64 The American / Winter 1980 Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021

Intellectual Access to : I. Provenance and Content Indexing Methods of Subject Retrieval

RICHARD H. LYTLE

Introduction AVAILABILITY OF COMPUTERS ENCOUR- ing access to archives. The first method, AGES visions of sophisticated informa- the Provenance or P Method, is the tion retrieval systems for archives. The traditional method of archival re- visions include national on-line subject trieval, based on principles of archives retrieval systems using controlled vo- administration and reference practices cabularies and the latest searching of . Subject retrieval in the P techniques. But the visions are shat- Method proceeds by linking subject tered by the paucity and poor quality queries with provenance information of subject access information about ar- contained in administrative histories chives: there is little to computerize^ or biographies, thereby producing and what exists is unimpressive when leads to files which are searched by us- subjected to the rigors of computer ing their internal structures. Informa- manipulation. Moreover, compiling tion in the pure or theoretically de- subject access information about ar- fined P Method derives only from what chives can be a very expensive process is known about the file—the activities of with uncertain benefits. The chal- the creating person or organization and lenge, then, is to design for archives in- the structure or organizing principles formation retrieval systems which are of the file itself. The second method, both powerful in subject retrieval and the Content Indexing or CI Method, cost effective. derives from librarianship but has been System design should be founded applied extensively to manuscript col- upon careful analysis of the character- lections, and, to a limited extent, to ar- istics of subject retrieval in archives chives. Subject retrieval in the CI and upon experimental testing of re- Method matches subject queries with trieval systems. The major purpose of terms from an index or catalog. In the this article is to contribute to subject pure CI Method, information is retrieval system design by describing gleaned by an indexer who examines two methods of gaining subject access the records; as the CI Method usually to archives. These methods are theo- is practiced in manuscript collections retically distinct, but in practice can oc- and archives, provenance-related in- cur as complementary means of gain- formation is not considered in index- Intellectual Access to Archives: Subject Retrieval 65

ing. Although P and CI Methods do exist indicates, defining the components of as pure types in practice, often they occur as the archives system is no simple task. complementary approaches within a given The body of materials included in Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 repository. The theoretical distinction is im- the system may be defined in part by portant for analysis, especially for reduction the characteristics of the materials of method characteristics to a form suitable themselves. The distinction between for quantitative study. archives and manuscript collections is The Provenance and Content In- becoming less important with the trend dexing Methods are described below. to apply archival techniques of control A companion article forthcoming in well beyond the archives of govern- the American Archivist1 reports an ex- ment or other institutions; certainly all perimental evaluation of the two meth- materials subject to control by archival ods. techniques should be included in the archives system, without reference to Definition of the Archives System traditional definitions. Moreover, the The following discussion is struc- archives system should include active tured according to the principles of records of government and other insti- systems analysis, including definition tutions as well as non-current records of the system and its components or selected for archival preservation; the subsystems, requirements on the sys- traditional dichotomy of retrieval pro- tem, and responses of the system. The visions for current and archival rec- object of the analysis is the totality of ords makes little sense in the modern archives and manuscript collections in era of records and information man- the United States; precisely which ma- agement. terials are included in the system is The materials in the archives system part of the problem, not a given. Data, may be distinguished from materials in admittedly inadequate, are derived other systems. , for example, are from the archives administration liter- included in the bibliographic system ature, the writer's experience, and and are properly distinguished from from the experiment and data analysis archives by their focused subject con- reported later. The purpose of the tent and their multiple copy distribu- analysis is more to take first proce- tion. Some materials, trade literature dural steps than to present results. collections, for example, and perhaps The archives system is made up of some manuscript collections, are not so (1) the body of materials contained in easily classified. These materials de- the system, (2) the users who make de- serve special attention in information mands on the materials, (3) the finding system design.2 aids used to access the materials, and Definition of materials in the ar- (4) those responsible for servicing the chives system may be achieved in part materials. As the following discussion by identifying the system's body of

' This article, and the subsequent one reporting experimental results, are based on my doctoral dissertation. See Richard H. Lytle, "Subject Retrieval in Archives: A Comparison of the Provenance and Content Indexing Methods" (Ph.D. diss., University of Maryland, 1979). My advisor, Dagobert Soergel, was a major influence on this work. Also, my intellectual debt to Richard Berner extends well beyond the cited references. 2 For details on trade literature, see Irene Travis, "Trade Literature at the National Museum of History and Technology," Special Libraries 70 (1979): 272—80. Trade literature is an excellent exam- ple of material which presents real problems for system definition. 66 The American Archivist/ Winter 1980

users. Although there are specialized kept mostly for security reasons, and users who present demands for certain archivists have seldom analyzed the in-

kinds of materials, there are many formation to describe users and their Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 users who ask questions which cross demands on the archives system. traditionally defined boundaries. A Some data have been gathered for common body of users constitutes the analysis, and a few institutions report primary reason for regarding archives regularly on their users. The National and manuscript collections as one sys- Archives has made at least one use tem. Moreover, active records should study,3 and the University of Illinois be included in the archives system be- Archives, for example, gives user sta- cause many users make demands which tistics in its annual report. But there cross this traditional boundary as well. are no synthetic studies, and virtually Two final components of the system nothing on the topic appears in the ar- are the finding aids used to access ma- chives administration literature. Gath- terials in the archives system, and per- ering and synthesizing existing data is sons who provide reference assistance. itself a major research task, and it is The term finding aid ordinarily is used likely that existing data would be in- only in archives (traditionally defined), adequate for cross-institutional stud- but for present purposes it is used for ies. What follows, then, is a very im- all access devices in the system, includ- pressionistic commentary based on ing card indexes for manuscript collec- experience, conversations, and years tions as well as administrative histories devoted to reading colleagues' annual and inventories for archives. Refer- reports. ence staff may include archivists and Genealogists probably have numeri- other information service personnel. cal superiority and item-use superior- Definition of the archives system, ity, although they may not consume a then, must be accomplished as a first corresponding proportion of refer- step for information storage and re- ence staff time. Service to record crea- trieval systems planning; failure to do tors—usually administrators—proba- so will undermine thinking about in- bly comes next, although this category formation systems for archives and is appropriate mostly to archives tra- manuscript collections. System defini- ditionally defined. Scholars probably tion is especially critical for effective come next, and within that group it is planning of a national information sys- presumed that historians predomi- tem. nate. Biographers rank high. A last category of user is the casual user, which might include everyone from a Requirements on the System chance walk-in to college undergradu- 1. Who are the Users? Most archives ates making a pass through the ar- require that clients register and give chives on an assignment. In recent some information about their re- years, efforts have been made to widen the user group to include, for exam- search; usually a record of materials ple, high school students. used is kept, as well. These records are

3 National Archives and Records Service, Office of the Executive Director, Planning and Analysis Division, "A Study of Users of the Records of the National Archives," unpublished, 1976. Available from the Planning and Analysis Division. Intellectual Access to Archives: Subject Retrieval 67

The distinction between administra- low number of responses and prob- tive and scholarly users is pertinent to lems with repository selection and re-

the traditionally defined archives part sponse, the results of this study cannot Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 of the system. Preserving the record of be generalized, but results tentatively government activities is the avowed indicate the following about users of reason for preservation of archives of archives and manuscript collections: the state, and the creator, the govern- (1) most users locate sources through ment functionary, receives high prior- colleagues; (2) users also locate sources ity in the archives system. It may be approximately equally via archivists' presumed that strong creator-orienta- suggestions, teachers' suggestions, ci- tion was a major factor in development tations in the literature, and repository of archival doctrine. As principles of guides, but all of these rank well below archives administration have been colleague's suggestions; (3) the Na- applied to a wide range of non-govern- tional Union Catalog of Manuscript Col- mental materials, the primacy of crea- lections is not extensively used; (4) the tor-oriented access has been perpetu- "most helpful" terms used when ated. searching a repository's finding aids are, first and overwhelmingly, proper 2. What Questions Do the Users Ask? names, and second, topics. Despite Given the difficulty of defining users these studies, however, no significant of archives and manuscript collections, data exist in analyzable form docu- menting users' demands on the ar- a description of what questions they 6 ask must be very tentative, indeed. The chives system. Finding Aids Committee of the Society What can, then, be said about users' of American Archivists has completed demands on the archives system? The two studies on the topic. The first, in following continues the impressionistic 1976, queried archivists about their commentary begun above. Requests opinions of users' access requirements.4 received for known documents are The second study, completed but not rather rare. Usually, these requests de- fully analyzed at this writing, collected rive from bibliographic citations by data directly from users.5 Because of a previous users; and, if the citation is

4 Kim Efird, "Report on the User Analysis Survey of the Committee on Finding Aids of the Society of American Archivists" (1977). Available from Finding Aids Committee of the Society of American Archivists. This study is mentioned because data were not collected directly from users. 5 Finding Aids Committee of the Society of American Archivists (untitled: raw data from data on a user analysis survey done by a subcommittee, 1979). Available from the Society of American Archi- vists. Only 108 useable questionnaires were received from eighteen institutions, and a large percent- age of these came from four institutions. 6 Since appraisal of archives is the process of determining what to preserve for posterity, the ap- praisal literature might be a source of information about the users of archives and their demands on the system. But the appraisal literature, to quote a recent manual published by the Society of American Archivists, is "disappointing, considering its major significance to archival practice." The manual itself mostly describes appraisal from the viewpoint of the creating organization and characteristics of the records. Due respect is paid to the principle that frequency and quality of use are important to appraisal, and specifically that use statistics should be employed. But beyond these general admoni- tions the author could not go; the present state of knowledge about use of archives will support no more. See Maynard J. Brichford, Archives and Manuscripts: Appraisal and Accessioning (Chicago: Society of American Archivists, 1977), especially p. 2. It is not desirable to attempt to base appraisal entirely on past usage; on the other hand, archivists should study past uses of archives as one criterion for appraisal. 68 The American Archivist/ Winter 1980

correct and if physical location desig- names only because he has learned nators are still valid, response is rou- that archives access techniques are

tine. more effective at retrieval by name Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 Requests specifying names of per- than by subject? Berner7 says that re- sons and organizations are said to con- searchers usually request access by stitute the majority of requests, and proper name because they have asso- probably this is so even if genealogists ciated proper names with activity, and are not counted. The strength or de- he further argues that proper name is gree of name specification varies and an effective means of subject access. the relationship of names in the query But these assertions have not been sub- to provenance-related information may jected to empirical test in an experi- be an important factor in retrieval. mental environment. Subject requests are presented fre- Browsing is an important character- quently to the archives system. A sub- istic of user behavior. The browser ject request may be broad or narrow; holds an uncertain image of his subject the breadth of a subject question will interest; he probably changes his no- be perceived differently in different tion of what he wants as his search pro- parts of the system: what appears a ceeds. Browsing behavior may be char- broad subject question to a specialized acteristic of historical research, and repository might well be a narrow indeed it may be characteristic of ini- question from the viewpoint of a na- tial phases of all research. Since the re- tional system. quirement to scan files may distinguish Users also may specify date, or may one archival retrieval system from an- qualify proper name or subject re- other, browsing must be considered in quests by date; most archival bodies system evaluation. Two cases of users and many manuscript collections are are identified: the user who knows easily described by creation dates, and what he wants, does not benefit from therefore date may be a very impor- scanning files and does so only when tant aid to searching. Requests by geo- required by the system; and the graphical area are common in some browser, who is uncertain of what he depositories, but probably are less wants and may benefit from scanning common across all institutions. Some- files. Apparently browsing and broad times requests by form are used as a subject questions are related, although means of limiting the search, especially there is no reason why a user cannot if the form of interest is exceptional in present a broad, precise subject re- the collection or the repository at hand; quest. Browsing should be distin- a user might surmise, for example, guished from the use of leads in rec- that the information he wants will be ords, where the user discovers the found only in diaries. location of other areas of the file to be Requests for records by proper searched; for example, in searching a name, geographical area, date, or form subject file, he finds a reference to re- may conceal a subject request. Does lated subject heading. the user really prefer to ask for docu- Users generally have modest recall ments by name, or does he use proper and precision expectations from the

7 Richard C. Berner and Gary M. Bettis, "Description of Manuscript Collections: A Single National System," College and Research Libraries 30 (1969): 410. Intellectual Access to Archives: Subject Retrieval 69

archives system. The check on recall— factors is a prerequisite to evaluation a measure which estimates the degree of system response to user demands. to which the system retrieves relevant Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 documents—is very erratic. A histo- 3. Problems in Defining Users and User rian may discover that a document he Demands on the Archives System. Al- had not seen invalidates his conclu- though archivists have not attempted sions in a recent monograph. An ad- to define the users of archives empiri- ministrator may discover that he has cally, they know that users change over repeated a mistake made five years be- time. Shifting research traditions bring fore, because the system failed to re- users with new interests. One illustra- trieve pertinent files. Often, recall fail- tive example is the growing interest in ures are less obvious—a scholar or women's history: since archivists had administrator has a "lead" which in- no way to predict the widespread study dicates that he has not seen everything. of women's history, they could not in- Probably most recall failures go unde- corporate this access point into their tected in the archives system. How finding aids. much the failures cost is not known. Another obstacle to definition of user Precision is a measure of the selectiv- requirements is users' internalization ity of a system. Most users of archives of system limitations: they request only are tolerant of low precision; they are what they have learned from experi- accustomed to poring over large quan- ence that they can get. A study of user tities of irrelevant documents. Their demands on the archives system might patience may derive mostly from ac- reveal overwhelming use of proper ceptance of system limitations, but name requests, but users might be bet- probably it is related also to a high rate ter served by subject access and might of browsing. Low recall and precision even prefer that approach if given the in the archives system in part may be option. Distinctions should be made ascribed to low use-rates of archives between what the user wants, what de- and low funding levels of archival in- mands he places on the system, and stitutions, resulting in inadequate find- what information he actually needs. ing aids. Little effort is devoted to Wants can be determined by canvass- preparation of files, with the result ing users; demands can be determined that considerable effort is required by recording users' search requests on when the files are searched. systems. But determining what the user Clearly, data are needed concerning needs is a difficult undertaking, re- users' demands on the archives sys- quiring in-depth studies of users and relating their information-seeking be- tem. Studies by the Society of Ameri- 8 can Archivists Finding Aids Committee havior to task or outcome. should be expanded. Subject questions The environment in which user should be studied in the context of the needs studies were developed, infor- educational level of the user, his func- mation services in mission-oriented re- tion (scholar, administrator, etc.), sub- search and development organiza- ject question generality, and the tions, differs radically from the browsing phenomenon. Sorting these environment of the archives system.

8 Maurice B. Line, "Draft Definitions: Information and Library Needs, Wants, Demands, and Uses," ASLIB Proceedings 26 (1974): 87. 70 The American Archivist / Winter 1980

Even within a given area of scholar- individualistic but essentially allied with ship, research needs are difficult to as- librarianship. The following discussion sess; the needs are diffuse and the describes these subsystem responses. users are unaccustomed to articulating Most of the discussion focuses on the P Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 their information needs. Many of these Method because its retrieval character- objections apply to general libraries istics have seldom been studied. The also, but the difficulties may be even examination of archival theory is very more severe in archives. Perhaps the selective, making no attempt at an his- techniques of user need studies are not torical review. applicable, or perhaps they are appli- cable only to a few archival subsystems. 2. The Provenance Method of Subject Another obstacle to definition of Retrieval of Archives. Archives adminis- users and their requirements on the tration doctrine emphasizes that rec- system is resistance by many archivists ords of organizations are created in to social and behavioral science tech- the course of human activity, and that niques, especially those applied in li- the resulting organic character of ar- brary and information services. Often chives is the key to their physical and this expresses itself in opinions that intellectual control. Provenance is im- something "cannot be quantified." A portant for control of archives for two more sophisticated version of resis- reasons: it can be used as a principle of tance is to point out that there are too series arrangement which reflects the many variables, known and unknown, structure and activities of the creating to permit practical outcomes from such agencies; and, since agency functions studies. While there are serious prob- determine the subjects of the records, lems in identification of the archives provenance-related information about system's users and their needs, empir- archives can be used to gain intellec- ical studies should be pursued as long tual control over bodies of records as they prove fruitful. Some adjust- without examining or indexing them. ments of attitudes within the archival The closely related principle of re- profession may be required. spect for original order is, in a sense, a corollary of provenance: if individual files were rearranged in violation of Response of the System the original order, the principle of 1. General. The response of the ar- provenance would be violated since the chives-manuscript collection system to organic quality of the records would be user requirements varies greatly across disturbed. Yet, the original order doc- subsystems. The response in public ar- trine should be considered separately chives is seen in the development of from provenance. Provenance per- principles of archives administration tains to the activities and structure of culminating in nineteenth-century Eu- the organization which created the rec- ropean theory and twentieth-century ord; it does not pertain to individual American practice. The response in file order. The principle of respect for original order does pertain to file or- the manuscript collection subsystem, at 9 least until very recently, has been highly der, within series.

9 Richard C. Berner, "Arrangement and Description: Some Historical Observations," American Ar- chivist 41 (1978): 179. Intellectual Access to Archives: Subject Retrieval 71

Archival doctrine claims more for rangement of archives to suit the needs the Provenance Method than an effec- of historians.10 The Provenance tive means of intellectual access. The Method permeates archival practice, Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 doctrines of provenance and original from creation of finding aids to search order state that information will be lost room reference service. if the organic quality of archives is vi- T. R. Schellenberg outlines a de- olated. Items in isolation from an ar- scriptive program as follows: (1) iden- chival body lose part of their meaning; tify the record unit to be described; the reason for this is that the file, not and (2) enumerate (a) physical quali- information in the items alone, is re- ties, to include cubic feet, medium, etc; lated to activity. Thus, the meaning of (b) substantive qualities, to include the a file depends on (1) the information government or other body that pro- in each document; (2) the information duced the record, the functions that derived from the context of docu- resulted in their production, and their 11 ments within a file (file order related subject content. The primary finding to creating activity); and (3) the con- aid to archives is the inventory, a shelf text of the files—provenance-related list of archives of a specified adminis- information such as that contained in trative unit or combination of units. organization charts and administrative Description relates the documents to histories. Most of the propositions of the activity that created them, or the archives doctrine seem reasonable purpose they serve. For example, to enough to experienced archivists. None describe dossiers the archivist explains of the propositions has been tested em- the action which created the dossiers— pirically. to what common matter they relate, or to what function of the administrative The principles of provenance and 12 respect for original order combine into body they pertain. Most inventories a creator-oriented and document-ori- in large archives specify only at a gen- eral level, usually the series. ented doctrine. Archives arranged to reflect organizational structure and in- Subject access in the Provenance terpreted in the context of administra- Method depends primarily on making tive history presumably respond to the the connection between a subject re- information needs of organization quest and provenance-related infor- functionaries: the creators, their asso- mation. If this step can be taken, the ciates, or their successors. Appraisal of archivist is on home ground; he has a archives and intellectual access to ar- set of functional statements which re- chives are founded on creator-ori- late to the structure of the organiza- ented analysis and document-oriented tion and hence at least partially to the description. Physical rearrangement, structure of the archives. This process sometimes proposed to serve users, has taken him to the record group and generally is rejected; the famous late ultimately to the correct file which he nineteenth-century Dutch manual, for can use, with assistance of folder lists example, specifically rejects the rear- and the like, in the search for relevant

10 Samuel Muller, J. A. Feith, and R. Fruin, Manual for the Arrangement and Description of Archives, translated from the Dutch by Arthur H. Leavitt (New York: H. W. Wilson Co, 1940, reprinted 1968). 11 Theodore R. Schellenberg, The Management of Archives (New York: Columbia University Press, 1965), p. 101. 12 Muller, Reith, and Fruin, Arrangement and Description of Archives, p. 121. 72 The American Archivist/ Winter 1980

documents. Most users cannot fully trieval within files is arrangement or implement this search process them- rearrangement. Files may be arranged selves, although they may try if the ar- to re-create an original order, to pro- chives has an institution-level guide. vide order where there was none, or to Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 The archivist as an intermediary is an provide access points different from indispensable part of the search pro- those of the existing order; in the last cess in the Provenance Method. two cases rearrangement is similar in The user is expected to conform to purpose to indexing. In the P Method, eccentricities of the P Method. The ex- archives which exhibit an interpret- tent of required conformity can be seen able, usable order must not be rear- in Philip Brooks's excellent , Re- ranged; although this principle is gen- search in Archives, the Use of Unpublished erally accepted, it should be applied Primary Sources.13 First, the potential with judgment. user is instructed to acquire back- Enhancement of P Method power to ground in his subject by reading sec- select files for searching has received less ondary sources. He should be espe- attention than have principles of ar- cially careful to acquire knowledge of rangement. In his Management of Ar- persons, organizations, places, and chives, Schellenberg suggests that P events. Researchers must accept the Method subject access can be im- fact that finding aids will not provide proved by describing collections with access by the desired subject. Indeed, respect to geographic area, broad sub- there is no means of accessing the en- ject field, and chronological period, tire archives system; the researcher uses using provenance-related information. his ingenuity to locate the repository, He recommends analysis with respect and then uses guides, other finding to function or activity—for example, aids, and the assistance of archivists to agricultural, business, or diplomatic locate collections within repositories. activities. Schellenberg suggests what While the P Method does accommo- amounts to a faceted classification, date subject searching by such tech- where the facets are human activities. niques as subject guides to the Na- Most important for the present analy- tional Archives, in general the user sis of the P Method, he states that sub- must understand the system—almost jects are to be deduced from prove- to the point of becoming an archivist— nance-related information, not from to retrieve archives effectively. analysis of the files: "Specific subjects, then, should be ascertained deduc- 14 3. Enhancement of Provenance Method tively " Schellenberg did not de- velop these ideas in the literature. Subject Retrieval. The most common 15 method of enhancing P Method re- Richard Berner has developed

13 Philip Brooks, Research in Archives (Chicago: University of Chicago Press, 1969). 14 Schellenberg, Management of Archives, pp. 139-40. On the same topic, but largely repetitive of sections in his books, is Theodeore R. Schellenberg, "A Nationwide System of Controlling Historical Manuscripts in the United States," American Archivist 28(1965): 409-12. 15 Richard C. Berner, "Archivists, Librarians and the NUCMC," American Archivist 27(1964): 401- 9; "Manuscripts and Archives, A Unitary Approach," Library Resources and Technical Services 9 (1965): , 213-20; "Observations on Archivists, Librarians, and the NUCMC," College and Research Libraries 29 (1968): 276-80; "Manuscript Catalogs and Other Finding Aids: What Are Their Relationships?" American Archivist 34(1971): 367-72; "Arrangement and Description of Manuscripts," Drexel Library Quarterly 11 (1975): 34-55; Richard C. Berner and Gary M. Bettis, "Description of Manuscript Collec- tions: A Single National System." Intellectual Access to Archives: Subject Retrieval 73

Schellenberg's ideas concerning P empirically and understood; then it Method subject retrieval. Berner ad- can be systematized and even comput-

vocates creative use of the subgroup to erized. Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 gain control over collections. For ex- 4. The Content Indexing Method of Sub- ample, when a congressman's papers ject Retrieval of Archives. Content Index- are received, Berner segregates them ing is the practice of accessing collec- according to separate activities of the tions by examining the content of the congressman's life, thereby gaining files. The traditional manuscript col- some immediate control. Berner's lection catalog is the classical applica- ideas parallel similar thoughts in the tion of the CI Method in the archives earlier development of provenance as system. Processors go through collec- a method of arrangement of archives tions, piece-by-piece and often without above the series level, and he is espe- regard to the logic of the collection, cially influenced by Oliver W. applying subject terms which appear Holmes's "Archival Arrangement— as subject added entries in a card cata- Five Different Operations at Five Dif- 16 log. The application of the CI Method ferent Levels." Berner further sug- in the archives system thus is item-ori- gests a ^ptemjn_w£icji general_su5ject ented; or, at least, it is not consciously h controj-j^established yigj)f "^^

16 Oliver W. Holmes, Jr., "Archival Arrangement—Five Different Operations at Five Different Lev- els," American Archivist 27(1964): 21^tl. " Frank G. Burke, "The Application of Automated Techniques in the Management and Control of Source Materials," American Archivist 30 (1967): 255-78; "Automation in Bibliographical Control of Archives and Manuscript Collections," in Dagmar H. Perman, ed. Bibliography and the Historian: the Conference at Belmont (New York: Clio Press, 1968); "The Impact of the Specialist on Archives," College and Research Libraries 33 (1972): 312-17. 74 The American Archivist/ Winter 1980

of the National Union Catalog of Man- which comes to mind; an example uscript Collections. (used here because of the experiment Content Indexing in principle is very reported later) is the Baltimore Region Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 flexible. It can include user-deter- Institutional Studies Center (BRISC). mined access points derived from past BRISC folder level indexing is an ex- user requests, and document-deter- ample of larger physical unit but vir- mined access points selected by the in- tually no use of provenance-related in- dexer. In the archives system, the Con- formation; while BRISC indexes tent Indexing Method certainly should records for relevance to urban studies, consider provenance-related informa- context in its P Method meaning is not tion. But historically it has emphasized taken into consideration. a rather narrow item-focus, to the vir- A manuscript collection (or record tual exclusion of provenance-related unit) is a larger physical unit but its de- information.18 scription may be based on fundamen- tally different intellectual foundations. On the one hand, a collection may be Summary: Two Dimensions of Access described on the basis of notes or in- to the Archives System. dex entries made during examination The preceding discussion has aimed of the contents of the collection, with at a theoretical description of P and CI little or no regard to provenance-re- Methods. The methods may be sum- lated information; in this case, collec- marized as two dimensions of subject tion level description amounts to a dis- access to archives: increasing use of tillation of indexing data derived from provenance-related context, varying lower physical level analysis. On the from no dependence (CI Method) to other hand, a collection level descrip- entire dependence on provenance-re- tion could be derived entirely from, lated information (P Method); and in- say, biographical data and knowledge creasing size of the body of materials of the organization of the collection in- to be accessed. volving little or no examination of con- A single letter cataloged on its own tents of the collection. A record unit merits or an item reference in an in- described in the National Archives dex employs virtually no provenance- guide, based on provenance informa- related context and represents the tion and knowledge of the structure of 19 minimum physical size. Folder-level the file, is an example of P Method indexing (or some other fairly small access to a large physical unit. aggregation of items dictated by the In any many cases, P and CI Meth- file) is the next level of physical size ods are mixed or used as complemen-

18 The recent manual by Kenneth Duckett is the latest representation of the CI Method approach as described here. See Kenneth W. Duckett, Modern Manuscripts, a Practical Manual for their Management, Care and Use (Nashville: American Association for State and Local History, 1975). 19 Where does proper name indexing fit into this conceptual structure? Personal and organization names are an important part of provenance-related information: who participated in the activity; what was the result of the activity? Proper names often are used to gain access to subjects, or may consist of most of the same terms as a subject request. Name indexing in the archives system is not peculiar to either method, but the means by which the proper name gets into the system may be classified by method. If the name occurs in provenance-related information or in the structure of the file, it is included in the Provenance Method; if it enters the system by examination of the content of the items, it is considered part of the Content Indexing Method. Intellectual Access to Archives: Subject Retrieval 75

tary techniques for archival control and Conclusion: A Strategy to Examine retrieval.20 For example, archivists pre- Provenance and Content Indexing paring inventories may look at docu- Methods ments in a file to some considerable ex- The primary reason for attempting Downloaded from http://meridian.allenpress.com/american-archivist/article-pdf/43/1/64/2048208/aarc_43_1_31031523737j16n7.pdf by guest on 26 September 2021 tent and note what they see, while many a precise distinction between Prove- CI Method indexers doubtless do in- nance and Content Indexing Methods clude provenance-related information is to prepare the research design for in their descriptions. In retrieval, ar- an experimental study of the two chivists or users may well employ both methods. Thus the foregoing discus- methods—perhaps the P Method to se- sion may be viewed as the thought pro- lect files for searching and some file- cess preceding an empirical study of level CI Method index to search within archival subject retrieval systems. The a file. Moreover, there is no reason to empirical study has a major objective believe that archivists must or should of further refining conceptualization choose between the methods; probably of P and CI Methods, as well as of pro- many repositories always will use them viding an evaluation of their relative as complements in searching. But rel- performance. The wider importance ative emphasis on the methods will of all of these efforts, as mentioned at have service and budgetary implica- the outset of this article, is to support tions for archival institutions, and design of better archival subject re- hence description of the qualities of P trieval systems. and CI Methods has practical as well as Experimental results will be re- theoretical interest. ported in the next issue of the American Archivist.

20 Interactions between methods also exist (and present special problems for experimental design). For example, one might enter a file via the CI Method and then proceed to search in that area of the file, in effect using the P Method. Leads from within files were mentioned above, page 68, in dis- cussing browsing. Leads within files may or may not be method-specific. Information within an item giving new background information is not method-specific, while file structure leads such as cross references definitely are P Method related; some cases resist definition.

RICHARD H. LYTLE is the Archivist of the Smithsonian Institution.