marxist premise that hard work in the factory goes against the workers’ class interests, he draws a distinction between the objective meaninglessness of work (as seen by the social theorist) and the workers’ subjective experience of this work.

      Willis’s (1977) Learning to Labor presents a similar argument, with even more powerful ethnographic evidence. Willis looks at how working class culture is reproduced in British schools, pointing out how the students’ counter-school culture prepares them for acceptance of manual work. Willis’s “lads” see themselves in opposition to “ear’oles” — the students who buy into the school’s agenda (and presumably later move on to clerical work). They see manual work as masculine, “rough” work that not everybody could do.64 They also see manual work as allowing them freedom of mind: requiring them to perform certain activities but allowing them to see those activities as they like.65

      This dissertation draws on ethnography of software developers — people typically classified as “professionals.” The image of work that it paints is clearly more similar to that drawn by Hughes and even Becker than to the one presented by Braverman (1974). I do not accept Braverman’s argument that professionals represent a disappearing breed and that technical expertise is concentrated in just about 3% of the labor force.66 (Though, they clearly do not represent the typical case, as one might infer from Hughes’s writing.) At the same time, I take from Braverman the key insight that even skilled work must be understood as a form of labor. To be more precise, I believe that it must be understood as often being a matter of both practice and labor.

      The Marxist critique presented a challenge of reconciling the cultural understanding of work with the realities of work as a “labor process.” This reconciliation is partly provided by Van Maanen & Barley (1984). Speaking primarily to the organizational studies audience, Van Maanen & Barley argue for looking at workers as embedded in “occupational communities.” They define an occupational community as:

a group of people who consider themselves to be engaged in the same sort of work; whose identity is drawn from the work; who share with one another a set of values, norms and perspectives that apply to but extend beyond work related matters; and whose social relationships meld work and leisure.

Like Hughes, they stress the fact that work is often carried out in bounded occupational groups, the members of which share culture and identity. Their paper elaborates this aspect of Hughes’s approach, mostly leaving aside (as does early Becker) the question of division of labor between the occupational groups.

      Van Maanen & Barley (1984) for the most part do not engage in an explicit debate with the Marxist literature.67 At the same time, they present an image of work that is largely aligned with that of Hughes and Becker, but reinforced against some of Marxist critique by explicit discussion of the organization. They also apply the concept of “occupational communities” to a many different types of work, including work that is typically classified as “unskilled” (e.g., janitors or the street sweepers of Benares) — thus going explicitly against Braverman’s claims.

      Van Maanen & Barley’s analysis incorporates the idea that members of an occupational community stand in a particular relationship with organizations that employ them, essentially recognizing that this relationship involves in part the sale of labor (though they do not use this term). To some extent (and only while employing a micro-sociological perspective), we can see parallels between what Braverman (1974) calls “the capitalist” and “the capitalist enterprise” and what Van Maanen & Barley (1984) call “the organization.” However, while Braverman would see the enterprise and the workers as being fundamentally in conflict, Van Maanen & Barley stress the complexity of this relationship. They see individual workers as participating simultaneously in two groups: the occupational community and the organization. They see this dual membership as often being a source of conflicts, which arise from the misalignment between the interests of the workers as members of the occupation and the organization. At the same time, they recognize that goals of the two may align to a substantial extent and that for many occupations the organization is a necessary evil, without which the occupational tasks cannot be practiced: “certain kinds of engineers often discover that their collective aims and identities can be satisfied only within large, heteronomous organizations where sufficient resources to pursue occupationally-valued ends are to be found” (p. 310).68

      Van Maanen & Barley draw on a theory of motivation that is rarely employed in Marxist thought. Like many other Marxists, Braverman (1974) seems to recognize two theories of motivation: people may engage in work for the sake of the products that their work creates, which they then use individually or collectively or they may sell their labor for money, which is then used to obtain products. Both theories are fundamentally economic. Van Maanen & Barley, on the other hand, recognize, following Hughes, Becker and others, an entirely different source of motivation. To quote Becker (1953) again, “the motivation or disposition to engage in the activity is built in the course of learning to engage in it and does not antedate this learning process” (p. 235). Engaging in work activities then cannot be understood without considering how workers see their work, and their perception of work cannot be reduced to economic interests, as sociologically-minded Marxists have come to recognize (Burawoy 1972, Willis 1977).69

Boundaries

      Van Maanen & Barley (1984) stress the importance of boundaries, which they define phenomenologically: “the first attribute of an occupational community is that it is composed of people who consider themselves ‘to be’ members of the same occupation rather than ‘are’ members of the same occupation” (p. 295). This focus on member-defined boundaries leads them to say that the relevant groups are often smaller than we might think at first:

Conventional labels typically represent the theoretical limit of an occupational community. Within this boundary, socially significant types… are sure to exist. […] Commercial fishing provides a useful example because within its boundaries are found several rather distinct occupational communities. “Traditional fishermen” recognize differences between themselves and “nontraditional fishermen” such as “educated fishermen,” “part-timers,” and “outlaw fishermen” […]. Even more important are the distinction made within types. Thus, in the port of Gloucester, Massachusetts, traditional fishermen divide themselves into two groups, Guineas and Greasers. (p. 296)

Van Maanen & Barley’s stress on the boundaries recognized by the members as defining the occupational communities, however, creates a problem, if we try to use the concept of “occupational communities” as an elaboration of Hughes’s “occupational groups.” A crucial feature of Hughes’s occupational groups is that members of a group get license to perform certain tasks by virtue of their membership. Since Hughes sees this license as something recognized by outsiders, the boundaries of the group as perceived by outsiders cannot be ignored. To practice medicine, one must be recognized as a doctor not only by colleagues but also by patients, nurses and pharmacists. While recognition by fellow members may often be a prerequisite to being recognized by outsiders, it is the recognition by outsiders that ultimately allows the member to take part in Hughes’s division of labor, engaging in tasks that have been allocated for the members of his occupation. We thus need a better theory of boundaries and legitimacy.

      We can find such a theory in the Science and Technology Studies (or STS) literature, which comprises another tradition that is closely related to the discussion of occupation (and especially technologically-intensive occupations). This tradition becomes prominent in 1980s, and later discussion of technical work is impossible without considering it at least briefly. The origin of STS can be found in two disciplines: philosophy of science and sociology. Philosophy of science provides the question around which most early STS work has revolved and the shadow of which can be seen in much of the later STS literature: what distinguishes science from pseudoscience. This is often known as “the demarcation problem” (see Gieryn 1983 for history of the concept.) Philosophy of science has typically attempted to provide a normative answer to this question: a method that could be used to distinguish science and pseudoscience. Popper’s (1959) suggestion that scientific claims must be “falsifiable” is an example of this approach.70 Kuhn (1962), however, critiqued Popper, drawing on detailed accounts of how science was actually done. By presenting a social critique of what up to then was a philosophical discussion, Kuhn opened the door to sociological investigation of demarcation practices, which focused on how demarcation decisions are made, drawing heavily on the notion of “social construction,” originally introduced by Berger & Luckmann (1966).71 Latour & Woolgar’s (1979) Laboratory Life, perhaps the most influential book in this tradition written in 1970s, provided a detailed analysis of how scientific facts are “created” in the lab.

      The STS literature differs substantially from studies of work conducted in science or engineering labs (such as, for example, Barley & Bechky 1994) in that STS is typically interested first and foremost in the science (or the technology), subjecting the work of scientists and engineers to careful study in order to achieve a better understanding of the result of their work. Work studies, on the other hand, often take work itself as the primary object of investigation. Consequently, studies of work applied to scientific and engineering occupations typically compare such work (at least implicitly) with other types of work. Despite those differences, the two traditions have often looked at the same types of work and some of the STS methods are quire relevant to understanding work broadly.

      STS has traditionally taken a critical approach to the people it studies: rather than practicing Weberian “sociology of understanding” towards scientists and technologists, STS has typically aimed to undermine their accounts of how science and technology are done.72 Such critical approach is a double-edged sword. On the one hand, STS literature teaches us how to cut through members’ ideology, especially in the difficult cases where the subject matter of their work is sufficiently complex that the researcher cannot fully understand it, and when some of the subjects may be of higher social status than the researcher, be generally seen as more knowledgeable and better qualified to speak for the rest of the members. I employ some of those tools in this dissertation, in particular when trying to look beyond the idea of software development as a “global” profession.73 At the same time, such critical attitude can incidentally cut off too much of the members’ experience, leaving us with overly rationalized and economistic explanations of their actions (see for example Latour & Woolgar 1979, Latour 1988a).74

      Gieryn’s (1983) notion of “boundary-work” is particularly valuable to understanding practice.75 Gieryn looks at the problem of demarcation as a practical problem for scientists. He writes:

Construction of a boundary between science and varieties of non-science is useful for scientists’ pursuit of professional goals: acquisition of intellectual authority and career opportunities; denial of these resources to “pseudoscientists”; and protection of the autonomy of scientific research from political interference. (p. 781.)

He then introduces the notion of “boundary-work” as “an ideological style found in scientists’ attempts to create a public image for science by contrasting it favorably to non-scientific intellectual or technical activities” (p. 781) and “their attribution of selected characteristics to the institution of science (i.e., to its practitioners, methods, stock of knowledge, values and work organization) for purposes of constructing a social boundary that distinguishes some intellectual activities as “non-science” (p. 782). While Gieryn looks at specific rhetorical tools employed by scientists in their “boundary-work,” the term has gained a broader circulation, denoting more generally the active (and often purposeful) construction of boundaries between groups. The term has obvious application to the discussion of occupational cultures: while Van Maanen & Barley (1984) stress Weberian “consciousness of difference,” the concept of boundary-work suggests that such differences are not something that the members are merely conscious off, but are something they actively construct.76

      Finally, some of the research in STS has addressed directly the issue of space. I leave this discussion for later, however, when discussing Latour in chapter 1.3.

Practice, Power and Peripherality

      Lave & Wenger’s (1991) Situated Learning is written in a somewhat different tradition from Hughes, Becker or Van Maanen & Barley, but has since been incorporated into the work studies canon and is often read through its prism. Like Becker (whose 1972 article on schools Lave & Wenger cite), Lave & Wenger look at a broader category than “work” or “occupation.” The book looks at learning, re-conceptualized as progressive engagement with “communities of practitioners” or “communities of practice,” which they define as follows:

A community of practice is a set of relations among persons, activity, and world, over time and in relation with other tangential and overlapping communities of practice. A community of practice is an intrinsic condition for the existence of knowledge, not least because it provides the interpretive support necessary for making sense of its heritage. Thus, participation in the cultural practice in which any knowledge exists is an epistemological principle of learning. (p. 98)

The concept of “a community of practice” shares important characteristics with Becker’s “groups” and Van Maanen & Barley’s “occupational communities.” It similarly involves a notion of a bounded group that carries out certain activities, made possible by the groups’ “knowledge” (Hughes’s and Becker’s “technique”) and a shared interpretive system (Hughes’s “culture,” Becker’s “perceptions and judgments,” Van Maanen & Barley’s “culture” understood as “a meaning system.”). Like Hughes, Lave & Wenger explicitly define communities of practice “in relation with other tangential and overlapping communities of practice,” though they do not explore this aspect of communities of practice in any substantial depth.

      The concept of practice provides a useful alternative to the notions of “work” and “occupation.” It first of all gives us a broader term, helping us see similarities between organized activities that we consider “occupations” and those that we do not. (Lave & Wenger’s examples of practice include “Alcoholics Anonymous,” which is clearly not an occupation in Hughes’s sense, but likely an “activity” and perhaps a “deviant group” in Becker’s terms.) While a similar shift is made by Becker, it is obscured by the focus on deviance. Lave & Wenger’s notion of However, “practice” is not merely a generalization of an “occupation,” but a multi-faceted concept with its own theoretical history. It would therefore help to pause to analyze it briefly.

      Oxford English Dictionary breaks the usage of the English word “practice” into four semantic subfields (after omitting obsolete and rare senses):

      (1) exercise of a profession (“began his law practice”),

      (2) actual application of an idea (“integrates theory with practice”),

      (3) customary or established conduct (“has become common practice”),

      (4) repetition for the sake of improvement (“hours of daily severe practice”).77

When used in social science, the term “practice” typically takes one of the first three meanings, which for the sake of simplification can be reduced to two dimensions:

      A. a system of procedures established within some group (most of 1 and 3 above)

      B. action contrasted with theory (subfield 2).78

Those two dimensions are related, but are at the same time distinct, with different history in social science, though sometimes used by the same authors. Lave & Wenger (1991), in particular, invoke both dimensions and it is important to understand how they combine them.

      A. “Practice” understood as a coherent set (or a system) of activities in which some group of people engages fits quite naturally with the cultural theory of occupation discussed so far. It corresponds quite well to Becker’s “activity” and can be seen as a superset of Hughes’s and Van Maanen & Barley’s “occupation.”79 Specific practices in this sense are defined by differences and boundaries. A “practice” is always implicitly contrasted with some other way of doing things (or simply not doing something). Additionally, within the tradition reviewed so far (and also for Lave & Wenger), a practice is always linked (and mutually constitutive) with a group. When people who engage in a practice are in contact with those who do not, engagement in the practice marks social boundaries and serves as a source of identity for the practitioners. Lave & Wenger invoke this sense of practice when they talk of practices as having histories and involving sets of relationships, and also when they use “communities of practitioners” as an alternative to “communities of practice.”

      B. “Practice” understood as a form of engagement with the world that aims to transform rather than merely understand it, has no explicit parallel in the work of any of the authors discussed in earlier sections, but is common in Marxian thought.80 Lave & Wenger (1991) invoke this sense of practice explicitly, and more generally as they argue for the primacy of action over theoretical knowledge. Communities of practice are communities of action rather than theory, defined by what people do rather than by what they know abstractly.81 The distinction between practice and theory becomes particularly import with the rise of “informational” understanding of work.

      Lave & Wenger’s (1991) discussion of practice uses the second dimension to explain the first: a practice, understood as a system of activities, persists in time because new people come to engage in it, and this entry must be understood as a matter of joining the actual activities, rather than absorption of the abstract knowledge that may be said to underlie the practice.82 Reliance on this second dimension of “practice” leads Lave & Wenger (1991) to look at how the practice relates to the material world and how a novice comes to engage in the activities that comprise it.

      The attention to the material aspects of the practice highlights the importance of tools. The activities carried out by the practitioners, typically involve transformation of certain objects.83 (Sometimes the objects are just the members themselves, as is the case of the Alcoholics Anonymous example, but this is the exception.) This transformation is made possible by interpreting the objects in a particular way, but it often also relies on artifacts, which “carry a substantial portion of that practice’s history” (p. 101). In many cases it is the access to such artifacts that limits entry for new members.84 On the other hand, since material tools are mobile in different ways than persons or culture, they may be important when we later discuss reproduction across space, as we will see later.85

      To understand how a novice comes to partake in the activities, Lave & Wenger (1991) also introduce the theory of “legitimate peripheral participation,” arguing that new members enter the practice by engaging in peripheral tasks, and moving to other tasks gradually. Such peripheral participation allows them to become members (at least tentatively) and to start absorbing the knowledge and culture without having the expertise that is expected of the full members and necessary for engagement in the more central tasks. The notion of legitimate peripheral participation is in some ways similar to Becker’s theory of social learning (1953), but it adds a crucial component of legitimacy. Since practice often requires a “license” (from other practitioners, the larger society or both), we must pay attention to how the novice obtains the initial permission to enter. When looking at the role of the university programs that train software developers, we must consider not only what the future software developers might learn there, but also the way in which such practices may allow the students to become legitimate participants in certain practices.

      The notion of legitimacy links the discussion of practice to the notions of power and access. Power relationships within the community as well as those of the larger society grant access to some participants and deny it to others. “Hegemony over resources for learning and alienation from full participation are inherent in the shaping of the legitimacy and peripherality of participation in its historic realizations,” write Lave & Wenger (p. 42). Furthermore, those who are admitted do not necessarily become equal to the old-timers upon this admission. Legitimate peripheral participation therefore “can be a source of power or powerlessness” (p. 36). In particular, “as a place in which one is kept from participating more fully — often legitimately, from the broader perspective of the society at large — it [peripherality] is a disempowering position” (p. 36).

      The theory of legitimacy, access and power, however, is underdeveloped in Lave & Wenger, and is critiqued by Contu & Willmott (2003, 2006), who label Lave & Wenger’s theory of power “embryonic” and hold Lave & Wenger in part responsible for the later “conservative” misreading of their book. In particular, Contu & Willmott (2003) argue that Lave & Wenger often seem to take for granted the cohesiveness of communities:

In Situated Learning and its subsequent refinement (Wenger 1998), the concept of community is ostensibly positioned on a conflictual terrain (Wenger 1998, Ch. 2). However, the condition of existence of such communities is located in harmonizing categories such as “a sense of joint enterprise relationship of mutuality shared repertoire of communal resources” (Wenger 1998, emphasis added). Community is conceptualized in a way that tends to assume, or imply, coherence and consensus in its practices. Such “unequal relations of power” (Lave and Wenger 1991), we suggest, glosses over a fractured, dynamic process p. 42) of formation and reproduction in which schisms and precarious alignments that are held together and papered over by unreflexive invocations of hegemonic notions including “community,” “family,” “team,” and “partnership.” By default, Lave and Wenger’s usage of “community” is complicit in the reproduction and legitimation of this hegemonic process. (p. 287.)

Contu & Willmott underscore “the danger of assuming consensus in communities of practice,” and urge their readers to focus their attention on the idea of “practice” rather than “community.”86 In my analysis of my fieldwork I specifically attempt to avoid this alleged shortcoming of Lave & Wenger, avoiding the term “community” except as a “native” term used by the participant, and treating it with caution then. Instead of assuming that software developers join a “global community” of software developers (as they often do describe it), I see them as engaging in a practice that requires negotiation of many power relationships. “Boundary-work” and other notions from STS (see above) become useful in uncovering the “construction” of the community.

      Lave & Wenger (1991) use “legitimate peripheral participation” as a micro-sociological concept: something that applies to the activities and position of concrete individual people. However, this notion can also become useful at a higher level of analysis, when considering the position of certain groups within the larger world of practitioners. We can draw certain parallels between the role of an individual “newbie” software developer who have just started working at a software company and the collective position of Brazilian software developers who have achieved legitimate but peripheral role in the larger world of software, and whose progression to more central participation involves substantial difficulties. Such position can similarly be “a source of power or powerlessness” (see Lave & Wenger 1991, p. 36). Such macro notion of peripherality has been expressed in some of the political economy literature (e.g. Evans (1979). I discuss such links in chapter 1.3.

Work as Situated Practice

      Julian Orr’s ethnography of Xerox technicians (1996) is perhaps the best known of the books answering Van Maanen & Barley’s (1984) call for ethnographic studies of occupational communities. At the same time, the book “puts the flesh of everyday life on Lave & Wenger’s (1991) idea of a community of practice” (Barley, 1996, p. xiii). Orr’s focus on “work as doing, as practice” (p. 10) is indicative of this fusion — and the source of the title of this chapter. (Note, though, that Orr appears to focus mostly on practice in its “common sense” meaning, as defined above, without indexing Marxian praxis.) It is important to note that this fusion is not undertaken by Orr (1996) explicitly. In fact, as Duguid (2008) points out, Orr’s book “studiously avoids” any mention of Lave & Wenger. The first publication that uses Orr’s ethnography to illustrate concepts from Lave & Wenger is Brown & Duguid (1991).87 This union is later blessed by Barley (1996) in his foreword to Orr’s book.

      Orr’s ethnography follows a small “community of practitioners” that he calls “the reps” — Xerox field technicians, who attend to photo copiers installed on client’s premises — and looks at their relationship to a specific organization and the organization’s clients. The book provides a rich account of work that helped not only illustrate some of the ideas discussed above (especially those of Van Maanen & Barley 1984 and Lave & Wenger 1991 — see Brown & Duguid 1991), but also allowed for a critique of such ideas. For example, Contu & Willmott (2003, 2006) provide their own interpretation of Orr’s observation, accusing Orr himself and also Brown & Duguid (1991) of accepting the management point of view. While I agree with Contu & Willmott’s call for a more critical analysis of Lave & Wenger (1991) and Orr (1996), I see this re-interpretation as first of all a sign of the richness of Orr’s ethnography and aim for similar richness in my own examples.

      While greatly admiring Orr’s ethnography, I also see danger in its richness. As Barley (1996) and Barley & Kunda (2001) argue, social science of work often relies on the available “images of work,” often outdated. Orr’s ethnography provided an image of work that helped replace the earlier images. The book greatly affected (both directly and through Brown & Duguid 1991) the discussion of work in the years that followed its publication.88 Such a discussion has consequently ignored some of the issues not captured by Orr’s image. It had a particularly strong effect on they way Lave & Wenger’s “communities of practice” are understood.89 Even the literature that extended the term “communities of practice” to virtual communities bears the mark of Orr’s image, often talking about such virtual communities as tight-knit and inward-looking.

      Orr’s ethnography, however, is based in a particular and highly peculiar place, Silicon Valley. Even on a clear day, such as the one that Orr describes in his first vignette (p. 15), someone driving through the valley sees a world with remarkably narrow horizons, bounded by hills on both sides. (It is a valley, after all.) Orr’s writing captures the imagery of that small world in great detail, but largely ignores anything that happens beyond the hills that bound it. The reps’ world appears to be fully contained within this narrow space. (The corporate headquarters of Xerox happen to be on the other side of the country, but this fact is hardly apparent in Orr’s book — and is perhaps simply not relevant. Orr mentions the fact that “San Francisco” service area stretches as far as Guam, but does not describe any practical consequences of this.)

      Geography does come up in one small passage, which gives us a glimpse of what Orr might have observed had he chosen a different place:

Frank began working with the corporation in Indiana, which may account for this reference to a Chicago Fix. Some districts get new machines before others, and some have accounts that will use the machines much more intensely than others. These districts encounter most of the problems before others will and must develop their own ways to fix problems until engineering and manufacturing catch up. Word of these innovations will spread, usually carrying acknowledgment of their origin. Such districts have a reputation among the surrounding districts as something like the big time, the cutting edge of the blunt instrument that is field service. Silicon Valley is such a district on the West Coast, Chicago is the same for the Midwest, and this may remind us that although service is intensely local, they do work as members of a multinational corporation, and what happens elsewhere in the corporate geography may matter to them too. (pp. 60–61.)

In other words, reps working in places such as Indiana do have to pay attention to practice occurring elsewhere — for instance in Chicago. Those working in Chicago or the Silicon Valley, on the other hand, must rarely be reminded about the rest of the world. What happens elsewhere, may matter to them, but does not warrant a discussion beyond this one passage.

      I aim to present a different image of work: one that follows Orr in attempting to show the importance of local work, yet also captures how, for many people, what happens elsewhere has tremendous importance on a day to day basis, and where “elsewhere” refers to places that are much further from them than Frank’s office was from Chicago. Such an image will help us extend the insights of Lave & Wenger (1991) to the new, globalized work.

      Both Lave & Wenger (1991) and Orr (1996) rely on the notion of “situated activity” to explain their projects. (Orr also uses the terms “situated practice” and “situated action.”) They do not use them in the same way, however, and the difference between the two gives us a chance to get a better understanding of the ways in which action can be “situated.” Orr, who explicitly attributes his concept of situated action to Suchman (1987), focuses on the way action is situated in its immediate physical context — hence the commitment to detailed descriptions of work and the places in which it occurs, as well as the stress on the importance of the reps’ physical presence near specific machine and their ability to hear their characteristic sounds. Lave & Wenger (1991), on the other hand, reject a similar notion of situatedness, comparing it to “naive views of indexicality” (p. 32). While their understanding of situatedness is not defined as clearly as Suchman’s and Orr’s, it is clear that the book shows less interest in place and minute details of activities. Instead, it looks at learning as situated in the larger context of practice — that is, situated in a social world of action and in history of the practice, rather than in physical space understood narrowly.90

      My approach to situatedness integrates Lave & Wenger’s and Orr’s, looking at the work of Brazilian software developers as situated simultaneously in two different contexts: the local one, similar to the one described by Orr, and the more abstract context of a practice based far away — in fact, the very place within which Orr’s ethnography is situated. I stress that we can only understand their practice if we look at those two contexts (and those two kinds of situatedness) side by side and examine the contradictions between them.

A Synthesis

      This dissertation takes as a point of departure the notion of “work as practice,” recognizing that work often involves simultaneously an economic relationship of employment and a membership in a community of people who engage in similar work and share certain technique, culture and identity. Work is thus subject to both cultural and politico-economic analysis, and the two must often be performed simultaneously.

      At the most basic level, practice can be understood as a system of activities in which a group of people engages. Members see some sort of meaning in the activity and participation in the activity forms part of their identity. Those activities may be recognized as valid and meaningful by some people who do not engage in the practice, though such people’s understanding of those activities is likely to differ from the member’s own understanding. The activities that comprise the practice and the group of practitioners often mutually define each other: the group is typically understood (by both insiders and outsiders) as a group of people who engage in the activities, and at the same time the activities are considered a part of the practice to the extent that the group engages in them.

      The meaning of the practice is typically defined in relation to other practices and the practice often has a place in a larger division of labor. Those relationships can be economic or non-economic. They can also be cooperative or antagonistic. A practice sometimes needs to be understood by its relationship to “client” practices; in other cases, it needs to be understood in its relationship to its “anti-practices.” For example, the practice of investigating drug traffic cannot be understood in separation from the practice of trafficking drugs. Cooperative relationship with other practices must not be understood as necessarily equal and fair. They may well be exploitative, and the balance of power between practices may vary widely.

      Practice typically involves transformation or manipulation of certain objects as well as transformation of people from outsiders into members. Successful manipulation of the objects (per group’s and its clients’ definition of successful) requires a particular perception of the objects and situations and application of certain techniques to them. This involves doing things outsiders usually do not do. The application of technique may presuppose access to tools, which embody certain aspects of the practices heritage.

      The novice’s entry into the practice requires acquisition of perceptions and technique, as well as access to tools. The perceptions and the technique can only be fully acquired by engaging in the practice together with the more experienced members. (Later development of perceptions and technique requires continuous engagement in the practice together with other members.) Such participation is often possible because the members recognize the need to bring new people and allow novices to engage in peripheral tasks that gradually lead them to fuller participation. It may also be possible because old-timer benefit from the work carried out by the novices. The relationship between old and new members can thus be unequal and exploitative, and the group of practitioners will likely more generally be characterized by a system of power relations.

      The need to engage in the practice together with other practitioners requires that the practitioner is recognized as a member (or at least as a member-in-training) by other practitioners. Member-defined boundaries are important for this reason. However, certain groups of outsiders may recognize specific groups of practitioners as legitimate providers or consumers of certain objects or services (further recognizing their right to engage in the activities that are involved in the practice). Since such recognition is done at the level of a group, being recognized as a member of the group by outsiders may have consequences for the individual’s ability to engage in the practice. While outsiders may often rely on the membership judgments made by the members themselves, we cannot assume the insiders’ and outsiders’ understanding of the group’s boundaries to be the same. Members engage in “boundary work” to bring outsiders’ understanding of group’s boundaries in line with their own.

      This synthesis provides a foundation of a discussion of work as practice, but suffers from an important flaw — failure to theorize the relations between practice in different place. The next chapter presents a slice of my own ethnography, which I hope will expose this problem. I then return to theory in chapter 1.3, presenting a theoretical account that tries to make up for this failure. Parts 2 and 3 of the dissertation then put more ethnographic flesh on the theoretical skeleton presented in chapter 1.3.

Notes