College and Research Libraries From Authority Control to Informed Retrieval: Framing the Expanded Domain of Subject Access Prudence W. Dalrymple and Jennifer A. Younger The expanding domain within which subject searching takes place creates new challenges in providing effective subject access. Three changes in the domain of subject access are described. The purpose of this article is to propose a broader framework within which to analyze barriers to effective subject access and to propose directions for research and action. Methods of facilitating subject access are categorized as end points on a continuum: authority control, controlling the results of indexing, and informed retrieval-informing the process of searching. Identification of these two categories signals a paradigmatic shift toward a reliance on both authority control and informed retrieval to facilitate subject access. D his article reexamines some of the methods and approaches taken to provide subject ac- cess, especially in light of the many technological changes that have occurred in academic libraries over the past several years. The approaches are not necessarily new; some of them were proposed several decades ago. However, just as Vannevar Bush's dream of a MEMEX seems to be embodied in today's sophisticated workstations, the growing availability of integrated infor- mation systems, networks, and artificial intelligence may indicate that the ideal solutions which were once posed only as hypotheses are being developed today and will be adopted tomorrow. Solutions to problems previously limited by tech- nical constraints can now be realized and brought to bear in today's environ- ment. A central problem in academic librar- ies today is subject authority control. Subject authority control has generally been considered the domain of catalog- ers rather than reference librarians, of technical services departments rather Prudence W. Dalrymple is an assistant professor at the Graduate School of Library and Information Science, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, and Jennifer A. Younger is an assistant director of Central Technical Services for the General Library System of the University of Wisconsin-Madison, Madison, Wisconsin 53706. The authors wish to thank our colleagues who provided thoughtful comments on earlier drafts of this paper; we particularly wish to acknowledge F. W. Lancaster and Pauline Atherton Cochrane for their valuable assistance in helping us place our ideas in a historical context. 139 140 College & Research Libraries than public services departments. In an online searching environment, the capa- bilities of keyword searching and Bool- ean operators have enhanced subject retrieval by freeing searchers from the necessity of knowing the exact form of a subject heading and giving them the power to combine the terms that bring together the concepts of interest. Before the widespread availability of online in- formation retrieval, indexers, not librar- ians and library users, used thesauri. With the vision of the electronic "library without walls" rapidly becoming a real- ity, however, it has become possible, even necessary, to consider issues of in- formation representation and storage to- gether with issues of information search and retrieval. One of the most challeng- ing issues in electronic information re- trieval (CD-ROM, online catalogs, and online information retrieval systems) is providing effective subject access to users. This article examines three dimen- sions along which changes in the do- main of subject access may be considered and describes current meth- ods of providing subject access by con- trolling the results of indexing as well as by informing the search process. This study also suggests an integrated frame- work through which subject access may be facilitated and proposes an agenda for practice, education, and research. The major purposes of this article are twofold. First, we propose that methods intended to assist in the searching pro- cess be grouped together and the group subsequently be called "informed re- trieval" methods. Informed or guided retrieval is defined as using feedback from an information system in order to improve retrieval results. A simple ex- ample of using such feedback is scan- ning the subject headings assigned to a book about a topic of interest, and then incorporating those terms in a subse- quent search. This concept of informed retrieval parallels that of authority con- trol, the establishment and use of con- trolled vocabularies in indexing. Second, we propose that an integrated frame- work be developed, one that includes both authority control and informed re- March 1991 trieval as equally important methods in providing effective subject access. Whereas practical and research efforts are currently under way, these efforts should be coordinated within a frame- work that considers authority control and informed retrieval as partners, rather than distant relatives, in improv- ing subject access. DEFINITIONS Subject headings have been most closely associated with catalogs, and de- scriptors with indexes. In their elec- tronic forms, indexes are commonly referred to as databases, while catalogs retain the name catalog, modified by the term online. In this article, we use the term "database" to refer to the online form of an index, and refer to catalogs, regardless of form, as "catalogs." Of course, both online catalogs and online indexes are databases in that they gather information units in a structured form to facilitate access. The terQ('~bj~access" refers to the ability to con uct a search for informa- tion on a topic by various means: subject headings or descriptors, keywords, or the name of something or someone thought to be associated with the topic. (Citation indexes facilitate the latter form of access.) Closely tied to this issue of subject access is the issue of authority control. Authority control has been re- garded as a service to library users, en- suring consistency in the choice and form of words or groups of words that are used to represent a specific concept in a bibliographic record. From the indexer's perspective, authority control means the process of matching the in- tended term with the terms prescribed by a thesaurus. When no match occurs, the process of adding a new term to the thesaurus can be initiated or the next best term may be selected. THE DOMAIN OF SUBJECT ACCESS Librarians are accustomed to thinking of subject searching as taking place in a number of locations and using multiple sources. Users routinely search sources From Authority Control to Informed Retrieval 141 such as library catalogs, printed indexes, and online databases in their efforts to find bibliographic citations for items pertaining to the topics of their interest. Traditionally, these sources not only are physically separate, but also are searched in distinct ways. An examina- tion of three dimensions of the environ- ment, or the domain as we have termed it, of subject access reveals accelerating changes that are transforming it, offer- ing new options to users and creating new challenges for librarians. The three dimensions to be examined are: (1) new online searching capabilities that differ from those of manual searching; (2) de- creasing visibility of the boundaries be- tween multiple subject indexes and catalogs, as access is provided through a single computer terminal; and (3) the diversity of approaches to providing subject access. NEW ONLINE SEARCHING CAPABILITIES The introduction of online public ac- cess catalogs during the decade of the 1980s has extended online searching ca- pabilities to catalogs. This development presented to users the options of key- word and Boolean searching, truncation, and cross-field access (title, author, sub- ject) to enhance their searching. In all of these approaches, the searcher now has powers that were, in manual systems, available only to indexers. For example, indexers created a keyword index in order for searchers to have access to key- words. Likewise, indexers created pre- coordinated subject headings as a means of combining topics, whereas today, searchers can coordinate the desired concepts themselves, as needed. The searcher's ability to postcoordinate key- words is increasingly important as on- line searching has become more widespread. While searching vocabula- ries (subject heading lists and thesauri) have generally been available both as searching tools and as indexing tools, searchers can now use the terms in both keyword and linear access modes, thereby increasing their control over the preestablished use of the terminology. Recently, Marcia Bates suggested that searching has advanced to such a degree that "online search capabilities them- selves constitute a form of indexing." 1 That is, online searchers can, at the time of searching, use Boolean operators to coordinate concepts in the way indexers do when assigning precoordinated sub- ject headings. These new search capabil- ities significantly change the kind of subject access available to searchers, giv- ing them powers previously available only to indexers and catalogers. DECREASING VISIBILITY OF INDEX CATALOG BOUNDARIES On more and more campuses through- out the country, databases produced by external suppliers are being purchased and mounted on campus computer net- works along with the local online cata- log. The "library without walls," in which faculty and students can search the library's catalog, and selected databases such as MEDLINE and ERIC from a single terminal in home or office, has all but erased the distinction be- tween bibliographic records produced by catalogers and indexing records pro- duced by abstracting and indexing agen- cies. These new search capabilities signifi- cantly change the kind of subject access available to searchers, giving them powers previously available only to indexers and catalogs. (The proliferation of optical disc sys- tems that resemble online systems has further clouded these distinctions and may contribute to users' difficulty in knowing exactly what they are searching through a particular terminal.) The once clear division between subject headings and descriptors-the former occurring on cataloging records, the latter appear- ing in indexing records-has also 142 College & Research Libraries blurred. On the one hand, subject head- ings, constructed as precoordinated sets of terms, were coextensive with the sub- ject matter covered in the book. On the other hand, descriptors were applied as terms representing one aspect of the item, with many descriptors assigned in order to cover the subject matter of the item as a whole. From the viewpoint of the searcher using Boolean logic and keywords, the precoordinate or postcoordinate construction of the "sub- ject access point" is largely irrelevant, provided the subject matter is ade- quately represented. The controversy surrounding the most recent edition of the Library of Congress Subject Headings (LCSffi in which the subject headings are presented in a thesaural format, as de- scriptors have always been presented, highlights this tendency to treat indexes and catalogs as different versions of the same thing, even from the viewpoint of construction. (Thesaural format means the careful linking of terms in a syndetic relationship designated by Broader Term [BT], Narrower Term [NT], or Re- lated Term [RT]. This format differs from the traditional subject heading list in which the relationships among terms were less rigorously constructed.) DIVERSITY OF SUBJECT SEARCHING APPROACHES Along with expanded searching pow- ers has come a multiplicity of ap- proaches to subject searching. The degree and power of subject access avail- able in any one system can no longer be described solely in terms of standard in- dexing policies and procedures. Evi- dence of greater diversity in methods for subject searching may be found in the user manuals provided with online cat- alogs and databases. These manuals doc- ument the methods adopted by a particular catalog or database with re- gard to retrieval algorithms, searchable fields, scope of coverage, source and control of indexing terminology, and depth of indexing, as well as the system features and interface characteristics of- March 1991 fered for online searching. The conver- sion of some indexes from magnetic tape to optical disc storage prompted some producers to redesign the database and to introduce variations in virtually all of the areas listed above. The CD-ROM ver- sions of such bibliographic databases as Science Citation Index, MEDLINE, ERIC, and Psycinfo are just a few examples in which there are differences in searching methods between the online and CD- ROM products. Thus, in addition to the new searching capabilities offered in the online and CD formats, other measures of diversity in the methods for subject searching have been introduced. THE IMPACT OF THE CHANGING ENVIRONMENT ON USERS In the face of these environmental changes, the problem becomes one of continuing to provide subject access in a shifting environment. Although effec- tive subject retrieval has been under- stood as an optimal balance between recall and precision, there are currently no standards which can be used to deter- mine when this optimal balance has been reached. Rather, effective retrieval results from specific search strategies carried out in individual situations and evaluated by individual users whose judgments of relevance may be influ- enced by a variety of factors such as time, money, information need, and previous knowledge. To examine the problem of providing effective subject access, the authors describe current methods of pro- viding subject access by controlling in- dexing results as well as by informing the search process, thus defining a con- tinuum on which methods established to assist in subject retrieval can be identi- fied and placed. CURRENT METHODS OF CON- TROLLING INDEXING RESULTS A thesaurus or subject heading list is the traditional approach to defining con- cepts and mapping the relationships among concepts in a field in order that From Authority Control to Informed Retrieval 143 material on those subjects can be repre- sented in a consistent way. From the vo- cabulary "at large," some terms are selected for use in the controlled vocab- ulary and references are made from terms not used, synonyms or near syn- onyms, to the selected terms. Many choices are made in the construction of a thesaurus; among them is the choice of whether a "top-down" or "bottom-up" approach will be employed in assem- bling terms; that is, whether the terms represent all knowledge in the field, re- gardless of the frequency with which they appear in the literature, or whether terms shall be determined by literary warrant, i.e., their appearance in the lit- erature of the field. Choices about the number of cross-references from popu- lar usage or new terms must also be made. Overall, considerable time and in- tellectual energy are brought to bear in the construction of a well-designed sub- ject authority list or thesaurus. The ex- penditure of these resources is justified in the belief that controlled vocabularies, because of their classing functions, are the primary means of facilitating recall and, therefore, provide a needed service to users. 2 While authority control exists as a ser- vice to users, in practice, users have had little or no input into the construction of thesauri. With the exception of consulta- tive groups representing specific subject fields, it is usually librarians or indexers who determine controlled vocabularies. However, Phyllis Reisner proposed in- volving users in the creation and main- tenance of searching vocabularies nearly three decades ago. 3 In the 1980s, Jean M. Tague proposed user-responsive subject headings and Marcia J. Bates described a "superthesaurus" in which virtually all terms would be available as a "front end" to assist users in finding their way through a series of cross references to the controlled vocabulary term.4 While these proposals have not been developed into working systems, other similar ideas have been implemented. Sara D. Knapp's early work in developing BRS/.TERM, a vocabulary database for searchers, is a "home-grown" solution to the problems of rationalizing natural language to the numerous controlled vo- cabularies that exist in multifile systems such as BRS.5 In an experimental inter- face at the University of Illinois, users may propose additional terms for inclu- sion in the controlled vocabulary. Librar- ians review these for permanent inclusion. 6 The personal HYPERCATalog project initiated at LIBLAB at Linkoping University in Sweden also has features whereby authors and users indicate in- dexing terms and leave a "usage trail" that subsequent searchers may follow. 7 (A usage trail is analogous to hiking through a meadow. As hikers select their paths, they leave their footprints, even- tually making a visible trail. Several trails may exist simultaneously, each having advantages in terms of access, destination, ease, efficiency, perspective, or resources. Subsequent hikers can se- lect an already available trail or strike out on their own. Similarly, expert searchers can establish pathways through the library's catalog.) The control of vocabulary started within single disciplines and databases. Users can search multiple databases cov- ering a number of distinct subject areas from a single station, vocabulary control becomes more complex. While some the- sauri attempt nearly universal scope (LCSH), others, such as Medical Subject Headings (MeSH), are more closely re- lated to a particular discipline. To facili- tate moving from concepts in one database to the same concept in another database, special tools such as the Cross- Reference Index, which links already-ex- isting terms in thesauri, have been developed. Through facilities such as DIALINDEX and BRS/TERM, major online database vendors such as Dialog and BRS enable searchers to ascertain the frequency of occurrence of terms across several databases to assist in term selec- tion. Still another method of rationaliz- ing multiple indexing languages is a prescriptive approach such as the Uni- versal Medical Language System (UMLS). In this project, indexers are de- veloping a specialized language to rep- resent concepts in a variety of 144 College & Research Libraries documents such as scientific articles, medical records, and research reports . CURRENT METHODS OF INFORM- ING THE SEARCHING PROCESS As methods of controlling the results of indexing through authority control change to meet current challenges, users' efforts to design an effective search strategy must adjust accordingly. Librarians can assist users in developing improved search strategies through bib- liographic instruction and system design. Because the bibliographic instruction approach is adequately described else- where in the literature, the authors do not discuss this approach here. 8 We do suggest, however, that bibliographic in- struction librarians possess a rich store of knowledge that could be incorporated into system design. Such systems would have as their objective effective subject access through "informed retrieval." As mentioned earlier, we define informed retrieval as using feedback from an in- formation system in order to improve retrieval results. The system provides feedback and integrates it into system design. However, the user controls it and can invoke it as needed . Informed re- trieval is not totally automatic and does not, therefore, perform exactly as a coun- terpart of an automatic indexing system. Instead, informed retrieval provides to the user greater control over the results of the indexing process. Therefore, in- formed retrieval constitutes a form of user-exercised "authority control." While informed retrieval is presently undefined as a categorical set of meth- ods, progress towards this end may be observed in the current environment of information retrieval systems. Some ex- amples are systems that display descrip- tors and ask whether the user wishes to see other documents indexed to these terms and systems that display graphi- cally, or otherwise, indications of the probable relevancy of documents within a set. Still other systems supply auto- matic truncation or "wild card" features to increase retrieval. However, unless the system informs the user that such a March 1991 feature is automatically invoked, it can be confusing and fail to truly "inform" the searching process. Users themselves often modify an in- formation retrieval system by creating commonly needed search strategies as macros or "hedges," storing them at the searching station and invoking them as needed. Two examples illustrate the use of hedges. Users searching across multi- ple databases create a hedge linking syn- onymous terms, thereby bringing together varied terminology. When hedges are used to retrieve materials on a concept too new to appear in a thesau- rus, and especially when they are shared among searchers, they serve as a kind of local subject authority control device.9 Informed retrieval provides to the user greater control over the results of the indexing process. Therefore, in- formed retrieval constitutes a form of user-exercised "authority control." Almost twenty years ago, F.W. Lancas- ter described many of the premises on which these ideas are based. They are reflected in the design of such early sys- tems as SUPARS, an online system for accessing Psychological Abstracts devel- oped at Syracuse University. 10 These ideas are closer to more widespread im- plementation because the technology is now available. However, librarians can- not expect users to search immediately with the same sophistication that index- ers have, particularly when these search- ers have not been educated in the principles of bibliographic control and access. Librarians must provide help, e.g., through vocabulary displays that make relationships among terms clear, leading users from unused to used ter- minology, allowing the user to conduct the search in natural language which the system then translates into the appropri- ate controlled vocabulary. Although early online systems were intended to be searched by end users, they failed to at- tract large numbers of users not because of the theoretical principles that under- From Authority Control to Informed Retrieval 145 lay their design, but because the technol- ogy was too primitive to support them adequately. Despite the introduction of programs to assist users in devising suit- able search strategies, truly effective in- formation retrieval by end users remained an elusive ideal. Since the early 1970s, further research into end-user searching indicates that end users are not performing particu- larly effective searches. Comparisons of the search results of search intermediar- ies with those of end users reveal that trained intermediaries retrieve a greater number of relevant citations than do end users. 11 A FRAMEWORK FOR FACILITATING SUBJECT ACCESS The preceding discussion leads the au- thors to suggest that several trends are converging toward the shared goal of providing effective subject access to users. This convergence exercises con- trol not only over the indexing process, but also over the searching process. In- dexing and searching affect one another in many ways: depth, entry vocabulary, and strategies employed. The traditional controlled vocabularies supplied, main- tained, and applied by librarians and indexers can be used to form the basis for user vocabularies. Additionally, in- formed retrieval provides feedback d ur- ing the search process, notifying searchers of their location in the database relative to the controlled vocabulary's structure, as well as their progress in locating relevant materials. In other words, the system provides feedback about where the searcher is lo- cated in the search, whether more infor- mation is available, and what can be done to retrieve it. We do not wish to diminish the im- portance of, or necessity for, controlling the indexing process. On the contrary, we believe that as long as humans or machines perform indexing, guidelines and controls are essential. The concept of informed retrieval, when combined with controlled vocabularies, is powerful in- deed. Two examples of this kind of com- bination are Knowledge Finder, a CD- ROM MEDLINE system for the Macin- tosh, and Autocat, an online catalog developed at Dickinson UniversityY Both use the power of controlled vocab- ularies and provide the user with an array of tools that can be invoked as needed when monitoring the progress of the search. Other examples no doubt exist The traditional controlled vocabula- ries supplied by librarians and indexers can be used to form the basis for user vocabularies. Current research at OCLC exemplifies another line of investigationY A recent study examined the extent to which user-entered terms correspond to the catalog's controlled vocabulary and de- · termined that up to two-thirds of user- entered subject terms do not match LCSH terms in exact, normalized, or key- word form. Further analysis of the searches will identify those subject terms that are close approximations of LCSH terms and will identify what system ca- pabilities are needed to assist users in approaching LCSH terms more closely. 14 To the degree that such research is suc- cessfully implemented and disseminab:rl, subject authority control will be one means of achieving optimal recall and precision in subject searching. Subject authority control and informed retrieval will be seen as complementary andes- sential components of providing subject access. While ideally we would like to see orderly progress toward effective sub- ject access through directed efforts within a set of theoretical principles, in fact the practical steps are isolated and a theoretical. A concerted effort to recog- nize, coordinate, and promote work would encourage quicker and more effi- cient progress. AN AGENDA FOR PRACTICE, EDUCATION, AND RESEARCH Directed changes in practice, educa- tion, and research are needed to achieve 146 College & Research Libraries effective subject access for tomorrow's world. Most agendas consist of lists of projects to undertake, but our agenda calls for action, particularly for a cooper- ative approach to subject authority con- trol and informed retrieval. The literature of library and information sci- ence is replete with studies calling for improved controlled vocabularies, bet- ter front-end software, and more effec- tive bibliographic instruction programs. Yet the interchange is often confined to librarians from similar functional spe- cialties. Thus, awareness of other work is limited, and its benefit and impact are diminished. 15 More efforts to bring to- gether in dialogue interfunctional teams and to foster projects involving subject authority control experts, bibliographic instruction librarians, and online catalog system designers are needed. For the same reasons that the Council on Library Resources promotes collaboration be- tween practitioners and researchers, an integrated approach toward enhancing subject access would, if adopted, bring together the makers of subject authority control systems, the designers, and the teachers of end users. While some of these changes are already under way, we anticipate that the pace of change must accelerate if librarians are to retain an active role in future decisions. In the practice of librarianship, the walls between technical and public ser- vices are beginning to disappear. Occa- sionally, the two divisions are merged; more often, individual positions take on responsibilities for both cataloging and reference. In one of the best known of these experiments, at the University of Illinois Urbana-Champaign Library, li- brarians reported acquiring a broader, holistic approach toward connecting users to the information they wanted. 16 Librarians who take on responsibili- ties for cataloging and reference describe the "informing effect" that results-an effect that is subsequently used to mod- ify both cataloging and reference prac- tices.17 The knowledge and skills used in · cataloging, e.g., a knowledge of the cat- alog and the subject classification, are of direct use in assisting users. Conversely, March 1991 a knowledge of user needs and priorities is an important element of decision mak- ing with regard to cataloging policies and priorities, a point made years earlier by R. C. Swank. 18 Therefore, from the practical experience of the Illinois librar- ians, the insights of the faculty and the authors' cataloging and reference part- nership, we suggest that subject author- ity control and informed retrieval form a logical coalition for providing subject ac- cess. We suggest that subject authority con- trol and informed retrieval form a logical coalition for providing subject access. For many years, curricula in library schools have reflected the traditional ad- ministrative division between public services and technical services in librar- ies. In so doing, the schools fail to estab- lish leadership in designing an integrated approach to effective subject access. Subject indexing, bibliographic instruction, and information system de- sign courses are taught with minimal attention given to the role that each func- tion plays in enhancing subject access. A first step toward achieving this ap- proach is integrating the teaching of the representation of information with the retrieval of information. Some library schools have already taken this step. The theoretical foundations for a uni- fied approach to research in subject ac- cess come from both within the field of library and information science and from other disciplines such as cognitive psychology, linguistics, and information storage and retrieval. One strength of this multifaceted foundation is that var- ious insights can be applied to the broad problem of creating effective subject ac- cess. The three parts of this problem are controlled vocabulary and subject au- thority; informed retrieval applied dur- ing the search process; and the potential of natural language processing to either enhance or supplant traditional human subject cataloging. From Authority Control to Informed Retrieval 147 While well-established research tradi- tion on the performance of controlled and uncontrolled vocabularies in the re- trieval process exists, many of these studies have taken place in a laboratory environment.19 Earlier research concen- trated primarily on the performance of indexing schemes and did not consider retrieval from catalogs. Only with the introduction of the online catalog have catalog-use studies looked more closely at the process of subject access.2° Further- more, much of this research failed to con- sider the effect of either the user or the system interface on the retrieval process. Recent research has attempted to exam- ine performance from many dimensions, taking into consideration the effect of the query, the retrieval technique, the searcher, and the nature of the search process. 21 We call for greater attention to the potential in these areas and for dis- cussion and research that involves re- searchers in other disciplines, librarians, and information scientists. By suggest- ing informed retrieval as a theoretical framework, we hope to enable librarians and information scientists working in both areas-indexing and retrieval-to coordinate efforts to produce truly excit- ing and informative systems. FUNDING AND SUPPORT A critical item on any agenda for ac- tion and research is the question of fund- ing: Who will pay? In times of limited resources when new monies are scarce, funds must come from more efficient uti- lization of present programs, from cost shifting, or from changed funding prior- ities. The bibliographic control community has borne a large part of the fiscal re- sponsibility for subject access. Catalog- ers at the national level, through institutional or agency support of na- tional libraries such as the Library of Congress, the National Library of Medi- cine, and the National Agricultural Li- brary, or through cooperative cataloging efforts promoted by the bibliographic utilities such as OCLC and RLG, have carried the burden for current subject access. While users have indicated that enhanced subject access is a high prior- ity, librarians have long stated their con- cerns for the restricted resources available to fund the development of subject headings. 22 Yet, cooperation among decision-making agencies has brought a much-needed focus to ad- dressing the problems of funding effec- tive subject access through cooperative cataloging programs and shared re- search. Recently, system designers and devel- opers, both in the private and public sec- tors, have picked up some of these costs. In some cases, they have also sought to recoup these development costs by mar- keting their systems directly to users, bypassing the traditional library chan- nels. Perhaps the scarcity of available money to fund research and discussion between the research and practice com- munities further reflect this desire for a return on investment. With regard to the role users play in bearing the burden of subject access, Lancaster suggested that users also contribute to the cost of achieving precision in information re- trieval because they spend time and ef- fort to prepare a search strategy, run the search, and refine the results. 23 The implications of cost shifting sug- gest that the bibliographic control and system designer community, as well as users, have a substantial stake in deci- sions affecting subject access. If so, it is even more important to unite the knowl- edge and expertise of these communities in an integrated approach to creating ef- fective subject access. As money contin- ues to diminish, both cataloging and reference librarians are turning to other ways of managing their bibliographic control operations. Using paraprofes- sionals, incorporating artificial intelli- gence and expert systems for both training and task performance, and sim- plifying cataloging are all attempts to cope with financial and staffing crises in both cataloging and reference areas.24 SUMMARY In the world of Charles Cutter, provid- ing subject access was ·a manageable task. In retrospect, it appears that a 148" College & Research Libraries readily distinguishable topical nature characterized publications, mostly monographs. Catalogers and users alike identified topics similarly. The informa- tion explosion, together with the plural- ism of today' s society, produces a complex and diverse set of publications to which people want subject access. If librarians agree that effective subject access is a realistic goal, and that author- ity control and informed retrieval are viable means for achieving that goal, they must begin to integrate presently fragmented efforts into a unified agenda. A great deal of time, talent, and money have been expended to achieve these goals. Now a clear path toward a unified framework is needed. Separate communities within librarianship-in- dexers and catalogers, reference librari- March 1991 ans, bibliographic instruction specialists, and system designers-have developed differing approaches to solving this uni- versal problem. While positive effects may be derived from differing ap- proaches, all approaches must contrib- ute to a unified approach in recognition of the constraints of scarce resources. The framework proposed embodies a comprehensive approach that includes authority control and informed retrieval as a holistic approach to subject access. Both representation and retrieval are in- tegral parts of a holistic approach to sub- ject access. Through the implementation of an agenda for action in practice, edu- cation, and research, librarians can en- sure a future of effective subject access conceived within a unified framework of authority control and informed retrieval. REFERENCES AND NOTES 1. Marcia J. Bates, "Rethinking Subject Cataloging in the Online Environment," Library Resources & Technical ServiCes 33:400-412 (1989). 2. Elaine Svenonius, "Uminswered Questions in the Design of Controlled Vocabularies," Journal of the American Society for Information Science 37:331-40 (1986). 3. Phyllis Reisner, "Evaluation of a 'Growing Thesaurus,"' paper presented at the Center for the Information Sciences, Seminar in the Information Sciences, Lehigh University, Bethlehem, Pa., March 10,1966 (Yorktown Heights, NY: IBM Watson Research Center, 1966). 4. Jean M. Tague, "User-Responsive Subject Control in Bibliographic Retrieval System," Information Processing and Management 17:149-59 (1981); Marcia J. Bates, "Rethinking Subject Cataloging in the Online Environment," Library Resources & Technical Services 33:400-412 (1989). 5. Sara D. Knapp, "Creating BRS/TERM, a Vocabulary Database for Searchers," Database 7:70-75 (1984). 6. William Mischo and others, "The University of Illinois at Urbana-Champaign," in Campus Strategies for Libraries and Electronic Information, ed. Caroline Arms (Bedford, Mass.: Digital Press, 1990), p.117-41. 7. Roland Hjerppe, "HYPERCAT at LIBLAB in Sweden," in The On-Line Catalog: Develop- ments and Directions, ed. Charles R. Hildreth (London: Library Association Publishing, 1989), p.177-209. 8. Hannelore B. Rader, "Library Orientation and Instruction-1988" Reference Services Review: 17:73-85 (Winter 1989); Maureen Pastine and Bill Katz, eds. "Integrating Library Use Skills into the General Education Curriculum," The Reference Librarian no. 24:passim (1989). · 9. Anne B. Piternick, "Searching Vocabularies: A Developing Category of Online Search Tools," Online Review 8:441-49 (1984). 10. F. W. Lancaster, Vocaular.y Control for Information Retrieval (Washington, D. C.: Informa- tion Resources Press, 1972), passim; Kenneth H. Cook, "An Experimental On-Line System for Psychological Abstracts," Proceedings of the American Society for Information Science 7:111-14 (1970). 11. William H. Mischo and Jounghyoun Lee, "End-User Searching of Bibliographic Databases," Annual Review of Information Science and Technology 22:227-65 (1987); Sara J. Penhale and Nancy Taylor, "Integrating End-User Searching into a Bibliographic Instruction Program," RQ 26:212-20 (1986). From Authority Control to Informed Retrieval 149 12. Prudence W. Dalrymple, "Knowledge Finder at Rockford: Evaluation of a CD-ROM MEDLINE System in a Community-Based Medical School Library ·and an Outpatient Clinic," in MEDLINE on CD-ROM: An Evaluation, ed. Rose Marie Woodsmall, Becky Lyon-Hartman, and Elliott Siegel (Medford, NJ: Learned Information, 1989), p. 97-116; J. Bechtel, "Developing and Using the Online Catalog to Teach Critical Thinking," Information Technology and Libraries 7:30-40 (1988). 13. D. Vizine-Goetz and K. Markey, "Subject Authority Control in Online Catalog Design," in OPACs and Beyond: Proceedings of a Joint Meeting of the British Library, DBMIST, and OCLC (Dublin, Ohio: OCLC Online Computer Library Center, 1989). 14. Karen Markey, Jon Drabenstott and Diane Vizine-Goetz, "Increasing the Accessibility of the Library of Congress Subject Headings in Online Bibliographic Systems," Annual Review of OCLC Research July 1988-June 1989, (Dublin, OH: OCLC Online Computer Library Center, 1989),p. 31-33. 15. Pauline A. Cochrane, "Improving the Quality of Information Retrieval-Online to a Library Catalog or Other Access Service .. . Or ... Where Do We Go from Here?" Online 5:30-42 (1981). 16. J. W. Williams, "The Decentralization of Selected Technical Services at the University of Illinois at Urbana-Champaign," Technical Services Quarterly 4:5-19 (1987). 17. Jennifer A. Younger, "University Library Effectiveness: A Case Study of the Perceived Outcomes of Structural Change" (Ph.D. diss., Univ. of Wisconsin-Madison, 1990). 18. R. C. Swank, "The Catalog Department in the Library Organization," Library Quarterly 18:24-32 (1948). 19. Ann H. Schabas, A Com arativ lu.at-ion of_th.e_Ret rifval Effectivene~ o.f.J'.~i.hmr.y_ o C ub ·ect Head in s and PRECIS Strin s or Com12.uter Searching...o.~C _QBJ.a-(London: University of Lon on, 1979),passim; Cyril W. Cleverdon, Report on the Test ing and Analysis of an Investigation into the Comparative Efficiency of Indexing Systems (Cranfield, England: Aslib, Cranfield Research Project, 1962),/assim. 20. Karen Markey, Subject Searching in Library Catalogs: Before an After the Introduction of Online Catalogs (Dublin, Ohio: OCLC Online Computer Library Center, 1984), passim. 21. Nicholas J. Belkin, Robert N. Oddy and Helen M. Brooks, "ASK for Information Retrieval: Part I. Background and Theory; Part II. Results of a Design Study," Journal of Documentation 38 no. 2-3:61-71; 145-64 (1982); Nicholas J. Belkin and W. Bruce Croft, "Retrieval Techniques," Annual Review of Information Science and Technology 22:109-45 (1987); Christine L. Borgman, "All Users of Information Retrieval Systems are Not Created Equal: An Exploration into Individual Differences," Information Processing and Management 25:237-51 (1989); Tefko Saracevic and Paul Kantor, "A Study of Informa- tion Seeking and Retrieving: Part I. Background and Methodology: Part II. User, Questions, and Effectiveness: Part III. Searchers, Searches, and Overlap," Journal of the American Society for Information Science 39:161-76; 177-96; 197-216 (1988); Prudence W. Dalrymple, "Retrieval By Reformulation in Two Library Catalogs: Toward A Cognitive Model of Searching Behavior," Journal of the American Society for Information Science 41:272-81 (1990) . 22. Joseph Matthews, Gary Lawrence, and Douglas Ferguson, eds., Using Online Catalogs, (New York: Neal Shuman, 1983), passim. 23. F. W. Lancaster, The Measurement and Evaluation of Library Services (Washington, D. C.: Information Resources Press, 1977), p. 145. 24. Sue Anne Harrington, "The Changing Environment in Technical Services," Technical Services Quarterly 4:7-20 (1986); Marjorie E. Murfin and Charles Albert Bunge, "Para- professionals at the Reference Desk," Journal of Academic Librarianship 14:10-14 (Mar. 1988); Beth S. Woodard, "The Effectiveness of an Information Desk Staffed by Graduate Students and Nonprofessionals (at the University of Illinois in Urbana-Champaign)," College & Research Libraries 50:55-67 (July 1989); Charles Fenly, Expert Systems . Advances in Library Information Technology, No. 1 (Washington, D.C.: Library of Congress Cata- loging Distribution Service, 1988), passim; International Federation of Library Associ- ations and Institutions, "Minutes of the Meetings ... September 1988," 54th General Conference, Sydney, Aug. 29-Sept. 3, 1988.