 |
 |
 |
Return to Contents
|
|
|
There is certainly no absence of comment, largely negative, on a phenomenon characterized as "academic overdose" in an article in The New York Times entitled "Where Information Is All, Pleas Arise for Less of It."[2] The article suggested that "[a]s the population of books and journals continues to explode, librarians complain that shelf space is running out and expenses are spinning out of control" and reported that "[i]n February 1988, Harvard Medical School issued new guidelines for tenure review, recommending that the faculty consider requiring no more than five published works for a candidate for assistant professor, seven for associate professor and 10 for full professor," a decision made against a background of expressions of concern that "the multiplicity of mediocre publications makes it impossible to sift out the ones that contain fresh ideas. The proliferation of books and journals seems to have narrowed access to information instead of widening it." More recently, Donald Kennedy, then president of Stanford, was quoted as saying that "[t]he overproduction of routine scholarship ... tends to conceal really important work by its sheer volume ... and is a major contributor to the inflation of academic library costs."[3]
Others argue that concerns about a putative decline in quality are unjustified, that they mask other, unstated concerns about how scholarly fields are defined and about changes in methodologies and perspectives with which many of those who express such concerns simply disagree. Two papers, among others, make the very important point that the increase in the number of journal articles and books is in part a function simply of an increase in the size of the professoriate. Any assumption about a decline in quality or an increased emphasis on research based largely on the increase in the number of items published may fail to take adequate account of the statistical fact that these two papers highlight. The per-person output, that is, may not be significantly higher now than before.[4]
On one important issue, however, there seems to be little disagreement: the basic problem, if indeed the phenomenon is a problem, results in large part from the nature of the reward system. Appointment to the professoriate and advancement within it are contingent upon scholarly output, as measured (qualitatively and quantitatively) by one's record of publication.[5] Moreover, the intense competition for places that characterized many academic labor markets in the 1970s and 1980s only intensified the pressure to publish.
We have no new insights to offer on this range of issues concerning the quality of scholarly output, and we can do no more here than acknowledge the importance of the debate. We must limit ourselves to the more mundane task of calibrating output and not attempt judgments as to whether more or less of it is valuable now than was the case in earlier days.
How precisely can we document trends in the production of books and periodicals and relate them to rates of increase in the acquisition of library materials? "Book industry statistics," Chandler Grannis wrote, "may ... be likened to a handful of wet spaghetti. They may be more or less digestible, even a bit nourishing; but they are messy, slippery, elusive, never tidy."[6] Though such statistics have improved in recent years, there nonetheless remain many inconsistencies and anomalies in the reporting. In some instances, for example, it may be that apparent recent increases in the number of titles published is simply a function of improved data collection: a larger percentage of the items produced is now being "captured."
Moreover, all but one of the data sets used here to measure trends in book-title output have an important limitation for purposes of this study: they count items, perhaps even a great many items, that academic libraries would not choose to purchase. The fact that these data include publications written by nonacademics does not necessarily limit their usefulness, since academic libraries purchase many such items. Their usefulness is limited because, in addition to such materials, they also include items of other types---mass market paperbacks in some instances, university theses and government pamphlets in others---that do not figure in the acquisitions practices of many libraries. These limitations notwithstanding, the available data do permit us to make some provisional observations about changing levels of book-title production, seen in relation to acquisitions.[7]
Our principal concern is the relationship between levels of book production and levels of library acquisitions.[10] Since 1912 the overall average annual rate of increase in volumes added gross at our Research 1 composite has been more rapid (3.2 percent) than that of total titles published domestically (2.8 percent). However, most of the growth over the past three-quarters of a century in the number of volumes added gross occurred during the single decade of the 1960s (fig.5.2). [small | large] This unprecedented expansion was presumably the result of systematic retrospective purchasing permitted by double-digit increases in the materials and binding budget and, perhaps more important, the founding of many new serials and thus an increase in the number of volumes available for acquisition in any given year. The slopes of the curves diverge dramatically after 1970, with the annual number of volumes added gross staying more or less constant while the number of book titles published continued to rise at a steady rate.[11]
While these data document trends in the entire domestic book industry for the past century, inspection of data for a group of fields more relevant to the acquisitions practices of academic libraries provides a sharper sense of developments during the past two decades (fig. 5.3)[small | large] .[12] The shapes of the two curves are quite similar, with the same peaks and valleys, albeit with publications in the select fields growing at a slightly slower rate during these decades than publications in all fields (2.0 percent versus 2.3 percent). During this same period, the number of volumes added gross at our All 24 library composite actually declined slightly---at an average annual rate of -0.6 percent (fig. 5.4) [small | large] . The steadily growing gap between acquisitions and titles published is evident in the figure.
It is also useful to look even more closely at publications disaggregated by particular fields, since shifts in the relative field-shares are relevant to the acquisitions practices of academic libraries. Just as university curricula are redesigned in response to changes in the nature of scholarship and the emergence of new fields of inquiry, so libraries are expected to be responsive to shifts in the relative shares of title output represented by particular fields, even while continuing at the same time to build collections in fields that previously represented a larger share of the total number of titles produced. Shifts in relative proportions also are important because the price of published materials can differ enormously by subject area (discussed later).
The numbers of titles in many of the traditional arts and sciences fields---such as biography, literature-poetry-drama, and art-music--- as well as in education have decreased as a percentage of total publications over the past two decades. (See table 5.1[small | large].) The most significant drop is in the share of publications represented by the combined field of literature-poetry-drama, which fell from 17.2 percent in 1970 to 9.1 percent in 1988. Concurrently, there were increases in the shares of many professional-applied fields, including business, law, technology, and particularly medicine. Between 1970 and 1988 there has been an overall shift of approximately 9 percentage points toward the professional-applied fields. Once again, it is important to note that these counts include many titles that academic libraries would not acquire. Insofar as they are indicative of general trends, however, they are illuminating and confirm our impression of "global" shifts in publication patterns.
This redistribution of publications by field has had a real impact on library budgets because of differential pricing. Those fields that experienced the most significant gain in percentage share of publications since 1970 are precisely those with the highest average per-volume prices of hardcover copies---business ($37.51), law ($50.85), medicine ($66.59), and technology ($65.26). (Science also had a very high average per-volume price [$66.91], and maintained a fairly constant but significant percentage share of about 9.5 percent.) The fields that experienced the largest losses in percentage share of publications had significantly lower prices---biography ($25.99), education ($33.55), literature ($30.85), and poetry and drama ($28.02).[13]
Some part of the redistribution of publications by field is related, albeit in a complicated way, to shifts in enrollment patterns at both the undergraduate and graduate levels. At the undergraduate level, the arts and sciences share of baccalaureate degrees conferred declined after about 1970, especially at comprehensive institutions, concurrent with increases in degrees conferred in preprofessional subjects.[14] At the graduate level, the number of Ph.D.s peaked in 1973 and then began to decline, particularly rapidly in the arts and sciences. Concurrent with this decline was an increase in the number of degrees awarded in professional fields---especially medicine, law, and business. There was also a shift in the composition of doctoral degrees awarded within broad fields toward more applied subjects; for example, engineering grew faster than the sciences between 1973 and 1988, and, within the social sciences, clinical psychology grew faster than anthropology.[15] This general movement away from the arts and sciences may well have had a dual effect on the book publishing industry---creating both a declining pool of potential authors in the arts and sciences and a declining demand for books in these fields.
Trends in book publishing over the last three decades are reflected
also in the experiences of university presses. The activities of this set
of publishing institutions are especially relevant for present purposes
because academic libraries are likely to purchase a large proportion of
the aggregate list of their publications. Data compiled from records maintained
by the Association of American University Presses (AAUP) show patterns
that are something of an amalgam of the trends shown here for all domestic
publishers and the financial histories of the universities that are homes
to most
of these presses (fig. 5.5) [small |
large] .[16]
The AAUP data also show a rapid increase in the number of titles published throughout the 1960s. Beginning in 1969, however, there was a period of very little growth at these presses, a pattern not seen in the Bowker data (which show continued increases until about 1979). This flattening is similar to that documented in the first section of this study with respect to library acquisitions and the number of Ph.D.s awarded nationally. The financial condition of university presses is in some cases linked to that of their parent institutions, and the difficult circumstances of the 1970s inevitably had an impact on the number of titles that they could afford to publish. Moreover, given that academic libraries constitute one of the principal markets for monographs published by university presses, restrictions on their acquisitions budgets are more likely to affect the activity of university presses than of commercial presses, which predominantly serve other markets.
Beginning in 1979, there was a substantial recovery in the number of titles published annually by these university presses, and rapid expansion has continued without abatement through at least 1988 (the last year for which we have these data); the number of titles published annually by the university presses increased by a factor of 1.75 over this eleven-year period (5.2 percent per year). The average annual rate of increase since 1970 for these university presses has also been more rapid (3.6 percent) than the rate over the same period for all domestic presses in our selected fields (2.3 percent), despite the fact that the domestic presses expanded over that entire time period while the expansion at the university presses did not really begin until 1979. Moreover, growth in the aggregate number of titles published by all AAUP member presses is understated by our data because we have excluded those presses that did not begin to report titles published until after 1963.
This extraordinary increase in titles published by university presses was probably due to a confluence of forces including greater availability of good manuscripts as a result of the shifting boundary between commercial and university presses.[17] The recovery of AAUP presses was probably influenced as well by the desire to respond to the emergence of new scholarly fields. (This latter development is particularly relevant to trends in scholarly periodicals and is discussed in greater detail later.) Significantly, the increase in titles published was not a function of a recovery in the purchasing power of acquisitions budgets, since, as we have seen, the average number of volumes added gross has actually declined since 1970. A direct comparison of titles published by university presses and volumes added by all 24 of our research libraries shows clearly that since 1974 new university press publications have far outstripped acquisitions (fig.5.6)[small | large] .
One other dimension of publishing activity must be considered in assessing the implications of recent trends in acquisitions by research libraries---namely, international publishing trends, since these libraries collect many materials published outside the United States. The United Nations Educational, Scientific and Cultural Organization (UNESCO) has collected data documenting international book production for several decades, and we have sought to construct a subindex that would be particularly useful for our purposes, based this time, however, on countries rather than on subject matter. We have aggregated publishing data for six Western European countries (France, West Germany, Italy, the Netherlands, Switzerland, and the United Kingdom) and combined them with comparable data for the United States.[18]
International production of books---predictably---dropped sharply between 1938 and 1944 and then began to rise again in 1945. A period of rapid expansion between 1945 and 1948 was followed by four decades of more moderate but remarkably consistent growth. (fig. 5.7[small | large] and Appendix Table 5.1a[small | large] and Table 5.1b[small | large]; data before 1950 are not presented on the figure because they are not complete for all countries.)
The average annual rate of increase in titles published between 1950 and 1988 in Western Europe (3.7 percent) was greater than the annual rate of increase in volumes added at our Research 1 composite during those same years (3.0 percent), but again these overall averages mask major differences in growth rates during specific time periods. Until 1970, the rate of increase in volumes added gross was much greater than the rate of increase in titles published in Western Europe. The rate of increase in volumes added gross peaked in that year and has even declined slightly since then, while the number of titles published has continued to increase steadily over the past two decades (fig. 5.8)[small | large] .
In this context what is especially important is the relative percentages of the total represented by output in the United States and in Western Europe, and these percentages are changing. In 1971 American titles represented almost 40 percent of the total book production in the United States and these six Western European nations. By 1988 the U.S. share had dropped by more than 5 percentage points to about 34 percent.[19] The decrease in the United States's share was due to a flattening in the number of titles published annually in this country concurrent with a continued increase in the number of titles published in these Western European countries. Moreover, with the exception of Germany, the countries that produce the greatest number of titles annually---the United Kingdom, France, and Italy---are also those that have had the most rapid rates of increase since 1970 (between 3.5 and 5.0 percent per year compared to 1.1 percent for the Netherlands and 2.7 percent for Switzerland). Germany, starting from a base larger than any other Western European country's in this data set, has had a more moderate rate of increase since 1970 (2.6 percent) but a rate still greater than that of the United States (2.1 percent). In brief, the number of titles published in Western Europe has been greater than the number published in the United States and has increased more rapidly (table 5.2).[20]
Disadvantageous rates of exchange during the late 1970s and the mid- to late 1980s exacerbated the consequences of these trends for the acquisitions practices of American research libraries. Between 1985 (when the Plaza Accord depreciated the U.S. dollar against other major currencies) and 1988, the exchange rates for the dollar fell against these Western European currencies by at least 60 percent. West Germany stands out as the country that published the greatest number of titles annually and for which the rate of exchange was least favorable since the early 1970s.[21] In all countries, however, the confluence of these two factors---rapid increases in book production and unfavorable rates of exchange---has had important consequences for American academic libraries that wished to maintain collections international in character.[22]
For present purposes what is most important is the disparity between the rates of growth in the number of titles published shown by all these data sets and the number of volumes added by our group of research libraries (table 5.3). During the 1960s the average annual rate of increase in book titles published was comparable to that of volumes added only for the university presses. During this period, university libraries still constituted the principal market of the university presses, and the expansion in higher education during those years surely was the main factor propelling this extraordinary expansion of title production.
Then, however, the number of volumes added yearly within this group of 24 libraries decreased between 1970 and 1982 at an annual rate of -1.4 percent while the number of titles published, domestically and internationally, was increasing at a rate of greater than 2 percent per year. Furthermore, the number of titles published (according to all the data reported here) continued to expand into the 1980s. While the contraction in volumes added gross by these libraries appears to have tapered, there is no evidence of any major recovery. It appears, then, that libraries have been able to respond less effectively and comprehensively to developments in the publishing industry---in particular, the steady expansion in the number of titles published since the 1960s---than at any previous time in the 20th century.
The potential consequences of these trends are evident. Some members of the library profession are concerned that as acquisitions budgets are less adequate to perceived needs and permit less comprehensive coverage of the world's output of books, there is a tendency to concentrate on core materials, with the result that library collections are perhaps beginning to resemble one another more than before and lose some of the variety that previously distinguished them and some of the richness that characterized the entire national collection.[23] To describe the significance of these developments in starkest form: as libraries are increasingly unable to respond effectively to increases in the numbers of book published, the national collection is characterized by less comprehensive coverage of the world's title output, and access to information, the "capital" of scholarship, may be said to be narrowing in this important respect.
For a variety of reasons, statistics on trends in the production of serials and periodicals are not as easily acquired as those for book production. Unlike monographs, which can be defined clearly albeit arbitrarily by a minimum number of pages,[24] there exists no clear definition as to what constitutes a serial. The Association of Research Libraries, for example, uses a fairly standard definition of a serial as "a publication issued in successive parts, usually at regular intervals, and as a rule intended to be continued indefinitely."[25] Nevertheless, there has been some uncertainty as to whether items such as government documents and monographic serials should be counted in this category. Also illustrative of definitional difficulties is the fact that between 1972 and 1974 ARL used this same definition but called the relevant category "Current Periodicals."[26]
Moreover, what is most relevant to research libraries is not the total universe of serials but the subset of scholarly journals, which are even more difficult to define. In a preliminary report to the Mellon Foundation one economist wrote, "...no existing definitions precisely distinguish between journals and `other periodicals,' much less, scholarly journals and `other journals.' Lack of agreement on proper classification criteria has led to an enormous range in estimates of the number of journals published: figures from less than 5,000 to upwards of 100,000 have been cited."[27]
Even if clear definitions of "serials" (or "journals") existed, counting would still be very difficult. "Publication," wrote Allen B. Veaner, "is a living thing, and trying to count its components may be as futile as attempting to number the cells of the human body."[28] This description seems especially characteristic of serial publishing since a single title may be "alive" for several decades or even centuries. Moreover, unlike books, serials may merge, split, or even change title during their lifetimes; publication of particular titles may be assumed by another publisher. In the past few decades commercial publishers, benefiting from economies of scale, have achieved a significant role in the academic serials arena and now publish, for example, the proceedings and other publications of academic societies that were once published by the societies themselves.
Despite these complications, we can document very general trends, particularly with regard to the timing of periods of expansion. This is important because the proliferation in the number of journals has been one of the primary sources of pressure on library budgets both directly and indirectly through effects on serials prices.
The largest serials database that exists is Ulrich's, which includes "all publications that meet the definition of a serial except general daily newspapers, newspapers of local scope or local interest, administrative publications of major government agencies that can be easily found elsewhere, membership directories, comic books, and puzzle and game books." The 30th edition of Ulrich's International Periodicals Directory lists more than 118,500 titles.[29]
We can compare the growth in this serials universe to the growth of current serials in our All 24 library composite beginning in 1972 (fig. 5.9)[small | large] .[30] The upward trend in the number of serials contained in the Ulrich's universe was particularly rapid during the 1970s. Beginning in about 1983, however, the growth appears to have tapered. The pattern for current serials was similar; however, the rate of growth was slower during the 1970s, and beginning in the early 1980s the curve was essentially flat. As the proliferation of serials continued, libraries did not---probably could not---respond with more serial subscriptions, and the gap between serials published and serials acquired began to widen. The trends displayed here are not independent of each other; the slowing of the growth in the Ulrich's universe of serials during the mid-1980s may be attributable, at least in part, to the declining demand for serials earlier in the 1980s.
Another way of looking at the question of the timing of expansion in serials publication is to take all periodicals currently available and graph their founding dates. This method does not take into account journals that have ceased publication, but it does give an idea of when the "continuing" journals were first published. We are able to use data from the Science Citation Index with the founding dates of the 1990 Source Publication list graphed in (fig. 5.10)[small | large]. [31] This figure shows that the real proliferation in science literature began in the 1950s, with the number of journals founded in that decade more than double that of the previous decade. This growth continued into the 1960s and 1970s, with 43 percent of the journals in this list founded in those two decades alone. The proliferation tapered somewhat during the 1980s, with the number founded almost returning to the 1950s level.[32]
Data on periodicals in the field of modern languages and literature may serve as a rough index of trends in the general availability of scholarly journals outside the sciences.[33] In this field there were substantial increases in the number of journals founded throughout the post-World War II era. The decade of the 1970s, however, stands out, as it alone witnessed the founding of more than 400 new journals (fig. 5.11)[small | large] . Although there was a pronounced slowing in the rate of increase in the 1980s, more than half of the titles currently available were first published during the last two decades.
This picture must be modified somewhat by the conclusions reached by Daniel Uchitelle, director of the Center for Information Services at the Modern Language Association, in a very recent unpublished paper in which he tracks the dates when journals in literature ceased publication throughout the 1970s and 1980s. When this information is combined with the more familiar numbers of new journals begun in this period, it appears that the total number of journal titles increased only gradually. Indeed, Uchitelle suggests that the number of journal cessations has been larger than the number of inceptions in recent years.[34]
The founding of so many journals in the decade when the academic labor market was at its weakest point may be attributable in part to a heightened pressure to publish and to the efforts of academics to seek new outlets for their scholarship. Also, it has been argued that the established journals are slow to reflect changes in scholarship and that the resultant founding of new journals is a function of the redefinition of scholarly disciplines. The proliferation of journals in the humanities is surely related in some way to the debates that have occurred in many of these fields (and in some of the related social sciences) about the virtues and limitations of various methodologies and theories.[35] These developments have served to transform the character of scholarly discourse and have led in some instances to the founding of new journals, as have changes in the content of various disciplines (the interest in the experiences of "nonelites," for example, and of women and members of various racial and ethnic minorities). The increase in specialization is yet another force at work here.
Such increases in the number of serials inevitably raise questions about a decline in the quality of scholarship, although the increases do not in themselves substantiate such concerns. One need not enter into arguments about quality, however, or about the virtues of various scholarly approaches to appreciate that the proliferation of journals has had important consequences for academic libraries. Even journals that in some quarters are considered less prestigious or whose methodological approaches are deemed problematic will contain some number of items of interest, and many libraries will want to continue to maintain serials collections that are comprehensive in scope. Nonetheless, despite the substantial redeployment of library acquisitions funds toward the serials budget (see fig. 4.7 [small | large]), the number of current serials acquired has increased only modestly. The obvious inference is that---as all librarians know too well---the prices of serials have also increased significantly, especially during the decade of the 1980s. We now turn to this difficult and complex topic.
[1] Carolyn J. Mooney, "In 2 Years, a Million Refereed Articles, 300,000 Books, Chapters, Monographs," The Chronicle of Higher Education 37 (May 22, 1991):A17. However, in a letter to The Chronicle (June 26, 1991), Ann Okerson, director of the Office of Scientific and Academic Publishing at the Association of Research Libraries, observed that Mooney's article failed to take account of the fact that articles typically have more than one author; Okerson suggested that in 1986 the average number of authors per paper was 2.98 and that the number of articles published in a two-year period is therefore closer to 326,000 (two-thirds of 489,000), or fewer than 200,000 per year.
[2] See the July 9, 1989 issue of The New York Times.
[3] Carolyn J. Mooney, "Efforts to Limit `Trivial' Scholarship Win Backing from Many Academics," The Chronicle of Higher Education 37 (May 22, 1991):A13.
[4] See Francis Oakley, "Against Nostalgia: Reflections on Our Present Discontents in Higher Education," National Humanities Center Newsletter 12 (spring/summer 1991):1-14, especially p. 5; and Henry W. Riecken, "Scholarly Publication: Are There Viable Options?" Draft for the Research Library Committee [of the Council on Library Resources], October 1989, 5-6.
On the question of quality and its relationship generally to increases in the number of titles published, see also two provocative recent articles on the editing of book manuscripts at commercial publishing houses: Jacob Weisberg, "Rough Trade: The Sad Decline of American Publishing," The New Republic 204 (June 17, 1991):16ff.; and Ted Solotaroff, "The Paperbacking of Publishing," The Nation 253 (October 7, 1991):399-404.
The decline in editing standards noted by Weisberg in particular, if it in fact exists, is certainly to be related to the phenomenon of overproduction. Any attempt libraries might make to be selective in their acquisition of titles published by commercial houses would involve them inevitably in a debate about quality; Warren J. Haas has argued that "[l]ibrarians cannot make qualitative judgments by themselves, but as senior administrative officers they have the responsibility to see to it that those judgments are made." ("Reflections/Directions," Council on Library Resources Reports 3 [February 1989]:2).
[5] On these questions, see Ernest L. Boyer, Scholarship Reconsidered: Priorities of the Professoriate, A Special Report (Princeton, New Jersey: The Carnegie Foundation for the Advancement of Teaching, 1990). Boyer (pp. xi, 12-13) describes a "narrowing" of "the standards used to measure academic prestige" in the post-World War II era: "professors were expected to conduct research and publish results. Promotion and tenure depended on such activity....[T]he research mission, which was appropriate for some institutions, created a shadow over the entire higher learning enterprise." In "Publish or Perish: The Troubled State of Scholarly Communication," Scholarly Publishing 22 (1991):131-142, Dennis Carrigan offers another, briefer history of the developments Boyer describes.
[6] Chandler B. Grannis, "1974: U.S. Book Industry Statistics: Titles, Prices, Sales, Trends," Publishers Weekly 207 (February 3, 1975):39.
[7] An additional measurement point we should make is that many sources of data on book production distinguish new titles from new editions of existing titles. Since one cannot assume that a library acquired a particular title when it was first published, our analysis makes use of the combined total of new titles and new editions.
[8] For the relationship of university press activity to the financial circumstances of parent institutions, see Ellen Coughlin, "Face of University Publishing Changed by Years of Adversity, Decades of Growth," The Chronicle of Higher Education 36 (June 27, 1990):A1.
[9] The data are taken from The Book Publishing Annual: Highlights, Analyses and Trends, 1985 edition (New York and London: R. R. Bowker Company, 1985), 127. See the various qualifications concerning the data in notes a-c of that source. The closer analysis of trends for the years 1970-89 offered below is based on data taken ultimately from the same source (the R. R. Bowker Company).
[10] For another analysis of this relationship, see Ann L. O'Neill, "Book Production and Library Purchases: Looking Beyond the Thor Ruling," Publishing Research Quarterly 7 (Summer 1991):39-51.
The categories that are the subject of comparison in this section have many dissimilarities, one of which is the fact that book production data are reported for the calendar year while volumes added gross are reported for the academic year. The analysis is meant only to provide a sense of the changing levels of book publishing as compared with acquisitions of academic libraries.
[11] An important caveat with regard to this comparison is that the data for volumes added gross include serial titles, while book production data (Bowker here and, later in the discussion, UNESCO and AAUP) include book titles only. However, the important point is that during the 1970s libraries failed to maintain rates of growth in acquisitions equivalent to rates of growth in book titles published annually, and the fact that serials (many of which were founded during the 1970s) are counted in the volumes added gross data simply serves to reinforce the notion that libraries have not been able to acquire a full collection of available materials.
[12] To define a relevant data set, we limited the publications in the select fields to publications in agriculture, art, biography, business, education, history, law, literature, medicine, music, philosophy and psychology, poetry and drama, religion, science, sociology and economics, and technology; we excluded titles in all other fields (fiction, general works, home economics, juveniles, language, sports and recreation, and travel).
While one could certainly challenge this particular choice of fields, there is evidence that supports this classification. Beginning in 1981, a distinction can be drawn between hardbound and trade paperbound only and "all hardbound and paperbound" (which includes mass market paperbound books). We report the latter figure here for all years to maintain consistency with figures before 1981, which were not so disaggregated. When the number of titles in the subject areas we excluded are deducted from the total count, the difference between the total numbers of hardbound and trade paperbound books only and of all hardbound and paperbound books almost disappears. In other words, mass market paperbound books---the kinds of items academic libraries would be unlikely to purchase---seem to be concentrated almost entirely in the fields that we excluded from consideration.
These data are taken from the relevant issues of Publishers Weekly. Until 1976, the count of titles published in a particular year was as of the beginning of the following year. Beginning in 1976, the count was as of midyear in the following year.
[13] Prices are as of 1988. See Chandler B. Grannis, "Titles and Prices, 1988: Final Figures," Publishers Weekly 236 (September 29, 1989):26.
[14] See Turner and Bowen, "The Flight from the Arts and Sciences," for a detailed analysis of the factors responsible for this trend.
[15] See Bowen and Rudenstine, In Pursuit of the Ph.D., particularly chapters 2 and 3.
[16] We use the acronym AAUP throughout this discussion to refer to the Association of American University Presses, not to the American Association of University Professors.
The data on titles published were obtained directly from the AAUP. For purposes of this study a subset of the total membership of the AAUP was selected. In order to define a stable population, the selection was limited to 51 constituent member presses that reported data in 1963 and more or less continuously thereafter: Arizona, Brookings Institution, California, Chicago, Columbia, Cornell, Duke, Florida, Fordham, Georgia, Harvard, Hawaii, Illinois, Indiana, Iowa State, Johns Hopkins, Kansas, Kentucky, Laval, Louisiana State, M.I.T., McGill, Metropolitan Museum, Michigan, Minnesota, Missouri, Nebraska, New Mexico, New York, North Carolina, Notre Dame, Ohio State, Oklahoma, Pennsylvania State, Pittsburgh, Princeton, Rutgers, Smithsonian Institution, South Carolina, Southern Methodist, Stanford, Syracuse, Texas, Toronto, U.S. Naval Institute, Washington, Wayne State, Wisconsin, Yale, Cambridge, and Oxford.
[17] Sanford Thatcher, director of the Pennsylvania State University Press, has suggested that some of the traditional distinctions between university and commercial presses and between the profiles of the authors each type serves may no longer be current and indeed may not have been current for some time. ("Scholarly Monographs May Be the Ultimate Victims of the Upheavals in Trade Publishing," The Chronicle of Higher Education 37 [October 10, 1990]:B2-B3.) Thatcher suggests that the decision of the management at Random House---to have its subsidiary, Pantheon Books, publish books that would sell more copies than had many of Pantheon's previous titles---would not disadvantage the kinds of authors whom Pantheon had traditionally served. These authors would have another outlet for their work, the university presses, which have been interested for some time in the kinds of "mid-list books" that typified Pantheon's output. Given that they can capitalize on the interest of this new class of authors, the university presses may be increasingly unwilling to publish titles with potential sales of 1,000 copies or fewer. Under such circumstances the authors ultimately likely to be disadvantaged by these changed circumstances are "scholars seeking publication of their monographs in fields where average sales are low."
Solotaroff ("The Paperbacking of Publishing") described the conditions at commercial publishing houses that led to the "shift in boundary" between the commercial and academic sectors of which Thatcher wrote. For other statements of general trends in academic publishing, see Ellen K. Coughlin, "Face of University Publishing Changed," A1, A8-A9; John F. Baker and John Mutter, "University Presses: Weighing the Options," Publishers Weekly 238 (August 2, 1991):12-15; and, most important, Herbert S. Bailey, Jr., The Rate of Publication of Scholarly Monographs in the Humanities and Social Sciences, 1978-1988 (New York: Association of American University Presses, 1990).
[18] This decision was based partly on the practical grounds that it was possible to construct for these countries a fairly continuous data set from the 1930s to the present. Also, titles published in these countries have been of particular interest to the research libraries with which we are concerned in this study. To be sure, as universities in the United States have redesigned their curricula to include course work on the intellectual traditions of other, non-European cultures, their library collections have changed accordingly and will continue to change. Nonetheless, we may assume that until fairly recently, titles published in Western Europe constituted the largest proportion of nondomestic titles purchased by most academic libraries. (For some evidence that Western European countries provide the largest numbers of nondomestic materials to academic libraries, see Sally F. Williams, "Construction and Application of a Periodical Price Index," Collection Management 2 (Winter 1978):329-344, especially p. 331, where it was observed that the United States, France, Germany, Great Britain, and Italy were the major sources of periodical titles acquired for the central research collection at Harvard).
These data are taken from the following sources: For 1937-49, Preliminary Statistical Report on Book Production in Various Countries (Paris: UNESCO, Statistical Service, 1951), 18, 32, 44, 47, 66, 71, and 74. For 1950-78, An International Survey of Book Production During the Last Decades, Statistical Reports and Studies, No. 26 (Paris: UNESCO, Office of Statistics, Division of Statistics on Culture and Communication, 1982), 49-55 and 58-64. For 1979-88 (France, Germany, Italy, Netherlands, Switzerland), appropriate issues of the UNESCO Statistical Yearbook (Paris: UNESCO, 1983- ). For 1979-84 (United Kingdom), ibid. For 1985-87 (United Kingdom), The Bowker Annual Library and Book Trade Almanac, 36th Edition 1991, comp. and ed. Filomena Simora (New York: R. R. Bowker, 1991), 445. For 1979-81 (United States), The Bowker Annual of Library & Book Trade Information, 30th Edition, 1985, comp. and ed. Julia Moore (New York and London: R. R. Bowker Company, 1985), 492.
The R. R. Bowker Company is the source of UNESCO figures for the United States and the United Kingdom for all years. The Bowker data presented here (hereafter Bowker international data) representing U.S. book production differ from the Bowker data presented earlier in the discussion of book production in the United States in that the latter do not include either university theses or government documents. Beginning in 1965, the R.R. Bowker Company added these additional categories when reporting counts to UNESCO to maintain comparability with reporting standards of other countries.
For 1988 (United Kingdom), the value was projected from the average annual rate of increase in the number of titles for the time period 1937-87. For 1982-88 (United States), data were imputed based on the relationship of Bowker (domestic) data to UNESCO data for the period 1965-81. For 1950, 1956, 1972, and 1973 (West Germany), the values were imputed on the basis of values of neighboring years.
Although the UNESCO data are disaggregated by subject area, the categories are defined in such a way that any of them might contain many titles of interest to academic libraries. One cannot, therefore, easily identify subject categories that are likely to be less pertinent in this context. The 1987 UNESCO questionnaire is included in Gretchen Whitney, "The UNESCO Book Production Statistics," Book Research Quarterly 5 (Winter 1989-90):12-29, especially 25-29. See the various issues of Publishers Weekly for the items included in the Bowker data.
[19] We discuss relative shares of U.S. and Western European book production only for the post-1970 period because there is reason to believe that 1970 marked the beginning of a period of better reporting. Accordingly, for purposes of this analysis 1970 perhaps should be considered the beginning of a new statistical series.
[20] Although we have chosen to show trends in book production in Western Europe and the United States only, there is also evidence that book production is increasing more rapidly (in relation to the United States) in other areas of the world as well, particularly Asia. Japan, for example, is among the top five publishing countries, and book production there increased at an average annual rate of 2.8 percent between 1970 and 1986. The rate of increase was particularly rapid during the latter 10 years of that period. (Data are from UNESCO: An International Survey of Book Production During the Last Decades, and relevant issues of the Statistical Yearbook.)
[21] Moreover, more recent events, including the shift from state-subsidized to private-sector publishing in the former East Germany and the merging of the two German economies, have dramatically increased the prices of German library materials.
[22] Data on exchange rates are from the International Financial Statistics Yearbook, 1990 (Washington, D.C.: International Monetary Fund, 1990). Monthly exchange rates of German marks, Dutch guilders, and English pounds per dollar from 1973 to 1988 are graphed in Economic Consulting Services, "A Study of Trends in Average Prices and Costs of Certain Serials Over Time," 13-15; and annual exchange trends between the English pound, the French franc, the German mark, and the Japanese yen from 1975-1988 are graphed in Ann Okerson, "Of Making Many Books There is No End," 26. Both sections are contained in Report of the ARL Serials Prices Project (Washington, D.C.: Association of Research Libraries, 1989). For general statements of the difficulties libraries currently encounter in attempting to acquire international materials, see "Research Libraries in a Global Context," prepared by ARL staff, December 1989; "Scholarship, Research Libraries and Foreign Publishing in the 1990s," prepared by ARL staff, March 1991; and Jeffrey J. Gardner, "What They Have and How We Might Get It: Son of Farmington?" Paper read at the Seminar for the Acquisition of Latin American Library Materials, San Diego, California, June 4, 1991.
[23] That acquisitions practices may be so described was suggested in personal conversation by Dale Flecker of the Office of Systems Planning and Research, Harvard University Library. His view is shared by, among others, members of the staff of the Association of Research Libraries (see the ARL staff paper entitled "Research Libraries in a Global Context").
The tendency for individual libraries to concentrate their acquisitions on a core set of materials may result in collaborative efforts across institutions to preserve the national collection. Such efforts are not without precedent. In 1947, under the leadership of the Association of Research Libraries, more than 60 research libraries participated in a collaborative effort called the Farmington Plan. The intent of the plan was for individual libraries to "take responsibility for collecting and cataloging material from specific countries and/or areas with the intent of building a distributed national collection of foreign materials ensuring coverage of all major areas." However, the effort was abandoned in the early 1970s, with its demise attributed to several factors, "including the lack of a mechanism to monitor the degree of implementation and development of the plan, budgetary constraints of the early seventies that led many libraries to turn inward in their collection development efforts, and real or perceived deficiencies in the services of bookdealers designated for the different parts of the plan" (Jeffrey J. Gardner, "What They Have and How We Might Get It: Son of Farmington?" 5).
[24] One example is the recommendation adopted by the General Conference of UNESCO in 1964 for the purposes of standardizing international reporting of book production statistics. The Recommendation defined a book as "a non-periodical printed publication of at least 49 pages, exclusive of the cover pages." UNESCO, An International Survey of Book Production During the Last Decades, 18.
[25] This definition also appears in Ray Prytherch, comp., Harrod's Librarians' Glossary, 7th edition (Brookfield, Vt.: Gower Publishing, 1990):561.
[26] Kendon Stubbs and Robert Molyneux, Research Library Statistics, 1907-08 through 1987-88; A Guide to the Machine-Readable Version of the Gerould and ARL Statistics (Washington, D.C.: Association of Research Libraries, 1989), 28.
[27] Lisa Lieberman, Roger Noll, and W. Edward Steinmueller, "Economic Analysis and Empirical Protocol for Examining Scholarly Periodicals Pricing," report submitted to The Andrew W. Mellon Foundation, June 7, 1991, 14.
[28] Allen B. Veaner, "Into the Fourth Century" (College of Information Studies, Drexel University, 1986), 9.
[29] Ulrich's International Periodicals Directory, 30th Edition (New Providence, NJ: Reed Publishing, 1991), viii. The Ulrich's data used here are taken from Ann Okerson, "Of Making Many Books," 13, 15-16. In this report Okerson uses Bowker-Ulrich's serials database to estimate how the serials universe has grown since 1971-72, from 70,000 titles to 108,590 at present. Okerson then graphs average ARL library serials holdings as a percentage of the serials universe, showing that there has been a substantial drop in the percentage of titles collected in research libraries.
[30] "Current serials" includes items other than periodicals and, more important, items "received but not purchased." A better measure of libraries' ability to acquire serials would be "serials purchased" only, but ARL did not begin to collect data for this category until 1986.
[31] The titles in this universe were taken from the Institute for Scientific Information's printed list of the source publications in the 1990 Science Citation Index. Information on founding date and country of publication for each title was taken from Ulrich's International Periodicals Directory. When this information was unavailable from Ulrich's, we used the OCLC database. In just a few cases information was available from neither source.
[32] These data also allow us to gain some idea of the number of science journals published abroad (which tend to be more expensive than journals published domestically). Of those journals in this list of source publications for which the country of publication is known (3,658 out of 3,680), more than half (55.5 percent) were published outside the United States.
[33] The MLA Directory of Periodicals lists titles in modern languages and literatures and the date they were first published, so that one can determine the aggregate number of titles available for acquisition in any given year. See Kathleen L. Kent, comp., MLA Directory of Periodicals, A Guide to Journals and Series in Languages and Literatures, 1990-91 Edition: Periodicals Published in the United States and Canada (New York: The Modern Language Association of America, 1990).
In this instance as in others before, the counts will include some number of items that libraries would not choose to acquire; the aim in compiling the Directory is to be comprehensive in a way that libraries in their acquisition practices cannot be.
[34] Daniel Uchitelle, "An Analysis of Data from the MLA Directory of Periodicals to Describe Patterns of Journal Publication in the Humanities."
[35] Such debates are reflected, for example, in large-scale changes in scholarly ethos, from the perspectives reflecting the influence of 19th-century German positivism, to the fundamental challenges to that optimistic vision of the potential of scholarship posed above all by deconstructionism, to more recent exhortations to attempt anew to "to negotiate the perceived differences between subject and object, reader and text, interpreter or describer and the external cultural and other structures." On this subject, see Neil Rudenstine's survey in William G. Bowen and Neil L. Rudenstine, In Pursuit of the Ph.D., Appendix F, from which the quotation is taken.
