Unifying our cultural memory: Could electronic environments bridge the historical accidents that fragment cultural collections?

in Information Landscapes for a Learning Society, Networking and the Future of Libraries 3, 1998. and presentation at UK Office of Library Networking Conference, July 1998.

David Bearman and Jennifer Trant, Partners, Archives & Museum Informatics, USA

(Section 3)

Purposes of the Actors and their Consequences

A model of the system of cultural documentation that takes to account the approaches to finding that documentation which can be taken by a user can be represented within this triangle.

• The User Seeks to Satisfy a Purpose

The first face of the triangle represents the users and their questions.

In the simplest model of user interaction with the system, users ask a question which they believe will yield an answer suited to their purpose. This question is directed to a repository, which interprets the question and answers to the extent that they can with their documentation. Repository staff may refer to reference resources to try to understand the question better. If the user is dissatisfied with the answer, they may also look in reference resources to find additional terms, related concepts and other clues to an improved formulation of the query. But in the most direct use of an information resource, the query/response dialogue is unmediated.

Typically, user profiles (in an anonymized form) and users' queries are collected for analysis. We are interested in who uses systems, what classes they might belong to, what resources they are reviewing, and what searches they make. Sometimes, repositories are also interested in whether or not these searches are successful. Too often these two tangible aspects of the user are all that is represented to a system.

However, between the user and the query is the users' purpose. This "purpose" reflects the actual problem that led to the question reflected in the query, and it is what must be satisfied if the interaction between the user and the system is to be perceived as successful. The user's purpose dictates both what information content and form will satisfy the user, but it does not express itself directly.

• Different Repositories: Documentation Purposes and Disciplinary Schemas

One of the major reasons why this simple model often fails to produce a result that satisfies the user is a mismatch between the documentation and control methods employed by the repositories and the users' sense of the questions. Documentation and control methods are the strategies for organizing knowledge employed by the publishers of the reference resources and the staffs of the repositories. These methods are optimized to serve the kinds of users each resource anticipates and the management requirements of the materials the methods document. Not surprisingly, these methods work less well for users with points-of-view and purposes for inquiry other than those traditionally served by the repository or reference resource, or considered in the design of an information management system.²

The second face of the triangle model represents the many types of repositories which collect various genres of cultural documentation. Genre is the critical aspect of this documentation; while each repository may also collect many different media, we emphasize their intellectual characteristics, rather than physical ones, because these are what leads to most of the barriers to providing access. The physical media may create conservation difficulties, require special care in copying or delivery, or require users to acquire special systems to use it, but each of these problems can be overcome (possibly though increased investment). Genres - which can be defined as information packages distinctive because of the manner in which their content is presented - include poetry, prose, song, paintings, motion picture, numeric data, etc.. More detailed prose genres distinguish, for example, between scientific articles, diaries, laboratory or field notes, and correspondence. Genres of motion pictures include, among others, feature films, documentaries newsreels, cartoons, previews, and nature films. Visual works may be paintings, sculptures, installations or drawings, and drawings themselves may be architectural, preliminary studies, cartoons.

Each genre has characteristics which are crucial to a correct 'reading' of its content. These characteristics can be described, and may be the subject of inquiry, but typically users are expected by repositories to understand the nature and limitations of the genres they use and to interpret their information content with a sophisticated understanding of these constraints.³ In addition, for many valid intellectual and historical reasons, different types of cultural repositories have developed distinctive documentary practices, even though they may hold the same genres. Rationales for diverging practices include different requirements for control, preservation and access, and the expectations of their primary user group, which may have a distinct disciplinary background, and therefor a particular tradition of using particular kinds of documentation.

Consider, for example, the treatment of a large body of diverse materials from a single estate. The business and personal records of the donor would be housed in the archives. Archives traditionally accession large collections of paper, which have been organized by their original owner to reflect the way in which they conducted their businesses and lives. These might be described using the ISAD(G).⁴ Archivists do not reorganize these collections internally, nor do they index the items within them, because their coherence is provided by the life and/or processes that they supported. The same individuals books would be given a library, that would treat each volume individually, making decisions about which, if any are duplicates, and which unique copies which require special care. Each work will be described separately, indexed by author, title and subject, and most likely catalogued using AACR2 in the MARC format.⁵ The volumes will be reorganized on the shelves with other volumes of similar subject matter. This reflects the perspective of library clients who seek individual titles for the content they hold, not for the fact that they were part of an estate. Artifacts from this estate might be accessioned by a history museum and treated in two quite distinct ways - they might keep together those artifacts collected purposefully "as a collection" and disperse into their holdings those artifacts which the individual collected in the course of every day life. If the individual was extremely important historically, they might also keep the items collected in every day life as a unit. In either case, they are likely to provide provenance as only one access point among many, and to catalog the items themselves at least superficially. If the collection of our donor was one of art, an art museum would describe each work separately, providing a number of access points including artist, title, medium and donor. However, it is unlikely that any two museums would treat the same collection in exactly the same way.⁶

These gross distinctions between archives, libraries, and cultural or art museums can be refined and extended to explain methods prevalent in slide collections, zoos, botanical gardens, local history societies and many other types of institution. In addition, we can overlay on these rough differences practices which relate directly to the specific types of clients served by the institution. Thus Winterthur Museum, an historical museum with a clientele of connoisseurs and graduate students in decorative arts and museology is notable for providing access to its collections by the artisans who made the artifacts, the designers, the owners and many other methods of concern to students of high culture. Most local historical collections would be happy to provide approaches by decade, by social class, and by types of objects.

Suffice it to say, at this point, that these purposes - that is the specific purposes of the curators and clients - are combined with the mission of the repository itself in determining methods and granularity of control. These perspectives affect such basic functions as the conservation and preservation of works themselves. Many art museums can provide an account of all the storage and exhibition spaces in which a particular work has resided over a number of years, and the number of lumens to which it was exposed. A very exceptional rare book library may be able to provide such data. Most libraries know how often a work was checked out, they don't know how frequently it was consulted and usually purposefully do not collect information about who used it. Archives on the other hand can usually account for all uses, and users, of a particular record group for many years.

Recording documentation about things serves many purposes, and institutions that do not share purposes tend to accumulate different data. The attributes of things recorded in one institution differ significantly from the attributes of these same types of things collected by other institutions. In the end, none of these may be the salient aspects sought by the researcher.

Part of learning the methodology of research is developing a familiarity with different kinds of documentation and control methods.⁷ Library user training in North America will teach you that library catalogs can be searched by author, title and subject term. Visiting an archivist, and getting oriented to a finding aid system will show that archival records can be searched by who owned them, and sometimes by who received them, and how they are organized. But special needs in subject-specific institutions extend these generalities.⁸ Film libraries can often be searched by names of actors as well as directors and producers, script-writers, cinematographers, etc. These differences in methods of documentation and control reflect the management requirements of the collecting organizations. But they may not support the needs of potential users for access to materials in these repositories.⁹

Organization of materials, physically and/or intellectually, is based on a schema or logical construct. The schema employed by repositories or reference resources of a particular type tend to be quite similar to each other - libraries like libraries, archives like archives, gazetteers like gazetteers, etc. Users also have schemas - art historians may focus on the visual output of individuals, archaeologists are interested in the items collected at a particular site, material historians study artifacts in use in a particular time. Indeed, the same user will employ many different schemas depending on a particular purpose. And, because users have been trained in research methods, and want to get results, they also employ schemas that they believe are used in the systems they are trying to use. In this case, they become victims of both the difficulty in translating between their own schema and that of the repository, and of any mistake they make in understanding the schema used by the repository.

• Knowledge Models and Reference Data Resources

The third face of the triangle model represents the many reference resources that a person could consult in search of an answer to their question. Reference resources could be used in place of, on the way to, or in addition to, primary materials in cultural repositories. Typically these reference resources answer questions about people, places, events, things, or ideas and provide definitions, alternative terms, citations to other resources, and pointers to parallel, similar or related concepts.

Users, quite rightly, are often catholic about what kind of repository holds the information they need in their research. They are aware that the same kinds of sources, or different kinds of sources about the same theme, can be found in many kinds of repositories, even if they are unaware of the particular accidents of fate, politics, and time that influenced what a particular repository holds. But they have learned that they cannot be indifferent to the tradition by which the evidence they seek is being managed, or they will not find it. As any reference librarian knows, they have learned to approach each institution by asking questions that do not directly reflect what they hope to learn, but rather what they expect to find. They try to translate their own model of the problem into what they imagine will be the language and organizational schema of the collecting institution.¹⁰

Needless to say, this process is fraught with pitfalls, and continually amazes the reference staff, who often do not understand why this seemingly indirect strategy is regularly employed. But the reference librarian's complaint - "why don't they just ask for what they want?" - reveals an ignorance of the limitations imposed by the schemes that "libraries" employ to describe their holdings. Users are often right in realizing that if they asked for what they wanted, in their own terms, they would not receive an answer.

It is relatively easy to imagine a simpler strategy that would produce a more precise result. The user would translate their own model into a common form of expression and ask questions against a documentary model expressed in that same form. But, are such translations possible?

Possibly. A more complex model of the relationship between users, reference resources and repositories, is one in which the schemas of the documentation (both the genres and the methods of control) are explicitly declared to a "knowledge model". The user's query criteria would also mapped to the knowledge model. The knowledge model thereby mediates between the points-of-view of the user and the resources. The cross-walks developed in the Dublin Core discussions are first generation versions of such tables of equivalent concepts in different documentation systems.¹¹

A shared knowledge model with common "categories" in queries also makes it possible to position reference resources between the user and the information repository. Mapping to the model enables the user's query to be deconstructed by the system and put first, and automatically, to the reference resources. In this first stage of query processing additional references, terminology and context can be included, to enable the query to succeed in repositories where, for example, more specialist terminology is used. The commonly modeled query can then be put to the repositories in terms of their known (declared) schemas. The result of this enhanced query can be expressed in terms of the user's schema, offering the user a much more complete and precise result, in a form that is easily understood.

In a very simple example of the use of model-mediated inquiry can be seen in a user request for primary documents about the author Mark Twain or maps of the city of St. Petersburg Twain was, of course, a pseudonym for Samuel Clements and most documentation about the life of that individual will be found under his own name. St. Petersburg, is a place that has been known by many names throughout history. Looking up the individual first in a biographical dictionary (or 'authority file' that collocated alternate names) and the city in a geographical thesaurus, historical atlas or gazetteer, would provide the alternative names that should be searched.

NEXT: The Goal of Integrated Access

PREVIOUS: Intellectual Barriers