Roger L. Presley, Column Editor
Serials Review, Vol:25(3) p.67-80, 1999
Interview with Dr. Eugene Garfield
Conducted on Wednesday, April 7, 1999
between San Antonio, Texas
And Philadelphia, Pennsylvania
Interviewer: Beatrice Caraway
Presley is Associate University Librarian for Resource Management, William Russell Pullen Library, Georgia State University, Atlanta, GA 30303-3202; e-mail:firstname.lastname@example.org: and Caraway is Serials Cataloger; Coates Library, Trinity University, San Antonio, TX 78212-7200; e-mail: email@example.com
Acknowledgement: The author would like to acknowledge the assistance of Carol Gill, Trinity University Library, who compiled the introductory information on Dr. Garfield using information found on his personal website : www.eugenegarfield.com
Introduction to Eugene Garfieldís Career
Eugene Garfield's name is one of the most readily recognized in the area of library and information science. He was born in New York City in 1925. He earned a BS in chemistry and an MS in library science from Columbia University, and a Ph.D. in structural linguistics at the University of Pennsylvania. Dr. Garfield's career in scientific communication and information science began in 1951 when he joined the Welch Medical Indexing Project at Johns Hopkins University. Funded by the Army Medical Library, the predecessor of the National Library of Medicine, the goals of the project were to examine problems of medical information retrieval and apply new methods to the indexing of bio-medical literature. One of the key objectives was to improve the currency of the Current List of Medical Literature through machine methods of compilation. The Current List of Medical Literature has since evolved into the present Index Medicus.
One of Dr. Garfield's contributions involved the revision of the subject heading authority list used to produce the Current List. The more than 30,000 subject heading terms were transferred to punched cards for machine sorting and the lists they produced became the first Subject Heading Authority List, the prototype for the Medical Subject Headings used to produce the present Index Medicus. The Welch Project also laid the foundations for the Medical Literature Analysis and Retrieval System (MEDLARS) and MEDLINE information services of the National Library of Medicine.
His work on the Welch Project led to Dr. Garfield's interest in how to help researchers stay informed about currently published articles and how to produce more timely and less costly access to published information through the use of machine generated indexes. In order to pursue these interests, Dr. Garfield earned a Master's degree in library science from Columbia University in 1954. After graduating, he became a consultant to the pharmaceutical industry and later formed his own company, Eugene Garfield Associates. In 1955 he produced a contents-page publication covering the social sciences and management literature with Bell Laboratories as the first major corporate client. In 1957 he began a similar service covering the literature of interest to pharmaceutical companies, Current Contents/Pharmaco-Medica, Chemical & Life Sciences, and the family of Current Contents publications was started.
In a 1955 article in Science (Garfield E. Citation indexes for science: a new dimension in documentation through association of ideas. Science 1955;122: 108-111), Dr. Garfield proposed a citation index for science literature. In 1959, the geneticist Dr. Joshua Lederberg began a correspondence with Dr. Garfield following up on the idea of a citation index (www.profiles.nlm.nih.gov). Their correspondence led to a meeting with the genetics study section at NIH and funding to produce a Genetics Citation Index which would include a multi-disciplinary index to the literature of 1961. When NIH and NSF declined to publish the index Dr. Garfield's company, now called the Institute for Scientific Information, began regular publication of the Science Citation Index in 1964. SCI distinguished itself from other indexes by being current, comprehensive, and multidisciplinary. Most importantly, the SCI uniquely indexed the references cited in the articles it indexed so users could take advantage of the associations and connections to earlier work that researchers made through the references cited in their papers.
Since 1961, Dr. Garfield's career has been marked by the continued development of new information tools for researchers. The new tools includeIndex Chemicus, Current Chemical Reactions, the now seven editions of Current Contents, the Genuine Article (rapid document delivery service), Social Sciences Citation Index, Arts and Humanities Citation Index, Index to Scientific and Technical Proceedings and Books, Index to Scientific Reviews, and others. In 1986, he founded The Scientist, a bi-weekly newspaper for researchers that is available in full text and at no charge on the Internet.
While he has been the Chairman Emeritus of the Institute for Scientific Information since 1993 [in 1992, ISI was acquired by Thomson Business Information, a subsidiary of The Thomson Corporation], Dr. Garfield continues to be active in scientific communication and information science. He maintains a heavy schedule of invited speeches and presentations at high-level medical, scientific, and information symposia and conferences. His topics have included science education, peer review, research evaluation, future trends in medical information and documentation, the economic and social impact of basic research, and other subjects [http://www.garfield.library.upenn.edu]
Beatrice Caraway: The uses of your citation indexes have ranged widely--on the once hand, as a tool for scholars to use in identifying relevant papers, avoiding duplications of research, assessing the multidisciplinary influence of their papers, perhaps sparking new ideas through unexpected associations. On the other hand, they have also allowed people to study the way scientific information is communicated, how the research output of different countries, institutions, and so forth compares. Thus, they lend themselves to use by historians and sociologists of science. Which of these two broad uses (a tool to aid actual research or a means to study that scholarly activity itself) is nearer and dearer to you?
EG: My initial objective in working on citation indexing was strictly for information retrieval. That was never artificially separated from my interest in its implications in doing historical research. I was interested in the history of science even when I was at Johns Hopkins. In the earliest papers we ever did, we pointed out that in order to do interesting history and other kinds of studies, weíd need to have effective information retrieval. This is why it was necessary. It wasnít very long after we got started that we realized how serious the bibliometric and scientometric implications were. Still, I always stressed in the beginning the use of the Science Citation Index (SCI) for information retrieval. It is unfortunate that itís been so successful, if I can put it that way, in science policy uses. A lot of people donít even seem to recognize that itís there for information retrieval. Now, I myself donít really even fully know how much use is made for either purpose. How are you going to measure that? I had to rely on the fact that people renewed their subscriptions. That told me something. And considering the price of this product, it indicated that some portion of the scientific community was using it.
The field of scientometrics, however, absolutely began with this tool. Itís a standard measuring rod for people who do bibliometric studies. And as a matter of fact, just the other day, I found an important paper thatís up on the Web saying that the Social Sciences Citation Index (SSCI), as far as it goes, is a standard measuring tool. Unfortunately, unlike the SCI, it still has a way to go in order to be universal, because it doesnít completely cover a lot of the local language material that some people would like to see.[Baldi, Stéphane, "Normative Versus Social Constructivist Processes in the Allocation of Citations: A Network-Analytic Model," American Sociological Review 63:829-846 (December, 1998)].
BC: So the SSCI has narrower coverage linguistically than the SCI?
EG: -- there are foreign language journals covered in the SSCI, but coverage is not as comprehensive.
In the beginning journal selection was "biased" towards the English language -- not American, but English. I published a paper recently challenging the allegation that there is an American bias in the database. [Garfield, E. " A statistically valid definition of bias is needed to determine whether the Science Citation Index discriminates against Third World journals," Current Science, 73 (8) p.639-641, 1997.
As I said, if there is any bias, itís an English language bias. Therefore, the bias works against material published in Chinese, Japanese, Russian, and the other difficult, shall we say exotic languages. But nobody really knows how much is published in those languages. Furthermore, any of that material can be cited in the journals covered in the SCI. The question is, how much is published in the English language that doesnít cite the most significant material in these other languages? Now in the social sciences itís somewhat different. In most countries much of their social science material will be published in the local language. Mind you .... there are foreign language journals covered in the SSCI, but coverage is not as comprehensive.
BC: I believe you once wrote that there should be more translation into English of this foreign research, more long abstracts in English, or more non-English-speaking researchers writing their papers in English. ("The Foreign Language barrier: Problems in Scientific Communication" [book review], Nature, 303:554 (1983))
EG: Yes. The question remains, however: Is that material of interest outside of those countries? For example, very few people outside of Scandinavian can read Norwegian. So if those journals are not translated or generously abstracted, few could read them. And publishing it in English is no guarantee that itíll get cited, but it may get read or noticed. It may not be cited but at least there will be awareness.
BC: We donít know either, do we, how many foreign scientists are reporting their research in English?
EG: Yes and no. We donít have precise numbers. But we certainly know the geographical distribution for the English-language material. Every significant country in the world is represented in the ISI database. In my recent AAAS talk--itís on my Web site [http://www.garfield.library.upenn.edu/papers/mapsciworld.html] -- I showed that only three to five percent of whatís in the SCI content now is in a foreign language. Most everything thatís significant is being published in English--both by scientists who live and work in English-speaking countries and also by those from other countries.
Editors of Chinese journals complain to me that itís difficult to get their best scientists to publish in their journals. Why? Because leading Chinese scientists prefer to publish in the leading international journals wherever they are published. English is the lingua franca today. The top journals in the world are published in English. There are few exceptions. And those are usually fully translated, like Angewandte Chemie.
BC: So the Chinese would like to have some means of enforcing publication of Chinese research in Chinese journals?
EG: The Chinese may have a legitimate complaint in that many of their English-language journals are not yet covered in the ISI and other databases. They publish a lot in English. They would also argue that the Chinese-language journals which include titles or long abstracts in English should also be covered. But again, thereís the question of relative value and interest worldwide, and of the level of quality. Local Chinese peer review is one thing, and international peer review is quite another. They are going through developmental problems similar to those that other countries have experienced. I mention China simply because thereís a conference being held in Beijing in November, sponsored by the International Conference of Scientific Editors, where these issues will be discussed. The local authorities have to convince their scientists, as other countries have, that they should publish in English in order to attract more foreign readers. But they pay a price for that.
BC: Which is?
EG: The local journals wonít attract the best research. The French and Germans have their vernacular journals. But the best research material from those countries is published in the international journals[Eugene Garfield, "The Impact Factor and Using It Correctly," Der Unfallchirurg, 48(2):413, June 1998. http://www.garfield/papers/derunfallchirurg_v101(6)p413y1998english.html.]
BC: Will you be attending that conference?
EG: I havenít made up my mind yet. Iíve been to China many times, so itís not a question of tourism for me. Iíve been to several similar meetings in Taiwan and other countries. The editors of journals in China would like us to tell them what they have to do to make their journals qualify for inclusion in the SCI. Getting covered by SCI is not simply a matter of cosmetic issues such as format, adherence to standards, or even publishing in English. There are many English language journals that are not of sufficient import or uniqueness to be covered in the SCI. The ISI database is a research database. The key to being covered is to publish what is new and significant. That is the missing ingredient in many journals. Every country has its practitioner-level journals in different fields. What do they have to say thatís new? Theyíre important, but only for the local scene. This in no way denigrates the value of these journals. Iíve often written about the need for national journals that serve the purposes of local education and dissemination of information, but that has little to do with world impact research. Some Indians have complained that ISI used to cover a lot more Indians journals. Gradually many were dropped because they were of such low quality. But Indian scientists publish much more in the international journals now. With advent of the Internet there is the possibility that everything electronic will be accessible on the Web. The Third World may be able to bypass the indexing and abstracting services. Then the reader will decide what is relevant.
BC: In regard to that, in one of your commentaries recently for The Scientist, you expressed hope that most researchers will eventually create Web sites that supply the full text of their papers and books. ("From Photostats to Home Pages on the World Wide Web," The Scientist, 14 (4): 14, February 15, 1999.)
EG: Absolutely. Iíve got my own. (Available at: http://www.garfield.library.upenn.edu) I think it will be fantastic if researchers create, up-date, and maintain their own Web sites. And there is evidence that is happening.
BC: What do you think the scholarly consequences of that development will be? The economic ones? The consequences for publishing?
EG: I wonít live to see the printed journal completely displaced as a primary means of communication in science and scholarship. But youíre seeing the advent of "dualization." Almost everything is going to be published in both electronic form and in print. The electronic version of these articles has an incredible advantage for information retrieval purposes and for easing the distribution of "reprints." But it doesnít change the basic nature of what is being published. It does facilitate better refereeing and feedback.
The electronic form offers other advantages. Many journals have had restrictive policies on the lengths of their bibliographies. JAMA and other journals made no bones about it. There are editors who say, "Unless itís absolutely essential, donít use more than fifteen references per paper." Now, whoís to say that fifteen is the right number? If they had said fifty, that might be different. Well, if you have a more liberal policy on bibliographic citation, then authors will be inclined to give the readers more background material. There may not be space for complete bibliographic backup in print, but they do have the space electronically -- not unlimited, but much more. In addition, there are supplementary materials that can be published electronically. That's already commonplace. But I do not see the complete transformation from print to electronic. Some things will come out originally in electronic form. Do you know about the Los Alamos database?
BC: Yes, the physics "e-print" or preprint database at the Los Alamos National Laboratory.
EG: What has been the impact of that preprint service on print publication? There is a lot of talk about this helping to change the publishing culture. But The Physical Review, the leading journal of physics in the world, published about fifteen thousand papers per year for the last two years. If there is any impact on the volume of printed publication, I don't see it yet. And this doesn't take into account all the other physics journals which may have published these preprints. So if the preprint database has an effect on the volume of published papers, it is not yet apparent. So far, all they did was to replace the old paper preprint service. Most other fields do not have a comparable tradition; the physicists have always had preprints. So now, if I am a physicist, instead of my mailing or faxing preprints, Iím sending an electronic version. That doesnít change the fundamental nature of publication. It makes it easier to facilitate refereeing, which is a key advantage of the electronic format: More readers can be involved in the refereeing process.
In the new electronic era some journals will fall by the wayside; others will get smaller and some will be purely electronic. But the large high-impact journals are going to remain in print, partly because they are the only trusted archival form and publishers, many of whom are professional societies. Depository libraries will play a greater role whether for the print or the electronic versions.
BC: You said that in your lifetime you expect to see a continuation of this dualization. Would you be willing to speculate beyond that, say thirty or forty years into the future? Do you think that print will persist?
EG: At that point, who could know? The technology may be so convenient that we may not even be able to tell the difference. Maybe at some point everything will become electronic. Some of us will have given up some of these really large printed journals. Itís easy to understand: Consider that the Journal of Biological Chemistry published thirty-five thousand pages last year -- about five thousand papers? The average scientist can no longer deal with that much printed material. Some now say, "This electronic version is good enough for me. I can print out the individual articles I want." But theyíll be paying for the same subscription. I asked Chuck Hancock, the Executive Director of the American Society for Biochemistry and Molecular Biology, how the electronic version has affected their print distribution. Every copy they donít have to print and mail saves money. A weekly issue of JBC is about 800 pages per week. You can just manage to hold an issue in your hands. But many readers still scan each issue. It may still be the most convenient way to scan rapidly. Most readers want their institutional libraries to subscribe. So it may be uneconomical to use print and even the old guard will use the electronic version. Economics will drive the shift. The cost of postage and paper is the main consideration. Authors continue to buy reprints, but in the future, they may only send someone a URL.
BC: Wonít the cost of printing and paper just be transferred to the subscriber, and away from the publisher?
EG: Yes, thatís true. Occasionally I read papers on the screen, but frequently I have to print them off to read at home.
BC: And the paper is easier on your eyes, too, not to mention being able to write marginal notes.
EG: Well, youíll be able to do that electronically.
Given the present state and rate of change in technology itís safe to say that every journal of significance, e.g., the thousand highest impact journals, will be available on the World Wide Web. Most of them will also have an electronic archive going back as much as 100 years. Full text searching and citation indexing will be routine, as will many other capabilities, such as historical mapping.
BC: You suggested earlier that from the very first issues forward, those most prominent journals would be digitized.
EG: I would expect so. The rate of change is so rapid in many fields, that it may not be necessary. Dana Roth [Senior Technical Information Librarian at the California Institute of Technology] questions the value of digitizing entire journal runs. He suggested that the ISI database be used to identify the most-cited articles, and only digitize those?" One way or another, back issues will be digitized--either the whole journals or selected articles cited above a given citation threshold. Roth told me that I said this in one of my old essays. I haven't found it yet. I'm sure he's right. I published dozens of lists of highly cited articles. So if ISI compiled a list of the 25,000 most-cited articles, that could serve as a core library to satisfy most requests for older material.
BC: You closed your 1979 work, Citation Indexing, with these sentences: "As long as scientists and scholars continue to use the instrument we call "papers" as a primary communications medium, citation indexing and analysis will continue to play an increasingly significant role in the management of mankindís knowledge and the processes by which that knowledge is produced. The future of the scholarly paper has often seemed in jeopardy. Whether the technological changes available to the next generation of scholars will undermine the role of the paper in the process of scholarship remains to be seen. My innate optimism gives me hope that it ill not, and that citation indexing will have an increasingly strong, positive influence on scholarship." In your estimation, what is the present state of the "paper" as a medium of communication? Do you believe that the "paper" will continue to be the unit of scholarly communication?
EG: Yes, I do. But Iíd like to see better-condensed papers. The lengthy paper is not always the best. In the social sciences I think people tend to be wordier. But in the scientific electronic archive, much of what you report can be included as attachments, graphic or otherwise.
The paper as the primary communication medium is certainly not in danger. Indeed, there is good evidence that the World Wide Web has strengthened the written paper. As I said earlier, there is no evidence that preprints have changed the role of print. The various societies still publish huge journals. Electronic preprinting has improved refereeing and information retrieval is much easier since full text searching is possible.
BC: And you say the electronic form strengthened the paper because it provides additional access, is that correct?
EG: Many people worry that the newer generation is not providing adequate documentation, that they are ignoring the older literature. If it is true, then anything that helps counteract that development is good. It is a lot easier for editors and referees to search the literature and say, "Oh, this was done twenty years ago," and easily back it up with a reference so that the older literature can be accessed.
BC: But the easier it is for them to do that, the better the chances are that they in fact will consult the literature.
EG: Absolutely. I think thatís what has changed things so much.
BC: Now thatís a new idea to me.
EG: There is only so much human energy. When you are refereeing a paper, youíve only got so much time. As you know, today, itís difficult to check out a citation. But what if it becomes very easy to pull up any reference in full text? You can quickly determine whether the author is citing something that is relevant or out of date or is misquoting or if heís left out a key reference, etc.
BC: In that electronic world, where researchers are publishing their papers and books on the Web, what is going to be the mechanism for tracking citations? How does citation work in the electronic world? Does it function according to the same principles at work in the print world? Or do new principles underlie citation in the electronic world? Iím thinking, for example, of the impermanence of any given version of an electronic paper. Doesnít this greatly complicate citation tracking and analysis?
EG: There will be linking systems that will permit you to track URL citations, either directly as you read a journal or even independently of that through a citation index search. You could do it through the ISI database, which is already on the Web and linked to full text articles. If a paper goes through many versions, each new version will be linked.
BC: Can you talk about that a little more. Itís hard for me to envision how you can track uses of an electronic document.
EG: By "uses" do you mean readership?
BC: No, I mean citations to it.
EG: You can do both. Today, people are doing the equivalent of citation tracking by recording the number of people who consult the URL.
BC: Yes, if there is a counter on the Web site to count the number of hits.
BC: But you donít have a record of what use Iíve made of it.
EG: What do you mean? Citation use?
BC: Yes. If Iíve cited your paper in my electronic paper, which is mounted on my Web site, you wonít have a record of that use unless you happen to read my paper on my Web site. And how do you find that electronic document mounted on my Web site?
EG: Well, I think that itíll take some time, but eventually all of these electronic documents will be linked in such a way that as soon as you cite another paper, the two papers will be linked. For example, right now itís already happening with High Wire Press. If you go into High Wire and look for an article, it will tell you where that article has been cited in the other High Wire journals.
BC: So thereís a web of links among their journals. Youíre suggesting that eventually there will be a web of links within all electronic documents citing each other, or rather linking to each other as they cite each other.
EG: Exactly. Of course, there will. Citation indexing has always been what I call "hypersearch," the same thing we are discussing now. It was more cumbersome in print. We called it "cycling." Now, hyperseach is a reality.
BC: But the citation in electronic documents will have to consist of more than the chronological and numeric designation, page number, and authorís name and title in order to be tracked or counted in a citation index. It will have to have an actual hyperlink. Do I make myself clear? If Iím writing an electronic document and I cite something that youíve written, but I donít make a link from what Iíve written to what youíve written in electronic format, how do you know that Iíve cited you?
EG: Well, when you put that document in electronic form, and it goes into a database such as High Wire, links will be made between citations and the papers cited.
BC: The editors, then, go in and make the links?
EG: The system does it. In other words, go into a document at High Wire and look for a paper that was published two or three years ago. It will say that this paper has been cited in such and such another paper. Well, those links were not created by each author. Just like the links to all the Medline records. Those links are created by a system. And provided that your original document was recorded somewhere, then there can be a link to it. Now if itís a document that never was recorded anywhere electronically, there canít be a link. But if you go back and start scanning in the old literature, you'll include all the cited references as links. Eventually theyíll do that, too.
EG: Well, theyíre doing it now. Thatís why you have JSTOR.
BC: A complete retrospective digitization?
EG: For the key journals, yes. JSTOR is going to do 100 years of Science. Theyíve done quite a few economics journals going back to the beginning.
BC: So links will be created as they digitize, so that there will be links to all the works cited?
EG: Right. And, as a matter of fact, you can have a hot link between a reference and another paper without having had that particular paper put up on the Web. If Iíve cited something, then it gets scanned in, you can automatically do the citation linking to that reference as long as itís in electronic form. You just have to combine it with everything that has that same digital identifier. Do you know what a CODEN is? The CODEN will bring together all the references to the same document.
BC: OK. Thatís becoming clearer now. I was thinking more in terms of URL addresses, but youíre talking about DOIs, the digital object identifiers?
EG: Well, we used to call them CODENs.
BC: Yes, the creation of Charles Bishop.
EG: I gave an historical talk at the ASIS meeting in November. I would have like to have included Charlie Bishop in that group. [See Serial Reviewís interview with Charles Bishop in v. ?, no. ? (?, 1999).]
BC: What about the problem of the impermanence of any given version of a paper in electronic format? Is that a problem?
I donít know how true that is of electronic papers. I think that the electronic are no less permanent than some library collections in print. People donít keep them forever.
BC: True. But what I am referring to here is that an author could make slight changes to his paper and make the newest version of it accessible, whereas somebody has cited an older version of his paper that is no long on the Web. Thatís the kind of impermanence I had in mind.
EG: I think that it is hard for you to visualize all of that being recorded, but there is nothing that prevents it from being so.
BC: Do you mean, making each subsequent version available?
EG: Yes. Itís like your citing the 1963 version of Derek Priceís book [Derek deSolla Price, Little Science, Big Science, New York: Columbia University Press, 1963, 119 pgs] and my citing the second edition, 1986 [Little Science, Big ScienceÖand Beyond, New York: Columbia University Press, 1986, 301 pgs.]. There is a difference between the two versions but is that always important? If there is, then you would be able to track it down. Just as there can be many editions of a book, there can be many editions of a paper. What difference is there between the two versions? Is it that important? If there were an important difference, you would have a means for tracking it down.
BC: But it is a lot harder to make new editions of a book than to make changes to an electronic paper. That is what I was thinking of. It is so easy to make a change every day, every week.
EG: Well, you can speculate about it. Itís not going to be a perfect system. Eventually, it may present problems. We will probably will have to live with the date that the cited URL was recorded. If somebody changes a version, then they have to record that.
BC: And isnít there something like a document stamp....
EG: ...that eventually may solve that problem. And then...a lot of these things youíre talking about may not warrant being preserved.
BC: I guess thatís always been true.
EG: You know what a palimpsest is. A lot of earlier manuscripts just "disappeared," but were revealed later. Due to the cost of parchments, scholars "erased" earlier writings but they often show up through the new writing. They wrote over the old papers. Underneath the palimpsests, you can see the traces of older texts. The eminent sociologist, Dr. Robert K. Merton, has written about the palimpsest syndrome at great length. [Merton Robert K., On the Shoulders of Giants, New York: Harcourt Brace Jovanovich, 1985, p. 218. See also Eugene Garfield "From Citation Amnesia to Bibliographic Plagiarism, Current Contents No. 23, pgs 5-9, June 9, 1980; reprinted in Essays of and Information Scientist, Volume 4. Philadelphia: ISI Press, pgs 503-9 (1981) http://www.garfield.library.upenn.edu/essays/v4p503y1979-80.pdf-80.pdf].
BC: So youíre drawing a connection between those earlier paper versions that were lost and earlier electronic versions that may be lost.
EG: Essentially youíre talking about the palimpsestic syndrome at a textual level. But Merton covers even more general forms of false attribution of original sources.
BC: You have said that editors and scholars "are primarily concerned as to whether authors have acknowledged intellectual debt. Such acknowledgement is widely believed to be fundamental to ethical scholarly behavior." You have expressed interest in the problems of false reporting, of plagiarism, and of negligence in citing, especially among younger scholars. There is much anecdotal evidence that the Web has blurred the lines for young people between their own work and the work of others. Does this represent an erosion in the tradition of acknowledgment of intellectual debt among the youngest scholars? [Give citation]
EG: All of this is anecdotal. There is no significant systematic evidence to show this. All generations ignore the literature to some extent. Itís a question of doing a proper educational job. If mentors teach their students that it is ethically wrong not to cite the literature, not to study the literature, then the problem will be neither more nor less than before. In fact, it may be easier now for editors to verify references. As I said, if itís easier, then maybe people wonít omit so much. I have to see the evidence for all these claims.
BC: I think the same claim is being made at the college level, too, with professors being worried that they wonít be able to identify plagiarism so easily in the electronic world. Others take the stance that itís much easier to identify plagiarism electronically than it is in paper.
EG: Let me tell you what amazes me. Almost daily I get an email from some completely unknown person who says, "I am doing a paper on such and such a subject." The reason they are writing to me is because they have found a reference to The Scientist -- through a crawler, not by going directly to our Web site. They send out blanket requests to anybody who mentions the subject they are interested in. There are certain hot topics that I get requests for all the time. We did an article, several years ago, about Jonas Salk. There are probably 100,000 kids in the country writing term papers about him. So what do they? Instead of going to a library to do their research, they think, "Iíll email anyone who has mentioned Salk. Heíll tell me all about Jonas Salk. He'll do the work for me."
BC: So theyíre actually requesting that you provide them more information on Jonas Salk? That you tell them all about him?
EG: Theyíre writing to me as an authority on Jonas Salk. What do I do? I reply with a simple question: Did you consult a library before you asked me? And if you did, did you search Medline or SCI or CA?
Young people are getting into this habit because, ironically, it often works! You can get a lot of information without doing much work. Itís unbelievable. I must say that in a similar fashion, I do the same --not exactly what those kids do. But those of us who use discussion lists are essentially working the communal brain. Iíve got a question Iím going to put out on the ChemInfo Listserv any day now. If I'm unsuccessful, I might use other listsl. I want to know of any author whoís ever published over a thousand papers. Thatís not an easy thing to look up, but knowledgeable people know names of such individuals, for one reason or another. Using this listserv with a thousand informed members maybe someone will remember. Sometimes Iíve asked a question and received an answer in a few hours.
BC: Your use of the electronic communications is much more judicious than that of the students that you just talked about.
EG: Some just use a crawler and locate an article in our archive. Others simply write because of my affiliation with a scientific institution.
BC: In a similar vein, you wrote "since authors refer to previous material to support, illustrate, or elaborate on a particular point, the act of citing is an expression of the importance of the material." I would like to ask you about what I call "citation padding"óscholars who cite each other. ("Iíll cite you if youíll cite me.) This sort of citation clearly is not made in order to support, illustrate or elaborate a point. Do you believe this is happening to any great extent? Is it more prevalent in some fields than in others? There seem to be no way that citation indexes can distinguish between the two categories. [Give citation]
EG: Again, this is anecdotal. There is no systematic evidence. Someone recently published a paper that proved just the opposite. [Baldi, Stéphane, "Normative Versus Social Constructivist Processes in the Allocation of Citations: A Network-Analytic Model," American Sociological Review 63:829-846 (December, 1998)].
They did a very large study demonstrating that the references cited in biochemistry papers are invariably relevant to the subject, whereas in the social sciences, less so. In the social sciences, there is so much disagreement. You donít have a standardized vocabulary, as it were, a standardized set of references that would automatically come to the fore for certain subjects. Itís a very important study.
BC: Only through a systematic study can one distinguish between those two categories of citation.
EG: Thatís right. It takes a lot of work to prove it.
BC: And a lot of intellectual analysis.
EG: In that same study, there was no evidence that they cited people because they were famous--in science, that is. In social science, it was somewhat different. Also, there was no evidence that they were citing their friends.
Now, self-citation is a separate issue. I always say that I was the greatest self-citer in the world, because when I wrote my essays in Current Contents, I always wanted to bring readers up to date on what we had done before, so I cited all the previous essays on that particular subject.
BC: Are mergers and acquisitions in publishing affecting scholarly communication?
EG: Formerly, a scientist might have a dozen different publishers to choose from. Today, the choices may be more limited. The library profession perceives that some large publishers are virtual monopolies. On the one hand, societies like the American Chemical Society have a dominant position. They do have some competition, since other publishers and societies publish chemical journals and information services. But too much concentration of power can be detrimental to change. In the past, the virtual monopoly position of the British Chemical Society, frustrated organic chemists in that country. This led to the foundation of Tetrahedron and other journals established by Pergamon Press and now part of Elsevier.
BC: I see. I wasnít aware that that sort of society monopoly had engendered commercial publishing in some areas.
EG: Yes. Youíve got to remember, when Tetrahedron was started, the Chemical Society wasnít able or willing to accommodate enough of the papers that people wanted to publish. In the fifties, there were thousands of papers published on steroids. They werenít responsive or fast enough. However, if these societies continue to stay on their toes and meet the needs of members, they will not provide opportunities for competitors to move in. Many journals are started by individuals who recognize the need for twigging a new journal. Even if the number of publishers is narrowed down, there will always be entrepreneurs who will recognize the need for a new niche. Some of these niches grow very rapidly, as was the case with Cell. However, Cell has now been sold to Elsevier. Librarians donít particularly like to be beholden to a few large publishers. In some subject areas, you don't always have a choice. There are certain journals that compete with each other, but when they reach a critical mass and become top journals, competition is not really a significant issue. Some journals get too large. At some point, they start branching out -- twigging -- and that may partially solve the problem. Publishers have to be ready and willing to jump in to accommodate the new twig.
BC: In the time we have left, I wonder if you might talk a little bit about The Scientist. Itís now in its thirteenth year and you are the editor-in-chief and president. Could you tell us something about your work today at The Scientist?
EG: I was the publisher until about a year and a half ago. The new publisher is Alexander Grimwade, a biochemist and former colleague of mine at ISI. He was in charge of the Atlas of Science which ISI published for a short time but unfortunately abandoned before it could break even. At The Scientist, I attend weekly editorial and executive committee meetings. I read every single of copy. Itís sent to me both electronically and in print. My editorial role is varied -- I scan a lot of literature and often suggest names and topics for commentaries or stories. We are always seeking novel ideas for stories.
BC: And are you still writing commentaries from time to time?
EG: As a matter of fact, I have one coming up on the topic of "prior publication." Many editors today claim that if youíve posted an article on the Web, it is prior publication. I believe thatís not a proper reason to deny somebody publication in print. (My editorial will appear in a special issue on science publishing.)
BC: For our readers who may not be familiar with The Scientist, could you take a moment to characterize the niche that that publication fills?
EG: Long ago I wanted to publish a daily newspaper of science. That history goes back thirty years, but I wonít go into detail. [http://www.garfield.library.upenn.edu/essays/v9p249y1986.pdf#xml]. The Scientist is a bi-weekly news journal of science research; it is not a primary research journal. Therefore, we do not publish original results of research papers. We select what is significant . We also report the human side of science -- science policy, employment, training, gender and race issues -- any human aspect of the scientific community. And, of course, we occasionally inject some bibliometrics or scientometrics, as e.g. in our very popular "Hot Papers" feature. We also like controversy. In our March 29, 1999 issue we asked, "Mixing Religion and Health: Is it Good Science [www.the-scientist.com]?" We solicit opinion papers and commentaries and report controversy. We depend on advertising for most of our support, both display and classified. The Scientist was the first publication to be continuously available full text in electronic format, free of charge. Everything in the print version, except display advertising, is included on the Web. Our full text archive covers ten years and we are in the process of converting the remaining years 1986 to 1989 to html format.
BC: This might be a good introduction to those of our readers who arenít familiar with it.
EG: They can access The Scientist free of charge. We also have a free bi-weekly table of contents service, which you can sign up for at our Web site -- www.the-scientist.com. Librarians can distribute it to as many readers as they wish or they can send us a list of email addresses. We welcome your readers to The Scientist!
BC: The Web of Science is ISIís newest product. What are the most important contributions of the Web of Science?
EG: I think the ease of moving rapidly from item to item makes it possible for users to do things that previously were more cumbersome and time consuming. There really is no reason to miss relevant material now. The SCI always involved what I call "hypersearch." The electronic version removes an enormous impediment to hypersearching. You could always do cycling with the print version but, as with any print index, it took a great amount of patience. Scholarship in the old days took perseverance--lots of manual copying. It was incredibly cumbersome. Any of the standard print tools were difficult to use. Using Chemical Abstracts and BIOSIS was the same. Itís a completely different experience to use these tools electronically. Now the Web of Science also makes it possible to link references to the full text. So it will play an important role in preventing duplicate research. It will also facilitate historical research in the natural and social sciences.
BC: Is the main difference between the Web of Science and the CD-ROM products the linkage to the full text?
EG: That will only add to its value when all journals are available in full text. But there is another very substantial difference. In the Web of Science youíve got a 25-year database rather than a series of one-year databases on CD-ROM. Searching one year at a time is very different from searching 25 years simultaneoulsy. A major qualitative change occurs, especially with respect to the "related records" function. You can now deploy that command across an entire 25-year file covering SCI, SSCI and AHCI as well. Itís unbelievably powerful. When you used the related record feature on CD-ROM for one year, it was very useful but limited. Now if you are interested in tracing the literature in a field that was developed over a ten-year to twenty-five year period, you can use bibliographic coupling (related records) to cover the entire period. Itís incredible. (Related records is just the ISI invented term for bibliographic coupling.) Similarly, being able to determine instantly how often each paper has been cited for a twenty-five year period is very powerful and facilitates citation analyses.
BC: Would you talk about bibliographic coupling?
EG: That term was coined by Mike Kessler at MIT back in the sixties. We often cite his paper for that [Kessler, M. M. "Bibliographic Coupling Between Scientific Papers," American Documentation 14:10-25 (1963)]. The distinction between bibliographic coupling and co-citation is an important one to make. Many people confuse the two concepts. Bibliographic coupling occurs when Paper A cites a group of papers in common with the papers cited in Paper B. If both papers had the identical bibliography, you would have 100 percent bibliographic coupling. Now thatís very different from co-citation, which means that Papers A and B are subsequently cited in common by other papers. The co-citation strength is a measure of the frequency that the pairs of papers are cited. The primordial paper on co-citation is by Henry Small ("Co-Citation in the Scientific Literature: A New Measure of the Relationship Between Two Documents," Journal of the American Society for Information Science 24:265-269, 1973).
BC: And if two papers have 100 percent bibliographic coupling, then they are very closely related.
EG: That might be a case of plagiarism! The ultimate proof of plagiarism would be that the person who published a paper changed the title and even some of the text but picked up all the references. Thatís an excellent way to detect plagiarism. If you want to uncover plagiarism, thatís one way to do it. Itís another variant on locating duplication of research. Thatís a feature I demonstrated years ago. ("Citation Searches Can Be Powerful Tools in Combating Redundant Publication," The Scientist, 7 (8): 1, or http://www.the-scientist.com/yr1993/apr.comm_930419.html)
BC: My last few questions for you are in a more personal vein. First, let me ask about your Transliterated Dictionary of the Russian Language, Philadelphia: ISI Press, 1979). The foreword indicated you would produce a second volume if the first were successful. (I noticed there are nearly 700 holding libraries in OCLC). Were you pleased with it? Was there a second volume? Does it still have value? Does it still interest you? Would you pursue it again?
EG: There has not been a second volume. There is a paperback edition, as well as the hardback. The paperback is still in print. It still interests me and I believe it still has value. If I found the right assistant, I would pursue it again. Youíve got to remember that I had an enormous amount of help from Robert Hayne who was my colleague for many years. He died many years ago. I had assistance from programmers who understood the project. I would really need to put a notice out on the Web to find someone to help me. I donít think I could manage it by myself since there is a lot of manual effort in locating the new terms to be added. And I suspect there is less Russian science being read today because so much is in English.
BC: It certainly seemed like a valuable tool.
EG: People who use it find it useful. Iím sure there are many situations where it can be useful. It needs to be expanded. The old words are still valid.
BC: [Dr. Garfield has lectured in French, so as a change of pace, I put the following question to him in French. He understood it very well, but modestly declined to answer in French. The question asks: I read with much interest your opinion about French linguistic xenophobia, in particular "Anglophobia," if I may call it that, and about the law proposed in 1994 by Jacques Toubon to limit publication of scientific research to French. You went on to say that, fortunately, that law did not pass, and that, more and more, French scientists and researchers use English. What has happened since then? Are efforts still ongoing to limit the use of English as a scientific language in France?] Puis-je vous poser une question en francais? Jíai lu avec beacoup díinteret votre opinion sure la zenophobie linguistique des Francais et sur la loi proposee in 1994 par Jacques Toubon. Vous avez ecrit que la loi etait, heuresement, tombee, et que de plus en plus, les scientistes et les chercheurs francais utilisaient líanglais. Quíest-ce qui cíest passe depuis? Existent-ils toujours des efforts pour limiter líemploi de líanglais comme langue scientifiqe en France? (See "Multilingual Capability Is Essential in the Global Science Community," The Scientist, 8 (18):13 or http://www.the-scientist.com/yr1994/sep/comm_940919.html)
EG: While there will be francophiles who will always complain about the overabundance of English, I need only remind them that the Scandinavians, Dutch, and Swiss, among others, are all essentially bi- or trilingual. But that has not prevented them from speaking their native languages and to teach them to their children. In one hundred years, French children will be speaking French. They will use a lot of English words and cognates; but there is nothing Francophiles can do to change that. How do you say Internet in French?
By the way, the question of what children speak has nothing to do with whether or not scientists will speak English in the future. Scientists will all speak English and write papers in English if they wish to publish in international journals. Most French scientists accept that. There are very few left who would even question that anymore. There is a chap in Mexico [Bracho-Riquelme, R.L., Pescador-Salas, N, and Reyes-Romero, M.A. "Bibliometric repercussions of adopting English as the language of publication," Revista de Investigacion Clinica 49(5):369-372 (September-October 1997)] who published a paper demonstrating that publishing papers in English or French in one of the French medium-level impact journals didnít make any difference in impact. I agree with that. The reason is that it is primarily local mediocre material. What difference does it make whether French authors publish in French or English in French journals? The French will be able to cite the paper whether itís in French or in English. The real question is how many people will read those articles? And if they read them, are they going to cite them? So, the reason for publishing in English is not necessarily that it is going to increase citation, but it certainly is going to increase readership and awareness. If it is high quality, then it will matter.
There are a lot of papers that ISI covers that I would love to read that are still published in French, Spanish or German. Some of them have abstracts, but many donít, or they are too brief. Many are editorials. I often write to the author, saying that I canít take the time to read in French anymore. I used to read French, German, and Spanish easily. Today itís very difficult. Some authors are kind enough to translate for me. Itís not that difficult for them. The English may not be perfect, but it is more than adequate. I sometimes try to use automatic translators, but they are cumbersome to use.
BC: The automatic translators, you mean?
EG: I would like to see a simple-minded word-for-word translation, using something like a spell checker. That would be useful. You should be able to feed the text in and get a quick word for word look-up. If you had that capability, you could do your own translation much faster. A human being uses the context to infer meaning, but the automatic translators canít. My problem is that my memory gets frayed. Reading a German article, I inevitably come across a word or phrase that I've forgotten, and then I have to get the dictionary. That's too time consuming. If I could just quickly look it up, preferably without having to key it in, it would be great.
BC: So if the paper were in electronic format, you could click on the word you were having trouble with and get a translation of that word. And context would do the rest.
EG: Exactly. Iím sure thatís feasible. I could scan the paper in and get it converted to ASCII text, and then have it translate only the particular words I need.
Going back to the issue of linguistic bias that we talked about before, it simply doesnít exist. There have been a lot of false or mythological statements made about this alleged bias. ISI doesnít say, "This journalís not important because itís in French." Weíve always covered French, German, and other foreign language journals. But how important are they in the scheme of international research? They are, however, useful in other context even to English reader. Any clinical article on a pharmaceutical product is important to the firm that sells it. But it may not add anything to the research front.
BC: Finally, I believe you know the great Jacques Barzun. Past the age of 90, he is just completing a book on the cultural history of the United States. He is a striking example of intellectual vitality and productivity. Like him, you are a man marked by great curiosity and inquiry. What are your thoughts about the intellectual life in your crowning years? Are you parsimonious with your time?
EG: I am probably not as focussed as I used to be. I let many things, like this interview, distract me from major projects. I have to depend on others to get certain things done. I gave up large-scale administration when I sold ISI. Dealing with personnel problems and and hundreds of employees was just too stressful. At one time I had fifteen people on the editorial staff of Current Contents alone. It takes a lot of support to do that level of citation analysis and editorial work. I loved doing it, but I had to stop when my staff was eliminated after the sale. The new owners just didn't want to spend that much on that form of pubic relations and user education.
BC: Now you are able to pursue the things that simply interest you.
EG: I could do what interested me before, but I needed many people to do it. Now, I have to select things I can get done with minimal help.
BC: Do you enjoy the writing you do?
EG: If I have the time. There are so many writing and other tasks that have to be done. If I had the time, I would like to respond to all the people who publish nonsense about citation analysis. But I also have to devote time to The Scientist and to the charitable foundation I set up when I sold ISI.
BC: And what foundation is that?
EG: The Eugene Garfield Foundation.
BC: And what is its purpose?
EG: Itís what we call a plain vanilla foundation. It provides support to various kinds of charities and to a limited number of research projects. It helps support travel grants to some conferences, like the international conference in Colima, Mexico on bibliometrics that is coming up on July 5-8.
BC: And does it support research in the areas of citation analysis?
EG: We have given small grants to a few scholars. That's not its main function. I serve on the Board of Temple University Children's Hospital and the University of Pennsylvania Library. We provide support for certain outreach programs.
BC: Are your papers presently at the University of Pennsylvania Library, or are there any plans for that?
EG: My oral history is at the Chemical Heritage Foundation [http://www.chemheritage.org/perkin/Garfield/garfield.htm]. Thatís where some of my papers probably will go. Penn is my electronic publisher as can be seen in the URL for my website. We have just put up the Lederberg-Garfield correspondence there
[http://www.garfield.library.upenn.edu/lederberg/list.html]. Some is also available at NLM on Profiles of Science [http://www.profiles.nlm.nih.gov]including a mural by the Huichol artist Emeteria Rios Martinez at the University Museum -- see my home page [http://www.garfield.library.upenn.edu]. More recently, ISI donated a mural by Jennifer Bartlett to the Penn Library[http://www.garfield.library.upenn.edu/essays/v5p207y1981-82.pdf].
BC: And what is the subject of that correspondence?
EG: He and I have corresponded since 1959. Heís the one that wrote to me asking what had happened to the idea of a citation index. I wrote back and told him the story. We wrote back and forth over the next several years. He was very helpful in steering me in the right direction. Eventually, I got to the point where I put in an application for a grant to the NIH. He was on the advisory committee, and has been associated with ISI for 35 years, and is chairman of The Scientist's editorial board.
BC: So that correspondence more or less narrates the genesis of the citation indexes.
EG: In a sense. The Web site will also include my correspondence with Derek Price or with others.
BC: I look forward to your putting that on the Web, because I love to read other peopleís letters!
EG: As a matter of fact, there is a very good book that you should know about. Itís a dissertation that was just finished by Paul Wouters, The Citation Culture, published by the University of Amsterdam Press.. [Paul Wouters, The Citation Culture, PhD Thesis, University of Amsterdam, March 1999]. Heís done a comprehensive examination of my correspondence files.
My youngest son is fourteen. It wonít be that long before he is away most of the time. When that happens I can devote more time to the Eugene Garfield Foundation and eventually I want him to play a leading role. We might encourage scholars to engage in citation-based research projects. I would love to see a Book Citation Index created. The Atlas of Science ought to be revisited by someone. Whether ISI will ever do it, I don't know. Most people still donít realize that I have no control over what goes on at ISI. I'm like an emeritus professor at the university who just occupies an office and uses the library and facilities.
BC: Do you have a moment to talk about the Atlas of Science? Iím not familiar with that.
EG: Atlas of Science was a quarterly series, later called Research Reviews, published by ISI. It was based on co-citation mapping of the science literature. There have been many papers published on mapping the world of science. References to some of them can be found on my Web site. Henry Small pioneered those mapping techniques [Dr. Small is Director of Contract Research at the Institute for Scientific Information]. With co-citation mapping, we were able to identify the emerging research fronts in science. Through that technique, we developed a series of volumes in pharmacology, biochemistry, animal and plant science, and immunology. The methodology enabled us to identify research fronts that needed to be reviewed. We commissioned individual authors to write the reviews. They were sent all the necessary bibliographic data and papers on the core of each front. The plan was to expand Research Reviews to cover the whole spectrum of science. The main point was to make sure that every emerging area of science was reviewed. The review literature is critical to the advance of science. There is lots of review literature published, but none is based on systematic coverage. Each topical annual review today may do a fine job. I am on the board of Annual Reviews. Their editorial boards have the highest reputations. Their selections are based on the deliberations of editorial committees who carefully decide what topics will be covered each year. However, their decisions are not based on the systematic objective examination literature implied by global mapping. Their periodic identification of the changing research fronts is subjective. The best of both worlds would be ideal.
BC: So you would like to see ISI take that up.
EG: Yes, but it requires an editorial operation. ISI has a lot of other projects they can launch. But if I had to make a choice between that, a book citation index, and other projects, it would be a tough decision. Naturally, I'd prefer all of them.
BC: Do you have a book in mind to write?
EG: My book on citation indexing, [Garfield, E. Citation Indexing -- Its Theory and Application in Science, Technology and Humanities, ISI Press, 1979, 274 pgs.] needs to be updated. I doubt that I could do that by myself. And there's a need for a book on citation analysis as well. I've got to focus on finding somebody to work with me on these projects. There are probably a dozen or more people who could do it. Or it might become a team effort.
A significant demand on my time at present is ASIS. As president-elect of ASIS, I've had to catch up on what has happened at ASIS in the past two decades. ASIS, I believe, is at a crossroads and needs to decide what itís going to be and where itís going.
BC: And is your goal to help them do that?
EG: Yes, of course. I confess I wasnít very active for some time. I agreed to run for president not realizing how much I was out of touch. ASIS used to be a much more pragmatically oriented group. It has become much more academic. It certainly is perceived that way. And these perceptions percolate down to the grass roots. Many people have left ASIS because they donít feel that they get anything that will promote their careers or that they can really use on the practical side. Information technology has advanced enormously and needs to be reintegrated with our existing body of information science. Iíve proposed changing the name from ASIS to ASIST
[http://www.garfield.library.upenn.edu/asist.html]. That by itself wonít change much, but it symbolizes the other changes that I think are necessary to increase membership. Giving information technology greater recognition will help make outsiders aware that most ASIS members are involved with information science and technology. Information science seems too far removed for a lot of peopleóeven the library schools are moving in the direction of emphasizing information science and technology. There are all sorts of other fragmented societies that have picked up on this: the informatics groups, LITA, etc. Everybody is doing information technology under a different guise. The irony is that ASIS has for many years published the Annual Review of Information Science & Technology. So why not include "technology" in the name of the society? There is a kind of schizophrenia there. Weíre also working on getting our SIGs act together. Weíve also got to get a lot more young people involved in leadership. I donít think they realize that weíd be delighted for them to take a leadership role. It used to be easier to get folks involved in the society's affairs.
BC: Well, I think you come back to the problem of time again. Itís such a time-consuming responsibility to be a leader in an important society.
EG: But there are lots of tasks that can be done that arenít that time-consuming. What is the reason people belong to societies? It used to be more idealistic. Today young people seem to ask, "What do I get out of joining" rather than "What can I contribute?" Joining used to be more associative. You joined to be among people like yourself and to promote the information society of the future. Well the future is now!
BC: Best of luck to you on your presidency. And thank you very much for devoting some of your time to the readers of Serials Review.