Return to Contents

CHAPTER ONE

INTRODUCTION

For centuries the library has been a repository of the written record and a powerful symbol of human intellectual achievement, but today, as perhaps never before, fundamental questions are being raised concerning its nature as an institution.

Libraries of different types serve different communities, of course, and it is important to say immediately that this study is concerned primarily with only one type: the research university library, whose special purpose is to support advanced scholarship and scholarly communication and the research activities of faculty members and doctoral students at Ph.D.-granting institutions. Undergraduates also make heavy use of these libraries, but the provision of services to undergraduates is not the distinctive purpose of the research library. (Chapter 3 contains a discussion of the differences in library expenditures at universities and at liberal arts colleges.)

The forces affecting research libraries are numerous and complex, and they are not easily described and calibrated. As indicated in the Foreword, this is in no sense a definitive study. Rather, this report is best seen as an initial examination of some of the relevant trends, intended to improve our understanding of the issues and choices that are emerging so rapidly. In the main, it is an attempt to collect in one place a considerable amount of both statistical data and information on technological changes affecting libraries. We have tried to assemble these materials in a way that draws together parts of the library puzzle that are often considered separately.

ORGANIZATION OF THE STUDY AND PRINCIPAL PROPOSITIONS

It may be helpful to think of the principal challenges facing research libraries today as falling under two broad headings, which correspond to the two parts of this study.

First, libraries must continue to acquire the books and periodicals necessary to maintain and strengthen collections in the face of both (a) a rapidly proliferating universe of published material that it seems desirable to collect and (b) rapidly escalating unit prices, especially for some journals. In almost all cases, university budget constraints have compelled research libraries to acquire an ever smaller share of the universe of materials from which they are accustomed to make selections. Part 1 of the study documents trends in the sizes of collections, rates of acquisitions, levels and categories of expenditures, publications of monographs and serials, and prices of these publications. Second, libraries must determine how to respond to the rapid emergence and development of electronic information technologies that permit one to envision radically different ways of organizing the services the library has traditionally provided. As currently configured, the library is a print institution whose essential characteristics have been determined by the technology of printing. That technology in turn entails a particular set of legal, economic, and cultural relationships among scholars, their universities, publishers, academic book vendors, libraries, and other participants in the process of print-based communication. A new technology with different characteristics will almost inevitably entail different arrangements among these participants---with far-reaching implications for scholarship, graduate education, and the economics of creating, publishing, disseminating, and using scholarly information. Part 2 of the study deals with these topics.

The analysis in Part 1 parallels in certain respects the study of trends in doctoral education sponsored recently by the Foundation.[1] The library has traditionally been the most important of the university facilities supporting advanced scholarship, at least in the humanities and related social sciences, and its continued vitality has been seen as critically important to the vitality of Ph.D. programs in those subjects and to the ability of institutions to support distinguished programs.

Because of this close association, it is not surprising that the data documenting historical trends in the growth of library collections resemble closely certain trends in doctoral education.

The number of volumes added to library collections increased rapidly during the 1960s, when university enrollments were expanding, new doctoral programs were being established, and considerable funding was available from external sources. The increase in the number of titles acquired was both a stimulus to and a consequence of the expansion of graduate education: a larger library collection supported the research activities of larger faculties and graduate student cohorts; larger numbers of active scholars, in turn, produced more scholarly works that research libraries then acquired. The widespread financial and programmatic retrenchments of the 1970s and the early to mid-1980s are similarly reflected in the library data studied here: rates of increase in expenditures for acquisitions were much slower in the 1970s and early 1980s than during the previous decade, and the inflationary trends of the period are reflected in the prices of library materials, which rose still more rapidly; the purchasing power of acquisitions expenditures was therefore diminished. The rate of increase in the number of volumes added slowed considerably and at many institutions was actually negative---that is, the number of volumes added in any particular year was lower than the number added in the preceding year. The moderately improved financial situation of the late 1980s permitted some recovery, though not to levels characteristic of earlier decades.

One perhaps surprising finding in Part 1 of this analysis is that library expenditures have not tended to increase more rapidly than other university expenditures over recent decades. On the contrary, the library's share of total expenditures has tended to decline in spite of the rapid increases in the prices of materials, especially serials. We suspect that this shifting relationship between library expenditures and total expenditures, which is remarkably consistent across various kinds of libraries and universities, reflects a judgment that there was simply no way in which constrained university budgets could accommodate the increases in spending on acquisitions that would have been required to maintain past rates of increase in acquisitions. The conclusion may well have been that, if it simply was impossible to keep up, it made sense to accept this reality and restrain the growth in expenditures for acquisitions in keeping with limits on the growth of overall university resources.

Although many of the challenges described here are universal and affect libraries of many different types, the ever-expanding size of the universe of published materials and the rapidly increasing prices of these materials are especially troubling to research libraries. Such libraries have traditionally aimed to be as comprehensive as possible in their acquisitions practices, to provide faculty members and graduate students with access to as much of the entire professional literature as can be acquired. The fact that university libraries, with few or no exceptions, are now able to respond less comprehensively than ever before to general trends in book production is widely regarded with anxiety, in that access to scholarly information may be narrowing.

A related concern is that pressure on acquisitions budgets will cause various research libraries to look more and more alike over time, as each ceases to purchase as many of the more esoteric publications and chooses rather to be sure that essential volumes are acquired. The consequence could be a decline in the richness of collections overall, not merely a decline in the range of holdings of any one library.

These broad trends raise deep questions concerning the viability of the traditional model of the library. The rapidly rising prices of materials, the continued increase in the number of items available for purchase, the fact that university libraries seem to be acquiring a declining share of the world's output, the impracticality of continuing to build large, costly, warehouse-type structures to shelve printed materials, thus replicating collections that exist elsewhere---these and other developments cause one to ask whether established practices, which are already eroding, can be continued for very much longer.

Part 2 of the study suggests that electronic technologies may permit different assumptions and practices to characterize scholarly communication in the future. In a sense, the technology of print demands that individual institutions build self-sufficient, comprehensive collections, in anticipation of user demand. There is no other way to ensure prompt local access to scholarly information, given that printing results in the production of material objects that must be purchased, shipped, classified, and shelved. Electronic technologies, in contrast, permit different practices; in principle, information in electronic form can be disseminated much more rapidly, and its storage is altogether different in kind.

Such characteristics have led many observers to suggest that the process of scholarly communication can now be based on a principle of access rather than ownership. More than ever before, libraries, or the institutions that may succeed them as they undergo redefinition, can envision building collections in collaboration with other institutions, specializing locally in certain kinds of materials and distributing resources among the members of a consortium. The ability to share materials readily in electronic form obviates the need for each institution to attempt to build a comprehensive collection, with all of the costly redundancies that such a model entails. These new technologies also have major implications for the allocation of expenditures among library functions and activities (see Chapter 4) and certainly for the space requirements of libraries.

We hasten to emphasize that the transition to any such alternative model will not be easy, for the reasons mentioned earlier: a particular technology, of whatever type, is joined to a set of economic and legal arrangements appropriate to it. As we discuss in Part 2, the technology of print fits hand-in-glove with copyright practices as they have evolved to the present day; moreover, the current model is sustained by a set of complex economic relationships among publishers, vendors, and libraries that have been carefully worked out over decades. New technologies, in contrast, will result in new sorts of relationships yet to be defined and agreed upon; it will not be easy to reach agreement on such matters.

Some observers believe that the new electronic technologies will have considerably more far-reaching effects, that their emergence signals the beginning of a fundamental shift in accepted practices governing the dissemination of ideas and even their development. According to that view, the institutions, practices, and forms of a print culture will undergo complete transformation or in some instances disappear altogether. The self-sufficient research library, the scholarly publisher, the printed book, the monograph, the learned journal, the process of peer review, and copyright practices---these and other familiar elements of the current system are all implicitly challenged by electronic technologies.

Electronic methods of disseminating information are at least as different in kind from print as print is from manual copying and may be much more so. We are not yet far enough along in the transition to a fully electronic environment to be certain of what new forms and institutions may ultimately emerge; but we may be certain, we would argue, that they will be very different.

DATA SETS, LIBRARY COMPOSITES, AND OTHER SOURCES OF INFORMATION

For the first part of the study, data pertaining to internal library developments were obtained primarily from the Association of Research Libraries (ARL), an organization founded in 1932 and presently located in Washington, D.C. For the academic years 1907-08 through 1937-38, the data were collected by James Gerould, who was university librarian first at the University of Minnesota and then at Princeton University. Following his retirement in 1938, members of the library staff at Princeton continued to collect the data through 1961-62, when ARL assumed direct responsibility.

The data are derived from an annual questionnaire. The version used in 1990-91 with accompanying instructions is reproduced in Appendix A. Despite minor inconsistencies, the ARL database is a highly reliable source and the only longitudinal data set of this scale in the library field. We have chosen to emphasize categories that reflect broad changes in the collections and patterns of expenditures over time. These categories also are presented in Appendix A.

Of the 107 university libraries that are current members of ARL, we have chosen 24---at twelve private and twelve public institutions--- for intensive analysis. These 24 universities were chosen to reflect the experiences of four broad sets (or composites) of libraries, which we call Private 1, Public 1, Private 2, and Public 2. Composite values of relevant variables were obtained for each set by calculating an unweighted average of the values for each component institution. The resulting values are intended to describe the experience of the "typical" library within each set.

These 24 libraries have been chosen to permit systematic analysis of broad developments over time within different sets of well-established research libraries. They are not meant to be representative of the universe of ARL libraries. For example, they include a disproportionately larger number of private universities and libraries that have been long-time members of ARL. We selected particular universities based on the consistency of their data over time, similarities to one another in terms of academic strengths, and our general knowledge of individual institutions. For our purposes it was especially important to have a fixed set of institutions, so that comparisons over time would not be distorted by the addition (or subtraction) of a particular library from the database. This is why trends in summary measures for the entire universe of ARL libraries are so difficult to interpret.

The universities included in the Private 1 and Public 1 composites are the older and for the most part larger of the ARL libraries (all were charter members of ARL).

Private 1 universities are the University of Chicago, Columbia University, Cornell University, Princeton University, Stanford University, and Yale University.[2] Public 1 universities are the University of California at Berkeley, University of Iowa, University of Michigan at Ann Arbor, University of North Carolina at Chapel Hill, University of Virginia, and University of Wisconsin at Madison.

The libraries included in the Private 2 and Public 2 composites tend to be smaller, and many are more recent members of ARL.

Private 2 universities are Boston University, Georgetown University, New York University, Northwestern University, University of Southern California, and Washington University in St. Louis. Public 2 universities are University of Florida, Iowa State University, University of Maryland, Michigan State University, Rutgers University, and Washington State University.

At some points in the analysis we found it useful to group the Private 1 and Public 1 universities into a larger composite called Research 1. And in much of the discussion, we group all institutions into a single composite called All 24 Universities.

Complete data were not available for every institution for the entire time period that we discuss (1912 through 1991). Discussion of trends prior to 1963 focuses primarily on the Research 1 universities. Discussion of developments after 1963 makes use of the full data set. In general, we discuss broad trends for All 24 Universities and then note any significant differences by sector.

One finding of this study, which can be generalized from the previous discussion of the relationship between library expenditures and total expenditures, is that the same trends tend to be found within all four composites. This suggests that the forces affecting libraries have been quite general, with the same waves washing over all these institutions. While idiosyncratic patterns of course exist (with some libraries, for example, almost surely operating more efficiently than others), these individual variations do not seem so pronounced as to obscure general trends---especially when it is possible, as it has been in this study, to work with averages for sets of libraries.

Data pertaining to university expenditures in general were obtained from the Higher Education General Information Survey (HEGIS), which is administered annually by the National Center for Education Statistics (NCES). These data were available for the period up to 1985-86. The data on expenditures after that year were obtained from the Integrated Postsecondary Education Data System (IPEDS). The data showing expenditures for college libraries, in relation to other expenditures by the same colleges, were obtained ultimately from this same source.[3] The definitions of the categories we used from these questionnaires can be found in Appendix A.

Internal developments---trends in collection growth and expenditures, described in Chapters 2 through 4---are then related in Chapters 5 and 6 to a series of important external developments: trends in book and periodical production, both domestic and international, and in the prices of library materials.

Data on domestic book production, which have important limitations described in detail at the appropriate point in the discussion, were obtained from two sources, the R. R. Bowker Company and the Association of American University Presses; data on international production were obtained from the United Nations Educational, Scientific and Cultural Organization. In each instance, subsets of the total were defined in an effort to confine the data to a more relevant set of fields. Data on periodical production in some representative disciplines were obtained from the Modern Language Association and the Institute for Scientific Information. Data on the prices of books and periodicals were obtained from the R. R. Bowker Company and various issues of the Library Journal and Publishers Weekly.

For the years since 1963 the longitudinal data sets were often subdivided into shorter time periods: 1963-70, 1970-82, and 1982-91. In this way we were able to pinpoint with reasonable precision exactly when various trends began to emerge and to document the "boom and bust" syndrome that characterized so much of the experience of institutions of higher education (including their libraries) during the 1960s, 1970s, and 1980s.

The second part of the study is very different in kind from the first. We shift from a quantitative mode of analysis to a synthetic one. Our objective in Part 2 was to summarize some of the more important statements concerning the alternative model of scholarly communication proposed by many observers and offer as clear a description as possible of some of the principal elements: the availability of bibliographic records in electronic form that provide scholars with information about the professional literature and collections housed elsewhere; the potential applications of electronic technologies to scholarly publishing; the possibility of the kinds of collaborative collection development that the availability of electronic material permits; the various cultural, economic, technological, and legal issues to be resolved before any such alternative model is viable; and the need for an adequate telecommunications infrastructure capable of moving large quantities of electronic text and data at high speeds.

Consistent with this shift in methodology, the primary sources utilized for Part 2 are quite unlike those used in Part 1. There were few data to be analyzed (apart from those pertaining to recent shifts in the composition of library expenditures, which are discussed in Chapter 4), and we relied instead on such materials as literature published by some of the major bibliographic utilities (the Research Libraries Group, DIALOG Information Services, Inc., and so on); articles in specialized but still "popular" periodicals (including especially The Chronicle of Higher Education); and unpublished memoranda and studies of various kinds. While a considerable amount of literature has been consulted, there are many other studies that we would have liked to review and incorporate into this analysis.

The nature of the issues treated dictated the nature of the sources used: the situation is changing very rapidly, and our objective was to ensure that our account of developments be as current as possible. In some ultimate sense that objective proved unattainable, and we cannot claim that our picture of the situation is anything other than a snapshot, taken in the spring of 1992. A snapshot taken subsequently might show a rather different picture. In fact, even as this introduction was being drafted, new articles of importance were appearing on various of the issues treated in Part 2. With this important proviso we feel that Part 2 nonetheless indicates the ways in which new technologies suggest a model for the library of the future that may differ sharply from the traditional one.

OTHER QUESTIONS

Before concluding these introductory remarks, we should say more about what this study does not do, about some of its most obvious limitations in terms of coverage. First of all, the study concerns universities of very different types, some of which have a large complement of graduate and professional schools, some of which do not. The presence or absence of such professional programs clearly affects the character of the library system, expenditures, the sizes of collections, and so on. The available data sets have been compiled in such a way that, in general, they cannot easily be disaggregated; as a consequence, data on medical, law, and business school libraries have not been subtracted from the totals.[4] To some extent, therefore, we are obliged to work with unlike entities.

Moreover, the study does not consider in any depth the different sorts of problems one finds in particular subject areas within the arts and sciences. Scientists and humanists make very different uses of the library, and the collections maintained for each "class" of scholar have different characteristics and pose distinctive problems. We have not specifically addressed these kinds of contrasting needs in any detail, though we do occasionally refer to them.[5]

Nor does this study consider the problem of book preservation. Many of the volumes housed at libraries of the type studied here are in danger of disintegration; they were printed on paper manufactured from wood pulp and as a result have a high acid content. The fact that we do not discuss the preservation problem should not be taken as an indication that we do not consider it serious; on the contrary, it must be regarded as one of the most important problems research libraries face. Our sense, however, is that it is one problem that is particularly well understood, at least in relation to the other kinds of problems identified here.

The preservation problem is also more clearly separable from the others, whereas those issues discussed here are in some sense more inextricably linked with one another and therefore must be treated as parts of a whole. Finally, there have been systematic, determined efforts underway for some years now that address preservation issues with visible success. That success is due in no small part to the activities of the Commission on Preservation and Access, ably directed by Patricia Battin, and other similar programs, notably that supported by the National Endowment for the Humanities.[6]

Considered together, as they must be, the panoply of questions, issues, and choices outlined in this study will define in no small measure the nature of the process of scholarly communication in the years ahead. There are implications for both graduate and undergraduate education, for the finances of higher education, for the publication process itself, and perhaps even for the ways in which some ideas are formulated, reviewed, and then revised. While a considerable amount of experimentation and learning as we go is inevitable, we are persuaded that the time is at hand for systematic efforts to define sets of alternatives, test their implications, and devise feasible modes of collaboration across sectors and among different types of entities.

The interconnections among both the questions to be considered and the institutional players are so strong that limited perspectives are likely to lead to unsatisfactory outcomes. The opportunity exists to rethink an entire set of relationships that, if reconstituted appropriately, can give libraries both new dimensions and an even more central role in the educational process than they have enjoyed in the past.

Endnotes

[1]William G. Bowen and Neil L. Rudenstine, In Pursuit of the Ph.D., Princeton, N.J.: Princeton University Press, 1992.

[2]An obvious omission is Harvard University. We did not include Harvard in the Private 1 composite because it is in a class of its own with respect to scale and would skew the averages for the composite. In 1989-90, for example, Harvard reported a total of 11,874,148 volumes held, while most of the universities in our Private 1 composite reported volumes held in the general range of 5 to 6 million (with Yale at the top of this list with holdings of nearly 9 million volumes).

[3]The college data were supplied by Anne MacEachern, research coordinator, Williams Project on the Economics of Higher Education.

[4]It would have been possible to separate data for law and medical libraries between 1978 and 1990, but this was not done.

[5]For excellent general statements on these kinds of issues, see the materials collected in Communications in Support of Science and Engineering, A Report to the National Science Foundation from the Council on Library Resources (Washington, D.C.: Council on Library Resources, August 1990). We are grateful to Warren J. Haas for sending a copy of this publication.

[6]For statements of the preservation problem and the efforts to work toward solutions, see the commission's many excellent publications, available from its office at 1400 16th Street, N.W., Washington, D.C. We are most grateful to Patricia Battin for providing a complete set and for useful discussions and correspondence concerning book preservation.

One aspect of the discussion of preservation does impinge directly on the other questions associated with electronic technologies---namely, the debate over the desirability of putting less emphasis on microfilming and more emphasis on the use of digital imaging.