The future of scientific journals: lessons from the past.

New technologies will soon bring fundamental changes to the process of scientific communication. To understand the path that these changes may take in the future, we need to take a careful look at the past. By examining the history of scientific communication, we can see how new technologies can interact with changes in communication forms. By looking at the complex roles that journals have traditionally played, we can better understand how and when journals may incorporate these new technologies. From these models we can project that electronic journals must meet the basic needs that print journals do, that they will initially maintain many of the features of traditional print journals, that their transformation may be driven by external forces, and that they will be slow in reaching their full potential.

It is clear to all involved--from the most astute students of electronic publishing to the most casual observers--that we are approaching a time when new information technologies will cause profound and elemental changes in scholarly communication. While these changes will eventually affect communication in all areas of scholarship, the sciences seem likely to be affected first. The scientific community has been receptive to the use of new technologies in everything from genetic sequencing to computer modeling of geological changes. Much of the infrastructure needed to support electronic journals is in place, with widespread availability of individual workstations, networks, and a technologically sophisticated user group. At the same time, other incentives--to decrease publication lags, reduce costs, and incorporate new kinds of data--are growing.

Despite these factors, the preeminent form of scientific communication--the scientific journal--has remained largely impervious to the force of new technologies. Peer-reviewed science jounals, presented to readers in an electronic form are still rare. It is true that the internal processing of the texts in science journals has been transformed by automation, yet the final version of these texts shows little evidence of these changes. It is true that bulletin boards and listservs abound on the Internet, yet true electronic journals are far less common, and refereed ones almost nonexistent. Recent estimates indicate that there were 110 scholarly electronic journals in 1991 and 240 in 1993.(1) While the recent rate of growth may seem impressive, the numbers are still low. What are the inhibiting factors that have prevented the full integration of new technologies into the scientific publishing process? What is the prognosis for the future? The answerws to these questions may lie as much in the past as in the future.

We frequently have heard comparisons of early electronic publications with early printed works, which retained many of the conventions used in manuscripts. Nevertheless, more detailed analyses of the advent of print technology, the long history of science journals, the sociology of science, and the study of scientific communication have often been ignored in developing projections about the future of science journals. What might be gained by such an examination of the past? First, we gain an understanding of how technologies can interact with new forms of communication and, second, we gain help in determining the essential functions traditionally performed by science journals. Both offer powerful models for understanding what the future may bring.

The Development of the Scientific Journal

The technological innovations essential to the development of the scientific journal were in place long before the journals themselves appeared in the mid seventeenth century. The most important of these, the introduction of print technology in the late fifteenth century, brought a wide range of changes to virtually every area of life. The widespread use of the printing press resulted in many changes in communication forms.(2) Included among these are a number of features of printed communication that we currently take for granted: the use of alphabetical order for organizing information, the title page, regularly numbered pages, punctuation marks, the indexing of individual works, and the ability to cite previously published works. Most of these features evolved from the greater standardization offered by printed works in comparison to manuscripts, which incorporated intentional and unintentional changes each time they were copied. These new features did not appear immediately but developed gradually.

The emergence of an efficient distribution mechanism was perhaps more important to the origin and success of the scientific journal (and to the development of its predecessors, the personal letter, and the non-scientific periodical). A widespread, reliable postal system was introduced throughout much of Europe during the late sixteenth century. Even here, there is not a clear and immediate cause and effect relationship between technological change and the development of the journal. David Kronick, in the introduction to his work on the history of the science journal, emphasizes this point, "technology itself had very little impact on the periodical's production and distribution. The processes of printing, paper-making and transportation remained remarkably stable throughout this entire period."(3)

Clearly other forces were at work in the mid seventeenth century when the first scientific journals originated. Primary among these were the profound changes taking place in science itself. It is no accident that the scientific journal arose in the midst of the scientific revolution. One of the principal features of this revolution was the development and acceptance of the experimental method as the norm for scientific investigation. Up to this time, science had largely consisted of debates over the virtues of various classical authorities, and expansions or interpretations of these classics, normally published in book form. With the advent of the scientific revolution, scientists were urged to abandon these "little books of men" and to turn their attention to the "great book of Nature."(4) Direct and structured observation of nature was to become the norm for science for the future.

Gradually the practice of observation was refined, and the experimental method emerged. One of the purposes of the new scientific societies established in the seventeenth century was to sponsor public demonstrations of experiments. At the same time the practice of private correspondence among scientists in Europe served to communicate additional experimental results. The experimental approach called for a reporting of small, discrete units of information, rather than in-depth development of broad topics.(5) This type of report was ideally suited to the format of the letter and later to that of the journal. Eventually private correspondence became institutionalized through the efforts of professional intelligencers such as Henry Oldenburg, who produced the first issue of a scientific journal, the Philosophical Transactions of the Royal Society of London, in 1665.(6) Of course journal publication eventually replaced book publication as the primary means of publishing in science, but this was a gradual process.(7)

The full development of the form of the scientific journal was also a slow process. The system of packaging information that resulted from this long evolutionary process continues even now to enhance the ability of journals to communicate effectively. Each journal has its own profile--a combination of focus, community of readers and contributors, and traditions. Based on an assessment of this profile, determined by scanning many titles over an extended period of time, readers determine the value of a particular journal title to their specific interests.(8) Each article, in turn, is presented in a traditional format. This format took close to 150 years to reach a truly mature form(9) and continues to be refined into the present century, with significant changes in the length of articles, the number of references, and in syntax, word choice, and organizational structure.(10) While the resulting structure may be made explicit in section headings and titles, it is often implied only by typography. This typography helps a reader glancing at a journal article to move quickly from one section to another--from abstract to references, back to methodology, etc. In fact, what little evidence we do have about the way people actually read journal articles shows that people do read in this way.(11) In addition, the visual image of the page helps readers to build a mental model that enables them to return quickly to specific sections.(12)

Other features that took years to develop include the structure of citations and the peer review process. The network of citations, which links together the individual pieces of scientific knowledge, was not present in the earliest journals: "Gradually researchers start to recognize the cooperative interlinking of their work ... Informal and irregular recognitions of debt occur throughout the eighteenth century, and in the nineteenth century modern citation practices start to develop."(13) Peer review, so critical to maintaining quality in journal publication, also evolved slowly. While the foundations of peer review were laid in the earliest days of the Philosophical Transactions, the concept was formally initiated in the mid eighteenth century.(14) Only gradually, over the next two hundred years, did scientific journals in a wide variety of fields accept it fully.(15) Even today there are important journals that do not follow this practice.

Functions of Scientific Journals

What do we know of the functions performed by science journals and their development? How can this help us to understand their likely course of evolution in the future? Publishing in printed scientific journals is so thoroughly embedded in the scientific process that it will not be given up easily. Electronic journals must, at the start, at least serve the basic functions that print journals have traditionally served. Once the transition has been made, new technologies may allow us to add new roles, to drop some of the traditional roles, or to fill them in intrinsically different ways.

What are these functions? On this point there is little agrement. Much has been said on this topic in the literatures of the sociology of science, philosophy of science, communication studies, and librarianship. Much of what has been said shows a great deal of cynicism.

While there is much overlap and ambiguity of functions, I would propose a model which includes the following functions, in order of importance:

* Building a collective knowledge base

* Communicating information

* Validating the quality of research

* Distributing rewards

* Building scientific communities(16)

Building a Knowledge Base

If we look at statements of scientists themselves from reminiscences and interviews, we find the frequently stated opinion that journals serve the most basic of all functions in science--the creation of published knowledge. The importance of research to the scientist is obvious, "Research as an activity comes to be 'natural' for [scientists]: they find it self-evident that persons should be excited by discoveries, intensely interested in the detailed working of nature, and committed to the elaboration of theories that are of no use whatever in daily life."(17) Less obvious, perhaps, is the relationship of publication to this central function of research. The job of the scientist is not only to produce knowledge, but to make it publicly available. Some observers even say that it is publication itself, not research, which is central to the scientific process,(18) that research is not complete until published,(19) or that scientific knowledge is defined by being published.(20) In informal conversations with scientists, this author has heard comments to the effect that scientific knowledge does not really exist unless it has been published. Suffice it to say that contributing to public knowledge is an essential function of science and that the central role of science journals is to create this collective knowledge base. In order to fulfill this role, journals must maintain scientific knowledge in a stable form that is publicly available and has the confidence of the scientific community.

While scientists do write in the hope that their contributions will be read, this is really less important than contributing to the public record. Current readers are desired, but future, or latent, readers may also be important. No one has stated this better than Johannes Kepler (though he was referring to a book and not an article), when he wrote, "I am writing a book either for my contemporaries or for posterity. It is all the same to me. It may wait a hundred years for a reader, since God has also waited six thousand years for a witness."(21) Clearly the function of contributing to a knowledge base serves the needs of authors as much as those of readers.

While the testimonials to this effect are numerous, this evidence is only anecdotal or philosophical. What harder evidence, if any, do we have of the importance of this role? Some evidence can be found in the standards for publication in the sciences as opposed to the standards in other areas of scholarship. Harriet Zuckerman and Robert Merton, in a major study of peer review, showed that the rejection rates in science journals were significantly lower than those in the humanities and the social sciences.(22) This can be interpreted as evidence of the tendency within science to publish work that meets some minimum criteria of accuracy and correctness in order to guarantee the construction of a complete knowledge base. As Zuckerman and Merton stated it, the editors of science journals prefer "occasionally to publish papers that do not measure up rather than to overlook work that may turn out to be original and significant."(23) While rejection rates may have grown in more recent years, it appears that the differences among disciplines have been maintained.(24)

Communicating Information

What of the reader in this model and the role of communication? Does it matter if anyone actually reads the individual articles that make up this knowledge base? Certainly most scientists publish with the expectation that their work will be read, confirmed, praised, cited, analyzed, and commented on in future works. This is how an individual work can become an integral part of the collective knowledge base. In this way, the functions of building a knowledge base and communicating information are intrinsically linked. Bruno Latour and Steve Woolgar, who studied a group of biomedical researchers using participant-observer techniques, noted, "Thus members of our laboratory regularly noticed how their own assertions were rejected, borrowed, quoted, ignored, confirmed, or dissolved by others. Some laboratories were seen to be engaged in the frequent manipulation of statements while elsewhere there was thought to be little activity. Some groups produce almost at a loss: they talk and publish, but no one operates on their statements."(25)

It is true that science journals, in their earliest years, did provide for interaction, or give and take on the part of active researchers. Perhaps the most famous example of this is the case of Newton's theories on optics. The resulting debate in the journal literature was so unpleasant that Newton became embittered and refused to publish in journals for the remainder of his career.(26) In this area journals have become increasingly dysfunctional in recent years as lag times in publication have increased.(27)

What of the importance of informal communication? The cynical viewpoint holds that most of the meaningful scientific communication takes place informally, and that journals are redundant, useful only for procuring tenure and grants.

It used to be that scientists learned about what their

colleagues did by reading the journals. Actually they

used to read books, then things moved so fast they read

only papers, then even faster so they read only letters to

the editor in the rapid publication journals. Now they

are moving so fast that they do not read but telephone

each other, and meet at society meetings and confer-

ences, preferably in beautiful hotels in elegant towns

around the world. They get by in what are now called

'invisible colleges' of little groups of peers.... These

groups are very efficient for their purpose, and some-

where along the line, people eventually write up their

work so that graduate students can read it and get to the

research front. By the time it gets published, however, it

is so old that all the good research juice has been

squeezed out of it.(28)

A number of studies over the years have looked at communication and information-seeking behaviors of both scientists and engineers. Virtually all have shown the significance of informal communication.(29) For our purposes, we need to highlight only a few of these. In 1961--1964, the American Psychological Association (APA) sponsored a ground-breaking research project, conducted by William Garvey and Belver Griffith, that outlined the linear process followed by a typicall research project.(30) It showed that the formal publication of research results accounted for only a small percent of the information communicated and that the first formal report normally followed the inception of a research project by at least 18 months. More specifically, more than half of the reports in core journals would be read by less than 1% of a random sample of psychologists and no research report was likely to be read by more than 7%.(31) On the other hand, 40% of authors distributed preprints and 62% distributed reprints after publication.(32) In the late 1960s Thomas Allen, using interviews, surveys, and diaries, studied the information seeking behavior of 33 engineering research groups working on 17 research projects. This study showed that informal sources accounted for 55% of the information seeking events and 74% of the time spent in information-seeking.(33) Diana Crane in her work Invisible Colleges demonstrated the importance of the informal social organization within a discipline in transmitting knowledge.(34) More recent data comes from such sources as the study reported at the 1991 Faxon Institute that surveyed 680 scientists working in the areas of chemistry, genetics, and computer science, and Jan Olsen's 1992 study of 46 chemists, sociologists, and humanists in two different institutions. The Faxon study showed that journals were ranked as the source used most frequently, but that consultations with colleagues were ranked next in frequency of use.(35)

It should be noted that electronic publication has already had a major impact on the conduct of this informal component of scientific communication.

On the brink of intellectual perestroika is that vast pre-

publication phase of scientific inquiry ... it has now

become possible to do all of this in a remarkable new

way that is not only incomparably more thorough and

systematic in its distribution, potentially global in scale

and almost instantaneous in speed, but so unprecedent-

edly interactive that it will substantially restructure the

pursuit of knowledge.(36)

If we look more closely at the literature on informal communication, several important points emerge. First, the amount of informal communication and its importance vary enormously from discipline to discipline. Second, its distribution is limited to a small circle of associates. Third, much of the highly-touted informal communication is actually about the formal literature. Finally, informal communication is qualitatively different from formal communication. Let's look at each of these briefly.

It is clear that the importance of informal communication varies widely from discipline to discipline. Several studies, from Allen's to the Faxon study have emphasized the greater reliance on informal communication by the engineering disciplines as opposed to the scientific disciplines.(37) Further details are provided in a study by Belver Griffith and A. James Miller that examined the characteristics of disciplines likely to rely heavily on informal communication:

High degrees of communication and organization are

associated with: (a) a limited number of institutions

having research facilities; (b) a single specialized organi-

zation containing most researchers in the field; (c) many

student-teacher relationships, especially if most re-

searchers have been trained by a single individual; (d)

long term commitments to research in the area; and (e)

the area being the principal research interest of most

researchers involved.(38)

It is also clear that informal communication is restricted to a relatively small group--although the advent of the Internet has certainly made significant changes in this area.(39) Traditional preprint distribution is limited to a few individuals--the APA study showed that on the average authors distributed only 10 preprints.(40) Crane's and Allen's studies showed a relatively small number of people in any discipline that are active in informal communication.(41) The cynical viewpoint is represented in this case by N. David Mermin's fictional character Prof. W. A. Mozart, who described this aspect of informal communication as "the undemocratic monopolization of cutting edge science by self-selected cliques through the proliferation of preprints as the primary publication procedure."(42)

Perhaps the most interesting finding of the studies of informal communication is that much of this communication is actually about the formal literature. Several studies have shown that in most organizations there are central people who are well acquainted with the literature and serve as important sources of information for others. Allen termed these people "information gatekeepers"(43) while the APA report used the term (somewhat outmoded and sexist) of "information man."(44) The APA report also gives us the concrete information that 13% of readers of any article will tell a colleague about the article.(45) This practice is, of course, institutionalized in the tradition of the "journal club," where faculty and graduate students meet on a regular basis to report on the literature.

Finally, it is clear that formal communication differs in important qualitative ways from informal. Even Richard Feynman, notoriously reluctant to publish his results, acknowledged in his Nobel Prize acceptance speech the difference between a formal publication and the informal communications which precede it: "We have a habit in writing articles published in scientific journals to make the work as finished as possible, to cover up all the tracks, to not worry about the blind alleys or to describe how you had the wrong idea first."(46) In a world in which researchers are barraged by information from all fronts, the process of selection and editing (including self-selection and editing by authors), peer review, and revision that go into the production of a formal journal article provide important filters for readers.

In fact, while studies show that informal communication is important, it appears that the art of reading and browsing journal articles is not yet dead. Olsen's 1992 study showed that chemists still relied heavily on the journal literature, with all respondents using journals at least once a week and more than half using them daily.(47) All the chemists browsed current issues of journals while fewer conducted more focused searches of the literature.(48) A 1990 Elsevier-sponsored study of 550 scientists worldwide showed that 50% browsed journals in their area of expertise weekly, and that 33% browsed on a regular, if less frequent, basis.(49)

Validating Quality

The role of the journal in validating the quality of research is related to the qualitative difference in formal and informal communications. Certainly researchers do not choose to publish formally everything they may report informally. Of the work that is published in less formal, complete form, as in letters journals, not all is brought to full fruition in a more formal report. One study showed that of all the reports published in Physical Review Letters, only 50% later appeared as full journal reports.(50)

What of the peer review process? If the rejection rates are so low, does it mean that the system for maintaining quality is not functioning? The evidence does not point in this direction. First, authors have to a large degree adopted the standards of the reviewers, exercising their own controls.

The experienced professional scientist seldom comes

into conflict with the referees of his papers, not because

he belongs to an inner conspiracy of mutual admiration

but because ... he has internalized the standards that

the referee is trying to enforce and has already antici-

pated most reasonable grounds for criticism; in other

words, he has already learnt to drive with due care and

attention.(51)

Second, even when papers are accpeted, the reviewers may suggest substantive (or stylistic) changes that need to be made. In some cases, reviewers catch important errors, as this scientist, interviewed by Warren Hagstrom, confesses, "A couple of times reviewers for the journals challenged my papers. In one case the reviewer was completely wrong, in the other case I was completely wrong.... In the latter case I was completely wrong and thankful to the reviewer."(52) For this reason, it may actually be dangerous for scientists to participate in unreviewed journals or events.

In recent years many have challenged the ability of journals to maintain quality in the face of increasing accounts of fraud and error. The scientific community, under heavy criticism from the public and the highest levels of government, has claimed that the practice of peer review prevents widespread fraud and deception. On the other hand, the increasing number of highly-publicized cases of fraud have made it clear that the system is failing in some important ways. As science has become more competitive in recent decades, the pressures on scientists to create fraudulent data have increased. At the same time, as science becomes increasingly fragmented, it becomes more difficult to find qualified reviewers. Reviewers often do not have the skills or take the time to double check the details of all reported results. In fact, while peer review can monitor accuracy and quality to a certain extent, the scientific community has always relied on a great deal of self-policing to control fraud. The entire system of peer review, based on a delicately-balanced combination of trust and volunteer labor, may be on the way to becoming dysfunctional, yet no new methods are readily apparent at this time.(53)

Distributing Rewards--Priority, Recognition, Tenure, and Grants

What of the role of journal publication in securing rewards--be they tenure, grants, or simply recognition? Again, there is a cynical viewpoint, this time expressed by Harold Wooster.

It turns out that the real problem is not in substituting for

journal publication, which is a fairly simple technical mat-

ter, but in substituting for the prestige arising from journal

publication, which is something else again. I have occasion-

ally proposed that this problem could be solved by author-

izing the central computer to issue Brownie points which

by universal academic convention would be fully substi-

tutable for items in personal bibliographies.(54)

Journals do serve to establish priority in research, and the protection of such priority claims is important to researchers. One of the ironies of the origins of the science journal is that the development of a system to safeguard claims of priority was a prerequisite for making the publication of results in the scientific journal possible. Prior to the appearance of the journal, scientists and technicians alike had carefully guarded their secrets until their claim to priority was well established and any significant material gains had been realized. With Oldenburg's establishment of the Philosophical Transactions, a public means of certifying priority in science was established.

Oldenburg's correspondents understood that the contents

of their letters were likely to be read or summarized at the

society's meetings and that abstracts of them would be

entered in its register. This provided a means for estab-

lishing priority in scientific discovery, which, along with

the recognition and status conferred by the leading scien-

tific society in Europe, induced many natural philosophers

to accept the new norm of free communication of scientific

information.(55)

W. Hagstrom, in his work The Scientific Community, has examined the importance of priority and recognition in some detail. Hagstrom's basic thesis is that scientists produce publications as gifts to the community in exchange for the rewards of recognition and acknowledgment. However, the rewards go only to those who are first. "The system of incentives in science does not encourage workers to devote their efforts to repeating past accomplishments when the record of such accomplishments is available in libraries."(56) At the same time Hagstrom argues against the importance of the rewards of money and position.

It is alleged that scientists publish, select problems, and

select methods in order to maximize these rewards. Uni-

versity policies that base advancement and salary on

quantity of publication sometimes seem to imply that

this is true, that scientists' research contributions are not

freely given gifts at all but are, instead, services in return

for salary ... an explanation of scientific behavior in

terms of extrinsic rewards is weakened by the fact that

many scientists in elite positions, whose extrinsic re-

wards will be unaffected by their behavior, continue to

be highly productive and to conform to scientific goals

and norms.(57)

By at least one measure--the competition for certain federal grants--it does appear that science in the U.S. has become more competitive in the decades since Hagstrom's work was published.(58) Most contemporary observers emphasize importance of the extrinsic rewards of grants and positions. "The reward system for scholars and scientists depends for now on traditional publication as a defining criterion for rank and status, with the real compensation for publication coming not from sales of the material itself but from the advancement in rank, salary, and prestige that publication makes possible."(59) One useful way of viewing this issue is provided by Latour, who outlines a "cycle of credibility" in which prestige, recognition, positions, and grants are intrinsically linked together.(60)

Building Scientific Communities

Finally journals often serve to cement together a group of researchers, an invisible college, in many ways. "The very existence of a journal implies a degree of sociability amongst those who subscribe to it. The hallmark of a new discipline is the establishment of a specialized journal catering to the scholarly needs of its exponents. It constitutes an act of solidarity and sodality, and polarizes the subject around it."(61) The introductory pages to the first issue of a journal or the promotional material for it often explicitly discuss this function. Thus in the introductory remarks to Vol. 1 of Cognitive Science, the editor explains, "Recently there has begun to grow a community of people from different disciplines, who find themselves tackling a common set of problems in natural and artificial intelligence.... The work of these researchers is converging toward a coherent point of view that is different from the focus of any of the current journals."(62) As such, the journal takes its place with other manifestations of group spirit found in meetings, conferences, and retreats, and, now, via electronic bulletin boards.

Conclusions

What are the lessons to be learned from this examination of the development of the scientific journal and its essential functions?

* Enabling technologies may not be sufficient to bring about major changes in communication forms. Technology had little to do with the origins of the scientific journal. It was the changes in science itself that drove the development of this new form of communication. At the present time new information technologies are at hand, and an efficient delivery mechanism is now available in the Internet. Nevertheless, it is the external changes, the changes in science itself, or aspects of the current system that are becoming dysfunctional, which may finally bring about the shift to electronics. Electronic journals that can make papers available immediately after peer review, that can eliminate backlogs, that can recapture some of the interactive qualities of early journals, or that can allow researchers to present new computer simulations will have an enormous appeal.

* New forms of communication are slow to develop and to take full advantage of new capabilities. The fully mature forms of journal articles and printed books took years to develop. We should expect the same from electronic journals. While the electronic journal of the future may bring exciting new features and capabilities and may dramatically change the way we interact with information, we should not expect it to appear full-blown overnight. We can envision an electronic journal that will allow us to incorporate hypertext links within documents and among different documents. We can imagine a journal that will incorporate video and sound. We can look forward to transferring a mathematical equation from a journal article to a local system and manipulating it or taking an algorithm from a journal and animating it. We might even expect the journal of the future to seek out its audience rather than vice versa. All of these features, and many more, will probably be part of the journal of the future. However, we should expect the earliest electronic journals to mimic their printed ancestors much more closely, just as the earliest books resembled manuscripts and the earliest scientific journals resembled personal letters. Once the user community has embraced electronic journals in a more familiar form, more advanced features may begin to become available. This will probably occur only after a lengthy and somewhat painful period of experimentation when we will have to deal with a multiplicity of forms and interfaces.

* Authors must have confidence in electronic journals' ability to serve as public knowledge. This central function of journals is thoroughly embedded in the scientific process and will not be given up easily. The peer review process and the validation of quality must be maintained. This may prove to be a difficult hurdle for electronic journals to overcome, because of their association with other forms of non-refereed publications presently distributed over the Internet. After an initial infatuation with the variety of information resources available on the Internet, many scholars are becoming disenchanted with the quality of much of this material.(63) True refereed electronic journals, while they will probably be distributed on the network, will need to disassociate themselves from many of the electronic bulletin boards currently available. At the outset, for these reasons, most successful electronic journals will probably be electronic versions of existing print journals that researchers already know and respect.

Scholars must also be convinced of the permanence and stability of electronic journals. They must feel that their contributions will remain through the ages, available to potential readers, and that these contributions will be protected from corruption, intentional or unintentional. They must also feel that the results of their research are publicly available, and not accessible only by a small elite.

* The information content carried in the structure of the current system must not be lost (at least until technology can provide alternatives). The grouping of articles into discrete journal titles with distinctive identities, and the format of individual articles, add structure and meaning to the body of scientific knowledge in subtle ways that are not yet thoroughly understood.(64) Electronic journals will need to maintain many of these structural elements. We cannot expect massive new databases of individual articles to be useful; journal titles will need to be maintained. We cannot expect readers to page through screens of plain ASCII text stripped bare of structure and of much of its meaning. The structure of documents serves to orient readers; readers must be able to move from one section of article to another easily, and they must be able to locate and re-read sections efficiently. For these reasons, page images, including graphics and typography, must be recreated on screen. This will be expensive and slow, and such images will need to be displayed on large-screen graphics terminals. These factors will limit the speed at which electronic journals will be fully accepted. Over the long run advances in technology, bringing totally new forms of searching, displaying, and linking information, may allow us to relinquish of some of these structural elements. Our whole way of viewing and interacting with information may someday undergo changes at the most essential level, permitting dramatically new forms to evolve.

* Electronic journals must be able to serve the social needs of sub-disciplines of scholars. The track record for electronic publications is already good in this respect. New communities are being built and maintained constantly on the networks. There is some chance that this function of journals may be taken over by the more informal publications already available on the networks.

The future of electronic journals will hinge on many additional factors that have not been touched on here, such as economics and copyright. Nevertheless, unless electronic journals can meet the most basic needs of researchers and readers that have been satisfied by print journals for almost 350 years, they will not be successful. When and how they will be able to do so is still an open question.

References and Notes

(1.)Ann Okerson, "Portrait of the Electronic Journals World" (paper presented at the International Conference on Refereed Electronic Journals, Winnipeg, Manitoba, October 1, 1993). The conference proceedings and other notes are being prepared for access via FTP from ftp.cc.umanitoba.ca/e-journal.

(2.)Elizabeth Eisenstein, The Printing Press as an Agent of Change (Cambridge: Cambridge Univ. Pr., 1979).

(3.)David A. Kronick, A History of Scientific and Technical Periodicals: The Origins and Development of the Scientific and Technical Press, 1665--1790 (Metuchen, N.J.:Scarecrow, 1976), 47--48.

(4.)George Sarton, Six Wings: Men of Science in the Renaissance (Bloomington, Ind.: Indiana Univ. Pr., 1957), 6.

(5.)Kronick, A History of Scientific and Technical Periodicals, 45.

(6.)The Philosophical Transactions of the Royal Society of London was actually preceded by several months by the first issues of the Journal des Scavans, published in Paris. However, because of the more general nature of the latter, the Philosophical Transactions is generally considered the first truly scientific journal.

(7.)Charles Bazerman, Shaping Written Knowledge: The Genre and Activity of the Experimental Article in Science (Madison, Wisc.: Univ. of Wisconsin Pr., 1988), chapters 4 and 5.

(8.)Jan Olsen, "Implications of Electronic Journal Literature for Scholars" (Ph.D. dissertation, Cornell University, 1992), 66.

(9.)For a thorough discussion of this issue see Bazerman, Shaping Written Knowledge, and Steven Shapin, "Pump and Circumstance: Robert Boyle's Literary Technology," Social Studies of Science 14 (1984): 481--520.

(10.)Bazerman, Shaping Written Knowledge, chapter 6.

(11.)Andrew Dillon, John Richardson, and Cliff McKnight, "Towards the Development of a Full-text, Searchable Database: Implications from a Study of Journal Usage," British Journal of Academic Librarianship 3 (Spring 1988): 37--48; and Olsen, Implications of Electronic Journal Literature for Scholars.

(12.)Olsen, "Implications of Electronic Journal Literature for Scholars," 52.

(13.)Bazerman, Shaping Written Knowledge, 139.

(14.)David Kronick, "Peer Review in 18th-Century Scientific Journalism," Journal of the American Medical Association 263 (March 9, 1990): 1321.

(15.)John Burnham, "The Evolution of Editorial Peer Review," Journal of the American Medical Association 263 (March 9, 1990): 1323--29.

(16.)For some of the many other models which have been proposed, see: AAU Task Force on a National Strategy for Managing Scientific and Technological Information, Appendix B, 1994; David F. Brailsford, "Adobe Acrobat--the Electronic Document Catalyst?" (paper presented at the International Conference on Refereed Electronic Journals, Winnipeg, Manitoba, 1993); Arthur Herschman, "The Primary Journal: Past, Present, and Future," Journal of Chemical Documentation 10 (Feb. 1970): 37--42; Marcel C. Lafollette, Stealing into Print: Fraud, Plagiarism, and Misconduct in Scientific Publishing (Berkeley, Calif.: Univ. of California Pr., 1992); and David Rodgers, "Maintaining Scholarly Quality in Electronic Journals" (paper presented at the International Conference on Refereed Electronic Journals, Winnipeg, Manitoba, 1993), Washington, D.C.: Association for Research Libraries, 1994.

(17.)Warren Hagstrom, The Scientific Community (New York: Basic Books, 1965), 9.

(18.)Bruno Latour and Steve Woolgar, Laboratory Life; The Construction of Scientific Facts, 2d ed. (Princeton, N.J.: Princeton Univ. Pr., 1986), 71.

(19.)Diana Crane, "The Gatekeepers of Science: Some Factors Affecting the Selection of Articles for Scientific Journals," The American Sociologist (November 1967): 195--201.

(20.)Shapin, "Pump and Circumstance: Robert Boyle's Literary Technology," 484.

(21.)Preface to Harmonice Mundi, Book V, quoted in Arthur Koestler, The Sleepwalkers (New York: MacMillan, 1959), 394.

(22.)Harriet Zuckerman and Robert K. Merton, "Patterns of Evaluation in Science: Institutionalisation, Structure and Functions of the Referee System," in The Scientific Journal, ed. Arthur Jack Meadows (London: Aslib, 1979), 112--46.

(23.)Ibid., 124.

(24.)Janice Beyer, "Editorial Policies and Practices among Leading Journals in Four Scientific Fields," Sociological Quarterly 19 (Winter 1978): 77.

(25.)Latour, Laboratory Life, 87.

(26.)For an interesting discussion of this issue, see Bazerman, Shaping Written Knowledge, chapter 4.

(27.)Olsen, "Implications of Electronic Journal Literature for Scholars."

(28.)Derek J. de Solla Price, Science Since Babylon (New Haven: Yale Univ. Pr., 1961), 126--127. It is interesting to note that these comments were made before the advent of the Internet.

(29.)An excellent, comprehensive review of this field can be found in F. W. Lancaster, Towards a Paperless Information System (New York: Academic Pr., 1978) and in the early volumes of the Annual Review of Information Science and Technology (Medford, N.J.: Learned Information, 1966-- ).

(30.)American Psychological Association, Reports of the American Psychological Association's Project on Scientific Information Exchange in Psychology, vol. 1--3 (Washington, D.C.: American Psychological Association, 1963--1969). The results of this study are also summarized in William D. Garvey and Belver Griffith, "Scientific Information Exchange in Psychology," Science (December 25, 1964): 1655--58.

(31.)American Psychological Association, Reports of the American Psychological Association's Project on Scientific Information Exchange in Psychology, vol. 1, p. 5--9. It should be noted that the field of psychology includes practitioners as well as academics, and that this would result in a lower percentage than that found in a pure research field.

(32.)Ibid. 5--9.

(33.)Thomas J. Allen, Managing the Flow of Technology (Cambridge, Mass.: MIT Pr., 1977), 71.

(34.)Diana Crane, Invisible Colleges (Chicago: Univ. of Chicago Pr., 1972).

(35.)Faxon Institute, An Examination of Work-related Information Acquisition and Usage among Scientific, Medical, and Technical Fields (Westwood, Mass.: Faxon Company, 1991), 9.

(36.)Stevan Harnad, "Scholarly Skywriting and the Prepublication Continuum of Scientific Inquiry," Psychological Science 1 (November 1990): 342.

(37.)Allen, Managing the Flow of Technology, and Faxon Institute, An Examination of Work-related Information Acquisition and Usage among Scientific, Medical, and Technical Fields.

(38.)Belver Griffith and A. James Miller, "Networks of Information Communication among Scientifically Productive Scientists," in Communication among Scientists and Engineers, ed. Carnot Nelson and Donald Pollock (Lexington, Mass.: Heath Lexington Books, 1970), 139.

(39.)David Stodolsky, quoting B. Reid, gives some astonishing figures on the size and growth of network discussion groups. "Comp.groupware had a readership of about 15,000 persons after its first month of operation." David Stodolsky, "Archiving Secure Interactions," Psychological Science 1 (November 1990): 353--354.

(40.)Garvey, "Scientific Information Exchange in Psychology," Science (December 25, 1964): 1656.

(41.)Crane, Invisible Colleges, and Allen, Managing the Flow of Technology.

(42.)N. David Mermin, "What's Wrong in Computopia?" Physics Today (April 1992): 9--11.

(43.)Allen, Managing the Flow of Technology.

(44.)American Psychological Association, Reports of the American Psychological Association's Project on Scientific Information Exchange in Psychology, vol. 1, p. 12.

(45.)Ibid., 245.

(46.)James Gleich, Genius (New York: Pantheon Books, 1992), 380.

(47.)Olsen, "Implications of Electronic Journal Literature for Scholars," 28.

(48.)Ibid., 30.

(49.)Elsevier Science Publishers, Information on Information (New York: Elsevier Science Publishers, 1990).

(50.)Joseph H. Kuney, "New Developments in Primary Journal Publication," Journal of Chemical Documentation 10 (Feb. 1970): 44.

(51.)J. M. Ziman, Public Knowledge; An Essay Concerning the Social Dimension of Science (London: Cambridge Univ. Pr., 1968), 115.

(52.)Hagstrom, The Scientific Community, 26.

(53.)For an interesting discussion of this topic see Lafollette, Stealing into Print.

(54.)Harold Wooster, "The Future of Scientific Publishing--Or, What Will Scientists be Doing for Brownie Points?" in The Scientific Journal, 67.

(55.)William Eamon, "Form the Secrets of Nature to Public Knowledge," in Reappraisals of the Scientific Revolution, ed. David C. Lindberg and Robert S. Westman (Cambridge: Cambridge Univ. Pr., 1990), 354.

(56.)Hagstrom, The Scientific Community, 69.

(57.)Ibid., 19.

(58.)Daryl E. Chubin and Edward J. Hackett, Peerless Science: Peer Review and U.S. Science Policy, SUNY Series in Science, Technology, and Society, ed. Sal Restivo (Albany, N.Y.: State Univ. of New York Pr., 1990), 26.

(59.)Anthony M. Cummings, Marcia L. Witte, William G. Bowne, Laura O. Lazarus, and Richard H. Ekman, University Libraries and Scholarly Communication (Washington, D.C.: Association of Research Libraries, 1992), xxv.

(60.)Latour, Laboratory Life, 207.

(61.)Ziman, Public Knowledge, 105.

(62.)Allan Collins, "Why Cognitive Science," Cognitive Science 1 (Jan. 1977): 1--2.

(63.)Faxon Institute, "Impact of Electronic Media on Scientific Research, Communication and Collaboration" (Westwood, Mass.: Faxon Company, 1992).

(64.)For a more detailed discussion of these issues see Olsen, "Implications of Electronic Journal Literature for Scholars."

Gale Document Number:A15989455