Designing Genres for New Media:
Social, Economic, and Political Contexts

Philip E. Agre
Department of Information Studies
University of California, Los Angeles
Los Angeles, California 90095-1520

This is a chapter in Steve Jones, ed, CyberSociety 2.0: Revisiting CMC and Community, Sage, 1998.

An earlier version appeared in The Network Observer 2(7) and 2(11), 1995.

Please do not quote from this version, which may differ slightly from the version that appears in print.

11300 words.


This chapter is a revised version of a manifesto for a course on conceptual design for new media, taught to communication undergraduates at the University of California, San Diego from 1996 through 1998.

1 Introduction

Portrayals of a digital future are too often monolithic: everything will bedigital, everyone will be wired, all media will converge into one, and the physical world will wither away. This kind of monolithic story is wrong, I think, and particularly unfortunate when it comes to the future of communications media. In fact, perhaps the most distinctive feature of the unfolding digital present is a proliferation of new media and new forms of communicative interaction: the Web, CD-ROM's, economical printing on demand, cellular telephones, messaging pagers, fax machines, MUD's, optical scanners, voice mail, and many other media have become widespread in recent times, and more will be marketed soon. Perhaps these media will undergo a shake-out, leading back to the relatively homogeneous days of yore. But more likely, I think, media will continue to multiply. Everybody's daily life will include a whole ecology of media; some of these will be voluntarily chosen and others will be inescapable parts of life in public spaces and the workplace.

As media proliferate and change, the task of designers becomes more difficult. By "designers" I mean to include everyone -- authors, composers, performers, public speakers, letter-writers, editors, and others -- who make decisions about the format and content of communications media, whether for others' purposes or their own. More indirectly I also mean to include the people -- librarians, publishers, book sellers, programmers, critics, anthologists, and others -- who operate the distribution channels that connect the producers and the users of media products. Designing for media, and particularly for new media where stable conventions have yet to be established, requires many kinds of effort -- research, experimentation, rational choice, iteration of prototypes, and learning from the work of others. Many skills enter into the process. In this chapter I want to focus on a single theme: the design of genres that fit into the activities of the audience one hopes to reach. Design for new media, I want to argue, requires some rational understanding of who is using the materials, what they are doing with them, and how they fit into an overall way of life. Such elaborate ideas about the audience might not have been necessary in the old days, when media were few and their uses evolved slowly. That is not so today, and it is not just possible but crucial for designers to learn what is known about the uses of media and to contribute to such knowledge themselves.

My analysis will be divided into four parts. I will begin by sketching some of the processes by which communities conduct their cognitive lives together. In the second part I will present a framework for media design based on inquiry into the role of genres in people's activities, followed by some examples. Putting these concepts to work will result in a vast space of potential genres and uses of media. The third part, therefore, will describe some of the economic forces that tend to select among these various potentialities, and the fourth part will sketch some of the democratic values that might guide concerned citizens and professionals in shaping the media infrastructures and policies of the future. Nothing is inevitable, not even in technology, and some of the choices we might make about communications media right now are much better than others.

2 How communities think

Every artifact is, in its own way, a collaborative construction of far-flung networks of people. Sometimes this is obvious: a car is built by hundreds or thousands of people organized by markets and hierarchies. But sometimes it is not obvious at all: words, sentences, conversations, speeches, memos, papers, and meetings are likewise "built" by enormous networks of people. Bakhtin (1981) described some of these phenomena as they manifest themselves in literary texts; this article offers some first thoughts on the machinery through which collective voices arise in real life.

Let us define a "community" to be a set of people who occupy analogous locations in social or institutional structures. This is not the ordinary use of the term "community", and it will take a moment to explicate it fully. First some examples. The following are all communities:

The "locations" in these examples vary widely. They are notable for their relationships: virtually every parking lot manager has a community of parkers and a police department to contend with; virtually every Republican candidate has a Democratic opponent to contend with; and so forth. Everyone might belong to a variety of different communities, and these communities can be defined in broader or narrower terms (the community of San Diego residents versus the community of California residents; drivers of Mack fire trucks versus all fire fighters; etc). The locations might correspond to formal institutional titles or they might not; but in every case they correspond to a relatively stable universe of structural relationships, and this is what makes them "locations".

We can readily observe some patterns among these communities. The community members have certain interests in common, as well as some interests that conflict. These shared and conflicting interests are "objective" in the sense that they are imposed by the institutions; this is a distinct from the question of how the people themselves understand their interests in a subjective way.

Another pattern is that the members of a community are frequently in ongoing communication with one another. This is clear enough when the members are routinely brought together into a shared physical space. But it is also true when the members' physical locations are distant. One purpose of clubs, unions, clinic meetings, Friday evening drinking groups, and professional societies is to bring the members of a community together periodically. It is by no means inevitable, however, that the members of a community will interact. Nor is it inevitable how they will interact. I will call these things the community's "forms of association". Observe that a community can exhibit elaborate forms of association without any two of its members ever being in the same place -- they can associate over the phone, through talk radio, through magazines, through the Internet, through the efforts of a small number of outsiders who carry news of one another from place to place, and so on. Numerous modalities of association may be combined in customary ways. Forms of association are contingent -- they could be different than they are -- and they are historical -- they arise through concrete processes that leave their marks. And, of course, they are relational -- they depend in crucial ways on the ensemble of relationships that constitute a given location.

This is very abstract, so let us make it concrete through a day in the life of a manager. Managers the world over have highly developed forms of association. These forms of associatio vary by country, sector, social status, and so forth, but they bear family resemblances due to the basic workings of bureaucracy. Managers live dangerous lives -- in some sense anyway. Credit and blame are constantly being assigned for large, complicated processes over which nobody has full control (Jackall 1988). Decisions must be made that depend on more information than any single individual could ever master, and they must be articulated and defended in terms that orient to a constantly shifting set of political alliances. It is not surprising, then, that managers exhibit a powerful orientation to the experience and thinking of others in their community. University campus parking lot managers are well aware of the practices on other campuses. They maintain an elaborate topography of these campuses; they know which ones are considered to be on the cutting edge of parking lot management. If one campus decides to try setting parking prices according to a market-based allocation mechanism, for example, then all of the others will be watching. This "watching" will be subserved by a variety of mechanisms, many of them well-institutionalized: consultants, newsletters, rumor mills, and so on.

In many ways this is a good system since it permits people to put their heads together. It is what Hutchins (1995) calls "distributed cognition": thinking that is distributed across a network of people rather than just being located in one person's head. If you ask how the University of Walla Walla made its decision about how to charge for parking spaces, you cannot formulate a serious answer without appealing to some collective construct such as the community of university campus parking lot managers. It is worth noting that this picture forces us to revise, or at least amend, the picture of the economics of knowledge in the newly influential work of Hayek (1945). Hayek depicted the economy as a sprawling network of people who know their own local conditions extremely well, all dealing with one another through the arm's-length mediation of the price system. But the reality is more complicated than this. Of course, people are arranged in some kind of social network. But they also put their collective heads together in ways that have more regularity than Hayek's extreme localism can admit. These collective minds do have their economics, and money certainly changes hands in the conferences and newsletter subscriptions that subserve the process, but this whole architecture is not just a sprawling mass, and it is not just a "spontaneous order" of localized market arrangements.

The analysis to this point also makes clear what the majority of Internet discussion groups are really for. Even those of us who use the Internet intensively have been influenced by journalistic representations, which have focused disproportionately on those areas of the net to which journalists can easily gain access -- especially Usenet. The concern with "rudeness on the Internet" derives partly from that bias. What this concern ignores is the thousands of discussion groups for professions and subprofessions and subsubprofessions of all sorts. These forums are where the real action is sociologically. Imagine if every community -- in the sense of the term I am using here: the people who share a certain institutional location -- had its own Internet discussion group. Of course, various factors will influence whether the people in a given community would actually benefit from an Internet discussion group:

Nonetheless, it is worth investigating how we might use technology to support the collective cognition of particular communities. First, however, it will be helpful to examine the physiology of collective cognition in greater detail.

3 The physiology of collective cognition

In my analysis so far, I have made it sound as though everyone in a community plays an equivalent role in the community's collective thinking. But this is rarely the case. The actual division of cognitive labor in a given community will depend on many aspects of its relationships and its history, and these will have to be studied concretely in each case. Nonetheless, some patterns do recur. The particular patterns I have in mind are driven by change in the community's environment and pertain to the role of innovators and leaders. Innovation in professions is often (but not always) surrounded by an ideology according to which brilliant individuals come up with "new ideas" through hard work, innate genius, sparks of creativity, and so forth, so that one might regard it as mysterious that someone else did not come up with the same "new ideas" years or decades or centuries earlier. My experience has been that it rarely works this way. Instead, innovators are people who (as common idioms put it) "see which way things are going" and "get ahead of the curve". As the world changes, everyone in a given community is going to face a common problem. And in practice, particular individuals position themselves as the thought leaders in relation to those changes.

The thought leader's role is to get on top of an issue: see it coming, gather positions and arguments about it, network with people who are relevant to it in various ways, and articulate it in terms that supply useful raw materials for individual community members' own thinking in their own situations. This, of course, is not an easy task, and the social capital that these people accumulate is usually well earned. The conditions that permit individuals to play this thought-leader role vary, and different people bring different strengths and strategies to the job. Networking helps; if you notice the same issue coming up repeatedly in conversations with community members, and if the community members do not realize how common the issue is, then that is an opportunity to do a service for both oneself and the community. Although original thinking on the matter is an advantage, all that is needed to be helpful is to assemble everyone else's thinking in some useful form.

Thought leaders accumulate capital in a variety of ways. Some of these are straightforwardly financial, for example the money they earn for books and magazine articles, as consulting fees, or through grants. But the capital can take other forms. Most especially, community members' willingness to commit resources to expose themselves to a synthesis of thinking on an emerging issue gives the thought leaders an opportunity to expand their professional networks in the course of organizing conference panels and the like. This social capital is then convertible into other forms of capital in a wide variety of ways, usually unforeseen but usually unsurprising. Professional communities have routinized much of this process: whole genres of writing and interaction are devoted to it and whole publications are often devoted to uncompensated articles written by people trying to establish themselves as thought leaders.

It is rare for anybody to be taught the "moves" through which one accumulates capital in these worlds, or the "moves" through which various kinds of capital are converted into one another in the course of a career. People go through whole careers without quite understanding the process, while other people have a highly cultivated instinct for it. Why is this? Part of the reason pertains to social class: if you watched your parents live their lives through the associational forms of distributed cognition through which thought leaders acquire capital, then you will probably grow up with a tacit awareness of the phenomena and a powerful head start in learning the skills.

But it is not all a matter of social class. Some people who did not grow up around successful professionals have good professional-skills mentors in college -- this is one purpose of public higher education. Others succeed in apprenticing themselves to masters of the craft in their jobs (Lave and Wenger 1991). Others get the idea in one world by working through rough analogies to processes of distributed cognition and capital accumulation through thought leadership in worlds with different class structures -- local politics, labor unions, social competition through parties and the like, street gangs and organized crime, civic associations, churches, support groups, lodges, social movement organizing, and so on. My point, though, is that not enough people ever get these things explained to them, and that this is a powerful and remediable force for social inequality.

In sketching the physiology of communities' collective cognition, I have tacitly opposed two extreme models: one in which all community members play the same role, communicating amongst themselves equally and symmetrically, and one in which a thought leader is the sole go-between among all the community members. The reality includes the vast range of associational forms through which community members circulate bits and pieces of thinking among themselves. In Orr's (1996) celebrated studies of photocopier repair technicians, these associational forms involved "war stories" about ugly copier repair problems. Business people engaged in public controversies circulate stories as well, but they do so within a different practice based on public relations; the stories are all crafted to provide support for an agenda of "messages" that the community (having done the necessary solidarity work within its own forums) wishes to get across to particular publics. (A "public", in the language of public relations, is precisely another community that stands in a specific structural relation to one's own: for example, a company's publics might include customers, regulators, neighbors, activists, journalists, and union officials.) I have referred to this circulation of structured interactional "stuff" as an "institutional circuitry" (Agre 1995). This circuitry is often partly professionalized, for example when an industry association sends its members a "manual" of facts and stories and quotes that they can use when articulating an industry perspective in one site of public debate or another. An institution's circuitry is defined by the genres that circulate within it; stories that photocopier repair people tell among themselves sound different from stories that managers tell among themselves because they serve different purposes -- that is, they are located differently in the larger system of institutional relationships.

This discussion of stories should remind us that, in speaking of community "thinking" and "cognition", I have lost sight again of the relational nature of the structural locations that define communities. Everyone lives in a set of institutional locations, and every significant situation that arises is defined (to some degree) in relation to these other locations. This is particularly clear in the case of an industry political voice. The repair technicians, likewise, spend much time discussing how to "fix the customer" as well as how to fix the machine, and this "fixing" is conducted through language -- language that nobody could fashion very well on their own, by pure improvisation.

People are often not aware of the extent to which the associational forms of their communities serve the purpose of fashioning a collective voice. They may not even be aware of the crucial role of these associational forms in gathering words for their own individual use back "home" in their interactions with their familiar environment of structurally related others. The associational forms, after all, probably serve other purposes as well, including plain old relaxation, the chance to "talk through" the feelings brought on by troublesome events, the exchange of mutually interesting facts (for example through "gossip"), and so forth. The fact is, though, that we are all members of communities that possess complex mechanisms for the collective construction of a voice. Our voices are not simply our own. This does not imply that we are all puppets who say what we are told -- such a system would not work anyway. Nor are we conformists who hide behind the average because it is safe -- though some such hiding is often prudent. Nor are we all conspiratists consciously plotting the most expedient utterances to use in manipulating others -- though clearly some of this does go on.

Complex as they sound, the mechanisms I am describing require little more than simple, disorganized self-interest: people trying to deal with their own local situations, discovering that others can provide resources that help with this, associating with them for mutual benefit, and then easing into the genres of interaction and the institutional mechanisms that formalize the process and help it scale up. In practice, of course, we inherit these associational forms and institutions from others -- in other words, we enter a given community's forms of association by being socialized into them. We may permit ourselves to be socialized because we see the cognitive advantages of it, or we might have other reasons for joining in. The bottom line, though, is that a community's institutional circuitry can grow quite complex without anybody ever understanding it, much less designing it.

It is important to discuss these things for many reasons. I have already mentioned one of them: that mastery of many communities' associational forms is unequally distributed, and this inequality helps reproduce other kinds of inequality in society. But another, more fundamental reason pertains to the nature of democracy. The roots of democracy lie in associational forms: people learn solidarity or division through their associational involvements; communities that can manage to think, speak, and act collectively can defend their interests much better than the ones that cannot; people who define their communities of shared interests in narrow ways will fare differently from people who define their interests in broad ways; communities that can form working alliances with other communities based on shared interests will fare better than those that cannot; and so forth. Too often we think of democracy in formal terms, as something that happens every couple of years at the ballot box. But democracy is something that happens all the time in society; it is the everyday process through which people negotiate their relationships with one another. Such negotiations may appear to our untrained eyes -- and the eyes of the law and the economists -- as historyless improvisations between isolated individuals, but they are not. Even if it were possible for isolated individuals to negotiate with one another, considerable advantages will accrue to whichever individual then decides to go off and get involved in a community of people who occupy analogous structural locations in society. The reason for this is obvious: by participating in such a community, an individual gets access to the thinking of many other people -- people who have probably faced similar negotiations already. As a general rule, any community that preaches against this broadly democratic conception of society will be discovered to practice it with terrific vigor. And anybody who can convince you to abandon your associational forms without abandoning their own will have an enormous negotiating advantage over you forever afterward.

4 Framework for media design

A worthy goal for design for new media, then, is to support the collective cognitive processes of particular communities. The principal object of design, I want to suggest, is the genre. I have already indicated some of the role that genres play in the institutional circuitry of a community's collective life. In this section, I propose to formalize the interconnections among these ideas to produce a robust analytical framework that supports the design process.

Let us begin with the concept of a genre -- that is, an expectable form that materials in a given medium might take. Here are some examples of genres:

romance novels
op-ed articles
IRS tax forms
scientific research papers
statistical tables
Romantic poems
"tagging" graffiti
classified advertisements
the blues
business memoranda
street maps
page of results from a Web search engine
conference announcements
corporate financial reports
boxing posters
railway timetables
sales pitches
action-adventure movies
shoot-'em-up video games
Notice several things about this list:

1. Genres can be defined more or less broadly. One may wish to focus specifically on research papers in biology, or on early blues, or on sales pitches for condominium time-shares, or on French companies' financial reports. An advertising campaign, for example, might be regarded as a small genre that inserts a range of elements into a recognizable shared frame.

2. Each genre implies a particular sort of audience and a particular sort of activity (Bazerman 1988). Who the audience actually is and what its members are actually doing are, of course, fairly difficult empirical matters. But romance novels and graffiti and financial reports do fit into people's lives in particular ways. (For romance novels see Radway (1984).)

3. Each genre also implies a relationship between the producer(s) and consumer(s) of the materials in question. The relationship may be a one-to-one personal or professional acquaintanceship, or it might be a one-to-many performer-to-audience interaction, or it might be mediated by institutionalized distribution channels. Interests may conflict, matters may be concealed, money may change hands, persuasion may be intended or disavowed, useful information may be conveyed, reputations may be gained or lost, and so forth. All of these aspects of the relationship will shape both the genre and the activities within which it is used.

4. A genre implies not a single document (or other communicative event) but a stream of them. Even if the "rules" of a given genre are never codified, past instances of each genre create precedents and expectations for the interpretation of subsequent instances, and this may create a pressure for future communications to conform to the pattern established by earlier ones. For this reason and others, genres permit people to seek out "more like that one", and they permit the establishment of efficient, familiar, habitual routines for using the materials.

5. The genre does not, however, fully constrain the ways that instances of it might be used. Financial reports might be read as if they were literary texts, IRS forms might be mined for poetic phrases, blues songs might be sampled to make hip-hop songs, video games might be played as if the goal were to avoid killing anyone, sales pitches might be solicited for use as sociolinguistic data, research papers might be interpreted as business plans, and romance novels might be read by those who hope that the heroine is going to blow off the guy and get a life.

6. Any given way of life will involve the routine use of several genres. Tourism involves guidebooks, menus, street signs, timetables, roadmaps, phrase books, and postcard notes back home. Genres sometimes imply one another, at least in the loose sense that they serve complementary roles in the same kinds of activity.

7. Genres change historically. The changes might be encouraged by regulation, by competition or influence from other genres, from changes in the lives of their users, from shifts to new media, or by the changing purposes of the people who are producing them. The changes might be decided consciously, evolve incrementally, or arise through the "natural selection" of markets and other mechanisms.

I focus on genres because they are the meeting-point between the process of producing media materials and the process of using them. Depending on your purposes, be they commercial or political or personal, you might wish to start your analysis with production or consumption. The point, in either case, is to cultivate an understanding of how the two halves fit together. Building on and summarizing the analysis so far, I want to formalize our understandings of the production and consumption of media materials under five headings: communities, activities, relationships, media, and genres. The abstractions provide a handy schedule of questions to ask in mapping out the relationships between people and media in particular situations. Analysis proceeds by enumerating all of the communities, activities, relationships, media, and genres in a given social world, and inquiring into their workings.

Communities. A community, once again, is a the set of people who occupy a given structural location in an institution or society. A community might have a stronger or weaker sense of itself as a community. It might or might not have its own organization, and might or might not meet as a group. But most communities engage in some degree of collective cognition -- the interactions through which they learn from one another's experiences, set common strategies, develop a shared vocabulary, and evolve a distinctive way of thinking. These interactions might take place through war stories, newsletters, rumors, conference speeches, philosophical tracts, music videos, management consultants, or bards who travel from place to place bearing news.

Activities. The life of every community includes shared forms of activity within a particular institutional logic. The community members do not necessarily follow a rulebook, although they might. The point, rather, is that the commonalities of their lives and goals and surroundings, together with their collective thinking about their situations and futures, tend to lead to similar patterns of activity. These include the activities through which particular kinds of media are used, but they include much more. In particular, it does not suffice to identify an activity such as "reading" without asking how this "reading" is part of some larger pattern of social practice. For example, students studying for an exam may apply procedures of "reading" that chop a book's contents into discrete, memorizable facts according to an economic calculation of which ones are likely to count for how much on the exam. This might be contrasted with the reading engaged in by American men reading the post-Vietnam war fiction that was inspired by characters like Rambo (Gibson 1994), with the way that Washington insiders read the newspaper to assess how yesterday's action is going to play back home, or with the business people who read the books of business gurus while asking "what concretely does this and this and this mean in the context of my own industry and firm?". Each kind of "reading" reflects a perfectly legitimate way of using a text within a particular system of culturally and institutionally organized practices (Scribner and Cole 1981). Note that "activities" include both the physical actions (sitting, writing, talking, looking, turning pages, pushing buttons, etc) and the cognitive and emotional processes (identifying with characters, wondering what the professor thinks is important, catching the allusions, etc). The genre needs to "fit" with the whole complex of "external" and "internal" aspects of the activity.

Relationships. The members of a community share a social location because they share relationships to people in other, adjacent locations. Thus temporary employees have relationships to temp agency managers, managers of firms that contract for temps, employees of those firms who have permanent status, and so on. Family farmers must contend with bankers, shippers, extension workers, their own family members, and so on. The lives of people in communities are similar largely because of the similarities of their relationships, and much of a community's shared thinking is concerned with these relationships. Farmers chat amongst themselves about their dealings with bankers, managers meet to discuss their dealings with the people they supervise, voters listen to poll numbers about their fellow voters' supposed views on the candidates, sales people read books by other sales people about selling things to buyers, and so on. Many of the characteristic activities of a community either directly involve these relationships (asking for a loan, writing a report, casting a vote, holding a meeting, and so on) or are heavily influenced by them (acquiring skills, gathering ammo, making oneself presentable, thinking about analogous relationships in others' lives, and so on). Relationships among particular individuals or institutions have life cycles: employment contracts and family relationships, for example, tend to pass through a more or less expectable series of stages within a given society, each with its characteristic issues and forms of activity.

Media. Media are the specific means of communication: telephone, television, CD-ROM's, video tapes, magazines, books, face-to-face conversation, drums, chalkboards, billboards, radio, clothing, and so on. People use media in activities, and the technical affordances of each medium conditions how it can be used. For example, it is difficult to carry a VHS playback system, it is painful to read a long text on a computer screen, radio is much for drivers than television, overhead transparencies can be projected better onto whiteboards than chalkboards, e-mail requires net access, face-to-face conversation requires travel, and so on. But media should not be confused with genres: radio supports both Top 40 programs and call-in talk shows, a magazine usually contains a stable mix of several genres among its contents, and the genres of face-to-face conversation include performance evaluations, party smalltalk, paranoid harangues, and accounts of one's research interests at a conference.

Genres. A genre, again, is a relatively stable, expectable form of communication. Genres are addressed to particular communities and fit into particular activities in the lives of that community's members. Of course, a given genre might be addressed to several different purposes simultaneously, or even to several different communities, but it stands to reason that a genre cannot be too many things to too many communities without diluting its usefulness for any one of them. It is probably best to identify a genre with a particular medium: a folk song goes through important changes in its transition from live performance to audio recording to music video. A novel might not change its words in the transition from paper to CD-ROM, but nobody really knows whether anyone has any use for a novel on a CD-ROM, or whether CD-ROM's need new genres that can participate in the activities for which the CD-ROM medium can actually be useful to the members of a particular community. It helps to think of a genre in historical terms as the product of an ongoing process of coevolution between its producers and consumers. Genres are effectively codesigned with forms of activity, even if this codesign process might be unconscious, haphazard, or even the result of conflict between parties with differing interests or worldviews. In particular, every genre implies a distinctive constellation of relationships: it is supposed to be useful to members of a given community, in activities whose forms and purposes are heavily influenced by relationships with the members of particular other communities.

I have sketched, then, an analytical framework consisting of communities, activities, relationships, media, and genres. My purpose is not to give precise abstract definitions to each of these terms. Instead, the framework is supposed to be useful in making sense of particular cases, whether for understanding what people are already doing with genres or for designing new genres. It can be highly illuminating to map out all of the communities, activities, relationships, media, and genres in a given environment. One might start this process anywhere one likes, for example with a community one hopes to assist or a medium for which one hopes to design new materials. I find it useful to start with particular genres. Having conducted an analysis of a community's existing genres and their place in its activities, it becomes possible to reason rationally about what sort of systems might usefully be designed. This approach resembles contextual design (Beyer and Holtzblatt 1997) and soft systems methodology (Checkland and Scholes 1990) in its wide-angle institutional analysis. It is relatively distinctive in its focus on collective processes of cognition, but its procedures for translating analysis into concrete design proposals are not as fully developed.

5 Some examples

When designing genres for new media, the slogan is: "do more". Pick a community, explore how existing genres fit into existing activities and relationships, and then consider how a new genre might "do more" for the people than the ones they already use. The new genre might, for example, be designed to ease certain functions (like searching or sorting or comparing or pooling group efforts) that the people now perform laboriously for themselves, or that they rarely perform because it is so difficult. For example, if you are working with reporters who must routinely produce documents that draw together information from several different sources, then you can provide the reporters with documents that draw together as many of those sources as you have access to. These documents would not simply dump the information in a pile, but would arrange it in a rational, intelligible form that creates and satisfies a stable set of expectations. I will return to the broader meaning of doing more for communities later on, but here let us consider some examples.

An example that arose during a recent workshop on these issues is the genre of art indexes. These are reference works connecting works of art to the authors who produced them. They are found on paper, but mainly today they are found on CD-ROM's. The communities that employ them consist mostly of students and scholars, and the relationships of these communities include teachers, critics, the artists themselves, the public of art enthusiasts, curators, and scholars in related fields such as literature. The activities that community members engage in include writing papers (which may be usefully decomposed into a variety of other activities), conducting seminars, and presenting talks. And the other genres produced or consumed in these activities include research papers, scholarly books of art criticism, student term papers, other reference works, class presentations, popular articles, and so on.

Given this background, it is possible to reason about how the art index genre might evolve. More detailed information might be required, for example what questions someone writing a research paper has in mind when opening an art index, what other questions they have in mind at other times, what later uses are typically made of the facts discovered in the index, how particular works of art are employed as examples in classroom teaching and the apprenticeship process of seminars, and so forth. One could spend a lifetime exploring these questions, but even a little such exploration will quickly provide the raw material for brainstorms about other applications of the genre or other genres that might fit into the activities of the relevant communities. For example, what kinds of reference materials might be invented to support the social processes of seminars? The media would probably need to be located in the seminar room itself, though it could have remote connections elsewhere. Perhaps it would be useful to brief specialized librarians ahead of time on a seminar topic so that relevant materials could be placed in a menu. Perhaps it would be useful to have a genre of visual presentations to support compare-and-contrast types of reasoning in seminar settings. And so on.

Evaluating such proposals is obviously not simple. The only real test of their practicability is to try them, hopefully through iterative prototyping (Bjerknes, Ehn, and Kyng 1987; Schuler and Namioka 1993). The design process itself will presumably lead to fresh discoveries about the real nature of the relevant communities, activities, and relationships, and it might even change them. Any such change will not be "caused" by the new genres, at least not in any simple sense. The changes will express latent potentials in the local social system, and they will be influenced heavily by the participants' own (shared or conflicting) understandings of the situation. The changes might settle into a new equilibrium, with genres once again fitted to activities that express relationships between communities. Alternatively the changes might continue, fueled by the social system's internal dynamics or by exogenous factors, including further innovations in media and genres of communication.

As another example, let us consider the design process involved getting a particular organization "on the Web" by creating some prototype Web pages. In my experience, most organizations try to jump directly to layout and graphics and bullets and hyperlinks, steering by an unarticulated sense of what they "like" without thinking through the issues in a strategic way. The relevant questions include:

These questions will have very different answers for different purposes -- that is the whole point. Some of the answers might be unknown, or they might be uneven across a given community, or they might change. Having at least sketched the answers to them, one is in a position to start designing and prototyping pages. The next step might be to sit down with some representatives of each user community, show them the pages, and get them to talk about their activities and the role of various media and genres within those activities.

6 Economic considerations

Discussions of new media are often framed in terms of "where things are going". The idea is to predict the future and then to accommodate oneself to it, hopefully to maximum advantage. This kind of reasoning leaves a great deal out. The future is not a deterministic outcome of a mechanical procedure; it is a human choice whose outcome may be constrained and biased but is not settled in advance. Only when we believe we have choices do we start articulating our values and figuring out how they apply to the situation at hand. Economic considerations help in understanding the practicalities of these choices, including the choices that other people are likely to make. They are only one part of the larger picture -- or at least they ought to be. Nonetheless, the vast range of potential applications of new media make the choices exceptionally difficult, and it will be helpful to take a broad range of considerations routinely into account during the design process. In this section I am going to describe some economic concepts that can influence the design of genres in new media -- or of genres that address new situations using old media.

A genre is a pattern, not a single document or event, and it implies a steady flow of materials that can play a definite role in the activities of some community. The economics of genres are the economics of this flow and these activities. Here are some issues to consider, some of which apply more directly to genres and others of which apply more directly to the media in which they are realized:

Fixed costs of distribution. One force for concentration in industry is the overhead involved in creating a network of distribution channels. Since this overhead must be recovered through sales of the stuff that passes through the channels, competition makes it necessary to fill the channels to capacity. Fixed costs of distribution include brand awareness through advertising, creating and updating policies about personnel and customer relationships, facilities and course materials for training personnel, product design costs, capital assets such as storefronts and vending machines and trucks, and so on. Newspapers have high fixed costs of distribution.

Marginal cost of distribution. Once the fixed overhead costs have been paid, what does it then cost to actually sell someone a product? This includes the manufacturing and shipping of a single unit, personnel time and paperwork to execute the sale, the rate of customer complaints and returns and other transaction problems, and so on. If the total of these costs is low compared to the fixed costs of distribution per customer then competitive forces will drive the industry toward monopoly until antitrust enforcement or countervailing diseconomies of scale set in. Information commodities tend to have low marginal costs of distribution because it is so easy to make new copies of an original (Baker 1997).

Fixed costs of consumption. What does it cost to become able to consume a particular kind of product or service? For information commodities these costs can be usefully classified into machines (to play records you need a record player), skills (to consume sheet music you have to learn to play a musical instrument), and content (to use software a hundred times you need to purchase it at least once). Machine costs tend to be associated with media, not genres, and skill costs tend to be associated partly with media and partly with genres (learning to play classical piano gets you halfway toward learning to play jazz piano). These fixed costs must be paid back across the particular occasions of consumption, which should hopefully be numerous. Some genres, like classical CD's and video games, are used in activities that entail using a given package of content repeatedly; others, like novels, are not. When content costs are high, it can make sense to rent (videotapes from Blockbuster) or share (books from the library) the content-bearing artifacts. All types of fixed costs of consumption can raise distributional questions when they are high, as with the case of "equity of access to the NII". This is particularly true when media that have high fixed costs of consumption (e.g., television or networked computers) compete against media that have high fixed costs of production (e.g., newspapers or books). As the latter lose their needed economies of scale and are forced to distribute their fixed costs among ever-fewer units, they will consolidate among themselves and may ultimately collapse. Those who cannot afford high fixed costs of consumption will be left without any service at all, since they can only consume a limited number of high-fixed-cost commodity streams.

Marginal costs of consumption. These include the price of the commodity itself (assuming it has one), but it can include a lot of other costs as well. These can include travel costs, wear and tear on bodies and machines, the risk of accidents, and the opportunity cost of not having done something else instead.

Specialization. Information commodities undergo two powerful economic pressures that push in opposite directions. It is well known, on one hand, that their high fixed costs of production and low marginal costs of production create powerful competitive incentives for distributing them to the largest possible audience. On the other hand, there often exists a pressure for specialization to particular communities, known to marketing people as market segments. This is obviously in part a question of genre: genres can often be tailored to the needs of more specific groups. Both pressures operate at all points in the market at all times. The balance between them can vary wildly, causing markets for particular products and genres to appear or disappear overnight. The emergence of a mass software market, for example, caused some categories of software to drop in price by two orders of magnitude. Content producers are developing a range of strategies to deal with these contending forces. Software can be tailored locally by setting a range of switches or through the purchase of utilities or add-on packages (as with the huge range of packages for use with Notes). Printed materials like books and brochures can be tailored locally as well through new technologies for economical printing on demand. This creates a need for genres for specifying a whole grammar of possible documents. Simple versions of this phenomenon include syndicated newspaper columns that include optional paragraphs that can be trimmed to fit space restrictions, as well as professors' "reading packets" assembled from a batch of chapters and articles from various sources. But much more complex versions are possible as well, all the way out to artificial intelligence techniques that design documents (like instruction manuals or advertising brochures) within a set of genre conventions based on elaborate symbolic representations to the uses to which they will be put.

Practicalities of duplication. Records can be copied to cassette tape and books can be photocopied, but neither process is particularly convenient. Copying software, though, is usually easy. People will be more likely to make illicit copies if their social network includes other members of the relevant community.

Time-critical nature of use. If the value of a commodity decreases rapidly over time then distribution costs will probably be higher. On the other hand, if an information commodity (like a stock price) loses its value quickly then illicit copying and sharing will probably be less prevalent.

Third-party costs and benefits of consumption. Television, radio, and print media advertisers subsidize the publications they advertise in because they expect to profit from your attention having been brought to their advertising. On a more subtle level, companies hire PR firms to "sell" friendly stories to the media because they expect to enjoy benefits if the story gets an audience. Even when no money changes hands, this is effectively a subsidy to the media because it saves them the trouble of digging up the story themselves, and media firms that do this enjoy a cost advantage, other things being equal, over their competitors (Gandy 1982). Third parties can also suffer costs from information consumption: rumors can cause harm to their subjects and trade secrets are worth less to their owners once they leak out. It follows in each case that effort will often be expended to suppress them.

Brand identity of the content stream. One might think about a magazine, for example, as branded content. A brand is a set of expectations and associations that a given community has about a product, and attaching a brand to one's content stream is a way of explaining what it is and enabling satisfied consumers to get "more like that". Newsletter editors, novelists, genre fiction publishers, concert promoters, television networks, record labels, booksellers, trade associations, and think tanks all try to brand the content streams they produce, with varying degrees of success. The ability to extract income from a content brand depends on the audience's ability to predict the qualities of each next unit of content before they buy it. The matter is particularly interesting in the case of brands established by distributors: television networks, booksellers, concert promoters, and so on. In some cases these brands are based on matters that go beyond the "content" narrowly speaking, for example how well-run the concerts are. In other cases they are based on the selection of materials to suit a particular audience, as in a special-interest bookstore or a magazine. Brands increasingly cross media boundaries; the "Lion King" brand, for example, is generating revenue across dozens of media. A content stream needs a brand whether or not money changes hands; a free Internet newsletter, for example, needs to build an audience over time, consisting of people who have read a few issues and are willing to read further issues on the expectation of getting "more like that". In all cases, the crucial thing (the basis for reckoning "like that") is how the stuff fits into the reader's life (its "use value" in one idiom), and that in turn depends on its relevance to that person's relationships and goals. Libraries and other public sources of information tend to fight against the logic of brands, and reasonably so, because their justification is based on serving the general public's needs, not the summed needs of a series of market segments. Yet it is hard to think about the general public in concrete terms.

Transaction costs. These are the costs of selling something to someone: finding customers/suppliers, free samples and browsing rights, negotiating the contract, dealing with later problems with the contract, collecting the money, keeping track of the money, getting the money to the bank, and so on. In the case of information commodities, these transaction costs can exceed the marginal cost of producing the commodity itself. As the cost of electronic transactions goes down, the contracts for purchase of information commodities may shift from a fixed per-copy price ($99 for a spreadsheet program) to a per-usage price ($0.001 per command that you type on the spreadsheet).

Compatibility and standards. As an economic matter, media industries are path-dependent because of effects deriving from the compatibility of different commodities (Arthur 1996, Farrell and Saloner 1987, Lemley 1996). VHS thrived and consigned Beta to a living death, but this was not because VHS was the better format. (Many think it was not.) Rather, each standard had high fixed costs of both consumption (for the player) and distribution (for the stock of the video rental stores), and VHS had a better alliance of content producers and distributors lined up. Microsoft Windows grows and grows, but not because it is the better operating system. In each case, an initial market advantage permitted a de facto standard to become embedded in the economy. People buy Windows because a lot of software exists for Windows; a lot of companies write software for Windows because a lot of people have bought Windows; people generally use just one operating system because of the high costs and low benefits of using more. Telecommunications industries in particular exhibit powerful critical mass phenomena. Everything needs to work the same way because everyone's equipment needs to be compatible with that of everyone they call; it is hard to introduce new categories of equipment when it is not useful unless most everyone you call is also using it.

All of these considerations should influence anybody who is considering the introduction of a new medium or genre. The really harsh effects operate more strongly on the media than on the genres, but in each case it is not sufficient for one's product to be "better" as measured in a vacuum. In particular, the considerations just listed include three arguments for believing that media industries tend toward monopoly:

1. Information commodities, to an even great extent than classical monopolies based on physical infrastructures, like utilities and railroads, have high fixed costs and low marginal costs of production. A company is rewarded heavily for having a large customer base because it can distribute its costs more widely, thereby creating huge barriers to entry.

2. When consumers make frequent choices among commodities whose qualities are hard to assess in advance, as with books and videos, brand identity counts for a great deal. A great advantage thereby goes to the organization that can amortize the high fixed costs of establishing a brand identity across a higher number of customers. (An advantage also derives from generating a higher number of media products under the same brand umbrella, but this can be accomplished through licensing once the brand has been established in the first place.)

3. Once proprietary standards become entrenched in the marketplace, so that compatibility effects create ever-higher barriers to entry for potential competitors, their owners can start to extract rents from a variety of other parties. Moreover, network externalities (benefits to individuals that derive from everyone else's choices) mean that dominance over a market tends to expand once it is established. Microsoft Windows is perhaps the worst possible case of these effects. Such situations can be prevented if a critical mass of customers (or, in some cases, other interested parties such as content producers for a prospective new medium) can exert bargaining power by acting in a coordinated way early enough to influence vendors' choices in the direction of nonproprietary standards and open architectures. This can happen if the customers are few and large, are able to cooperate, have a high degree of understanding of the issues, and are thinking ahead -- conditions that are rarely all met.

7 Political considerations

Economic reasoning about the media easily gives the impression of a seamless, impenetrable logic that neither requires nor permits dialog with political concerns. Yet the most significant questions surrounding the emergence of new media pertain precisely to their role in encouraging or discouraging democratic values. To be sure, technologies do not straightforwardly determine political cultures. A given technology can be appropriated in a variety of different ways, and technologies always coevolve to a certain extent with the institutional structures around them. In the past it has seemed sufficient to inquire about such things one medium at a time -- "what is the effect of television on democracy?", "need radio have evolved into a centralized medium driven principally by advertiser sponsorship?", and so on, or else to pose the issues in terms of an orderly whole called "the media" or "the press".

The rapid proliferation of new media, however, may call for a new type of analysis. Digital networks such as the Internet are so flexible that it is practically impossible to imagine the range of architectural choices that lie ahead. Indeed, the Internet is capable of simultaneously supporting a considerable range of facilities, each of which would count in normal times as a separate medium. These media might in turn support a wide range of genres, which might fit into people's lives in a wide variety of ways. To reason about the political values that such technological developments might support or inhibit, we must return to basics and pose the general question of the role that communications genres as such play in the life of a democracy.

It is useful to pose this question specifically in terms of genres because, as I have suggested, genres tend to imply and be implied by forms of activity within communities. Of course, communities engage in numerous forms of activity, some of which have greater significance for democratic values than others. Perhaps the most democratically significant activities are the ones I sketched at the outset, the forms of association through which communities conduct the collective cognitive processes of sharing experiences, maintaining memories, conducting conflicts, and building solidarity with regard to the other communities to which it is structurally related. It is obvious that, in society as we know it, some communities have more effective means for engaging in these kinds of group thinking than others. A core democratic value, I would suggest, is broad access to the means of collective cognition.

What are the conditions of collective cognition? In some cases they might include physical meeting spaces, and it may be important for these spaces to serve a range of other functions in addition to formally organized discussions. In other cases they might include the existence of a viable community publication such as a newsletter or newspaper. It probably matters whether the community is dispersed geographically and whether its members can travel. It probably matters whether other communities, such as employers, derive benefits from the shared thinking of the community's members, as in the case of many professional associations. It probably matters whether the community's members have some way of accumulating capital by serving as thought leaders. All of these considerations, in turn, depend upon the economics of genres, travel, careers, professions, real estate, and much else. It follows that someone who wishes to design genres of communication that support democratic values must assess a larger and probably quite complicated picture. No single solution will fit all purposes.

Nonetheless, some general considerations do apply broadly. Much attention has been focused on one of these: the relationship between producers and consumers of content within a given genre and medium. Broadcast television and the Internet are frequently held out as opposite extremes in this regard, in the sense that anybody with a computer and some basic skills (admittedly significant fixed costs of both production and consumption) can create content for the Internet, but hardly anybody can create content for television (and only under a great mass of constraints). But the degree of symmetry in the producer-consumer relationship is not wholly determined by the technology. Marketing considerations are significant as well: the power to create a coherent brand image across a coherent segment of the population is also the power to concentrate enough capital to gather facts, pay writers, support travel to the places where news can be gathered, maintain the most attractive production values, and so on. Whether the future brings 500 channels or 50 or 5000 may depend as much on the market logic of segmentation as on the physical capacity of the medium.

At the same time, the experience of fanzines teaches different and more appealing lessons. The 1980's fanzine genre was adapted to numerous aspects of the music-centered youth culture from which it emerged. Since the genre created a stable set of expectations, people could decide that they were interested in 'zines as a category, and this enabled mechanisms such as Factsheet 5 to arise to spread knowledge of them to a definite audience. Broad access to desktop publishing and photocopying provided basic production methods, and the genre incorporated the properties and limitations of these methods as part of its visual language. But fanzines did not operate in isolation from other genres; to the contrary, they coevolved with the genres of popular music upon which they were explicitly predicated. A fan who found something of value in a particular band, whether on the radio or through clubs and cassette circulation, could employ the fanzine network to join a community with a common language and common concerns. Mainstream elite discourse may reduce bands like Metallica or Hole to stereotypes, not least because the arbiters of this discourse do not participate in the relevant communities' activities and thus cannot comprehend the genres that are adapted to them. But such bands do sometimes provide occasions for serious social discourse among their adherents, largely through fan publications and the other spaces where fans find one another.

Another broadly relevant issue, already mentioned in my enumeration of economic factors above, is the role of third-party costs and benefits in a community's collective cognition. The problem is not that a community might be influenced by outside voices and opinions and pleas. The problem, rather, is the practice of simulation among practitioners of public relations. Most of the magazines that serve as the primary forums for interest-communities such as car and sports enthusiasts, for example, are thoroughly corrupted by the influence of advertisers and other interested parties on the editorial copy. Even general interest publications such as newspapers rely on information subsidies from a wide range of interested parties. In each case, the journalistic voice of the publication is shaped in covert ways by the interventions of interested parties whose messages would not have the same credibility if openly owned up to. These effects continually throw into question the notion of an authentic community voice. How can the channels of a community's collective cognition be designed to be immune to these types of corruption? One straightforward solution is to make them cheap, so that outside subsidies are not necessary. And the rapidly decreasing cost of communications bandwidth ought to contribute to the emergence of inexpensive channels of group thinking such as Internet mailing lists. But the bandwidth for distributing digital material is not the only cost of producing a publication.

One final set of broadly applicable issues concerns the infrastructure of a political organization. A modern organization such as the ACLU, Planned Parenthood, or the Christian Coalition employs a broad range of genres of communication for its internal operations, chapter and member relations, campaign mobilizations, networking with related organizations, recruiting, training, and so on. These genres might include fact books, pitch letters, member newsletter articles, op-eds contributed to newspapers, legislative briefings, support materials for lobbyists, meeting announcements, and many recurring types of phone calls. One of the innovations of GOPAC under Newt Gingrich was the development of a range of additional genres, such as issue-focused conference calls and training videotapes sent to candidates for use while on the road between campaign stops. Likewise, electronic mail can permit an organization to hold fewer meetings (particularly of committees) by doing much of the work online. The point is not that e-mail substitutes for meetings by any fixed proportionality, but rather that groups can explore which parts of their collective work can be performed in which medium, postponing until the physical meeting those interactions which must be conducted face-to-face. This exploration is precisely the evolution of genres of communication. As usual, each genre fits into the broader patterns of activity in the individuals' lives and the life of the organization, and participation in the genres is a skill that is acquired and in some way transmitted to others. To consciously design these genres of communication is precisely to design the social relationships of the organization and the values that these relationships reflect.

It is hard to generalize any further about these matters. Few fixed rules or lessons may exist. The important thing is to use the proliferation of new media as an opportunity to completely rethink the place of communication in our lives. We are all designers in our daily practice of communicating -- in the small ways in which we innovate and evolve the relatively stable genres of our mediated interactions with others. But we are equally dependent upon the professional designers who have the resources and skills to map out the broad systems of community relationships within which genres of communication live. This is why it is so important for the broadest public in a democracy to become conscious of, and choose, the values that inform professionalized design. For this design work provides some of the central conditions for the extension of democracy in the future, or else its decline.

* References

Agre, P. E. (1995). Institutional circuitry: Thinking about the forms and uses of information. Information Technology and Libraries, 14(4), 225-230.

Arthur, W. B. (1996). Increasing returns and the new world of business. Harvard Business Review 74(4), 100-109.

Baker, C. E. (1997). Giving the audience what it wants. Ohio State Law Journal 58(2), 311-417.

Bakhtin, M. (1981). The Dialogic Imagination: Four Essays, translated by Emerson, C. and Holquist, M., Austin: University of Texas Press.

Bazerman, C. (1988). Shaping Written Knowledge: The Genre and Activity of the Experimental Article in Science. Madison: University of Wisconsin Press.

Beyer, H. and Holtzblatt, K. (1997). Contextual Design: A Customer-Centered Approach to Systems Designs. Los Altos, CA: Morgan Kaufmann.

Bjerknes, G., Ehn, P., and Kyng, M. eds. (1987). Computers and Democracy: A Scandinavian Challenge. Aldershot, UK: Avebury.

Checkland, P. and Scholes, J. (1990). Soft Systems Methodology in Action. Chichester, England: Wiley.

Farrell, J. and Saloner, G. (1987). Competition, compatibility and standards: The economics of horses, penguins and lemmings. In Gabel, H. L., ed, Product Standardization and Competitive Strategy. Amsterdam: North-Holland.

Gandy, Jr., O. H. (1982). Beyond Agenda Setting: Information Subsidies and Public Policy. Norwood, NJ: Ablex.

Gibson, J. W. (1994). Warrior Dreams: Paramilitary Culture in Post-Vietnam America. New York: Hill and Wang.

Hayek, F. (1945). The use of knowledge in society. American Economic Review 35(4), 519-530.

Hutchins, E. (1995). Cognition in the Wild. Cambridge, MA: MIT Press.

Jackall, R. (1988). Moral Mazes: The World of Corporate Managers. New York: Oxford University Press.

Lave, J. and Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. New York: Cambridge University Press.

Lemley, M. A. (1996). Antitrust and the Internet standardization problem. Connecticut Law Review 28, 1041-1094.

Orr, J. E. (1996). Talking About Machines: An Ethnography of a Modern Job. Ithaca: ILR Press.

Radway, J. A. (1984). Reading the Romance: Women, Patriarchy, and Popular Literature. Chapel Hill: University of North Carolina Press.

Schuler, D. and Namioka, A., eds. (1993). Participatory Design: Principles and Practices. Hillsdale, NJ: Erlbaum.

Scribner, S. and Cole, M. (1981). The Psychology of Literacy. Cambridge, MA: Harvard University Press.