College and Research Libraries Research Patterns and Research Libraries: What Should Change? Peter S. Graham A paper by Francis Miksa has suggested that because research patterns have changed, the ser- vices provided by research libraries must change to match them. But the research patterns must be evaluated as well as described. Before major changes in research libraries are initiated, we must define our goals and relate our activities to them as well as to present research pat- terns. Research libraries need not adapt to all research patterns. If it walks like a duck, talks like a duck, looks like a paper, I will begin by reviewing and dis- duck-it must be a duck.-Proverb cussing Miksa' s major points. In what fol- Why a duck?-Chico Marx lows I then hope to suggest the jumping- off point for questions such as, Should How are research libraries responding to the needs of today's researchers? How should they respond? These questions beg many more, not the least of which is, Why do we do what we do? Early in 1987, Francis L. Miksa, an OCLC Visiting Scholar and library school professor, gave an important speech to a conference of re- search library directors which OCLC has produced as Research Patterns and Research Libraries. 1 In this paper he suggests that re- search patterns have changed signifi- cantly since World War II and that re- search libraries must change in certain ways as well. Miksa' s discussion is concise, stimulat- ing, and coherent. One of his essay's great merits is its clarity of presentation about truths in university research that libraries need to contend with. They are not, how- ever, the only truths, as will be developed here. In addition, his conclusions need to be debated. They speak only of procedure and do not address underlying assump- tions and values of our profession; yet it is there that the debate must begin. In this libraries only react to changes in their en- vironment, or attempt to act upon that en- vironment? If research patterns and re- search libraries are not in harmony, what should change? What are our goals? A VIEW OF RESEARCH PATTERNS Miksa' s Research Patterns and Research Li- braries rests on the contrast between what research activity is presently like and what research libraries presently provide; some specific recommendations are presented. His first section describes late twentieth- century research patterns, which are sub- sumed under three headings. 1. Research method as much more quantita- tive: Research has become more formally analytic than it was in the past. The re- search process is highly structured, quan- titative methods are more generally em- ployed, and the computer has become central. 2. Professionalization of research: Research has become a socially accepted occupa- tional status. Where the rewards used to be recognition of scholarliness, now they are PeterS. Graham is Associate University Librarian for Technical and Automated Services at Rutgers Univer- sity, New Brunswick, New Jersey 08903. 433 434 College & Research Libraries the more worldly forms of status: group leadership, travel, grant funds, and salary. Research is now preeminently a team pro- cess, driven by funding agencies, and pri- marily applied (rather than pure). 3. Research as information flow: Research, because it is increasingly quantitative and analytic, increasingly uses data from other studies (and passes data on to more), rather than using and passing on formal written documents. Communication be-- tween team members is increasingly infor- mal and quick, with concomitant growth in the role of gatekeeper. Because most re- search is applied, most disciplinary boundaries are blurred. Reports of re- search now serve funding and status func- tions, often reducing their value for fur- ther research and contributing to the sense that a complete literature review is not necessary. This theme of the shifts in research methodology was also developed for li- brarians several years ago in an important book by Charles Osburn, Academic Re- search and Library Resources. 2 In similar terms, Osburn carefully described the shift in research emphasis, especially since World War II, and how this shift has affected collection development in re- search libraries. Some of his points in- cluded the enormous effect of govern- mental policies and funding on research structures and goals, the transfer of scien- tific organizational patterns to the social sciences, and the increased demand for re- search to solve specific social problems. A VIEWOF RESEARCH LIBRARIES Miksa' s paper characterizes two views historically held by research libraries as having come to hamper their proper per- formance today. First is the assumption re- search libraries tend to make of "the cen- trality of the idea of a universe of knowledge." 3 Second is the assumed ''chief task'' of the current research library, which ''in relationship to research is to pro- vide access to documents. " 4 Both descrip- tions of research library views are impor- tant, and the insights drawn from them are stimulating. Both of them, I believe, need further discussion before we can accept or reject these characterizations. July 1989 Research Libraries and the Universe of Knowledge Miksa amplifies his first point, that of re- search libraries assuming a universe of knowledge, by saying that The very core of a universe-of-knowledge point of view is the assumption that all knowledge, whether old or new, is by its very nature a sin- gle, cohesive, interwoven whole. Knowledge is, in fact, a unity, a superstruc- ture of all things known that, whether evident or not, is naturally ordered or classified in terms of its branches or disciplines, subbranches or subdisciplines, etc. 5 New knowledge is, in this view, not only a direct product of the old but necessarily · becomes a part of the overall unity of knowledge. "Research has become nondisci- plinary. That is, disciplinary struc- ture is simply not at issue in the re- search process. 11 In a stimulating passage, the essay de- scribes contemporary research (in the aforementioned patterns) as fragmenting the world of knowledge. ''Research has in fact become increasingly problem- oriented rather than discipline-oriented.'' Where some have called much modem re- search interdisciplinary, what has really occurred is that research has be- come nondisciplinary. That is, disciplinary structure is simply not at issue in the research process .... Research has become incredibly selective, pragmatic and throw-away in its uses of knowledge. Another way to say this is that research does not need to be aware of the idea of the universe of knowledge to proceed. Re- search does not even much need the universe of knowledge organized at all .... 6 This is a brilliant description, and much of it is apparently true. The truly unfortu- nate aspect of the argument is that it ac- cepts the new nondisciplinary research behavior as a norm and urges that re- search libraries accommodate to it. The earlier sympathetic description of the uni- verse of knowledge could easily lead one Research Patterns and Research Libraries 435 to think that Miksa considers this view correct and valuable; but without a de- murring word, he appears ready to dis- card it in order to serve those for whom or- ganized knowledge is not a value. Whose Knowledge Is in Which Universe? There is a prior problem with the argu- ment at this point; it is in the description of the universe-of-knowledge point of view that is ascribed to research libraries. In fact, the idea of human knowledge as a unity, "a superstructure of all things known,'' does not hold up well, either philosophically or in practice. I should rather say that it holds up philosophically only in nonreligious thinking, for of course it is fundamental to (for example) Buddhist or neo-Platonic thought that the universe is one in all respects, and that our knowledge of it must mirror that oneness. The modem western philosophical tra- dition is considerably more humble; we hear little of philosophies that claim to en- compass all knowledge and to explain all phenomena (even serious Marxists have been quiet on this score for some time). The physicists, to take science as an exam- ple, acknowledge that classical scientific method fails before fundamental struc- tures of energy and matter: we simply don't know (and evidently can't know) what happens in any given interaction, and must project with the aid of statistics. There may yet be a Grand Unified Theory; Einstein thought there was, and Hawking thinks there is, but there was a long stretch in between when the idea was be- littled. The deconstructionists, to take the humanities, tell us that our understanding of what we read is fundamentally flawed and probably solipsistic; whatever the au- thor meant is not likely to be what we read. Right or wrong, this approach does not tell us that knowledge is a unity. In terms familiar to librarianship, the proposition of a structured universe of knowledge can be seen to be in difficulty if we consider such disparate evidence as Thomas Kuhn's work and creationism. Kuhn's Structure of Scientific Revolutions makes clear that, far from all knowledge being a unity, knowledge leads to an un- derstanding of disunity: as knowledge is accumulated in science, it more and more leads to a realization of the disparity be- tween what is known and what was previ- ously understood. 7 Eventually the dispar- ity becomes intellectually too much to bear and a paradigm shift occurs, sometimes with great intellectual and even political upheaval. A whole body of what was con- sidered knowledge becomes "known" to be error, or at the least to be interpreted quite differently. To understand Kuhn's approach is to understand the ephemeral- ity of worldviews and of a structured uni- verse of knowledge. Libraries often, it is true, have collected as if there were a struc- tured universe of knowledge; given the changes in worldviews over the genera- tions, the result is that in our house there are many mansions. At a more practical level, the recent de- bate in a popular library journal on where to classify creationism points up the way in which libraries do not today simply re- gard knowledge as a single unity. Knowl- edge is in the eyes of the beholder: there are creationists who deny the Darwinian view and the rest of us who regard cre- ationism not as a matter of knowledge but as a way of thinking. Each, however, de- serves and requires its place in collections representing human thought and culture. So do materials on geocentrism, spoon bending, racism, and the occult; the pa- pacy and atheism; scientific socialism and supply-side economics; catastrophism and Gondwanaland. An unfortunate con- clusion can be drawn from the proposition that human knowledge is a unity on which libraries focus; it is that libraries should collect only the truth. Miksa tells us he has done considerable work in classification theory. 8 It may have led him astray. Dewey and the other nineteenth-century synthesizers gave us mechanisms which purport to encompass human knowledge. The misleading com- prehensiveness of our classifications, and of our subject thesauri, can lead the un- wary to see a theoretical structure behind them; in the end these are only attempts, brave and helpful but inevitably futile, to describe what has come over the intellec- tual transom in all its marvelous variety. The fact remains that research libraries have collected comprehensively whether 436 College & Research Libraries or not their librarians have had a view about a unified structure of knowledge. It is correct to indicate that this breadth of collection has been a distinguishing char- acteristic of the great research library in the past. As suggested earlier, Miksa ap- parently is ready to replace this distinction in order better to serve those who wish to work on the particularities of the moment. Research Libraries and Documentation The second view within research li- braries that is said to hamper their perfor- mance is their commitment to providing access to documents. Librarians are ap- parently misled in two ways: first, they er- roneously believe that documents are a primary information input in the research process. Second, librarians mistakenly be- lieve that they need only supply access to documents, whereas they should some- how be providing the documents them- selves. This combination of errors has, for the past decade, led us down the primrose path of online catalogs, bibliographic net- works, and the distributed national li- brary.9 "Teamwork and collegial networks, particularly in an electronic mode, are causing researchers to use infor- mal reports and only the data hom prior studies rather than published research outcomes." Part of the analysis is persuasive as to why much current research is not oriented toward documents. First, teamwork and collegial networks, particularly in an elec- tronic mode, are causing researchers to use informal reports and only the data from prior studies rather than published research outcomes. Second, present re- search is oriented toward a specific prob- lem at hand, obviating the need to review the literature fully if a solution to the prob- lem can be found without doing so. Third, and most important, 'I the most urgent need facing the research informa- tion flow process is that of information management at the point of use.'' In other words, there is too much information. In- July 1989 . formation management means a filtering and organizational process; and (in a statement with appalling implications for the state of contemporary scholarship) 11 the chief difficulty here is that the re- searcher . . . has neither the time, the in- clination, nor the skill to do this .... 1110 It would have been helpful to have had a more consistent use of the term document. In the discussion of research patterns it reasonably appeared to mean the repre- sentation of information in any of a variety of forms, including electronic; but when Miksa is criticizing the libraries' orienta- tion, he says that II documents are, after all, physical materials. 1111 Having set up this straw man, he then describes libraries as only concerned with providing access to physical objects that they can count and control (i.e., printed volumes). This fairly assesses the mind-set of some librarians, but no longer of the majority of leading re- search librarians. For years now many of use have been providing information re- trieval services and experimenting with telefacsimile, not to mention countless re- ferrals to other sources for information not held locally. On campus after campus li- braries and computer centers are meeting to discuss how to use the institution's computing facilities and information skills to provide patrons with the research infor- mation they need-not just documents- regardless of the form in which it ex- ists.u,13 For most research librarians a document is no longer defined simply as a physical item owned by a given library, though it certainly helps to know where a copy might be. A more interesting discussion of what libraries should now provide might be centered around a more precise defini- tion of document-one that doesn't de- pend on physical manifestation. The issue has come up recently in a number of fo- rums of librarians grappling with how to provide machine-readable data to college and university patrons. The differences expressed on this issue represent in minia- ture the issues facing librarianship at large. For example, participants in theRe- search Libraries Group (RLG) experiment funded by the Pew Foundation met at San Antonio in 1987 to share experiences. One difference between institutions was what Research Patterns and Research Libraries 437 machine-readable data files (MRDFs) were included in their catalog. Some insti- tutions only included MRDFs that were fully documented and distributed by a rec- ognized agency; on the other hand, some included work in progress and recently developed programs by local faculty members. Increasingly, there has been an under- standing that electronic information can include the same full range of unpub- lished, prepublished, and published work as print materials. One proposal that makes sense has been for libraries to pro- vide access only to electronic information represented by published work, much as we have traditionally done for print (and other nonprint) materials. The definition of publication must be rediscussed to take into account new circumstances, but surely it will include matters such as for- mal dissemination, multiplicity of copies, consistent or standard format, self- documentation, authorship responsibil- ity, and the nature of the publishing agency. (Other influential practitioners, such as Peggy Seiden of Carnegie-Mellon University's software project, think that such a definition is unnecessarily con- straining.) It certainly makes sense for a library to provide access to prepublished data in some way (e.g., the RLG joint project with the Modern Language Association). Miksa is advocating that a research library has a responsibility to provide access to gestating research in the same way it does for what have classically been the prod- ucts of research. Such a position has impli- cations for the acquisition and preserva- tion of the intellectual record. There are also serious implications for the resources that must be diverted from other library purposes, for the patron and for the li- brary professional, and for the research process itself. Here are issues that need discussion in and beyond the profession, and that will shed light on whether docu- ment provision is tnily a limiting feature for present research libraries. ARGUABLE CONCLUSIONS The conclusions of Miksa' s paper derive from ''the need to adopt a revised per- spective about research patterns'' in order ''to shape technology and research library goals." 14 First, to replace the view that there is a universe of knowledge, libraries and librarians must change their approach to match the world of research. If research is indeed nondisciplinary, but research li- braries continue to assume the interrela- tionship of disciplines implied by the uni- versality of knowledge, then libraries are working at cross-purposes with those us- ing them. Libraries (and librarians) should, therefore, focus their resources to ''more closely match the research actually being done rather than ideal coverage of the entire structure of knowledge" or of any of its sectors. One approach is to move to demand-driven acquisition in place of long-range collecting plans. 15 Second, to replace the emphasis on doc- ument access, library personnel must rad- ically change the work they do in order to perform "information management" for research projects. Libraries should, in some active way, be providing the texts themselves, not simply access to the texts. And not only should the library mecha- nisms provide full texts, but librarians should become part of the research teams. Given a view of present research as team oriented, working most often on applied topics with the need to eliminate un- wanted material as much as to find neces- sary information, librarians should be sec- onded to the teams to do this work: ''I . . . mean distributed personnel whose task · must be done in the context of the research activity, not simply personnel who look in on research efforts now and then.' ' 16 These are recommendations to change fundamentally the way research libraries and research librarians work: research li- braries should eschew being libraries of record, and research librarians should take on entirely new tasks. But properly to found such recommendations, there must first be a discussion of goals and their con- comitant values; to my perception these are absent from Miksa' s paper. Goals and values must now be introduced. The first thing that must be said is that Miksa' s picture of research is very de- pressing. We see an academe full of nar- row opportunists who seldom generalize as they maneuver for the main chance. The universities are apparently full (and 438 College & Research Libraries only for the past few decades, not before) of researchers too timid to spread their wings fully and who depend excessively on each other as they work to ameliorate problems of little lasting import. Re- searchers as a class are either ignorant or untrained in reviewing the literature of their own field as an integral part of their own work, and they do not think it neces- sary. Is it true? I doubt it. In this account I look in vain for figures like Paul Oskar Kristel- ler, Lewis Thomas, Alice Ostriker, Rich- ard Feynman, Richard Ellman, Stanley Schoenbaum, James Watson, Susan Gu- bar, Oscar Handlin, Barbara McOintock, or E. P. Thompson. Is it really true that our civilization's scholars have neither ''the time, the inclination, nor the skill" to do their own work? ''As research librarians we must take the stand that it is indeed important for national research collections to be built that aim at comprehensiveness, whether or not we call it an 'ideal cov- erage of the entire structure of knowl- edge.' " Related to this depressing view of re- search is that it is an undifferentiated one. The term is always research not some re- search. The term is never qualified in the paper: apparently there are no differences between past and present research, be- tween scientific and humanistic, or be- tween fundamental and applied, much less between good and bad. The result is that in the essay research loses a descrip- tive sense and becomes prescriptive. The essay implicitly becomes one of prescrib- ing what research ought to be, in place of its purported intent to describe research in order to define library goals. (The words scholar and scholarship, by the way, are ab- sent from the essay except where the con- cepts are described as obsolete; but they are not, and we as librarians have a re- sponsibility to help assure that they do not become so.). But let us assume the picture to be, in July 1989 the main, true. Is it right? Is this the kind of research that should be done? More im- portant, is it the only kind of research that should be done? And which researchers should we librarians aim best to serve, dedicated as we are to preserving, orga- nizing, and providing the human heri- tage? Clearly we need to serve all researchers, even bad ones, as best we can. That is not to say, however, that we must be passive components of the research process. For example, as research librarians we must take the stand that it is indeed important for national research collections to be built that aim at comprehensiveness, whether or not we call it an ''ideal coverage of the entire structure of knowledge.'' Libraries should not be market-driven institutions. While we need to take cogni- zance of current patterns of use and schol- arship, we must also provide for the fu- ture and for the presently unfashionable. If some current researchers do not wish to use this cornucopia of knowledge, opin- ion, experience, and data, then that is their choice. We can have confidence that there are those who will; and even if there were not, we know that times change and that decades are a short term in the life of the mind, and therefore of libraries. Miksa argues that we in the research li- braries, and the researchers, are working at cross-purposes: "rather like two per- sons working in the same room, perhaps at tables near each other, supposedly on the same project, but without being fully aware of what each other is doing."17 His proposal to harmonize the two calls for li- braries to change, to "more closely match the research actually being done." But from his description, it seems at least as desirable that instead, some research processes change. I also do not think it appropriate for li- brarians to make up for the shortcomings of some current researchers by becoming handmaidens of their research teams. Li- brarians honorably share, with those who make use of knowledge, a division of scholarly labor: we serve to acquire, orga- nize, preserve, and make accessible the body of human knowledge in all its re- corded forms; we do this for scholars, who Research Patterns and Research Libraries 439 make the proper use of it. We have trained ourselves in valuable skills which re- searchers aren't required to have, and to discard these skills would be to break a link in the chain of knowledge transmis- sion. As an intellectual citizen, I am left pro- foundly uneasy by a description of re- searchers both unwilling and unable tore- view their own literature without our help. If the work is mechanical enough to be delegated, professional librarians should not be doing it-we need to do the difficult work of selection, organization, and preservation in the institutional re- pository, for no one else will. If the win- nowing and searching is substantive enough, then the researcher should do it-or the label researcher loses meaning. A competent scholar in most fields should feel displaced and aggrieved by someone else doing what is fundamentally part of his or her job in "filtering" the appropri- ate information. WHAT SHOULD WE BE DOING? If I do not agree with the propositions that research libraries should narrow their collections to the work of the moment, and that research librarians should be- come information management team members on specific projects, what do I advocate should be our role? While we cannot expect directly to change the re- search patterns that are the result of great social and economic forces, we can have personal and professional effect. Both li- brarians and other scholars should be writing, in their journals and more broadly, about the value to society of maintaining the full human record in ad- dition to the value of instantaneous avail- ability of selected portions for selected scholars. Some of us may even wish to join Miksa and Osburn in asserting the im- balanced emphases of current research in the terms that they use, but with the aim of changing them rather than adapting to them. In our libraries we need to argue for, and make, policy decisions that provide there- search context that we believe is correct. This means relying on our professional knowledge and values as well as our sense of history, and not ceding only to the de- mands of the moment. There is a role here for bibliographers with faculty members and departments, and for senior library administrators in the highest councils of our universities. We will find allies among faculty and administrators who believe that there is value in research beyond spe- cific investigation of particular problems. At times we must make controversial decisions in collection development, pro- vision of service points, separation or combination of collections, and allocation of staff resources. The result may be, for example, collection decisions that favor potential future research rather than spe- cific current work; or staffing decisions in favor of preservation of existing materials rather than direct assistance in online searching, even as we make electronic in- formation available in a variety of forms. We need to understand the electronic information revolution so that we inte- grate it into our long-standing goal of pro- viding broad information availability. If we ignore it, then we demonstrate that our heads are in the sand and we will have no credibility with researchers who need what it provides. If we respond only to the siren calls of immediacy, then we will lose touch with society's requirements for us to organize and preserve knowledge for the future. We must become clearer about our re- sponsibility to support at least two clien- teles: today' s scholars and tomorrow's. In the past what we did for the future scholar (collect, catalog, and preserve) more or less did the job for the present scholar. For years this has been less and less true as the immediacy of some literatures, and the immediate needs for them, have in- creased, while libraries have continued to provide best for monograph-oriented pa- trons. The wide availability of electronic information puts this contrast into high re- lief. Many research librarians are helping us to address this dual role; more of us should be. What we must not do is give up one clientele for the sake of the other. As Shirley Echelman says, So here we librarians stand, Janus-like, guard- ians of the accumulated knowledge of the past and guides to the information upon which fu- 440 College & Research Libraries ture knowledge will be built .... Without abro- gating our responsibilities to our collections, we must harness the technical means . . . to the search for knowledge. 18 We need to argue persuasively and ra- tionally for more resources to do these new tasks. Since we will partially fail, as there are never enough resources to do all of what needs to be done, we must be pre- pared to make hard choices about what we can do and what we can't. 19 These choices must be made in the context of our long- term goals, and not only to accede to the more pressing demands of our momen- tary clienteles, persuasive though they may be; we can hear from our future clien- teles only through our own professional- ism. At the outset I said that Francis Miksa' s July 1989 stimulating paper needed to be discussed in terms of the underlying values and as- sumptions of our profession. He is right that what research libraries are doing no longer seems as it once did to match what researchers are doing. In the course of my appreciation of his perspectives, and my disagreements with his conclusions, I have introduced some value judgments to begin the discussion of why it is that we do what we do and what our goals are. As re- search librarians, we need to reformulate our goals and values and to debate them, for there will be disagreement. Many of us need to search again, through the clash of such debate, for conclusions that will help us guide our profession and therefore to shape future scholarship. REFERENCES AND NOTES 1. Francis L. Miksa, Research Patterns and Research Libraries (Dublin: OCLC, c1987). Text of an address given at the Fifth Annual Conference of Directors of Research Libraries in OCLC, Mar. 30, 1987. (Available upon request from OCLC.) 2. Charles Osburn, Academic Research and Library Resources: Changing Patterns in America (Westport, Conn.: Greenwood, 1979). 3. Miksa, p.8. 4. Ibid., p.10. 5. Ibid., p.8. 6. Ibid., p.9. 7. Thomas S. Kuhn, The Structure of Scientific Revolutions (Chicago: Univ. of Chicago Pr., 1962). 8. Miksa, p.16, 19. 9. Ibid., p.9-10, 13. 10. Ibid., p.ll. 11. Ibid., p.10. 12. Examples of research libraries orienting toward electronic information are many and include those of the following universities: Carnegie-Mellon, Vanderbilt, Georgia Tech, Columbia, Rutgers, Pennsylvania, California, Florida. For some descriptions see, among others, Caroline Arms, ed., Campus Networking Strategies (n. p., Digital, 1988), p.4, 70, pass.; and recent issues of the EDUCOM Bulletin. 13. For examples of some of the best thinking by librarians, see David Lewis, "Inventing the Elec- tronic University," College & Research Libraries 49:291-304 Ouly 1988); and Patricia Battin's articles, "The Electronic Library-A Vision of the Future," EDUCOM Bulletin 19:12-34 (Summer 1984) and "Crossing the Border: Librarianship in the Information Age," The Harvard Librarian 19:8-10 (Sept. 1985). See also B. B. Siman, "Machine-Readable Data Files in the Library," Technicalities 6:6 Oune 1986); and Richard W. McCoy, "The Electronic Scholar: Essential Tasks for the Scholarly Commu- nity," Library ]ournal110:39-42 (Oct. 1, 1985). 14. Miksa, p.ll. 15. Ibid., p.12. 16. Ibid., p.13. 17. Ibid., p.12. 18. Shirley Echelman, "Why do Academic Libraries Get Such a Bad Rap?" Library ]ournal113:39-41 (Oct. 1, 1988). 19. For a perceptive discussion of the importance of making resource choices, see Paul Metz, "Special Pleading versus Self-Discipline in the Financing of Public Services," Journal of Academic Librarian- ship 14:208-13 (Sept. 1988).