Transparency At Different Levels of Scale: Convergence between Information Artifacts and Social Worlds

Susan Leigh Star

Geoffrey C. Bowker

Laura J. Neumann


Library and Information Science

University of Illinois, Urbana- Champaign

501 E. Daniels.

Champaign, IL 61820

1 August 1997

This paper was supported in part by the NSF/ DARPA/ NASA Digital Library Initiative under contract number NSF 93- 141 DLI, and by the NSF with a grant for research on classification and infrastructure, contract number 9514744. Our thanks to people who read drafts of this work and discussed the ideas herein, especially Ann Bishop and the DLI Social Science Team and members of the Graduate School of Library and Information Science’s proseminar.

  1. Introduction

    I: How do you keep up with your field?

    R:. . . [T]he field that I work in, and since I have been in it for so long and since I have trained a lot of the people in it, I know practically everybody in the world who is working in it. I am also on the editorial board of a number of journals, so in a lot of cases I see things before they are even in print. In my case I probably make very good usage of both of those things. I am in correspondence with a lot of people and they tell me what they are doing, just because I know them... and I go to meetings, to quite a few. This is where I meet my colleagues and we talk. I probably am almost never in the situation where I am having to do a search in an area where I don't know anything about the field. I may never.

    (Engineering professor of 40 years)

    Problems undergraduates have with the library system:

    * Journals are scattered all over campus;

    * The article I need is always ripped out;

    * I don't know how to phrase things just right, so I don't get what I want;

    * The professors won't let us write papers from the abstracts alone;

    * I'd take local, irrelevant articles over remote, relevant ones.

    (from a focus group with undergraduate students)


    "Not all the work that has made the ICD more applicable has been done internally through modifications to the list. Indeed, one background factor that has had a great impact has been the convergence of international bureaucracy. What we mean by this is that throughout this century in general people have become more and more used to being counted and classified. . ."

    (Bowker and Star, 1994: 202)


    As information systems are used by more people, and permeate more of our working and leisure lives, scholars studying the human side of computing must scale up various concepts traditionally seen as individual or psychological. For example, in discussing how human-computer interaction has changed as a field, Grudin notes that it has gone from individual-screen to groups-groupware, and now logically must seek to extend beyond the group into the wider social sphere (1990). When large groups of people are using a widely distributed system, old notions of "one person, one terminal" as the basis for design are inadequate. Bannon (1990) makes a similar point in discussing his ‘pilgrim’s progress’ away from the individualized, cognitivist notions of HCI. Kling and Scacchi’s classic notion, the "web of computing," seeks to expand beyond the single user terminal and formal program stereotype to locate design in a larger context of usability, the workplace and its social networks (1982). King and Star (1990) apply the scale-up problem to decision support systems, noting that the scale-up from decision support (DSS) to group decision support (GDSS) to organizational decision support (ODSS) is a complex one, involving issues of social justice, multiple interpretations, and adjudication of conflict across social boundaries. Within the field of computer-supported cooperative work (CSCW) and the design of large-scale information systems such as digital libraries, scale questions are at the heart of design. What is the proper unit of analysis for design -- group, organization, community, network? What does a community interface, or system, even look like?


    In this paper, we parallel these and other lines of scale inquiry about information systems to ask similar questions about the concept of transparency. A system is transparent if the user does not have to be bothered with the underlying "stuff" driving it, just as an automobile is transparent when one sits down, turns the key, and drives off without the foggiest notion of how internal combustion works. Total transparency, like the ultimate "killer app", is of course an ideal type, nowhere to be found in the real world (everyone’s car stalls now and then, or has its little idiosyncrasies requiring a deeper involvement with the tool). We define transparency here to include the idea of relative usability, and computer tools and interfaces which are basically ready to hand for some group of users.


    Scaling of transparency means that we must have an idea about:


    · for whom (and when and where) is a particular tool transparent;

    · what happens when levels of transparency differ across involved groups;

    · how does something become invisibly usable at the individual, community, and larger levels of scale, and with what differences in process and design content?;

    · how are newcomers to a community taught how to make a tool, interface, or retrieval system transparent?


    We seek in this paper to provide a description of information interaction with which to understand the continuum of usefulness of information organization and access. Transparency and ease of use are products of an alignment of facets of information resources and social practices; each of these facets are interrelated and in motion. In scaling up transparency, we wish to avoid fallacies of universalism; of an idealist concept of pure information devoid of locale or of the politics of knowledge; or a scenario which begs questions of material resources and infrastructure. Our research has been geared toward the real world use and design of information systems. It is always embodied, always historical, and always embedded in infrastructure.


    One impetus for clarifying a scaled-up idea of transparency comes from changes in the information landscape itself. In the current wave of building and research on digital libraries and information infrastructure, we find places and possibilities where new kinds of arrangements are being forged for information organization and access. Pockets that have been semi-autonomous until now are getting interconnected, including widespread federations of information repositories. One such example is the Illinois Digital Library Initiative. One of the goals of this project is to create seamless interconnection of federated repositories. Also, traditional models of information retrieval are being turned completely upside down; they can no longer presume an individual user, an information intermediary such as a librarian, and a repository with discrete documents which are then returned to the workplace or home. Information (documents and multimedia) is being fractured and distributed in different ways across the work-/play- scape. Thus a fuller understanding of people, their activities, and the scaling of transparency is especially crucial for a social informatics of digital libraries, but applies as well to other complex and large scale information systems (Bishop and Star, 1996).


    1. Transparency Results from the Convergence of Social Worlds and Information Artifacts

    We will explore in this paper the use of the concept of convergence for understanding scaled up transparency. It is a concept designed to bridge between social worlds (including membership and practices) and information artifacts "Social world" is a term in sociology first coined by Anselm Strauss (1978b). It refers to a group of people joined by conventions, language, practices and technologies. It may or may not be contained in a single spatial territory; in the modern world, it typically is not. It is cognate with the notion of community of practice (Lave and Wenger, 1992), and with reference groups. It was coined for social analysis in order to speak to strong ties which are neither family nor formal organization, nor voluntary association, and which may be highly geographically dispersed. Examples of social worlds are stamp collectors, rock climbers, activity theorists, and socialist feminists. Social worlds may segment (divide into subgroups) or intersect (have partial overlapping concerns); the degree to which they engage members or are voluntary is variable (Star, 1994). We use the term information artifact to mean a wide array of tools, systems, interfaces and devices for storing, tracking, displaying and retrieving information, whether they be paper or electronic or other materials.


    Convergence as defined in this paper is the double process by which information artifacts and social worlds are fitted to each other and come together. On the one hand, a given information artifact (a classification system, a database, an interface, and so forth) is partially constitutive of some social world. That is to say, the sharing of information resources and tools is a dimension of any coherent social world - be it the world of homeless people in Los Angeles sharing survival knowledge via street gossip, or the world of high energy physicists sharing electronic preprints via the Los Alamos archive. On the other hand, any given social world itself generates many interlinked information artifacts. The social world creates through bricolage, a (loosely coupled but relatively coherent) set of information resources and tools. People without houses also log into the Internet, and physicists indulge in street gossip at conferences - as well as engage in a whole set of other information practices (Palmer, 1996). Put briefly, information artifacts undergird social worlds, and social worlds undergird these same information resources. We will use the concept of convergence to describe this process of mutual constitution.


    In order to recognize this mutual structuration, we employ a term used by Elfreda Chatman (1985, 1991, 1992) - "information world." This term refers to all spheres of information resources employed by an individual, organization, institution, or other group in order to solve problems, learn, play, and work. They may be formal as in libraries and databases, but they are also informal, as in asking one's family and friends for opinions on health matters, or how to evaluate a particular book or film. Such resources are also material, in the sense that they are enabled and constrained by time, transportation, and material cultural practices. Information artifacts and social worlds converge to form information worlds in Chatman’s sense.


    We will through three examples explore how convergence occurs. First, we discuss how convergence is experienced by individuals as academic researchers. Second, we examine a study of nursing classification wherein nursing researchers are seeking to make their classification of nursing activities part of the social worlds of professional medical practice. Finally, we describe the convergence that subtends a global, well established information collection and analysis system administered by the World Health Organization- the International Classification of Disease. This is convergence that goes beyond a single occupation, nationality, or locale; in fact, it consists in the end of a number of intersecting information worlds. It speaks to the ways in which multiple overlapping systems support each other (that is, in the case of convergence; it could of course go the other way with divergence and incompatible structures), across significant space and time.

    Through each case, we will examine the operation of key feedback processes which underwrite the (ever partial) fitting-together of social worlds and information artifacts. We will then discuss how the concept of convergence can be of use in developing the integration of sociological and informational analysis towards which much recent work in library and information science has pointed.



  3. Convergence and the Individual: Membership and Transparent Information Retrieval Go Hand in Hand

    I: How do you find information or references?

    R: Often a paper comes across my desk, it is an archaeological dig here. . . simply because I will review articles or journal submissions and something seems interesting and I will file it. Or something seems interesting in one of the publications I subscribe to. . .

    (Professor of 12 years)

    Sometimes a person walks into a library (or other information system) and it's just OBVIOUS how to find things. This is the person who can put their fingers on things, who sits at the metaphorical center of the social web, who can tweak a vast complex and mature system of social networks to "get stuff." This is a person for whom the system is quite transparent.


    At the other end of the spectrum is the person who sees the information system as confusing, chaotic, insurmountable, unusable; who tries to follow given directions and misses by a mile and a half. Much professional socialization concerns moving from this lost state into the state of obviousness or naturalness. Analysis at this individual level asks how people adapt to, and work around, formal structures when they don't fit; how they put together and maintain networks which "feed" their work; and how to understand the interaction between vernacular and formal systems of information and classification. For example, professors will point to articles they review or items colleagues send them as more important information resources than the library.


    Convergence for the apprentice professional is a trajectory: the shaping of the individual so that they see themselves as having the set of information needs that can be met by their new social world's information resources. As the individual moves along this trajectory, they both learn a field that they can "put their mark" on and are changed by their socialization so that their original "information needs" are not the same as their final ones. For the individual, this feedback process, especially the ways in which their own information needs change, is frequently invisible. People often see themselves at the end of an academic question as asking the same "big dumb question" (Linde, 1997) that they asked going in, without recognizing that the constitution of the question has thoroughly changed through its adaptation to their relevant information and social worlds.


    Thirty-eight research scientists and students were interviewed for a large information infrastructure building project over a three year period. The goal of these individual interviews and focus groups was to learn more about information finding behaviors of research scientists and students as well as exploring their use of information systems. In line with established research findings in user studies of information retrieval, most of the senior-level respondents did not often use formalized information systems (Taylor, 1991; Garvey and Griffith, 1980, Pinelli, 1991). These researchers got the information they needed through an entire suite of other information seeking activities, such that they rarely found themselves going out and searching for anything. It was often difficult for these people to describe how they had access to items that they needed, these were simply ready to hand and a part of their immediate environment- transparent.


    The means for making this information readily available is the stuff of one's location at a particular point in the academic trajectory. It involves insider, membership status in an academic community- the fact of having been through the apprenticeship process. In this process, one becomes attached to the installed base or the extant infrastructure of academia. For example, membership in academia requires conference attendance, an amount of discussion with colleagues, membership in various professional organizations. One publishes in certain key journals and tracks career paths through key institutions. In the digital library interviews, those who had become professors had had to fulfill the taken for granted criteria (a specified number of classes, exams passed, thesis completed). This necessitated learning the map of the field, including salient research areas, acceptable methods and methodologies, and proper vocabulary. They had to identify a research area of interest and learn what in the field was relevant and important, and so on. Collaborations, friendships, peers and colleagues all add to the size and strength of the student's growing social network.


    With full membership, the professors information world takes shape through the convergence of information artifacts and social world membership and practices. Library services copy and send on needed articles; on-line archives send automatic reminders of new work in the field. In our study, the professors reviewed articles as well wrote them; often they had become journal editors. Further along the career trajectory, they had participated in professional governing bodies or perhaps served in a funding agency to shape future research trends.


    Through all of these levels of involvement, channels were established along which information flowed. All of these processes of professionalization and socialization into a social world enabled the convergence with information artifacts. The workings of the profession into which academics enter solidly embeds them in a community of exchange of information which happens (and is expected to happen) on both formal and informal levels. In most ways, knowing many members of one's research community and talking to them frequently is more important than reading professional journals (Garvey and Griffith, 1971; Crane, 1972; Wolek and Griffith, 1980). Many respondents emphasized that the idea is to know what will be in the journals before they are released.


    Throughout their careers as students and in the process of becoming professional, researchers focus their interests, which are represented in collegial networks, in the piles and files of papers kept, in research groups, and in journal publications. Thus people set up the means, artifacts, and processes that were necessary for social world/information convergence. The further one travels in the career line and into the more specialized reaches of an academic specialty, the more intense the convergence becomes. As Latour (1987) notes:

    There is a direct relationship between the size of the outside recruitment of resources and the amount of work that can be done on the isolated specialist is a contradiction in terms. Either you are isolated and very quickly stop being a specialist, or you remain a specialist but this means you are not isolated. (p. 152)

    This mutual reinforcement and the seeming seamlessness of the webs of attachments of many researchers can also be their "Achilles' Heel." One researcher pointed out that she did indeed rely almost entirely on her network of contacts and subscriptions, and if those resources had holes in than she would probably never know it. Reliance can become total (Swanson, 1986). In this way, too tight a convergence becomes over-determination; it becomes a rut that cannot be broken free of. Overly tight coupling between a social world and its information artifacts can be seen as a powerful routine that closes off other possibilities of finding information or using the imagination because they are not part of the routine. Bates (1979) describes how these over-trodden paths of thinking can be broken specifically for finding information. Campbell’s essay on the "fish-scale" approach to interdisciplinary knowledge also intuits the value of partial overlaps and a looser coupling between social worlds and information (Campbell, 1969).


    For the individual, while convergence can produce transparency and ease of use, the opposite often occurs. This frequently is the experience of newcomers to a social world, or those for whom the formal information systems fail to convergence with the criteria of membership. This results in fracture and frustration- a plight often felt by undergraduates. This person often has no clear academic home, she may not even have a "major" field of study. She does not know which the pertinent journals are, who the main people are, nor even the language to use to learn these things. As we were told in focus groups about what undergraduate's information needs are, "I just need five references. Any five references. I want to get those as quickly and as painlessly as possible." It does not matter who wrote the article, when, or why. This situation is less one of apathy than one that is typical of many undergraduates' relationship to the field of study. Those markers and the community gestalt which define quality, relevance, and usefulness are absent and thus irrelevant. If the new scholar learns and becomes a member of a particular community through participation, as described above, these community standards and needs become their own and at the same time become attainable and sensible and eventually, transparent.


    As Lave and Wenger (1992) describe, "learning the ropes" of a new community requires what they call legitimate peripheral participation. This is not only, or even primarily obtained through watching, but through gradually being made a fully participating member of a community through experience. In this way a new member comes to know what being an insider in that community is all about. The academic system is structured ideally to provide that space to participate and learn for incoming students: to cause a convergence between the student's developing academic, social, and artifactual worlds and available information worlds. In practice, of course, as Lave and Wenger (1992) and Eckert (1989) point out, the ideal is often not realized. Learning, or information acquisition, are treated as if membership were not an important factor, or as if students were not juggling multiple complex memberships. The trajectory to convergence, when traveled, is rarely smooth.


    We will now turn to a second example of convergence, from the point of view of a community of nurses. In this example, we will see the same feedback process occurring through which the nursing profession, by attempting to integrate into the professional world of medicine is shaping and being shaped by the information world it is becoming part of.


  5. Community and Convergence: Crafting and Aligning Information Systems with Community Practices

    I attended the American Nursing Association Database Steering Committee in June and a recommendation coming out of that meeting and I'm assuming making its way through the ANA channels is that ANA needs to put forth a major lobbying effort towards nursing representation at the National Committee on Health and Vital Statistics. The ANA has spent too many efforts on CPT [Current Procedural Terminology, used for encoding billable procedures] when it is clear (at least to me) that NCHVS is where the lobbying efforts should be focused so that nursing procedures can eventually be included in the procedures section. Nursing has not been able to crack the inner circle there...

    (from a news group discussing nursing interests and classification)
    In this section we examine a social world and its attempts to create convergence between information systems (specifically the classification of work and its embodiment in accounting systems and research protocols) and clinical practice.

    In their Nursing Interventions Classification (NIC), a group of nursing administrators are attempting to produce a standard list of everything that nurses do in the course of their working day - from changing a bedpan to giving an injection to telling a joke (McCloskey and Bulechek, 1996). Why do they need a classification system? NIC's developers argue that nursing has traditionally been literally `anomic' (without laws): in the sense that no body of knowledge has been built up which can be moved from place to place (cf. Turnbull, 1997 and Latour, 1987 on transporting local knowledge). Julius Roth (1963) makes this point with respect to tuberculosis classification: until a common classification of patient status had been introduced, no patient could carry a single description of their condition with themselves; and it was almost impossible to share information between institutions - to build a community or shared language for comparison purposes.


    The central argument the nursing researchers make is that they should be represented in integrated hospital information systems, from which they are currently absent. They need this for two reasons: first to act as a basis for scientifically demonstrating the value of nursing within hospitals (so that it can be properly `accounted for' and therefore receive proper funding and recognition) and, second, to act as a basis for the development of nursing knowledge: "Documentation of care with standardized classifications allows for the integration of research and practice" (McCloskey, 1996: 14). Each of these goals - institutional and scientific - aims to describe the way things are; each in practice also involves the operation of a principle of convergence - which involves fundamental changes in nursing practice and accounting. Looking at this example, we will see how the information needs of the nurses are fitted to the professional worlds they are trying to integrate with, at the same time as these worlds are themselves modifying information systems so as to represent nursing. What is at stake is the creation of an information world, where the practices of nurses can be transparently represented in both hospital accounting schemes and in comparative nursing research.

    NIC brings many threads of nursing practice together, into one, legitimated form of representation. The creators of the classification have attempted to include the many various practitioners and specialists in nursing, as well as nursing educators and students. It will no longer be enough, argue nursing informaticians, for nurses to act locally within the context of their home institutions. In order for a body of nursing knowledge to be built up, they will have to act the same as other nurses in other institutions. When NIC wins the day, then they will have to perform sets of NIC interventions specifically. This is further supported by the inclusion of NIC in nursing classroom curricula. Thus the information tool (the classification system) will change both nursing practice in the field and nursing training - the users will be disciplined to both act and represent their actions in NIC form.

    Further, the language of NIC will have to fit with the language of other sciences:

    The most frequently focused upon language is that developed by medicine and the numerous natural or physical sciences that contribute to medicine's scientific knowledge base. This language is focused upon most frequently by nurses, because of medicine's long history of dominance in nursing, its relative specificity about physiological phenomena, its apparent measurability, and its familiarity. It is also valued for its social status as a scientific language, used by a politically powerful professional group, physicians. (Kritek in Werely and Lang, 1988]: 25).

    What this has meant is that nurses have, in building their information infrastructure, emulated the successful research apparatus of other disciplines. Nursing work is deliberately made, through the classification, to be subject to scientific testing. The goal is on the one hand an informational one - to create manipulatable data by ensuring comparability between nursing interventions carried out at different local sites. But on the other hand, and crucially for our analysis, it is at the same time an institutional one - to make nursing 'look and feel' like other sciences (and in particular medicine) so as to be able to tap the same set of accountability and accounting information practices that the other sciences have so successfully developed to justify their funding. They are working to make their classification align with medical informatics and medical standardized languages by making alliances and integrating themselves in other formal medical language systems.


    The nurses promoting NIC clearly see that their classification must become integrated with the rest of medical informatics at this higher level. They argue that NIC must also become a part of not only the daily practice of nurses, as well as nursing education, but it must also be accepted by doctors and other health care professionals. They have in front of them the specter of a failed classification system - one which developed without the feedback from the wider setting which characterizes convergence. In the case of a classification of nursing diagnoses (North American Nursing Diagnosis Association), "When other professionals hear us use NANDA terms, they about laugh us out of the hall because some of them [NANDA terms] don't seem to be directly related to what we are talking about at all!" (Interview with NIC developer, 1993).


    Nursing interventions need to be countable in the same accounts package as medical interventions or pharmaceutical interventions. Nursing must be described in the same ways as other activities in the medical setting:

    Without a unified language system that is integrated into other health care standard initiatives, the nursing profession will not be able to use the standard language developed in collaborative health care initiatives. Nor will the unified language be capable of being integrated with other developments in clinical practice, reimbursement formulae, or case simulation models without a collaborative health care initiative (McCormick, 1988: 176).


    In general:

    It is not enough for nurses to accurately record what they do- they must record it in a language acceptable to physicians and administrators, and productive of a new, higher status. (Berg and Bowker, 1996).


    The feedback process between information needs, classification schemes, and the wider professional communities with which nursing seeks to align, produces a trajectory whose end result is both a nursing informatics that can integrate with wider medical informatics and a nursing practice that will be sensibly altered to permit this integration. The convergence here is between information tracking systems and the social worlds of practicing nurses -- a new kind of nursing information world, more broadly defined and strategically situated.


    As in the case of the individual academic discussed above, then, professional communities too find themselves struggling with transparency. Here the issue is not so much one of membership, but of aligning information systems and practice in such a way that the emerging information world is strategically situated amongst other information worlds.


    The fledgling academic must achieve transparency through learning and professional socialization; the social world here achieves its transparency through a series of translations and negotiations with its neighbors. In neither case is the picture entirely deterministic - both the individual and the social world have the ability to reconfigure the world(s) that they are adjusting to: particularly as they become more mature. Howard Becker’s analysis in Art Worlds (1982) notes the difference between a naive and a maverick artist as precisely constituting this relationship to the social world. Naive artists are not in dialogue with the reigning artistic conventions; mavericks play with them and subvert them. Thus we saw the academic becoming a reviewer or journal editor; we can see nursing informaticians developing positions of influence within the wider informatics field.


    A curious erasure of the learning and negotiating process often happens over time. In both the individual and the community case there is frequently a belief that all that is really happening is that the original questions and goals are now being served by the information worlds that they are now linked with. The achievement of transparency becomes naturalized -- this is just how the world is. For the mature researcher, it is easy to forget barriers and blockages facing the newcomer; as social worlds converge with wider-scale information systems, the categories come to seem entirely natural, rather than negotiated (Bowker and Star, 1994; Becker, 1994). In each case the operation of the feedback process of convergence has underwritten the mutual fitting of a changed individual/community and a new information world.


    In the following section, we will examine what convergence looks like when social worlds are stitched together and overlap in yet larger information systems. The International Classification of Disease (ICD) was first developed over one hundred years ago, and today is used by such various groups as epidemiologists in national health institutions and individual health practitioners in small villages around the world.


  7. Convergence and Infrastructure: Linking Multiple Social Worlds with Wider Scale Information Systems

    Alain Desrosiéres (1988) has shown beautifully how census breakdowns of the populations of German, France and England have remained closely tied to the history of work, trade unions and government intervention in those countries. We suggest that as the ICD "naturally" becomes more universally applicable, this is partly the result of the hidden spread of western values through the application of our own bureaucratic techniques. These techniques appear rational and general to us, but when looked at in detail prove highly contingent.

    (Bowker and Star, 1994: 202)


    We will now move further out from a focused social world, nursing, and sketch how multiple overlapping information worlds are generated and supported through wider scale convergence. At this overview level, we are concerned with the development of very large scale information systems which are shared by many professions - for example census data shared by economists, policy analysts, government agencies; or biodiversity databases shared by environmentalists, biologists, community groups and so forth. What emerges is that the same principle of mutual feedback between the information artifacts and the social worlds occurs: the wider collection of social worlds gets changed in just the same way as the individual research or the professional community did in the process of the development of a new information world. Transparency at this level begins to merge with the very notion of infrastructure -- that which is ready-to-hand across a wide scale of operation. Of course, it is a problematic notion, given the complexity of transparencies in such heterogeneous mergers (Star and Ruhleder, 1996, discuss this at some length).


    Convergence here fits together the representation of knowledge and the phenomena which are being represented, through highly distributed knowledge systems. Consider for example job classifications in Europe: there are large differences, as Desrosiéres (1988) points out, between "professional" in England, and "cadre" in France. Each classification system has been incorporated into a number of instruments of government (such as the census) and into a series of organizational decisions (for example the organization of trade unions). Each has spawned its own particular varieties of statistical knowledge: epidemiologists track stress, say, by socio-professional category; sociologists structure their surveys using those same categories; government policy makers then adopt the findings to target specific groups. While there is no "in principle'"difference in social and political organization in England and France, two different information worlds have historically reified two incommensurable subsets of the population: and these subsets have become ever more real and identifiable subsets as a result of this convergence. Is stress differently distributed in England and France? The question becomes virtually inseparable from the structuring of the information artifacts as well as from the practices of the different social worlds employing them. When they converge within nations, two different transparencies may emerge between countries. If they are standardized across national boundaries, the transparency will extend further.


    To explore this relationship between knowledge and infrastructure further, we will turn our attention briefly to the International Classification of Disease, tenth edition. The ICD is one of a series of international classification systems that were developed in the late nineteenth century - these ranged from classifications of labor to freight packages to ears. (The shape of ears, originally thought to be linked to types and degrees of criminality, became enshrined in the classification of people in the collection of vital statistics and immigration.) Each scheme involved the creation of a new international governing bodies together with the development of an information world in which those bodies could work. And many of the new classifications became strongly entrenched in information gathering practices - for example, Bertillon's classification of ears is still today the basis for the 'three quarter' photograph displayed on US green cards.


    In the case of the ICD, the series of overlapping communities that needed access to the same information world (in order to co-ordinate their own professional activities) included - but was by no means restricted to - the following:

    Epidemiologists -- who needed comparable data on diseases in different countries in order to discern patterns in outbreaks and thence deduce causes;

    Public health officials -- who needed to be able to patrol the boundaries of the nation and keep out infectious immigrants; as well as monitor health conditions at home;

    Health insurance companies -- who needed to be able to file data on diseases using doctors for information collection purposes (and so needing to use a classification system that doctors would understand)

    Census officials -- who needed to create comparable international data sets in order to serve the needs of the multiple professions that drew on the census.


    No one of these communities had priority in the development of the ICD.


    To the contrary, the final classification as it has developed has entrenched flaws from every perspective. Let us consider two examples. First is the number of diseases that are coded for. In order to get doctors (or, in outlying areas, lay personnel) to fill in death certificates using the code, there must be a restricted number of disease categories. This feature of medical practice is a blight for epidemiologists, for whom a nuance in a particular classification might well make a large difference in the context of a specific inquiry. Through this case, however, we can see the operation of the same principle of convergence that we have observed above. Over the past 100 years, doctors have been trained in medical school into use of the ICD throughout their practice - they use it in accounting forms, personal histories and death certificates. They have fitted their practice so that they can support the use of a developing ICD (which has grown consistently in scope and size and so become more epidemiologically useful). Our second example is the use of the ICD in the Third World. There is frequently a complaint in developing countries that the ICD is not relevant to their interests - when people are dying before they can be `accurately' diagnosed, what purpose is there in propagating the system, they ask. They argue that the ICD does not represent their own diseases. Against this view runs the Foucaldian (1991) observation that it is precisely the ability of governments to count people, things, diseases which is a feature of recent "development."


    The key point for us is that as the ICD gains a representational foothold in developing countries so at the same time (and in pace) does Western medical practice, which itself is tied to the ICD; and so ultimately does the ICD become the most apt descriptor of disease. The disciplining of representational practice onto a single classification scheme is tied directly to the spread of allopathic medicine - and where allopathic medicine is fully practiced, the ICD provides the best description of diseases. A continuing problem for the WHO comes in trying to integrate traditional systems of medicine -- for example, acupuncture has no category of disease per se, but a model of imbalances of forces in the body. It is extremely difficult to map the one onto the other in any simple sense without violating someone’s ontology. The ICD comes as a package. It would be a mistake to see it as 'just' a classification system that can be applied to as is to the Third World: it marks the convergence of a social world and an extended system of information artifacts, including death certificates, software, and the epidemiological record keeping of the WHO.


    Thus the development of the ICD can be seen as the mutual inter-adjustment of multiple information worlds in tune with the mutual inter-adjustment of multiple social worlds. Given the complexity of the processes involved, the resultant large scale information infrastructure is of course not seamless. However, it operates as efficiently and reliably as it doubtless does precisely because of the continual feedback process of convergence. In the case of large scale infrastructure, attachment to the installed base, affiliation with multiple communities, and compatibility with other systems of knowledge is necessary for success (Star and Ruhleder, 1996). As Bowker and Star (1994) have pointed out, the ICD's success, as seen by its almost universal adaptation and use, is based as much on the general westernization, industrialization, and bureaucratization of the world as the classification itself. The ICD is a central point in an international system of interlocking classifications and standardized medical languages as well as a key structuring agent in medical linguistics generally.


  9. The Ontological Status of Convergence

    We have looked briefly at three examples of convergence, from the perspective of the individual and information retrieval, a professional social world and its codification practices, and large scale information infrastructure involving multiple social worlds. We have witnessed some similar processes in each case, mediated by developing information infrastructure: the world comes to look as if the convergent description of it is accurate, and "natural." The degree of transparency achieved in each case depends on location, strategy and politics. As one scales up, transparency becomes more subject to contention arising from the heterogeneity of the participating social worlds; once achieved, it acquires considerable inertia and even coercive power. This finding is akin to Star's handling of anomalies in Regions of the Mind (1989) and Latour's analysis of scientific networks in Science in Action (1987), and at the level of large scale control, with Foucault’s original sense of discipline.

    At the same time, the co-creation aspect of convergence is central to our argument here -- while transparency may seem wholly natural, it is never inevitable or universal. In this we follow Yates and Orlikowski’s use of Giddens’ structuration model: structure and agency are dialectical co-creations (Yates, 1989; Orlikowski, 1991; Orlikowski and Yates, 1994; Davies and Mitchell 1994 use a similar approach).

    There is another powerful sociological tradition that has looked at similar processes as a fundamental to social stability: the functionalist tradition. For the purposes of dealing with the intertwining of information and social worlds, we can trace this tradition back to the path-breaking essay by Durkheim and Mauss (1963) through Evans-Pritchard to Mary Douglas' own innovative (1986) work on `how institutions think'. Durkheim and Mauss argued that at the root of primitive classifications is a classification of social relationships: these are then projected onto the world at large and read back into social discourse through the intermediary of myth. Thus for Durkheim and Mauss it is not surprising that stories about stars, the coyote, the origin of fire, and so on `converge' on the same moral: since they are in fact the same social story told in different `registers' (to borrow Levi-Strauss' (1973) phrase).


    Mary Douglas (1986) felt no qualms about bringing the lessons of Durkheim and Mauss right into the center of our own social organization of information. Her position is simply that: "How a system of knowledge gets off the ground is the same as the problem of how any collective good is created" (45). Social institutions, she explains, are about reducing entropy, and:

    the incipient institution needs some stabilizing principle to stop its premature demise. That stabilizing principle is the naturalization of social classifications. There needs to be an analogy by which the formal structure of a crucial set of social relations is found in the physical world, or in the supernatural world, or in eternity, anywhere, so long as it is not seen as a socially contrived arrangement. When the analogy is applied back and forth from one set of social relations to another, and from these back to nature, its recurring formal structure becomes easily recognized and endowed with self-validating truth. (48).5

    Douglas is at her strongest when she unflinchingly argues that basic classificatory judgments are social in the sense that they are created, maintained and policed by institutions: "Nothing else but institutions can define sameness. Similarity is an institution" (55). The relevance of her position in our discussion of convergence is clear: she is adopting the position - close to our own - that the development of a single information world in any particular setting is not about the enlightened discovery of the truth about the world but is rather a statement about the consolidation of social institutions and information systems. For us, it is this consolidation which explains some people's transparent use of information infrastructure, and others' inability or lack of desire to do so.


    Where Douglas, like Durkheim and Mauss, is weak is that she reifies `the social’ and sets it apart from technology (or information systems). She falls prey to Latour's (1993) arguments against all great divides which engender determinism. On the one hand, she defines society as apart from the existence of things-in-the-world; but then for a causal argument, she has trouble setting `society' as determinant on the one side of an equation and `the world' or `knowledge’ on the other. There has never been room for co-construction in this form of functionalist argument. A related weakness is that, as with many functionalist positions, she makes stability too much the norm. What needs explaining is not only information convergence within a pocket of stability but also the lack of convergence (and the preservation of an outsider status) within multiple overlapping institutions. As sociologist Everett Hughes reminds us, we need always to keep in mind that "it might have been otherwise." Behind functionalism is the position that we all belong to one professional world, one social class, one ethnicity and so forth - and that the alignment of these memberships into a coherent social institution acts as unproblematic guarantor for information convergence and thus transparency. Those for whom this knowledge is not transparent are deviants. (Illustrative of this point is functionalist sociologist Robert Merton’s categorization of scientific discovery (novelty) as a form of deviance! (1973))


    For all these criticisms, our understanding of convergence is enhanced by Durkheim, Mauss, and Douglas' work. It reminds us that convergence is a result of the consolidation of social institutions. The information science literature also lends studies of the multiple paths and means that information converges along- such as colleague networks, personal collections, community practices (cf. Lancaster, 1995; Bishop and Star, 1996; Pinelli, 1991; Aloni, 1985; Garvey and Griffith, 1980). Network analyses have shown the importance of interconnections (Haythornthwaite, 1996; Wasserman, 1994; Scott, 1992), and the actor network perspective (Latour, 1993; Callon, 1989; Latour, 1987; Callon, Law, Rip, 1986) has demonstrated the importance of linking the non-humans to the previously human centered networks. To these, we are adding that convergence is a process in which status, cultural and community practices, resources, experience, and information infrastructure work together to produce transparency within an information world. Convergence is fully situated and is not universal, nor is it exclusive. In times of change, it also has a certain fragility. It is a situation, for the most part, of privilege.


    From a conservative point of view, such "obviousness" or naturalness would mean successful integration with the normative structure of social life. We do not wish to recapitulate the recent history of American sociology -- people belong to many communities of practice, and participate in many information worlds. Success is relative and partial. For any given information world, the degree of convergence is not a matter of "proper" socialization or internalization of the correct norms. Rather, people balance conflicting requirements from different information worlds; they may have good reasons (such as more important memberships) for NOT undertaking pathways to membership or convergence (Becker, 1967); or they may be excluded on other grounds such as gender, age or race. So the situation of transparent, obvious, easy information retrieval on a routine basis for work and play is in some ways a special case, and in any event, in pure form is an analytic exercise.


    1. Conceptualizing Lack of Convergence

    Because social worlds and information artifacts often do not converge, it is important to conceptualize failures of transparency as well as the successful formation of information worlds. Again, turning to the foundational work of Durkheim in sociology, the term "anomie" seems suggestive of those situations where there is a failure to converge.

    Anomie in common parlance is thought to mean something like "at loose ends." The Oxford English Dictionary lists a range of definitions, beginning with a disregard of divine law, through the 19th and 20th century sociological terms meaning an absence of accepted social standards or values. Most sociologists associate the term with Durkheim, who used the concept to speak of the ways in which an individual's actions are matched, or integrated, with a system of social norms and practices. In his famous Suicide (1897), Durkheim accounts for the lower suicide rate among Catholics by noting the stricter, more binding and more clearly specified sets of norms and practices affecting individuals, by contrast with Protestants. When such a map for behavior is present, he argued, the person is less likely to experience the angst of normlessness, less likely to internalize a situation with no structure, and therefore less likely to kill themselves. Durkheim’s work has been widely criticized on empirical and methodological grounds. As well, we have the same criticisms of the functionalist notion of mismatch as of convergence.

    At the same time, the word does capture the sense of what happens with a profound mismatch between social world and information artifacts. Certainly the dispirited undergraduates we interviewed about library usage were experiencing anomie in the commonsense meaning -- at loose ends.


    Durkheim also formally posited anomie as a mismatch, not simply as the absence of norms. Thus, a society with too much rigidity and little individual discretion could also produce a kind of anomie, a mismatch between individual circumstances and larger social mores. Thus, fatalistic suicide arises when a person is too rule-governed, when there is, in the words of Ricoeur (1988), no free horizon of expectation.


    Information anomie may arise when there is an absence of social standards, guidance or values -- accepted as useful by those successfully working with the system. It may also arise when the system is overly rigid, allowing for no tailoring or customizing to individual needs in a place of practice.


    The term anomie can carry a lot of undesirable historical freight, such as the notion that there is one right system and anomie is a deviant falling away from that perfect recipe; similarly, the notion that there are "good users" and "bad users," the latter of which are anomic (or deviant). Again, we do not wish to over emphasize any centralized, too- rational or functional a priori visions of any given system. Rather, the emphasis here is on the notion of mismatch, a lack of convergence and on the senses of frustration, being lost, even desperation which may result when there are mismatches between a system and those who work with it. At higher levels of scale, it may mean the exclusion or invisibility of whole social worlds or classes of people -- institutionalized opacity, rather than transparency (Star and Strauss, in press).


  11. Related Work in Library and Information Science

    Library and information science research has dealt with most of the components or itemized examinations of the contents of the information worlds to which we are referring. Descriptions of how these components work together, and analysis of longer term interaction of information resources and social worlds is becoming more important in the field, as social informatics emerges as an increasingly important area. In this section, we will describe how our model of convergence can be used to situate the social informatics dimension of LIS research.


    Many important studies in this field are centered on items such as libraries, journals, and other formal information systems. They are grounded in a model of a "user" first conceiving of an information need, then going out and getting that information (Lancaster, 1995; Bates, 1989). The role of the information scientist (or intermediary) is in between those two points, interpreting the information need, creating and utilizing document surrogates, using or building a system to go and retrieve an item that will satisfy the user's need for information. While bringing attention to the often overlooked role of the user, research usually questions the circumstances influencing the utility and effectiveness of different forms of document and system use, representation, and retrieval under different circumstances (Covi, 1996, Bates, 1989).


    The view these studies take is often segmented along different axes, such as exploring the use of specific document forms, or the use of single systems among particular communities (Lancaster, 1995; Baym, 1995; Pinelli, 1991; Bizot, Smith, and Hill, 1991; Aloni, 1985; Warden, 1981; Crane, 1972; and others.) Covi (1996, in press) critiques the more segmented view as a "closed rational, or bounded database perspective." Specifically, she defines this as a perspective that reviews the use of a system as if that system were not situated in other resources on and off- line as well as community norms and other social practices. She calls for more use of the open and naturalistic inquiry which has recently begun to appear. This integrated, naturalistic view as a departure for LIS research is less common, although growing in interest and importance in the LIS community (Bishop and Star, 1996).


    We seek in the model presented in this paper to include and add to the scope and perspective of current understanding of information use. Indeed, they are complementary and supportive to what we are saying; for example, the many studies that have demonstrated the importance of colleague interaction for information exchange (Lancaster, 1995; Garvey and Griffith, 1980; Crane, 1972; Wolek and Griffith, 1980). Others point to the significance of personal collections and resources (Soper, 1976, Chatman, 1991, 1992). Also, the importance of physical proximity in use of resources has been well documented for some time (Waples, 1932, Hertz and Rubenstein, 1954; Rosenberg, 1966; Palmer, 1981; Pinelli, 1991; Gould and Pearce, 1991). All of these point to the importance of and describe the significance of various elements in information worlds, which we are then drawing together.


    Further complementary research is found in the strong tradition of studies of communication patterns examining the social networks of individuals. This particular thread of social network analysis "is an approach and set of techniques used to study the exchange of resources among actors" (Haythornthwaite, 1996: 1). It traces exchanges between actors and finds patterns of relationships of exchange. These relationships are synthesized into networks. These networks can be examined for patterns of exchange of information. Relationships between "nodes" are measured and evaluated for several properties, including strength, intensity, direction, and content (Wasserman, 1994; Scott, 1992). Hesse, Sproull, Kiesler and Wellman have all indicated how computer and social networks interpenetrate in creating these relationships (Hesse, et al., 1993; Finholt and Sproull, 1990; Sproull and Kiesler, 1991). The development of the strength of these ties points to the communicative processes associated with convergence; further research into the similarities and differences between social networks and social worlds would further specify these links.


    The advent of very wide scale networked information systems has opened many new discussions about communities and information. The possibility of information acquisition, retrieval, and tailoring at the community level (including finding communities on the Internet, designing for support for communities rather than individuals) is an exciting one. Its success depends in part on the ability to design to, and account for, the multiple levels discussed here. Not surprisingly, we know the most about the individual level of information retrieval, access, and use; some about the community or occupational level, particularly in the areas of scientific communication; and we are just beginning to learn about infrastructural convergence as a problem in design. Hanseth and Monteiro (1996; Hanseth, 1996; Hanseth, Monteiro and Hatling, 1996) have begun a series of investigations into information infrastructure and standardization that addresses many of these issues. They argue in particular that the tensions between standardization and flexibility in the creation of information infrastructure is a key design factor, and that analysis of institutions supporting the creations of standards and standardized classifications is key to successful design in the networked computing environment. This point receives support in Bowker and Star (1994; 1997) and Star and Ruhleder (1996).


    Our goal is to tie these many factors together into a holistic and dynamic model. As other researchers (Bates, 1989; Covi, 1996, Covi and Kling, in press) have pointed out, search and retrieval as a model is too simplistic, rational, and mechanical to describe the dynamics of how people, groups, infrastructure come together as resources. They are calling for study of the use of digital libraries in terms of open natural systems and a social world perspective; looking not only at the use of the system in terms of a bounded world and rational actors, but rather at the physical and temporal workplace, the professional field, and the occupational niche of the user (Covi, 1996; Covi and Kling, in press). To this we are adding the idea that not only are these variables crucial in understanding the use and usability of digital library systems or other information systems, but also the amount of convergence or the lack of convergence a person, community, or multiple communities are experiencing. We are concerned not with cataloging the contents of information worlds or with describing use of individual units therein, but rather how all of these interact dynamically.


    But in addition, we are also sketching out a larger framework. Search and retrieval are not compartmentalized tasks, but rather, are products of social processes and relationships. How an individual or a community or multiple communities are linked is the result of the situatedness of each. Information is not a static product that is "out there" waiting to be found, but is rather a construct of particular situations at specific points in time. Convergence is a process by which things come together through social process such as socialization into a profession or the formation of a community with the co-construction of information artifacts at different levels of scale.


  13. Conclusion

We know that there is no universal scheme by which information may be made easily available to everyone. Some systems of information organization are widely adopted, such as that of the Library of Congress or the telephone book. However, as Berman (1981) illustrates for the Library of Congress Subject Headings, these same headings may carry with them systematic exclusions of points of view of ethnic minorities; they overemphasize Western religions and do not well represent common names for everyday objects. Similar points hold for other attempts to organize and classify the world of knowledge. Current attempts to catalogue and classify the chaos of the internet come nowhere near a universal organizational scheme; if anything, they are proliferating towards "islands of indexing," despite powerful search engines and limitless hypertext.


Rather than perceive this state of affairs nihilistically, we have in this paper seen it as an opportunity to understand something about the relationship between people, history, information, and material and social order. While it is true that no one scheme has yet worked for everyone, some schemes work reasonably well for many people at particular times. If not absolute, there is sometimes relative transparency at all levels of scale. In our work, studying designers and users of information, we see "clear and present differences" between people who are very comfortable within some world of information, who are "plugged in" to a range of sources and are able to find all the information they need easily; and those who are intimidated, confused, and limited by their relationships to that same information. The differences are subtle, and to some degree are hidden by models of individual user needs and concepts such as levels of expertise (as transparency unmodified by time or place).

This has given us some insight into the locale of those on the outside, those experiencing a gap between their social world’s practices and formal information systems. These differences between insiders and outsiders are neither trivial nor easily understood. Sometimes there are differences in career stage which explain relative ease and familiarity - an old hand will more likely move easily through a world of information than a novice, as we have noted. But not always, if that world itself is shifting and changing rapidly, and training situations favor students over professors. Sometimes there are insider/outsider differences of the sort noted by Berman - systematic exclusions of members of a certain group, or such differences in access to materials and infrastructure that something like a digital library is a current impossibility. Sometimes other memberships are more salient (Eckert, 1989).


The greatest issue for the creation of information systems is that these processes are invisible to traditional requirements analysis; they can only be seen through the developmental analysis of information worlds. Convergence often masquerades as naturalness. It is not, but is rather the result of a particular configuration of relationships between information and social order. Convergence is not interoperability, but rather a layering up of solutions and conventions, and standards. Usability is an emergent property which is not currently addressed in this fashion. The reason why some things are obvious to some people and not to others has to do with the connectedness that some people and not others experience and the memberships that some have and not others.


The challenge in understanding and recognizing these differences is to design to them. This does not mean creating "simple" versus "complex" interfaces to systems, but rather recognizing the fundamentally different processes of searching and use germane to different information worlds. Recognizing differences in needs implies building systems that actually do different things. Unfortunately, many systems today do not and often cannot take these different needs into account. The new relationships and points of access being built and developed now benefit only a small group of people at a particular place in their careers. Systems are now being linked with hopes of interoperability without taking into account the trajectories of information on a systemic or infrastructural level. More than ever, the need for interdisciplinary teams for design and use are called for to addressed the multifaceted nature of information -- a social informatics.



Allison, Paul D. and Scott Long. 1990. "Departmental effects on scientific productivity." American Sociological Review 55: 469-479.

Aloni, Michaela. (1985). "Patterns of Information Transfer Among Engineers and Applied Scientists in Complex Organizations," Scientometrics, 8: 279- 300.

Bannon, Liam. (1990) "A Pilgrim’s Progress: From Cognitive Science to Cooperative Design." AI and Society 4: 259-275.

Bates, Marcia J. (1977). "Factors Affecting subject Catalog Search Success," Journal of the American Society for Information Science. 28 (3): 161- 169.

Bates, Marcia J. (1979). "Idea Tactics," Journal of the American Society for Information Science. 30 (5): 280- 289.

Bates, Marcia J. (1989). "The Design of Browsing and Berrypicking Techniques for the On-line Search Interface," Online Review. 13 (5): 407- 424.

Baym, Nancy. (1995). "From Practice to Culture on Usenet," Star, Susan Leigh, (ed.) Ecologies of Knowledge: Work and Politics in Science and Technology. Albany, NY: SUNY Press.

Becker, Howard. (1967). "Whose Side Are We On?" Social Problems 14: 239-247.

Becker, Howard. (1982). Art Worlds. Berkeley, CA.: University of California Press

Becker, Howard. (1994). "Foi por acaso": conceptualizing coincidence. The Sociological Quarterly 35: 183-94.

Berg, Marc. (forthcoming). Rationalizing Medical Work- Decision Support Techniques and Medical Problems. Cambridge, MA: MIT Press.

Berg, Marc and Geoffrey C. Bowker. (1996). "The Multiple Bodies of the Medical Record: Towards a Sociology of an Artifact," to appear in The Sociological Quarterly.

Berman, Sanford. (1981). The Joy of Cataloging: Essays, Letters, Reviews, and Other Explosions. New York City, NY: Oryx Press.

Bishop, Ann P and S. Leigh Star. (1996). "Social Informatics of Digital Library Use and Infrastructure," in M. Williams (ed.), Annual Review of Information Science and Technology (ARIST), Vol. 31. pp. 301- 401. Medford, NJ: Information Today, Inc.

Bizot, Elizabeth, Nancy Smith, and Thomas Hill. (1991). "Use of Electronic Mail in a Research and Development Organization," Advances in the Implementation and Impact of Computer Systems, 1: 65- 92.

Bowker, Geoffrey C. and Susan Leigh Star. (1994). "Knowledge and Infrastructure in International Information Management: Problems of Classification and Coding," in Lisa Bud (ed.)Information Acumen: The Understanding and Use of Knowledge in Modern Business. London: Routledge.

Bowker, Geoffrey C. and Susan Leigh Star. (1997). "How Things (Actor- Net)work: Classification, Magic and the Ubiquity of Standards," Philosophia, in press.

Bowker, Geoffrey C. and Susan Leigh Star. (1997). The Quiet politics of Voice: Classification, Infrastructure and Boundary Objects. Submitted to MIT Press.

Brown, John Seale and Paul Duguid. (1994). Borderline Issues: Social and Material Aspects of Design. Human- Computer Interaction, 9: 3- 36.

Callon, Michel (ed). (1989). "La Science et Ses Reseaux: GenesË et Circulation des Faits Scientifiques," Paris, France: La Découverte, Conseil de l'Europe: Unesco.

Callon, Michel, John Law, Arie Rip (eds). (1986). Mapping the Dynamics of Science and Technology: Sociology of Science in the Real World. Hampshire, England: Macmillan Press.

Campbell, Donald. (1969). "Ethnocentrism of Disciplines and the Fish- Scale Model of Omniscience," in Muzafer Sherif and Carolyn Sherif (eds.), Interdisciplinary Relationships in the Social Sciences. Chicago, IL: Aldine.

Chatman, Elfreda A. (1985). "Information, Mass Media Use and the Working Poor," Library Sciences Research 7: 97- 113.

Chatman, Elfreda A. (1991). "Channels to a Larger Social World: Older Women Staying in Contact With the Great Society," Library Sciences Research 13: 281- 300.

Chatman, Elfreda A. (1992). The Information World of Retired Women. Westport, Connecticut: Greenwood Press.

Covi, Lisa. (1996). "Social Worlds of Knowledge- Work: How Researchers Appropriate Digital Libraries for Scholarly Communication." American Society for Information Science, Midyear Proceedings: 84- 100.

Covi, Lisa and Rob Kling. (in press). "Organizational Dimensions of Effective Digital Library Use: Closed Rational and Open Natural Systems Models." Journal of the American Society for Information Science, in press.

Crane, Diana. (1972). Invisible Colleges: Diffusion of Knowledge in Scientific Communities. Chicago, IL: University of Chicago Press.

Davies, Lynda and Geoff Mitchell. 1994. "The Dual Nature of the Impact of IT on Organizational Transformations." Pp. 243-261 in R. Baskerville, O. Ngwenyama, S. Smithson, and J. DeGross, eds. Transforming Organizations with Information Technology. Amsterdam: North Holland.

Desrosiéres, Alain. (1988). Les Categories Socio- Professionnelles. Paris, France: La Découverte.

Desrosiéres, Alain. (1993). La politique des grands nombres : histoire de la raison statistique. Paris : Editions La Découverte.

Douglas, Mary. (1980). Evans- Pritchard. Brighton, England: Harvester Press.

Douglas, Mary. (1986). How Institutions Think. Syracuse, New York: Syracuse University Press .

Durkheim, Émile. (1897). Suicide: A Study in Sociology. New York City, New York: Free Press.

Durkheim, Émile and Marcel Mauss. (1936). Primitive Classification. Chicago, IL: University of Chicago Press.

Eckert, Penelope. (1989). Jocks and Burnouts: Social Categories and Identity in the High School. New York, NY: Teachers College, Columbia University.

Finholt, Tom and Lee Sproull. 1990. "Electronic Groups at Work," Organization Science 1: 41-64.

Foucault, Michel. (1991). "`Governmentality' in The Foucault Effect," Chapter 4 in Graham Burchell, Colin Gordon, and Peter Miller (eds.) Studies in Governmentality With Two Lectures by and an Interview With Michel Foucault. Chicago, IL: University of Chicago Press.

Garvey, William, and Belver Griffith. (1980). "Scientific Communication: its Role in the Conduct of Research and Creation of Knowledge," Key Papers in Information Science. White Plains, NY: Knowledge and Industry Publications.

Gould, Constance and Karla Pearce. (1991). "Information Needs in the Sciences: An Assessment," Program for Research Information Management. Mountain View, CA: The Research Libraries Group, Inc.

Grudin, Jonathan. (1990) "The Computer Reaches Out: The Historical Continuity of Interface Design." In: Chew, Jane; Whiteside, John, eds. Proceedings of the CHI ‘90 Conference on Human Factors in Computing Systems; 1990, April 1-5, Seattle, WA. New York, NY: ACM Press. Pp. 261-268.

Hanseth, Ole. (1996). "Information Technology as Infrastructure." Ph. D. Thesis, Department of Informatics, Gøteborg University, Sweden. Report 10, Gøteborg (Gothenburg) Studies in Informatics.

Hanseth, Ole and Eric Monteiro. (1996). "Inscribing behavior in information infrastructure standards," In Dahlbom, B., Ljungberg, F., Nulden, U.; Simon, K; Sørensen, C. and Stage, J. (eds.) Proceedings of the 19th Information Systems Research Seminar in Scandinavia, pp. 293-332, Gøteborg, Sweden, August, 1996.

Hanseth, Ole, Eric Monteiro and Morten Hatling. (1996). "Developing information infrastructure: The tension between standardization and flexibility," Science, Technology and Human Values 21 (4): 407- 426.

Haythornthwaite, Caroline. (1996). "Social Network Analysis: An Approach and Technique for the Study of Information Exchange." Presented at ALISE conference, 1996.

Hertz, David and Albert Rubenstein. (1954). Team Research. Boston, MA: Eastern Technical Publications.

Hesse, Bradford W., Lee Sproull, Saa Kiesler and John Walsh. 1993. "Returns to Science: Computer Networks in Oceanography," Communications of the ACM 36: 90-101.

ICD- 10. (1992). International Statistical Classification of Diseases and Related Health Problems. Tenth Revision. 3 Volumes. Geneva, Switzerland: World Health Organization.

King, John and Susan Leigh Star, "Conceptual Foundations for the Development of Organizational Decision Support Systems," Proceedings of the 23rd Hawaiian International Conference on Systems Sciences. Vol. 3 Washington, DC: IEEE Press, pp. 143-151, 1990.

Kling, R. and W. Scacchi. 1982. "The Web of Computing: Computing Technology as Social Organization." Advances in Computers 21: 3-78.

Kritek, P. B. (1988). "Conceptual Considerations, Decision Criteria and Guidelines for the Nursing Minimum Data Set from a Practice Perspective." In Harriet H. Werely and Norma M. Lang (eds.) Identification of the Nursing Minimum Data Set. pp. 22- 33. New York: Springer.

Lancaster, F. W. (1995). "Needs, Demands and Motivations in the Use of Sources of Information," Journal of Information, Communication and Library Science. 1 (3): 3- 19.

Latour, Bruno. (1993). We Have Never Been Modern. Cambridge, Massachusetts: Harvard University Press.

Latour, Bruno. (1987). Science in Action: How to Follow Scientists and Engineers Through Society. Cambridge, Massachusetts: Harvard University Press.

Lave, Jean and Etienne Wenger. (1992). Situated Learning: Legitimate Peripheral Participation. Cambridge: Cambridge University Press.

Lévi-Strauss, Claude. (1973). From Honey to Ashes. New York, NY: Harper and Row.

Linde, Charolette. (1997). Personal Communication. Knowledge Work Workshop, Cleveland, OH, June 19- 21.

McCloskey, Joanne and Gloria Bulechek (eds). (1996). Iowa Intervention Project- Nursing Interventions Classification (NIC), Second Edition. St. Louis, Missouri: Mosby Co.

McCloskey, J. M. (1996). "Standardizing Nursing Language for Computerization," in Mary Mills, Carol Romano, Barbara Heller(eds.) Information Management:in Nursing and Health care. West Dundee, IL: S- N Publications, Inc.

McCormick, Kathleen A. (1988). "A Unified Nursing Language System," in M. J. Ball, K. J. Hannah, U. Gerdin Jelger, H. Peterson (eds.), Nursing Informatics: Where Caring and Technology Meet. New York, New York: Springer.

Merton, Robert. (1973). The Sociology of Science : Theoretical and Empirical Investigations. Chicago : University of Chicago Press.

Orlikowski, Wanda. (1991). "Integrated Information Environment or Matrix of Control? The Contradictory Implications of Information Technology." Accounting, Management and Information Technology. 1: 9-42.

Orlikowski, Wanda and Joanne Yates. (1994). "Genre Repertoire: The Structuring of Communicative Practices in Organizations," Administrative Science Quarterly 39: 541-574.

Palmer, Carole. (1996). "Practices and Conditions of Boundary Crossing Research Work: a Study of Scientists at an Interdisciplinary Institute," Ph.D. Dissertation, University of Illinois, Urbana- Champaign.

Palmer, E. S. (1981). "The Effect of Distance on Public Library Use: A Literature Study," Library Research, 3: 315- 354.

Pinelli, Thomas. (1991). "The Information- Seeking Habits and Practices of Engineers," Science and Technologies Libraries. 11 (3): 5- 25.

Ricoeur, P. (1988). Time and Narrative. Vol. 3. Chicago, IL: Chicago.

Rosenberg, Victor. (1966). "The Application of Psychometric Techniques to Determine the Attitudes of Individuals toward Information Seeking." Masters Thesis, Lehigh University.

Roth, Julius A. (1963). Timetables: Structuring the Passage of Time in Hospital Treatment and Other Careers. Indianapolis, IN: Bobbs- Merrill.

Scott, John. (1992). Social Network Analysis: A Handbook. London: Sage Publications.

Soper, Mary E. (1976). "Characteristics and Use of Personal Collections," Library Quarterly, 46: 397- 415.

Sproull, Lee and Sara Kiesler. (1991). Connections: New Ways of Working in the Networked Organization. Cambridge, MA: MIT Press.

Star, Susan Leigh. (1989). Regions of the Mind: Brain Research and the Quest for Scientific Certainty. Stanford, California: Stanford University Press.

Star, Susan Leigh. (1994). "Misplaced Concretism and Concrete Situations: Feminism, Method and Information Technology," Working paper in the Gender, Nature, Culture Network Working Paper Series, Odense University , Denmark.

Star, Susan Leigh and Karen Ruhleder. (1996). "The Ecology of Infrastructure: Problems in the Implementation of Large- Scale Information Systems," Information Systems Research. 7 (1): 111- 134.

Star, Susan Leigh and Anselm Strauss. In press. "Layers of Silence, Arenas of Voice: The Dialogues between Visible and Invisible Work," Journal of Computer-Supported Cooperative Work.

Strauss, Anselm L. (1978a). Negotiations: Varieties, Contexts, Processes, and Social Order. San Francisco, CA: Jossey-Bass.

Strauss, Anselm. (1978b). "A Social World Perspective." Studies in Symbolic Interaction 1: 119-28.

Swanson, Don R. (1986). "Undiscovered Public Knowledge," The Library Quarterly. 56, 2: 103- 118.

Taylor, Robert S. (1991). "Information Use Environments." in Brenda Dervin and Melvin Voight (eds.) Progress in Communication Science. Norwood, New Jersey: ABLEX Publishing Corp.

Turnbull, David. (1997). On With the Motley: The Contingent Assemblage of Knowledge Spaces. University of Melbourne, Ph.D. Thesis.

Waples, Douglas. (1932). "The Relation of Subject Interests to Actual Reading," Library Quarterly, 2: 42- 70.

Warden, Carolyn. (1981). "User Evaluation of a Corporate Library Online Search Service," Special Libraries. April: 113- 116.

Wasserman, Stanley. (1994). Social Network Analysis: Methods and Applications. Cambridge, England: Cambridge University Press.

Weech, Terry and Herbert Goldhor. (1984). "Reference Clientele and the Reference Transaction in Five Illinois Public Libraries," Library and Information Science Research 6: 21- 42.

Wolek, Francis and Belver Griffith. (1980). "Policy and Informal Communication in Applied Science and Technology," Key Papers in Information Science. White Plains, NY: Knowledge and Industry Publications.

Yates, JoAnne. (1989). Control through Communication: The Rise of System in American Management. Baltimore, MD: Johns Hopkins University Press.