College and Research Libraries The Continuing Debate over Academic Branch Libraries Leon Shkolnik To centralize or decentralize: that is the question. Librarians have debated the organization of the academic library for the last century. This article analyzes both sides of the debate, placing each in its proper historical perspective. The author presents prospects for the future organization of the library in light of current trends and technological developments. ver the past century, librari- ans have debated the organi- zation of the academic library. Two distinct schools have de- veloped, each advancing logical and persuasive arguments. The debate con- cerns one vigorously contested issue: Should the academic library be central- ized in one main building or should it be decentralized into several branches based on differing divisional schemes? For the purposes of this paper, the terms departmental and branch library will be used interchangeably. Both sides of the debate will be examined and some views for future developments presented. The literature distinguishes three types of branch libraries. First, there is the professional library, such as a law or medicine library. In general, few librari- ans believe that these distinctively spe- cialized libraries should be incorporated into the main library. The second type of library is the undergraduate library, which is separated from the main branch because of the nature of its clientele. Many professionals believe that the un- dergraduate library should provide a multitude of services that cannot be of- fered in a modern research library. Fi- nally, there is the departmental library, a library established to serve a university academic department. This type of li- brary fuels continuous debate and will be the major focus of this paper. The concept of the distinctive depart- mental library separate from the main library building grew out of the seminar movement in late nineteenth-century Germany. German faculty members found it preferable to use their own co1- lections of books in their graduate semi- nars rather than rely on the resources of the existing, but antiquated, library sys- tem. Their books were usually more up to date and were kept in the faculty members' offices for easier access. Charles Kendall Adams first introduced this Eu- ropean concept in the United States when he used it in his English constitu- tional history class at the University of Michigan. 1 From this beginning, the practice of using a private collection to teach a seminar led to the development of the departmental library. The reasons for the early acceptance of departmental libraries are many and varied. Louis T. Ibbotson observed that the departmental library started as a pro- test. In the late nineteenth century, Ameri- can education was at a critical juncture. Education had greatly expanded, and Leon Shkolnik is Head of Reference, D'Youville College Library, D'Youville College, Buffalo, New York 14201. 343 344 College & Research Libraries that expansion created a severe need for books-a need the university library could not meet. Ibbotson believed that the library was "asleep" and that "on the whole it had never been awake and was very slow in waking." 2 In Ibbotson's opinion, the late nineteenth century was an American education renaissance to which the university library was indif- ferent. It awoke only after the depart- mental library had usurped its place on campus.3 In 1929, Mary C. Venn observed that university libraries, basically collec- tions of classical texts and commentaries, were out of step with the educational pro- cess. Endowments, permanent book funding, and comprehensive policies of book selection were all lacking. Departmental libraries were the natu- ral outgrowth of the days when libraries were mere prisons for books.4 Librarians were afraid to let the books out of their sight, an attitude that destroyed any idea of service to the community. As Venn wrote, "A librarian protected her books as a lioness her cubs."5 While local con- ditions determined the extent of decen- tralization, clearly the growth of education in the late nineteenth century, the concepts brought from Germany, and the woeful state of university libraries combined to create the departmental library system.6 By the end of the nineteenth century, the issues concerning decentralization of the university library had developed into their modern form . In 1895, Zelia Allen Dixson, an associate librarian at the University of Chicago, stipulated certain ad vantages and disadvantages of decentralization. With few exceptions, these arguments still dominate the liter- ature. Dixson stated that the major ad- vantage of decentralization was that it would allow the student of a certain dis- cipline to become familiar with the bib- liography of that discipline-"like a workman with his tools."7 This familiar- ity would encourage individuals to pur- sue original research and study because access to resources was so convenient. The disadvantages of such a system were two. First, departmental libraries could be physically distant from each other, which could cause innumerable July 1991 administrative problems. Second, de- centralization would tend to result in a narrowness of study, thus defeating the purpose of a total "liberal" education.8 Clearly, the notion of interdisciplinary study is hardly new; it was recognized and debated a century ago. In 1898, George H. Baker, a librarian at Columbia University, observed the many difficulties inherent in a decentral- ized library system. The cost of materials would grow immensely. Duplication of works would be nece·ssary to meet the needs of various departments. The ad- ministrative costs would overburden even the most affluent of universities. Additional staff would be needed to op- erate the many branches efficiently. Se- curity would be a problem; the danger of book losses would grow as the number of libraries grew. Finally, Baker noted the inconvenience of access to a collection for a person not from that department. 9 The debate over branch libraries has centered on one conflict: Librarians wish to centralize libraries to increase administrative control, while the faculty opposes centralization. William Warner Bishop of the Poly- technic Institute of Brooklyn observed that a major issue of decentralization would be the sense of possession, own- ership, and control of each departmental library.10 Each department would con- sider its library its personal domain, in- dependent from the whole. The faculty would not be willing to give up this source of prestige and control for the benefit of the entire university system. Bishop contended that any benefits of a decentralized system would be attained at the expense of economical and effi- cient supervision and direction.n How, for example, would a departmental li- brary cope with the complexities of cata- loging, classifying, inventorying, repair ing, and binding materials?12 These turn- of-the-century librarians clearly deline- ated the issues in the centralized versus decentralized academic library debate. As the twentieth century progressed, however, both sides found additional evidence to strengthen their positions. In the twentieth century, the debate over branch libraries has centered on one conflict: librarians wish to centralize li- braries to increase administrative con- trol, while the faculty opposes central- ization because it means a decrease in their authority, control, and prestige. The library literature is replete with articles presenting the pros and cons of each sys- tem. While these arguments for and against departmental libraries seem dia- metrically opposed, they both stress that local conditions are always the major determinant for the organization of the entire library. A small college has less need for branches than does a large uni- versity. The debate, therefore, really per- tains only to the largest and most research-oriented academic communities. DECENTRALIZATION The major point favoring a decentral- ized system concerns access to materials. Proponents of decentralization argue that literature on specific disciplines should be located near places of instruc- tion and research. In other words, im- mediate accessibility is the most ilnportant feature in the use ofbooks Y A faculty member at a university will walk no further to a library from his office than he will to his car. 14 Libraries should be organized to offer information in a way most useful to patrons. In 1983, Hugh C. Atkinson wrote, "We are not really in the business of running librar- ies, we are in the business of providing library service."15 Atkinson also stated that staff members in smaller libraries are happier and that any group consist- ing of more than ten to thirteen members will become bureaucratic and lose sight of the goal of service. 16 The decentralization argument is based on the theory that use of a collec- tion is directly related to access to it Y That is, ease of use is as important as access. Proponents of decentralization argue that because the sheer bulk of the collection is distributed throughout sev- eral branch libraries, a closer librarian- Academic Branch Libraries 345 user relationship can develop. The li- brarian would become a subject matter specialist who could deal more effi- ciently with the unique problems of a . particular discipline. Discussing the huge collection at Harvard University, Douglas Bryant argued that even utiliz- ing a . branch of one million was more convenient than using a general collec- tion of seven million.18 The decentralization argument is based on the theory that use of a collection is directly related to access to it. Advocates of decentralization also be- lieve that branch libraries result in a closer librarian-faculty relationship, which leads to greater faculty support of the library. When faculty members be- lieve they have a stake in the growth and development of the collection, they are more likely to provide support than if they view the library as one great and distant monolith. This support helps de- velop greater endowments. Cooperation between librarians and faculty benefits the librarians' mission to disseminate in- formation to users. Teaching faculty be- lieve that in a departmental arrangement they have greater input into book selection and acquisition, which, in turn, provides a more effective system of collection de- velopment.1 9 Other advantages of decentralization, as discussed by Snunith Shoham in 1982, include a flexible loan policy designed to meet the needs of the department, more personal attention from the staff, more direct service to a particular group, more collection responsiveness to users, and speedier searches.20 In general, a decen- tralized structure, ideally, would re- spond more directly to the needs of the users. Because this is the goal of the li- brary in general, decentralization, sup- porters argue, necessarily would be the preferred arrangement. Generally, older and larger libraries are more apt to be decentralized. 21 Rob- ert R. Walsh believes there are two main causes for this fact: The sheer bulk of the 346 College & Research Vbraries collection necessitates some kind of de- centralized plan, and small offices or lab- oratories adding to their collections often create small libraries. If the main library has room to expand physically, branches can be avoided, but this is often not possible. When funds are available, branch libraries seem to be the logical solution for the problem of space and the needs of a growing library clientele. 22 Many other points of debate exist. Pro- ponents · of decentralization hold that separation of the library from the labora- tories and the classrooms is not cost ef- fective. It is not efficient to have highly paid faculty members walking between their classrooms or laboratories and a distant central library when all their ma- terials could be close at hand.23 In addi- tion, decentralizationists postulate that collection development improves when subject specialists do the choosing. De- partmental libraries foster a feeling of pride among both faculty and librarians. A centralized library will often overlook problems on which a specialist-oriented library would focus; a decentralized sys- tem can arrange hours of service, refer- ence hours, circulation, reserve readings, and all other services to meet the unique needs of the department. By concentrat- ing on service, a departmental library becomes more valuable to the patron than the service provided at a general reference desk. 24 Branch library proponents dismiss the issue of interrelation of subject areas. They believe that little overlapping of subject materials exists and that only an occasional nondepartmental user would have any need for the departmental li- brary. The final argument for decentral- ization involves the location of the library within the university community. Propo- nents of the branch system stipulate that libraries placed close to classrooms bet- ter link instruction and research, inspir- ing scholars and allowing students to identify with their specialities.25 At larger schools with graduate and profes- sional departments emphasizing re- search, branch libraries provide better, more direct service with little increase in cost.26 Shoham dismisses the notion that July 1991 branch libraries increase cost because of duplication of materials and staff. He argues that these costs would be incurred in increased service in the main library. Space would still have to be found and personnel hired to meet the needs of the university. 27 Shoham also concludes that users prefer accessibility to the greater completeness of the collection. In summary, the proponents of a de- centralized library system consider ease of access and a perception of greater ser- vice to the user as the primary reasons for branch libraries. The problems of ad- ministration, while recognized, are con- sidered secondary to the basic goal of providing service. Among the major pro- ponents of such a system are faculty members, who view branch libraries as a means of perpetuating their influence in library administration. Because their concern is not library administration, they consider it of lesser importance; use and service are the main goals of the library, in their view. Branch libraries are an entrenched institution on the cam- puses of large research universities. The possibility of these branch libraries vol- untarily giving up their local power is remote. CENTRALIZATION In 1986, Robert A. Seal stated that "for the most part, the history of branch li- braries in the twentieth century has been an effort by librarians to centralize facil- ities and materials." 28 The early litera- ture notes major problems concerning a decentralized library pattern. In 1929, Mary Venn, a reference librarian at Ober- lin College, observed that the interests of departments were so closely allied that duplication was necessary. 29 Venn felt that centralization is necessary to care for books more economically and efficiently, to provide safety from fire and theft, and to promote inter- departmentalism.30 Efficiency through uniform catalog and classification sys- tems outweighs any gain from a decen- tralized system. Thomas D. Watts of the University of Texas attacked the branch library concept as leading to a "fraction- alization" of knowledge.31 He saw branch libraries as being inconvenient because users would have to go to many different locations . In addition, particu- lar collections would become isolated from the rest of the library, making them almost inaccessible. Watts also discussed the additional costs in staffing, security, and collection development, and he ar- gued that a decentralized system hinders communications between departments. He concluded that the existence of branch libraries results in loyalty of librarians to a given branch instead of to the univer- sity library as a whole. Each department would desire its own branch library, which would destroy the notion of unity of knowledge.32 Centralization is necessary to care for books more economically and efficiently, to provide safety from fire and theft, and to promote interdepartmentalism. The major advantages of a centralized system are administrative. It is generally agreed that departmental libraries are an administrative nightmare. Problems of control, coordination, and communica- tion are massive.33 University librarians have always been opposed to seminar libraries, but did nothing about them until the situation became so critical that it required action. The immediate prob- lem was the cost of duplicating the col- lection. 34 Departmental libraries seemed to operate without regard to administra- tive costs. Centralization of manage- ment produces economies of time and money and provides more convenience to the greatest number of users. 35 Louis Ibbotson recognized the trend toward centralization as early as 1925, when he stated that greater control was needed to facilitate efficient administration and to avoid duplication.36 The early arguments for a centralized system-efficiency, cost, and security-are still central and unas- sailable. N. Orwin Rush, director of the library at Aorida State University, ob- served in 1962 that if more money were spent on operating departmentallibrar- Academic Branch Libraries 347 ies, less would be available for acquiring more books. 37 He found eight important disadvantages of a decentralized sys- tem. In his opinion, separate collections waste time, cause administrative prob- lems, duplicate materials, increase costs, result in less effective service, complicate security problems, affect hours of ser- vice, and retard interdisciplinary study. 38 Any advantages gained through depart- mental libraries, such as greater conve- nience and access for the researcher, cannot overcome these eight disadvantages, Rush contended. The ad verse effects decen- tralization would have on the general uni- versity patron and to interdisciplinary studies far outweigh any possible ad- vantages.39 Robert Miller identified seven consid- erations in the debate regarding central- ization, concluding that in at least four of the categories-cost, interrelation- ship, efficiency, and educational signifi- cance-a centralized system was clearly preferable. Cost seems the most obvious factor; a centralized system avoids the problem of duplicating materials, either through multiple purchase or through copying. Also, one main building com- bines staff, uses staff more efficiently, or reduces staff where necessary. There is no need for a separate catalog, a reserve desk, a reference desk, and other techni- cal services for each branch library.40 Over- all, a central library provides greater accessibility to the entire collection, not only for users, but for staff, creating bet- ter service and a more standardized sys- tem. Standard hours of service, circulation regulations, reference services, and inter- library loan are but a few of the benefits of a centralized system. Also, materials are more secure if they are all in one location. 41 Two of Miller's considerations-in- terrelation of subject field and educa- tional significance-are closely akin. He believed that a centralized collection en- hances use by patrons in diverse aca- demic fields. While patrons would not walk to a different library to use a de- partmental collection, they are more apt to do so if it is convenient to the rest of the collection. Miller also stated that a 348 College & Research Libraries centralized library provides a common meeting ground for all students and fac- ulty, helping create a feeling of fellow- ship and a unity of knowledge, which are cornerstones of a university commu- nity. A centralized structure shows that the university administration considers the library a major part of the academic apparatus in which scholars from all dis- ciplines can converge and communicate. As Miller observed, "The day of the nar- row man must soon be over .... A central library serves to remind us of the miracle of print and its social importance in higher education." 42 Miller concluded that centralization is the only choice for a small college-decentralization would only cause headaches. For a large uni- versity with an accent on research, de- centralization is possible, but only as local conditions warrant and as the bud- get allows. 43 . DISCUSSION The arguments on both sides of the decentralization debate have remained constant throughout this century. The fundamental argument is one of control: Who will control the library-the librar- ians or the faculty? In many instances, the branch library has become little more than a relic of a system of librarianship that may have had meaning in the aca- demic community of the nineteenth cen- tury, but serves little purpose now. Because they thought departmental li- braries would be difficult to abolish, the board of directors of the Association of College and Research Libraries (ACRL) issued eight guidelines in 1975 to be used in determining the need for a branch library: 1. Mission of the parent institution 2. Campus geography 3. Enrollment patterns 4. Space requirements 5. Financial conditions 6. Instructional techniques 7. User expectations and demands 8. External pressures.44 The ACRL board believed that an anal- ysis of the need for a branch library should consider the requirements of the branch's primary clientele, as well as the July 1991 entire academic community. The analy- sis should compare any possible benefits of a branch library with the current li- brary situation. The comparative analy- sis must then be viewed in light of the philosophy and mission statement of the library.45 Any branch library must fit into the larger framework, objectives, and programs of the university. The ACRL guidelines included four general principles that should be consid- ered in all cases: 1. Responsibility for the management of all libraries rests with the library administration. Services are best when centralized. 2. The staff of the branches should report to the main library adminis- tration (not the departmental fac- ulty). 3. The goals and objectives of the branch should be clear and up- dated as needed. 4. All branches should have an oper- ating manual that delineates their raison d' etre. 46 The ACRL directors concluded that a branch library exists solely for the bene- fit of the users. The quality of the branch depends on its responsiveness to the needs of the community it serves. Once this responsiveness becomes question- able, so does the reason for having a branch library. Divisional libraries are a possible so- lution to the controversy. This plan would consolidate departmental librar- ies into three interdisciplinary libraries: science, humanities, and social science. The consolidation would allow for a de- gree of interdisciplinary study and would ease the administrative burden of smaller libraries because technical ser- vices and administration would remain centralized. As early as 1942, Lawrence Thompson observed that a subject-ori- ented library system would solve many of the problems of the departmental sys- tem. 47 While this plan encountered much criticism from faculty unwilling to relin- quish any influence in the administration of the library, many librarians believe that the divisional plan will eventually replace the departmental system.48 In 1986, William E. McGrath provided a quantitative study of the centralization debate. He created five distinct disciplin- ary clusters based on circulation of ma- terials in one discipline by users in another. These clusters included: hu- manities and arts, social sciences, life sci- ences, engineering, mathematics and science, and business and quantitative. 49 McGrath stated, "As a compromise be- tween complete centralization ... and com- plete decentralization ... it makes sense to consider derived clusters." 50 The clus- ter system puts departments into logical groups, and the clusters themselves re- flect familiar relationships among de- partments. McGrath argued that if a central library is outgrowing its main building and some sort of divisional scheme is necessary, this quantitative clustering could provide a guideline for such divisions. He concluded that whether the administrative decision is to centralize or decentralize, the decision can be more realistically supported by an analytical understanding of the collec- tion, its components, and how these components relate to each other. 51 The literature thoroughly discusses the advantages and disadvantages of de- centralization. For the past century, both sides in the debate have presented the same arguments, and some basic axioms have emerged. For instance, rising costs in library service are fundamental in de- termining if and to what extent decen- tralizati