In: Wavelength, The Journal of Science, Society, and the Media. Univ. of the West of England, Bristol UK. Issue 10, febr. 1995

Journal's end: scientific communications in the electronic future

Joost Kircz

Most scientific results are announced to the world on the pages of the research journals. The papers which appear there follow a strictly prescribed form, roughly speaking the "method, results, conclusion" structure which we learn at school. But in the light of the new electronic media, the non-linear texts which these facilitate and the overwhelming quantity of publications which the new technologies will only aggravate, it is perhaps time to look for new ways of presenting research. A project at the University of Amsterdam, called "Communications in Physics", is examining the make-up of the modern research paper in order to suggest better ways to structure future publications.

Some history and figures

With the emergence of modern science in the seventeenth century, the need for well-organised, stable and reliable outlets for new research became a necessity. The first scientific journal Le Journal de Scavans was published in Paris in 1665 by Denis de Sallo de la Coudraye. Just three months later, this was followed by the publication of the Philosophical Transactions by the Royal Society of London.

The correspondence between Henry Oldenburg, the first editor of Philosophical Transactions, and Robert Boyle reveals that they were concerned with all the issues of modern scientific publishing. We find remarks like: "The Society alwayes intended. . . ye registring of ye time, when any observation of Expt is first mentioned. . ."(the priority claim); "Whereby the honor of ye invention will be inviolably preserved to all posterity. . ." (the intellectual ownership claim); and "as farre as they may, not doubting of ye Obeservance of ye Old Law, od Suum cuique truere" (proper referencing).

It is well known that the invention of the printing press gave rise to an explosion of information. This was already being seen as potentially problematic by the close of the eighteenth century. A comment in the Neues Medicinisches Wochenblatt fuer Aerzten on the occasion of a new publication in 1789 warned that, "This is truly the decade of the journal, and we should seek to limit their number rather than to increase them, since there can also to be too many periodicals".

In the three hundred years since those first scientific publications, the art of journal publishing has become a mature industry. Today the total world output of scientific publications can be estimated at about one million a year.

If we assume that we started three hundred years ago with one article, then the number of published articles has doubled every fifteen years. This implies that the total number of articles is now roughly twenty million. If nothing happens to disrupt this trend, by the time another fifteen years has passed we will be producing two million papers a year and the total reservoir of articles to be read will have grown to more than forty million.

From the sheer volume of research output, we are clearly dealing with an information overload.

Finding your way

Even in the early days of publishing, all sorts aids were developed to help readers navigate the sea of literature: the first abstract journals appeared around 1714; the first citation indexing, which emerged in law studies, dates back to the early 19th century. According to F.R. Shapiro, the first true citation index appeared in 1860 and dealt with Californian legal cases.

If we analyse the tools a reader needs to find her/his way in the published literature, we can distinguish three types of indicators:

  1. The internal indicators. This is all the information which appears with the document, e.g. author's name(s), institute's name, address, country, all the bibliographic information (e.g. journal name, volume and page numbers, publisher), as well as the title, the abstract, and, last but not least, all the words of the text. All this information can be listed and subsequently loaded in a database. With the appropriate search languages, we can try to find the desired document, provided that we know who and what we want.
  2. The external indicators. These are identifiers which are added later on by specialists to help the reader find the material they want. As all authors write their texts to tell the world about what they think is important at that particular time, many links to other research are not given. The indexers of abstracting services try to add keywords, registration numbers, classification codes and so on, to the documents to make retrievability more "neutral" and less jargon-dependent.
  3. The transmittal indicators. These indicators do not lead the reader to the actual document, but to another one. We are dealing here with the references at the end of the article. They only use the original document as a carrier to point to other documents.

In all cases, however, only concrete questions can be dealt with. All (on-line) indexing services are notoriously difficult to use if we deal with questions which are ill-defined in the context of that particular service. This has resulted in the emergence of a completely new kind of librarian - the information intermediary.

With the arrival of new electronic possibilities, the situation will only worsen. Now, not only is all published material available (in principle at least), but also all the 'grey' literature and preprints. The popularity of computer fileservers coupled to academic networks creates a hype for speed and availability but without tackling the fundamental problem of consumability.

New media, new methods

For the new technology to be worthwhile, new ways must be found of navigating through all the information. In the Communications in Physics project, we try not to follow the technology from behind but to analyse what conceptual changes in communication might be designed, given the intrinsic potentials of modern technology. In a way, we try to define system requirements for the future.

Before exploring ways in which communication can be reconceptualised, we must recognise the implicit ways in which our present day communications are structured. It is tempting to expand here at length on the cultural changes that took place during the transition between orality and literacy, as many of them can serve as metaphors for our time. But, to be brief, the main result of that cultural revolution was the establishment of the single document: a portable essay, telling a complete story with an introduction, a digression, and a conclusion; i.e. an exemplification of the individual results of an individual research project.

In an electronic environment, where (again in principle) everything could be linked to everything else, the need¸ for self-contained documents is less obvious. For that reason, the "Communications in Physics" project will analyse what kinds of information are contained in a scientific publication (and also what kinds are left out), to see if these could be relocated in a number of linked documents.

The project is an interdisciplinary collaboration between the Physics and Astronomy Faculty and the Arts Faculty of the University of Amsterdam. To start with, we will take a coherent sample of articles in molecular physics spanning about fifteen years of related research.

Two PhD students will analyse this corpus. A theoretical physicist will try to define heuristically which types of information are provided (like machine descriptions, theoretical assumptions, mathematical approximations, etc.), as well as what was left out (e.g. the reasoning why a method or material was abandoned in the research group's next article), or was impossible to publish (sound, colour, etc.). A linguist will apply a well-developed rhetorical analysis to the same articles, in order to make an inventory of the argumentative structure and the rhetoric of the works.

Information Modules

Of course, we will try to combine and assimilate both attacks on the corpus of articles. Ultimately, we would like to present a new template for scientific communications: a set of conceptual modules of undefined length which will each have their own characteristics (their own kind of information). This way we can get rid of repetitive information. For instance, the rationale for doing some research does not need to be repeated every time, but can become a pointer to another module, perhaps even one written by somebody else. Only the real additions to scientific knowledge need remain, like new measurements and their experimental or theoretical settings.

If we had such a system of collections of conceptually different modules, we would then also have a new way of searching and retrieving information. Along with the three types of indicators mentioned above, we could ask the electronic library to supply us with only those parts (modules) of research which interested us. Instead of getting full documents which we would have to browse through from start to finish to find the information we wanted, we could go direct to, for instance, the descriptions of the apparatus. The modular approach would make full use of new techniques such as hypertext, creating clusters of related information and stratifying our scientific communications more effectively than in the classical 'linear' essay.

Joost Kircz is associate publisher of the physics programme at Elsevier Science and is leader of the 'Communications in Physics' project at the Physics and Astronomy Faculty, University of Amsterdam.

The 'Communications' in Physics project is supported by the Physica Foundation, the Royal Academy of Science and Arts, The Royal Library, Royal Dutch Shell Research Laboratories, and the publishers Elsevier Science.

Back to the Communication in Physics Project home page with frames
or without frames .

Contact if you have problems with the server.
Last modifications on: 3-5 1996