Achieving Maximal Value

from Digital Technologies

in Scholarly Communication

by

Charles E. Phelps, Ph.D.

Evaluation of Individual Faculty, Departments, and Schools. Unfortunately, there remain pockets of academia where scholarship is evaluated by the pound rather than by its quality. I know of major research universities where schools within them have fixed criteria for numbers of journal publications before a person will be recommended for tenure. The review processes of major governmental funding agencies (NIH, NSF) have limited (e.g., to two pages per investigator) the length of the researcher's publication list that can be included in a proposal, which helps some, but there still remain major incentives to increase the number of publications. Even the prestigious National Academy of Science evaluation of the quality of doctoral programs⁽¹⁰⁾ contains indexes of the number of publications per faculty member, as well as the more pertinent number of citations per article. A few universities have adopted approaches that get away from this problem, for example, by seeking only five articles chosen by the candidate for tenure that will be used in the evaluation. This has the important (and to some, unhappy) effect that reviewers of the material actually have to read the manuscripts, rather than just counting publications on a CV.

Property Rights to Intellectual Property. Perhaps the most important aspect of this problem is that we, as universities, have totally abandoned any property rights to the work produced by our faculty. Our faculty produce the scholarly work that the journals publish, they provide the editorial boards and referees that evaluate submitted manuscripts, we buy and make available the journals when they are published, and take responsibility for archiving and providing access to the body of scholarly work.⁽¹¹⁾ In short, the large bulk of the published work -- and almost all of the intellectually important work in the world of scholarly publication -- is undertaken by our faculty, yet they routinely turn over all property rights to the publishers of the journals in exchange for benefits of publication. Our ability to use this work in our own teaching and scholarship through "fair use" laws is somewhat limited, such at (for example) even using work we have originally authored in "course packs" for our own students often requires a copyright fee.

The certification implied by publication is by far the most important benefit, in my view, in addition to the editorial improvement, as well as the distribution achieved by physically publishing and mailing copies of the journals. As David Shulenburger has detailed in a companion discussion to this work (see note 1 regarding his NEAR proposal), there appear to be significant and -- importantly -- achievable and acceptable modifications to the usual total assignment of property rights that would help shift the balance of benefits and costs more to the institutions of higher education that produce the original work. The issue is not whether the journals provide valuable services -- they do without question -- but rather whether the terms of trade are appropriate.

III.. Externally Focused Efforts

In addition to the internal steps that institutions of higher education must take to help resolve these problems, a number of externally focused approaches are now appearing on the scene that each hold important promise for helping to resolve part of the problems we confront. I wish to be clear that none of these efforts is intended to solve these problems alone, nor should we expect them to do so. It is also important to note that, in general, the efforts and ideas now appearing on the horizon mutually complement each other, rather than standing in conflict. Indeed, some cannot work without the presence of others.

To understand the role of the various efforts now underway, we need to review the functions performed by the current system of journal publication. These functions are so linked together in people's minds that they often find it difficult to think of separating them, yet it is precisely the new capabilities provided by digital technologies that allow us to consider their separation. The key functions performed by old "paper" publication processes include:

evaluation and certification of quality

editorial improvement

distribution

indexing and search

archiving and retrieval

Every discipline contains within it a hierarchy of scholarly journals, well understood within the field, such that the journal in which a manuscript is eventually published contains important and reliable information about the quality of the work. Authors understand this hierarchy, and normally begin their submission process by aiming a manuscript at a high quality journal, perhaps stretching a bit in quality. Serial rejections and re-submissions eventually lead most manuscripts to find a home in some journal, sometimes in the fourth or fifth iteration. Where it lands signals its quality. Virtually the entire editorial process to accomplish this important step is carried out by faculty of the major institutions of higher education in the US -- our faculty -- serving either as editors, editorial associates, or referees, a point to which we can later return.

Once a manuscript is accepted for publication, a round of editorial improvement takes place (which has indeed begun during the prior review process). This involves both academic editors' input and technical "editing" to bring the manuscript into compliance with the journal's standards of quality and appearance. Most of this work, and the subsequent typesetting and proofreading, primarily involves technical staff paid by the journal.

Upon publication, journals are distributed to individuals and libraries (subscribers) who have paid for the right to acquire a copy. The US Postal Service subsidizes this process through low cost postal rates for educational material.

Libraries receive and catalog these journals, bind them into annual volumes at the end of the year, and these catalogs form one basis for scholars to learn of possible valuable sources of material for their own research.

And finally, by controlling the environment, replacing missing or defaced copies, and general protection of the resource, the libraries of our colleges and universities collectively provide the method by which archiving and retrieval of journals takes place.

Any system relying on electronic media to replace paper journals must replace each and every one of these functions with a process viewed at least as desirable by the faculty carrying out the research involved. If the system fails on any of these accounts, the natural consequence will be that scholars use the new medium and in addition submit their manuscript to a traditional journal. This is not to say that every addition to the world of scholarly publishing need carry out all of these functions, but rather that the collection of new additions must reliably provide this set of functions before the previous system will be given up.

Publishers of traditional paper journals realize that they have in their hands an enormously valuable property -- the intellectual property rights (copyright) to the scholarly work that they publish. This property right has become increasingly valuable as the world of digital information increases in its scope, including the ability to communicate, exchange ideas, search data bases for resources and references, and the like. The very production of Ph.D. trained scholars who carry out research has accumulated through the decades since sputnik in a process that pushes much further into the hierarchy of academia the demands for high quality resources to support scholarly work, further adding to the demands for journals and related material. Publishers of journals, both for-profit and not-for-profit (mostly, in the latter case, learned societies) have taken advantage of this increasing demand by raising prices at rates that, at least in some cases, are far exceed increases in costs. David Shulenberger's work shows that in recent years, for profit publishers have, on average, increased their prices at a rate fourfold that of inflation.

Cost-Affecting Responses

Many ideas have emerged in higher education to deal with these increases, some helpful, some harmful, some nearly suicidal. At the suicidal end of the spectrum lie formalized policies from some universities that their libraries shall automatically receive budget increases necessary to match price increases posted by journal publishers. One might as well put on a deer costume and go out in the forest during hunting season! This simply invites publishers to raise prices without constraint, and if every college and university in the US followed these policies, publishers would drive prices to levels that we would need telescopes to find. At the opposite extreme, some universities have informed specific publishers that the total spending on their journals is fixed by current spending, so that if they raise prices, subscriptions will be canceled so that publishers' revenue remains fixed. Since the marginal cost of an additional copy of the journal represents a trivial amount of the overall costs of production, this has the potential for stemming publishers' price increases, particularly if the policy were more widely adopted.

Another innovation that both promises to stem costs and to improve access is the emergence of collaborative arrangements of various sorts. These consortia develop a common catalog, facilitating inter-library exchange, and at least offer the promise of reducing the multiplicity of copies purchased at every college and university.⁽¹²⁾ In addition, numerous regional buying consortia have emerged to capture scale benefits in purchases of both hard copy and electronic data base information. There remains a major educational effort with faculty of our higher educational institutions about the benefits and costs of using such services.

Fundamentally, however, the issues of journal pricing can only be resolved by systematic and widespread introduction of vigorous competition into the world of publishing, competition that has not emerged until recently in part because the producers of scholarly manuscripts (faculty) and their institutions (universities, both individually and collectively) have left the world of scholarly journals entirely to other entities, both commercial and not for profit (learned societies). One key to understanding this is the degree of specialization that has naturally emerged within academic disciplines and consequently in journal publications. Sub-specialization commonly leads to creation of a "field" that can only meaningfully support one journal. There, the commercial interests of the publisher coincide with the academic interests of the scholars. Once a sub-field is established, if the volume of scholarly work increases to the point where it could support two or more journals in the same field, the most likely response will have the original sub-field divide into two or more sub-sub-fields, each with their own journal. The scholars find this appropriate because it focuses their attention on a more limited spectrum of journals, and they also find more ready publication of their work in more highly specialized journals. The commercial aspects of publication also embrace and support this outcome, both through the natural economies of scale of journal publication and also because of the de facto monopoly position that this creates for the journal.

Particularly the commercial sector has made strong and successful efforts to expand the realm of journals that they publish and control, and have succeeded through time in establishing an intellectual stronghold in our worlds of scholarly communication by creating excellent editorial boards, attracting (in part, through lack of viable alternatives) our best scholarly work, and hence staking out the high ground in the intellectual hierarchy of our disciplines. Having achieved that high ground, they are now in a position to exploit the economic value of that reputation, and they do so by raising prices far above production costs, making journal publishing a highly profitable enterprise. They have also succeeded in stifling competition, not only through provision of excellent support for the editorial boards that they have recruited (often the best minds in the field) but also by such tactics as requiring five year non-competing clauses of editors who sign on, or simply by buying the competition.

Ultimately, to introduce effective competition into the world of scholarly publishing, the institutions of higher education must either separately or collectively create effective alternatives that serve all of the functions now provided by print journals, and make these alternatives at least as attractive to scholars (both in their roles as author, referee, and editor) as current journals do.

Effective components of successful competition need not each provide every function now performed by paper journals. New digital technologies allow the decoupling of these functions in ways that make entry into the market easier and cheaper. We must take full advantage of these new technologies wherever possible.

Electronic Journals. One obvious medium of competition comes through electronic journals. Electronic journals arise at a phenomenal rate. Thousands now exist in various format and degree of formality, some as replicas of paper journals, but many as entirely new vehicles for scholarly communication. These can and often do serve the same functions of paper journals, as outlined previously. However, these journals are often viewed with suspicion by authors and their colleagues, either because the portfolio of authors or the editorial board is less famous than those of competing print journals, or because they distrust the permanence of the system (ephemeral archiving), or both. Indeed, access to some electronic journals is bought on a year by year basis, so that the subscription must be continued indefinitely in order to maintain access to earlier years of the journal's history. Put differently, some of these journals do not sell access to previous years' archived work separately from the current-year subscription.

The ARL has introduced an innovative system to encourage electronic journals known as SPARC.⁽¹³⁾ This endeavor offers an important and promising vehicle to introduce further competition into the world of scholarly publishing.

.

Dissemination, Retrieval, and Searching and Indexing. New electronic media improve on some of the functions that paper journals perform, in some cases, hugely. "Servers" such as the system operated at Los Alamos National Labs by Paul Ginsparg, initially for high energy physics, and now for a wider spectrum of physicists, provide almost instant access to new material for scholars around the world; and the capacity to search not only on title, author, keywords, and abstracts, but on whole text, provides massive improvements in ability to find relevant information for scholars in fields served by these resources. While initially limited to pre-prints, this resource now carries, in most cases, copies of manuscripts identical in content (if not format) to those published by the journals.

Existing paper journal publishers clearly recognize the threat from this environment. Some threaten to withhold consideration of manuscripts that have been previously posted on web servers, invoking rules against publishing material that has been "published" previously. The ability of a journal to succeed in this strategy obviously depends on its position in the academic hierarchy: the most sought-after journals are more likely to succeed, since achieving publication in those journals confers important prestige upon the authors. Publishers have also tried (but immediately failed, at least in the case of the Los Alamos server) to deter use of this system as a competitor to their own services by insisting upon removal of a manuscript from the e-server once it is accepted for publication, or (in a later fallback strategy that also failed) actually printed. The Los Alamos server succeeded in breaking these rules because it was in such widespread use within the community of high energy physics that carrying through with the threats would have eliminated all source material for the paper journals.

A proposal from Jane Ginsberg, a law faculty member at Columbia University, would offer the same capability for a much wider array of disciplines. This proposal, known as HYPATIA (after an early Egyptian philosopher, mathematician, and astronomer) has not yet been implemented.

JSTOR and related servers now actively provide access to electronic versions of material previously published in paper format. Any of these generic retrieval functions can be served either directly from a file server or through dissemination of archival technologies (such as DVD).

Ultimately, the permanence of these digital resources remains a concern to some scholars. Some view the matter as trivial, arguing that multiple copies (mirrored sites) provide effectively guaranteed backup of any particular digital resource. Others note that the changes in computer technology that inevitably occur, including shifts in the basic processors, operating systems, and storage media create a systemic problem for archiving. Anybody who questions the importance of this matter might try to find a computer that still reads a 5.25" floppy disk, let alone simply trying to read more modern media (say, 3.5" disks) written in older versions of current software. (Few people realize this, but even such widely used software as Microsoft Word or Corel's WordPerfect are not necessarily backward compatible across versions, particularly for preserving important information such as formatting and spacing.)

Various scholarly groups, including ARL and the Council on Library and Information Resources (CLIR)⁽¹⁴⁾ have important efforts underway to learn more about issues of archiving in the digital age and to disseminate information that has become known.

Decoupled Certification

Ultimately, many observers of the process of scholarly publication feel that the most important step in achieving effective competition for existing journals is to create an alternative mechanism to provide the refereeing/certification process now provided uniquely by the editorial boards of print (and occasionally, electronic) journals. What has become apparent is that digital technologies allow the complete separation of the certification process from the other publication processes (distribution, indexing, archiving). Thus, the Digital Networks and Intellectual Property Management committee of the Association of American Universities (AAU) has begun a series of discussions designed to learn how to bring into existence a set of editorial boards that will perform only the refereeing function, leaving to other mechanisms the distribution and archiving. Whether this approach succeeds or not remains a completely open question.⁽¹⁵⁾

The potential advantages of decoupling appear in several areas. First, it may introduce more competition into the business of journal publication, especially in more highly defined sub-speciality areas where scale economies preclude the introduction of a new paper journal. Put simply, it is easier and less costly to just engage in the certification process than to do that plus undertake the subsequent publication process itself. Second, it will encourage scholars to use digital media for communication (once having achieved the desired certification of quality) as their final method of "publication." This will both enhance the shift towards electronic communication and -- as a secondary consequence -- reduce the reliance on the paper journal world, and hence ultimately the acquisition costs of libraries. I wish to be clear here that the process envisioned in AAU discussions relies completely on supporting the development of various "certification" or editorial bodies, but does not envision that the AAU or any other single organization would actually undertake such work. The goal here is not to establish a "monopoly" certification agency, but rather to support and encourage numerous groups to engage in such work in their own fields.

Despite the potential gains from achieving independent (decoupled) certification, some important issues stand in the way of successful introduction of this approach, including (a) finding mechanisms to support the costs of operating editorial boards (and refereeing systems); (b) inducing credible and "important" figures to serve as editors in this role, rather than in one leading to ultimate publication; and (c) finding parallel mechanisms to provide the other journal-like functions of dissemination, indexing and search, and archiving.

On the last of these points, a variety of mechanisms appears feasible. Easiest and lowest cost is for authors of manuscripts to post their work (in a locked and electronically certified version) on their own web site, so that the editorial board doing the certification can produce a virtual journal by simply posting a table of contents with links to the appropriate web sites. The problem with this, of course, is the ephemeral nature of the postings and links: any user of the Web commonly finds broken links or sites that no longer exist, and the tracking and updating of these links (and maintenance of functional copies of the manuscripts) is not a casual undertaking. This concern alone for lack of permanence makes Web-posting alone an unlikely permanent solution to the problem of scholarly communication, although it may well serve as a good "first step" in the process. The other important issue is that the current journal editorial process does provide for valuable improvement in the quality of scholarly writing, through editorial work, style and format improvement, and the like. Some fields find this more important than others, but any world of future electronic publishing may well involve some form of editorial improvement (before certification, potentially) or some form of archived (paper) publication for the most "important" works published electronically, the latter set obviously being more likely candidates for substantial editorial work and improvement.

A more complete process would link an e-server to the system, providing systematic oversight of the electronic source of the material, much as the Los Alamos server now does for the community of high energy physics. If e-servers and editorial boards are matched one-to-one, this simply becomes a system of e-journals. However, if a common site for the servers can be established to service a wide array of editorial boards, then the process of entering into the editorial fray is simplified. Thus, parallel development of mechanisms to support independent editorial boards and to provide e-server capabilities for distribution and archiving may be important.

Several paths for providing this electronic "server" access seem feasible. Without implicating individual organizations (and I wish to be clear that I have not discussed these issues with any of the named organizations), it seems tractable to expand a JSTOR-like environment to serve as a repository for manuscripts not yet published elsewhere. HYPATIA would serve the same function. The capabilities inherent in the ARL's SPARC system to support e-journals could also be adopted to provide the e-server function for other editorial boards. Ultimately, a wholly decoupled certification process can completely unlink the certification process from the distribution and archival process. And finally, if a national resource is established following David Shulenberger's proposal for a National Electronic Archival Repository (NEAR) the capabilities of that system would also easily serve as repository for manuscripts certified by editorial boards as envisioned in the AAU's decoupling proposal.

IV. Conclusions.

Incorporation of digital media into the realm of scholarly communication will inevitably occur, but the path can vary considerably by time, depending on the efforts and specific steps taken by the participants (including institutions of higher education, learned societies, and commercial publishers). Initially, at least, the digital and paper worlds will surely coexist, but eventually the digital world may will quite likely become dominant. Some steps will speed the transition, others will hinder it. The ultimate equilibrium can also look very different, depending on the steps taken by various parties. In particular, the costs to higher education can differ significantly, depending on how property rights to intellectual innovations (scholarship and research) evolve, and equally important, the benefits in terms of widespread and easy access can differ greatly across various equilibria.

The evolutionary path towards widespread (if not complete) inclusion of digital media for the exchange of scholarly information will occur more rapidly if higher education institutions and organizations, including AAU, ARL, NASALGC, and others, working where possible in concert with learned societies, move the process towards more favorable outcomes. Specific steps to take include a series of changes in the internal practices of our universities and colleges, e.g., in tenure and promotion review, in library procurement practices, and the like, as outlined in Section II of this paper. A series of external steps are also available, including support of consortial efforts (such as OCLC) for cataloging and inter-library loans, development of consortial buying for journals and electronic data bases, support for development of e-journals (such as ARL's SPARC), support for electronic archiving capabilities (such as JSTOR and research fostered by CLIR), and changes in the nature of intellectual quality certification (such as the proposed decoupling mechanisms discussed in Section III).

Ultimately, the long run equilibrium outcome hinges most strongly on the assignment of property rights to intellectual work and the degree of competition in the areas now undertaken by paper journal publication. Currently virtually every scholar signs over all property rights to a manuscript in order to get it published. In exchange, the author receives a series of benefits, most accruing directly to the author (through professional reputation enhancement), including the certification of the manuscript's quality, its editorial improvement, distribution, indexing, and archiving. Virtually all of the costs of this system are paid either directly or indirectly by institutions of higher education (and to some extent, for profit organizations, in some fields of inquiry), through the implicit or explicit subsidization of the editorial process (editor and referee time), journal purchases, and their indexing and storing in our libraries. This disconnection between the incidence of benefits and costs requires change.

Two types of changes will be necessary to bring this system into line with the purposes of universities and colleges. First, we must find ways to introduce competition into every phase of the process that journals once performed as a bundled effort -- quality certification, editorial improvement, distribution, indexing, and archiving. And finally, as a necessary step in creating appropriate competition, we must regain at least partial ownership of the property rights to the intellectual work that we collectively produce as the institutions of higher education in the US and around the world.

1. http://www.arl.org/arl/proceedings/Shulenburger.html

2. http://scilib.ucsd.edu/sio/guide/prices/perpage1.html

3. Zoomorphology, published by Springer. Of the 10 most expensive journals on their list, 6 were published by Gordon and Breach.

4. Journal of Biological Chemistry, published by Am. Society of Biochemistry, Astrophysical Journal, published by the University of Chicago Press, and the Proceedings of the National Academy of Science, published by the National Academy of Science.

5. http://www.wisc.edu/wendt/misc/costben/costben.html.

6. My own field of interest -- the economics of health care -- provides a good example of this process. When I began my work in this area 30 years ago, virtually all of the pertinent work was published in mainline economics journals such as the American Economic Review, Journal of Political Economy, Quarterly Journal of Economics, and the like. Second tier economics journals from regional societies sprung up to expand the scope of publication in economics generally, and occasional work appeared in journals of labor economics, public finance, or related "wide interest" journals, Then came the Journal of Health Economics, Health Economics, and the Journal of Risk and Uncertainty, not to mention journals of health services research such as Health Care Financing Review, Medical Care, Inquiry, and tangentially related journals in public finance, taxation, welfare economics, labor economics, regulation and industrial organization, and even medical journals such as Medical Decision Making that are all pertinent to my work.

7. I will borrow another example from our library at the University of Rochester. Four years ago we began installation of the Voyager electronic catalog and search engine system. We were the first US research university to actually bring Voyager to operational status, although the system is now in widespread use, and has recently been adopted by the National Library of Medicine and the Library of Congress. This replaces and earlier computer-based technology and eventually the historic card catalog, and does much more as well, but the library's operational costs have not declined. This is not a comment about Voyager, but about technology in general.

8. This paper was written during a cross country airplane flight on a laptop computer with computing capabilities immensely greater than those available to entire universities just a a few decades ago. The current machine weights 3 pounds, has 7 hours of battery life, a 255 MHZ Pentium chip, and a 2 Gbyte hard drive. All experience with computing power during the past decades has followed Moore's Law, stating that the costs of achieving a given level of computational capability halve every eight years.

9. In saying this, I wish to be clear that I am not advocating a complete turn-over of library acquisition funds to departments. Two serious concerns block me from adopting such a strategy. First, some journals span multiple departments (economics and business; biology and medical school, philosophy and medical ethics, chemistry and chemical engineering; chemistry and physics, etc.) Second, current faculty will not necessarily reflect appropriate inter-generational issues; continuous collections of periodicals have value for future generations of scholars that current faculty will not necessarily represent in their current-year decision making.

10. Research-Doctorate Programs in the United States: Continuity and Change, National Academy Press, 1995.

11. The indexing function previously carried out by libraries is now generally purchased from external vendors.

12. OCLC (see http://oclc.org) and the CIC (http://cic.net/cic.cic.html) are examples of collaborative consorial approaches.

13. http://www.arl.org/sparc/index.html

14. See http://clir.org for details.

15. A more complete description of the logic underlying this approach and the potential problems appears at http://econ.rochester.edu/Faculty/Phelps/Phelpspaper.html.