See sidebar for Communication Institute for Online Scholarship (CIOS) cost structures.

Marshall McLuhan said that people understand technological innovation "through the rearview mirror"; that is, that new technologies are comprehensible to the extent that they can be placed within the framework of existing structures of sensibility. Nearly a decade has passed since we first argued the benefits of modeling electronic resources in academe on the structural features of the disciplines they are intended to serve (Stephen and Harrison, 1993a, 1993b) and during that time McLuhan's maxim has guided our work as designers of electronic services for the communication field. Here's the story: in 1986 we launched an online information system for the field of human communication studies (human interaction, rhetoric, journalism, speech, mass communication, etc.). In building that network resource center for communication scholars and students, we took for granted the importance of disciplinarity, but we never fully imagined how adherence to this principle would pay off for the communication field fifteen years later. We feel more strongly than ever that there are significant advantages to a disciplinary approach to electronic services supporting advanced scholarship and higher education.

In the sciences and humanities, "the discipline" is the operant organizational construct. Yet only a few large-scale projects have evolved with a specifically disciplinary focus. There are of course many examples of useful Web resources designed to serve particular academic subjects, but very few examples of broad, multimodal services and archives of significant depth that encompass the boundaries of an entire discipline without exceeding them. The Web site of the National Institutes of Health (NIH) provides an example of the breadth and depth we have in mind. NIH's services link large and diverse collections of resources designed specifically to support and enhance the wide-ranging needs of medical research and practice. In surveying the range of well-known database services for scholars, we note that many either opt for interdisciplinarity without providing sufficiently fine categorization to allow scholars to focus on issues as they are manifest within a particular discipline, or they provide resources covering only a single modality of a discipline's scholarly communication (e.g., article abstracts, e-mail discussion archives, collections of links to other Web sites, collections of pedagogical resources, etc.). This article discusses some of the limitations inherent in global approaches to services for scholars and educators, introduces our model of intensive disciplinarity, and sketches some of the unforeseen benefits yielded by this model in the suite of services we provide for scholars and students in the communication field.

Disciplinarity Versus Multidisciplinarity

An academic discipline is a collective of professional scholars and students sharing interest in a set of intellectual problems and utilizing a common set of procedures for their resolution. Members of disciplines are frequently associated with various university departments and their curricula. In a more inclusive conception, Whitley (1984) refers to "the intellectual field," rather than the discipline, noting that an intellectual field may be constituted more broadly, including scholars working outside universities, especially within centers of research in government and the private sector. Whitley notes that the intellectual field holds the critical authority to legitimize the activities of its members, controlling systems of reputation, merit, and social power. Crane's (1972) discussion of "invisible colleges" refined our understanding of the nature of scholarly disciplines in her exploration of the sociometric relations of diffuse scientific communities that exist within as well as across traditional academic boundaries. None of these conceptions is necessarily incompatible with the others. However one defines a discipline — concretely as the occupants of university departments and majors, more generally as an abstract and diffuse institutional identity and system of rewards and powers, or sociometrically as the emergent outcome of networks of micro-level interactions between scholars — the fact is inescapable that scholars do identify with and orient strongly toward particular groups of individuals with similar backgrounds and intellectual interests. Scholars adopt group-specific normative beliefs and values about professional practice, speak to similarly oriented scholars with a specialized argot, and tend to view others not similarly oriented as outsiders, cousins of nearby-but-well-differentiated scholarly cultures. Few scholars would find it problematic to respond to a request to report their disciplinary affiliations, and most feel that disciplinary affiliation provides an important intellectual home base and a critical context of relevance for their work.

Unfortunately, we have seen little of the structure of the disciplinarity community in electronic services. Prominent database vendors may provide opportunities to search within extremely broad, ad hoc divisions of resources, but almost never within the boundaries of disciplines as such boundaries are understood by scholars themselves. ISI's citation databases, for example, allow searchers to target materials in the arts and humanities, in the social sciences, or in the sciences, but no finer resolution is provided. This type of multidisciplinary categorization, which is not uncommon in scholarly databases, can be a powerful benefit, but the level of resolution provided in such databases is inadequate for scholars who wish to trace an idea as it has been manifested within a disciplinary culture. For the most part, scholarly literatures exist within disciplines and comprise the critical record of a discipline's historical evolution as a body of inquiry and a community of scholars. Though particular intellectual questions are often addressed by more than one discipline, today's electronic resources tend to give inordinate emphasis to multidisciplinary exploration. As well, many electronic archives have only recent materials, which is unfortunate because disciplinary communities are historical entities with traditions of intellectual interest extending in some cases more than a century and in most cases for many decades.

When an electronic service for scholars is expansive, multidisciplinary, and ahistorical, it is not possible to pursue questions within the disciplinary frameworks that may have originally given them sensibility. For example, the composition and structure of the network of pivotal concepts associated with the core theoretical term "persuasion" would be considerably different in journal article databases devoted exclusively to the communication field, the field of psychology, and the field of sociology — despite the fact that these three disciplines are closely related and that they have on numerous occasions drawn upon each others' work. Key discipline-specific tracks of research on persuasion are evident in each literature, and related scholarship is handled differently in the pedagogical texts of the three fields. There are hundreds of other key concepts that appear commonly across the literatures of many disciplines, but which are nevertheless construed differently, ascribed different significance, and woven into different webs of theoretical relationships by each disciplinary community.

Although standard online search tools are useful for cross-discipline research, they are not sensitive to the local disciplinary cultures of scholars and students. A keyword system appropriately representative of the field of women's studies would not be the same as keyword systems developed to map fields such as psychology, communication, or political science, even when there is significant overlap in the catalog of keywords within each discipline. For each discipline, terms have their own range of meaning, their own emphasis within the discipline's literature, and their own patterns of conceptual co-occurrence. Yet, since the literatures of these fields are usually lumped together in multidisciplinary archives, one rarely encounters useful, keyword-driven interfaces. Keywording may be used, but only haphazardly and globally. Consider what is lost: in a database with an intensive focus on a discipline, it is possible to guide novices through a literature using a system of menus of prominent concepts and prominent theorists organized to represent the structural relationships between ideas and their proponents as such relationships are understood by scholars in the field. This is impossible in a global database because such conceptual networks are discipline specific; in fact, such conceptual networks are the sine qua non of disciplinarity itself.

As multidisciplinarity increases, search results may suffer from the "Yahoo! syndrome" — huge collections of tangentially related materials. A search for "network" in ISI's "Current Contents Connect" extended file, limited to the social and behavioral sciences collection, yields thousands of articles that, upon examination, go far in demonstrating the many disciplinary-specific twists the term can take as it is employed in one literature or another (e.g., television networks in communication, social networks in sociology and psychology, sales networks in business, and computer networks in computer science). Similarly a search for "programming" in an eclectic database yields studies of television from the communication literature, studies of computing from information science, studies of high-school curricula from the education field, and studies of brainwashing from the psychological literature. In the extreme case of the Internet's indiscriminantly eclectic databases, attempts at building assistive technologies to guide searchers are likely to do only little better than the Ask Jeeves site which, in its assumption of a completely undifferentiated archive and a completely undifferentiated mass audience, responds to plain language queries such as "What is the relationship between facial expressions and deception?" with results grouped under the headings: "Where can I learn about the spa treatment facials?" and "How can I find out if that special someone likes me?" The Northern Light search engine, with its connection to UMI's archive of scholarly articles, has potentially deeper relevance for scholars, but it does not offer a way to target searches within disciplinary boundaries and, at least in the communication field, its archive has little depth and is exceptionally narrow, covering less than one fifth of the titles in field's professional periodical literature.

Yahoo!, Ask Jeeves, Northern Light, Google, and the like are for-profit ventures seeking a mass audience in their effort to obtain revenue through advertisements displayed with search results. Presumably, as long as large numbers of site visitors find something of interest in these services, the sites will generate sufficient advertising revenue to satisfy the need to cover expenses. Subscription services have a harder task, as they must demonstrate sustained relevance and value for particular audiences. However, regardless of whether a service is provided with or without subscription charges, services provided under a for-profit model must contend with a natural pressure to seek relevance within the broadest audience possible, and it seems likely that this has sometimes been responsible for extravagant claims for coverage — even within databases specifically targeted for the higher education and research community.

It is not unusual for for-profit vendors of scholarly databases to entice subscribers with claims of coverage of extremely large numbers of titles in their article archives when in actuality they offer comprehensive and routine coverage of a significantly smaller number of titles accompanied by only occasional and selective coverage of other titles for which coverage is claimed. For instance, a recent flyer announced release of a Web-based linguistics database claiming coverage of 300 journals back to 1985, in all more than 15,000 abstracts. Since the average academic journal publishes fifteen articles per year (quite a conservative number), each year of coverage for a database of 300 journals should yield approximately 4,500 article abstracts. Accumulating 4,500 abstracts per year for fifteen years should yield more than 67,000 article abstracts. The linguistics abstracts database has abstracts for only 15,000 articles. What happened to the other 52,000 articles? Even allowing that some proportion of the 300 journals "covered" have not been in publication as long as 15 years, claims such as these are misleading and potentially dangerous to scholars who need to know that they've searched a particular set of periodicals exhaustively. It seems likely that the linguistics database provides occasional or selective coverage from a large list of titles, but comprehensive coverage of a significantly smaller core set. Such a system may enhance profit margins for database publishers when institutions subscribe without full knowledge of the nature and limitation of coverage. Subscribers may fall prey to the illusion of breadth, not realizing that their searches are not exhaustive of a discipline's literature. To the extent that this happens, this practice does a significant disservice to scholars and to the advancement of knowledge.

"Before the Web was invented the Comserve service inaugurated one of the first peer-reviewed scholarly electronic serials"

Academic disciplines are discernible microcultures, communities with traditions of intellectual activity and normative behavior that in many cases extend for generations. Disciplinary members share assumptions about intellectual priorities and narrative accounts of the discipline's evolution, and their shared support of the discipline's forms of procedure perpetuate the discipline's traditions and its system of status conferral. Central to each discipline is its own periodical literature, and disciplinary members are usually thoroughly aware of the prestige hierarchy within that set of journals and annuals. It would of course be ridiculous to suggest that disciplines are univocal enclaves of agreement, and we are certainly not doing so here. To the contrary, scholarly disciplines are most frequently hotbeds of contentious argument. Nevertheless, the cultural variance between disciplines is far greater than the cultural variance within them.

It is precisely because disciplines are such distinct cultures that electronic systems designed to speed scholarly communication, such as Paul Ginsparg's preprint server in high-energy physics (Ginsparg, 1994), may be revolutionary in particular fields but completely irrelevant in many others (e.g., within most of the social sciences and humanities, where speed of distribution of preprints has virtually no bearing on the advancement of knowledge). Similarly, e-mail discussion lists may play a vital role in one field but are never used in another, and article archives that emphasize recent materials suffice in particular fields while including older materials may be the only appropriate model for other fields. Electronic services need to be designed differentially and should deploy technologies selectively in service of the varying scholarly practices that define different fields. The disciplinary community is everything and it is our belief that significant benefits would accrue if this insight, translated into a guiding principle of design, were to be more fully exploited among today's electronic services for the research and education community.

Building the Services

In contrast to those services choosing to emphasize multidisciplinarity, in 1986 we sailed on the opposite tack, designing as a nonprofit operation a network-accessible set of electronic services tailored to the specific traditions of education and scholarship for scholars whose work converged on the study of human communication (media and broadcasting, rhetoric, speech communication, human interaction studies, mass communication, and journalism). As scholars with deep roots in that discipline, we understood its boundaries, its culture and normative framework, and the range of its literature as perceived by insiders, and we thought we knew something about the likelihood of adoption in our field for particular forms of electronic scholarly services.

In that pre-Web era, our services, collectively known as "Comserve," were provided free to those with access to Bitnet (a 1980s academic and research computer network) and the Internet. Document delivery and query results were provided by e-mail and by interactive messaging for those users whose access permitted it. Our services consisted originally of a repository of educational resources solicited from members of the discipline — papers, curricula, bibliographies, and more. As well, we mounted an interactive directory of scholars and students and deployed more than 25 e-mail discussion lists for students and scholars in the traditional divisions of the field (e.g., organizational communication, rhetoric, interpersonal communication, health communication, mass communication, political communication, gender and women's studies, etc.). We modeled the organizational structure of the discipline's principal professional societies. More detail on the design of our services is available in Harrison and Stephen, 1992, 1995; Harrison, Stephen, and Winter, 1991; and Stephen and Harrison, 1988, 1993a, 1993b, 1994, 1996.

In 1990, before the Web was invented, the Comserve service inaugurated one of the first peer-reviewed scholarly electronic serials, The Electronic Journal of Communication / La Revue Electronique de Communication, which continues today in its twelfth volume as an official journal of the Communication Institute for Online Scholarship (CIOS). CIOS is the not-for-profit organization that evolved from the Comserve project to ensure the future of its accumulating services and resources. Shortly thereafter, because the communication discipline has never been well served by existing print or electronic indexes, CIOS began to construct and deploy searchable electronic indexes to the literature of the communication discipline. Several news services announcing disciplinary events, new books, and academic positions were rolled out over the years and special forums were created to support electronic collaboration among scholars pursuing special projects. CIOS systems also supported some of the first experiments in collaborative computer-mediated education.

During the early period of our project, nearly every service we offered was an innovation within our field and we wrote the software for each one. Custom software was essential to ensure that our services had an appropriate fit and maximum appeal within our discipline. The services that flourished most readily were those that naturally amplified or extended traditional disciplinary practices. Those that did not, such as a preprint distribution server that we rolled out in the early 1990s, were met with quiet indifference. A series of rapid changes in inter-university computing and network technology in the first half of the 1990s caused considerable disruption and vexation (Stephen and Harrison, 1996) as we shifted from our original deployment on a Bitnet-based mainframe system to redeployment on microcomputers relying principally on Internet and Gopher technology. No sooner had we finished the transition to a Internet/Gopher-based technology then that technology was supplanted by the World Wide Web. For the third time in only a few years, we were forced to revise 150,000 lines of programming code that make up the software for CIOS services. Although some aspects of the original Comserve e-mail-based system are still supported, the primary interface for CIOS services is now the World Wide Web. Adding to this, in the early 1990s we had devoted significant resources to the creation of custom software for delivery of electronic journals (Stephen, Harrison, and Silvestre, 1995) and for delivery of our "ComIndex" (Stephen, Harrison, and Silvestre, 2001) electronic index to communication serials. The sweeping adoption of the World Wide Web required us to quickly redeploy these resources using Web technology. Although we found the continual need to revise services to meet rapid technological change a burden, the shift to a Web-based service did cause rapid increases in the use of CIOS services. Access grew from approximately 150 to 300 contacts a day (via e-mail and interactive messages) in the mid to late 1980s to 20,000 to 30,000 times a day through the Web, a figure we judge to represent substantial impact, given the relatively small size of the communication field.

Growing the Resource Archives

As we reached the new millennium, CIOS services had amassed significant archives of disciplinary resources. The suite of e-mail discussions has distributed more than 14 million copies of some 54,000 original messages contributed by disciplinary members and the file archive (referred to as "the resource library") has accumulated more than 5,000 entries. Archivesof all these resources are incorporated automatically by the CIOS's full-text indexing systems. Furthermore, our journals-index system grew to become the field's most comprehensive index to communication-related serials, now with 36,000 bibliographic records covering 83 titles from approximately 1970 to 2001. In 1997, under Tim Stephen's direction, CIOS launched the ComAbstracts database, which has grown to include 13,500 keyworded bibliographic and abstract records covering more than 50 journals from roughly 1985 through their most recent issues.

Shortly after the launch of our World Wide Web interface, CIOS initiated two new projects that each have increased substantially the variety and depth of resource materials archived by CIOS. The first of these new services is a full-text database of Web pages harvested from Web sites in communication-related academic departments and from sites of relevant professional societies. This service is a kind of mini-Yahoo! Web portal in which search results are of guaranteed relevance to the communication field. The other new initiative has been the creation of the Electronic Encyclopedia of Communication project, the goal of which is to underwrite the creation high-quality pedagogical materials focused on particular topics of disciplinary relevance. These materials are produced by experts in the field especially for the Encyclopedia. Revenues generated from institutional and individual subscriptions to CIOS services pay an executive officer for the electronic encyclopedia project and fund a program of small grants to help scholars create content for the sites. This recent initiative has resulted thus far in two projects that have progressed to completion, a presentation featuring the ideas of communication theorist Marshall McLuhan and another devoted to patient/physician interaction. Other content modules in various stages of peer review or active production feature persuasive communication, conflict resolution, and communication and disabilities. We anticipate funding numerous other such projects.

The Electronic Journal of Communication began its twelfth volume in 2002. It has been a great success. As with other early innovative experiments in electronic scholarly publishing, the initial period of EJC's publication was shrouded by questions about the viability and acceptance of electronic journals within the research and education community. As time has passed, early predictions of eventual acceptance have been borne out. Although EJC requires institutional or individual subscription, it was accessed 90,000 times over the last year. If this does not rank it as the most accessed scholarly periodical of the communication field, it is certainly among those with highest access. EJC is published under a peer-reviewed, focused-issues format with individual issue editors; contributors and issue editors routinely include the field's influential scholars. EJC has pioneered a progressive model of copyright that permits scholars wide use of their own texts and the right to re-publish articles as long as the original electronic publication is acknowledged. This model has resulted in republication as books of two collections of EJC articles. Individual articles from EJC are archived on CIOS's central servers with the entire text of each article included in a searchable full-text index to the journal.

Bridging Fragmentation

As our services have matured and our data archives have grown and diversified, we have appreciated increasingly the potential contribution of CIOS to not only support communication pedagogy and research but to better integrate areas of the communication field. As a not-for-profit organization specializing in electronic data services, CIOS services are considerably more advanced than the rudimentary Web resources mounted by the field's scholarly societies. The societies perform an array of essential services for the field that fall outside the mission of CIOS; however, they are not well positioned or equipped to craft and launch electronic applications. Hence CIOS has undertaken numerous projects on behalf of communication-related scholarly societies in the United States and abroad. One important recent effort has been to create software that permits a professional society to replace its annually produced paper-based membership directory with a searchable directory mounted on the World Wide Web. CIOS created, hosts, and maintains such a directory for the International Communication Association and is now in the process of inviting other communication-related societies to make similar use of CIOS services without charge. Our goal in this effort is facilitate the creation of a comprehensive and universally accessible online directory of communication scholars encompassing as much of the field as possible. Not only will societies save money and searchers have more powerful and convenient access to data, but eventually we will tie these directories to other CIOS data archives, providing student and faculty searchers with up-to-date contact information for authors of papers and pedagogical and research materials that are located through CIOS's electronic search services. This information may facilitate the direct exchange of reprints and other pedagogical and scholarly resources.

"The Idea Monkey randomly pairs vital concepts in a set of generic research questions"

In a similar way, CIOS has invested substantially in efforts to help many of the smaller professional organizations that have neither the expertise nor the funds to bring their journals online. At this time CIOS has created electronic archives for back issues of two print journals and will bring seven to nine more titles on line in the next year. Although the conversion process is indeed expensive, the major impediment to this effort is not so much cost as tradition. Professional societies are frequently uncertain about how to respond to offers to bring their serials online and timid about experimenting with new models of access. Disciplinary tradition is a formidable force of inertia, even in cases in which it works against the field's own interests.

Another obstacle in this and other efforts we've undertaken is the apparent perception that publishing projects gain stature and credibility when mounted by for-profit publishing companies — a variation of the same base of prejudice that mitigated against the acceptance of electronic serials in the early years of the World Wide Web. In some areas of the academy, scholars have by tradition coveted association of their publishing efforts with the imprimatur of presses deemed to be prestigious by various opinion leaders. It may be that a press's interest in bringing a text to market reflects the press's expertise in a subject and its dedication, interest, and inherent sublished products to market. Such perceptions may or may not be valid but what is certainly true is that, as an industry, the commercial presses — in their failure to moderate access costs for scholarly serials or to adopt copyright policies friendly to scholarship and education — represent at this time a paramount impediment to the realization of the full potential of electronic data services in the research and education community. Scholars and professional societies who continue to sign away copyrights to for-profit publishers in many cases are banishing critical and irreplaceable texts to an uncertain future, perhaps to utter obscurity. With a growing body of other agencies (cf. Magner, 2000), we at CIOS believe that the academy should make significant efforts to change current copyright practice and the self-destructive attitudes supporting it.

Learning from the Database

Recently, as CIOS archives have accumulated, Stephen has begun conducting statistical explorations of CIOS collections (Stephen, 1999, 2000, 2001). The long-range goal in his work is to assist in the development of more intelligent interfaces to textual databases for the field. The first steps in this effort have made use of the 36,000-record bibliographic database to identify key terms in the literature of the field and to study patterns of co-occurrence among them. We expect this work to help us construct a conceptual road map for the field that will guide searchers in CIOS archives.

To this end Stephen developed a series of automated filters for the titles of articles from the bibliographic database. The filters eliminate punctuation and insignificant or common words (e.g., "and," "or," "but," "not"), eliminate words that are used so frequently that they lose meaning in this context (e.g., "communication," "research," "analysis"), reconcile international spellings (e.g., "color" and "colour," "organization" and "organisation"), reduce all remaining words to their root forms (e.g., "know," "knowledge," and "knowledgeable"), and identify and contract to singular tokens hundreds of pre-identified multi-word phrases in common use in the communication field's literature (e.g., "sex differences," "content analysis," "spiral of silence," "prime time"). The words are also filtered to equate synonymous terms using a customized thesaurus representative of the field's unique patterns of language. The set of filtered tokens that remain are then studied to identify patterns of co-occurrence and the distribution of terms by textual sources and historical epochs in the field.

As an example, Figure 1 presents a tree diagram resulting from a cluster analysis of the portion of the conceptual relationships derived from a subset of 1,917 articles addressing issues in television research. The items listed along the side of Figure 1 are those forty-seven significant concept terms with the highest frequency of occurrence in the data set. The terms are listed in their "tokenized" form; that is, as they appear after filtering. Stephen (2000) has described the clustering process as similar to the formation of planetary systems in which the force of gravity draws dispersed matter into objects with substantial mass and then arranges them in relation to one another. In Stephen's clustering work, the conceptual objects are placed within a kind of "conceptual universe" that is defined by the field's patterns of use of these key concept terms. The "gravitational force" connecting conceptual elements is the frequency of co-occurrence of pairs of concept terms. Hence terms that co-occur frequently (e.g., "political" and "election") are taken to be conceptually linked, and those that do not co-occur (e.g., "newspaper" and "parent") have no "gravitational attraction."

The clustering process initially identifies concepts of especially strong co-occurrence and then places the remaining concept terms in proximity to the cluster centers and to each other so that terms are placed closest to those with which they co-occur more frequently. A dendogramatic representation depicts the proximal relationships between all the concept terms. In most cases, those terms that are grouped together in the cluster analysis have displayed a tendency to appear together in the discourse of the discipline as represented in the corpus of texts under study. The resulting dendogram is limited in that it is only a two-dimensional representation of a kind of multidimensional universe of concept relationships. Despite this limitation, if the textual corpus is carefully constructed and the analysis is targeted at the level of broad themes, the results of such a cluster analysis can and do frequently illuminate conceptual structures within the field's scholarship that might otherwise be difficult or impossible to discern. The payoff in this is that for a field that has never developed a canonical system of keywords for organizing access to its texts, this kind of dynamic empirical procedure can provide such a system.

Although a detailed analysis of the television data is beyond the purpose of this article, a look at Figure 1 suggests that a large division in the television literature separates those concepts representing issues in the television industry from those representing television's impacts on people and social relationships. A finer level of discrimination divides the industry variables in five subclusters addressing television news and programming, television's relationship to other broadcast media, television's role in coverage of political elections, international differences in media policy and practice, and commercial advertising. A finer analysis of the second large group suggests three subclusters addressing television and family interaction, television's structuring of perception (especially for adolescent viewers), and prime-time portrayals of minority groups in dramatic productions.

This demonstration illustrates the value of the intensive disciplinary focus in CIOS electronic services. As more discipline-specific data has accumulated, scholars are increasingly better positioned to extract and integrate information about the field: its structures of information, its conceptual relationships, its range of interests, and its historical evolution. This information is now being used to report on the field's patterns of conceptualization (Stephen, 1999, 2000, 2001) and as a basis for constructing and improving the CIOS search interfaces. The latest CIOS interface, "Communication Concept Explorer," was unveiled in 1998. Work is ongoing on its successor. The Communication Concept Explorer takes any word or concept of interest and generates a list of significant concepts that co-occur frequently. The user can then select from the list of co-occurrent terms and jump immediately to several of CIOS's textual databases and execute a search for the selected pair of terms. For example, a novice communication student searching for source materials for a class paper in a mass-media course can learn by entering the term "television" that it has been connected to the term "news" in 371 articles, to "children" in 215 articles, to "programming" in 157 articles, etc. Deciding to pursue information about the relationship between "television" and "children," the searcher selects an option that executes a search in CIOS's ComAbstracts (article abstracts) database for that pair of terms. The resulting collection of article abstracts allows the searcher to appraise quickly the dimensions of the field's research on television and children and to generate a bibliography of articles for detailed study. The Concept Explorer system is in essence a dynamic, empirically generated keyword system for the communication field in which is contained rudimentary information about the strength of relationship between vital theoretical concepts. It was made possible by the collection of an intensively focused archive of discipline-specific data.

"The Idea Monkey" is a related interface technology that works with an inventory of vital concepts from the communication literature. The Idea Monkey randomly pairs vital concepts in a set of generic research questions: How is X related to Y? How do factors of X impact the experience of Y? What are the components of X that are relevant to Y? The relationship between the two randomly selected terms is rated for the originality of the pairing. A pairing of concepts rarely paired in the field's literature is rated high on originality. A pair of concepts frequently seen paired in the literature is rated low on originality. For example, if the Idea Monkey selected "adolescence" and "power" — two high-frequency constructs from the communication literature — the result might be: How is adolescence related to credibility? How do factors of adolescence impact the experience of power? What are the components of adolescence that are relevant to power? The originality rating for this pairing of concepts is high because they appear together infrequently in the field's literature. The Idea Monkey is designed to leverage our empirically derived knowledge of the field's conceptual structures to identify research questions that are potentially fresh and novel. Once the user has executed the Idea Monkey function sufficiently to have found a pair of concepts of interest, the interface allows the user to immediately search for the terms in various CIOS databases.

The Concept Explorer project has informed another approach Stephen is using to construct assistive interfaces, a semi-automated keywording system. As we incorporate more scholarship from adjacent fields into the CIOS system, we have become more aware of the specialized use of language in the discipline. We are now testing a new keywording system on the ComAbstracts database. Abstracts in the ComAbstracts database have always had key words, but until recently the keywording process was manual and somewhat ad hoc: disciplinary experts read each abstract and, using rough guidelines, applied as many terms as they deemed relevant to tag each abstract record. The new system developed by Stephen capitalizes on his previous studies of the structural relationships between key concepts and phrases in the formal communication of the field. The new computerized procedure examines text associated with each article abstract, normalizes the language, and searches for collections of prominent phrases and terms appearing in approximately 105 high level categories of scholarship in the field. Whenever a term or phrase is matched, the term or phrase is added to the keywords field of the record and the name of the high-level category is added to a new field labeled "metaterms." Although the procedure requires manual correction and cannot yet run as a completely automated task, it imposes a constructive discipline on the keywording process. This is already paying off by producing database records with a greatly enhanced level of systematization and a vastly improved level of inter-record consistency.

Discussion

As the 21st century begins, the authors of a majority of all scholarly articles and pedagogical materials ever published in the communication field are still available to talk to others about their work — what concepts are encompassed, where related items have been published, who else has contributed, etc. Authors of most articles relevant to that tradition can be found working in academic departments or chatting in the halls of the national and international conferences, and are usually available to provide background on a particular line of study. Because of this, the evolution of particular literatures has not been overly difficult to track. However, this situation is in the initial stages of dramatic change and the pace of change will increase rapidly in the immediate future.

Since the 1970s, the number of communication journals and annuals that are actively in publication has more than tripled. A minimum of two new serials begin publishing every year. No university library subscribes to all of the publication sources currently active (we estimate there are roughly 100 active scholarly communication journals and annuals) and many subscribe to only a handful or have radically cut back their holdings by canceling subscriptions. Even if libraries subscribed to all the journals and annuals that are relevant to the field, the sheer volume of scholarship pouring out of the field annually is now so large that no scholar can afford the time to peruse the entire corpus of the field's scholarship manually. Although one might turn to a colleague for help in tracking research literatures today, within only a few years virtually the entire generation of scholars who contributed substantially to the communication field's modern definition in the 1960s and 1970s will be in retirement and with their retirement the field will lose many of its living guides to its scholarly legacy.

The field has already begun to cope with this onslaught of change by relying increasingly on electronic databases for literature reviews and for locating other kinds of scholarly and pedagogical resources. This practice will expand as the field's literature expands, generations of primary authors retire from contact with students and colleagues, libraries continue the trend of paring down existing subscriptions and resisting subscription to new journals, and the proportion of computer-literate scholars in the field continues to expand. Electronic databases are clearly useful, but to serve the field well they need to be constructed in ways that are sensitive to the discipline's intellectual values, traditions of thought, and habits of language. Aside from CIOS services, multidisciplinary databases are the only electronic databases providing support for the communication field. Multidisciplinary databases have considerable value, but in their attempt at cross-disciplinary breadth such databases may inadvertently submerge a field's intellectual culture. Today the communication field's literature is tracked most effectively by its core of senior scholars whose extensive disciplinary knowledge connects conceptual threads, understands their historical evolution, and preserves information about priority, nuance, and direction within the field's literature. Without the development of discipline-specific electronic resources that incorporate this wisdom, the retirement of these guides could well leave the communication field disconnected from its intellectual roots.


SIDEBAR: CIOS Model of Operations

Communication Institute for Online Scholarship (CIOS) services operate under the direction of Tim Stephen, who oversees technical development, and design issues, and is responsible for operations. CIOS's Electronic Journal of Communication project is managed by Teri Harrison. Both Tim and Teri have professorial positions at the State University of New York at Albany and work on the project on a voluntary basis. In recent years, Tim has negotiated half-time release from his university position to provide more concentrated time for CIOS. CIOS relies on independent contractors for numerous small jobs in data entry, software development, etc.

All funding for the project derives from revenues from four sources:

  • Personal CIOS memberships ($45/year) that permit individual scholars to use CIOS Web services with a personal password. The majority of the journals indexes are available to CIOS/Comserve associate members only. (See CIOS membership details for more information.)
  • Sales of the ComIndex bibliographic database ($50 for a personal copy through $340 for a departmental or library copy and up to $1000 for a campus-wide copy). This database is a citation index (author, title) to approximately eighty-three titles in communication — 36,000 records from roughly 1960 to 2001. (See ComIndex desktop software for more information.)
  • Basic institutional affiliation ($250 to $500), which confers campus wide IP-based access to most CIOS databases; and
  • ComAbstracts subscriptions (at average $1,100 for institution-wide access via IP recognition).

Currently CIOS has approximately 400 institutional subscribers to various service packages and sustains approximately 300 individual members.

As of the writing of this article, CIOS is beginning to revamp its hardware, obtain new custom-built full-text indexing software, and more than triple the number of professional societies' journals whose back issues are on line. The object is to phase in new services and improve support for CIOS's expanding business as the income permits. We are a true nonprofit — we truly aren't in this for the money, but as a service to our field.



Tim Stephen is Professor of Communication at the Department of Communication at the University at Albany, SUNY, in Albany, NY. He is also president of the Communication Institute for Online Scholarship (CIOS) and serves on the editorial boards of numerous scholarly serials. In addition to his research in the history of interpersonal relationships, he focuses on innovation in text analysis, interface design, and the development of new electronic tools for scholarship. You may reach him by e-mail at stephen@albany.edu.

Teresa M. Harrison is Professor of Communication and Chair of the Department of Communication at the University at Albany, SUNY, in Albany, NY. Her research and teaching focuses on computer-mediated communication, with a special emphasis on community applications of information technologies. Most recently, in a project funded by the National Science Foundation, she is working with a team of computer and social scientists together with representatives of government and community organizations to design and develop a community information system and study its diffusion over time. She is the managing editor of the Electronic Journal of Communication and, with Timothy Stephen, is co-editor of the SUNY Press Series on Computer-Mediated Communication in Work, Education, and Society. You may reach her by e-mail at harrison@albany.edu.


References

Crane, Diana (1972). Invisible colleges: Diffusion of knowledge in scientific communities. Chicago, IL: University of Chicago Press.

Ginsparg, P. (1994). "First steps towards electronic research communication." Computers in Physics, 8, 390-396.

Harrison, T., Stephen, T., & Winter, J. (1991). "Online journals: Disciplinary designs for electronic scholarship." Public Access Computer Systems Review, 2, 18-28.

Harrison, T. & Stephen, T. (1992). "Online disciplines: Computer-mediated scholarship in the humanities and social sciences." Computers and the Humanities, 26, 181-193. [doi: 10.1007/BF00058616]

Harrison, T., & Stephen, T. (1995). "The electronic journal as the heart of an online disciplinary community." Library Trends, 43, 592-608.

Magner, D. (2000). "Seeking a radical change in the role of publishing." The Chronicle of Higher Education. June 16, A16-A17.

Stephen, T. (1999). "Computer assisted concept analysis of HCR's first 25 years." Human Communication Research, 25, 498-513. [doi: 10.1111/j.1468-2958.1999.tb00458.x]

Stephen, T. (2001). "Concept analysis of the communication literature on marriage and family." Journal of Family Communication, 1, 91-110. [doi: 10.1207/S15327698JFC0102_01]

Stephen, T. (2000). "Concept analysis of gender, feminist, and women's studies research in the communication literature." Communication Monographs, 67, 193-214.

Stephen, T. & Harrison, T. (1988). "Bitnet and Comserve: Electronic resources for teaching and research." Communication Education, 37, 81-84.

Stephen, T. & Harrison, T. (1993a). "Comserve: An electronic community for communication scholars." In Ann Okerson (Ed.). Scholarly publishing on the electronic networks: Proceedings of the second symposium. Washington, D.C., Association of Research Libraries and Association of American University Presses, pp. 53-58.

Stephen, T. & Harrison, T. (1993b). "Online disciplines: Building electronic scholarly communities." Media Information Australia, 67, 71-76.

Stephen, T. & Harrison, T. (1994). "Comserve: Moving the communication discipline online." Journal of the American Society for Information Science, 45, 765-770.

Stephen, T. & Harrison, T. (1996). "Assessing the costs of technopoly: Constructing scholarly services in today's network environment." In T. Harrison & T. Stephen (eds.) (1996). Computer networks and scholarly communication in the 21st century university. Albany, NY: State University of New York Press.

Stephen, T., Harrison, T. & Silvestre, P. (1995). DISPLAY & MARKUP: Software for Producing and Displaying Electronic Scholarly Journals. [Computer software]. Rotterdam Junction, NY: Communication Institute for Online Scholarship, Inc.

Stephen, T., Harrison, T. & Silvestre, P. (2001). ComIndex: An electronic index to communication serials. [Computer software]. Rotterdam Junction, NY: Communication Institute for Online Scholarship.

Whitley, Richard (1984). The intellectual and social organization of the sciences. Oxford: Clarendon Press.


Links from this article

Ask Jeeves, http://www.askjeeves.com

ComAbstracts database, http://www.cios.org/www/absrch.htm

Communication Institute for Online Scholarship, http://www.cios.org

Electronic Encyclopedia of Communication, http://www.cios.org/encyclopedia/enindex.htm

Electronic Journal of Communication / La Revue Electronique de Communication, http://www.cios.org/www/ejcmain.htm

Google, http://www.google.com

International Communication Association, http://www.icahdq.org

ISI's citation databases, http://scientific.thomson.com/products/categories/citation/

ISI Current Contents Connect, http://scientific.thomson.com/products/ccc/

National Institutes of Health, http://www.nih.gov

Northern Light, http://www.northernlight.com

Yahoo!, http://www.yahoo.com