Program for Cooperative Cataloging

Task Group on AACR2R

Cataloging from the piece in hand in an electronic age:
a position paper

Melinda Reagor Flannery and Sherry Kelley
with the assistance of the other members of the
PCC Task Group on AACR Revision

Through the library catalog, the descriptive cataloger has organized resources collected and stored by the library so that these resources and other objects like them (e.g., editions, translations, versions, works by the same author) can be retrieved for use. Historically, catalogers described physical, immutable objects they could hold in their hands. Through the catalog records, library users could find these objects and, what is more, find them again at any point of need. Though different kinds of objects have appeared at different times and in different settings, cataloging activity has always been limited to the subset of objects

  1. owned or housed in at least one library and likely to be retained there, and
  2. of a sufficiently high value to library users that the labor of description is both warranted and economically feasible.

Many acknowledge that the current permeation of digital systems and resources throughout our culture is fully as revolutionary as that attributed to the printing press. The popular press and other media are wrestling publicly with such basic questions as: what will be available, for how long, to whom, at what cost, to whose profit. At a time when resolution of these questions is in the balance, libraries must decide their stance toward digital resources and reword their cultural mission in light of those resources. In other words, how will an electronic object, available either remotely or locally, be included in the totality of library resources and how will catalogers represent it in the catalog so that the library user can continue to find it again and again?

Traditional library cataloging developed in the world of the printed book and gradually evolved to accommodate non-book formats as libraries expanded their collecting policies both to include objects in non-book formats and to value these objects highly enough to catalog them. The 1908 Catalog Rules, Author and Title Entries (American edition) published by the American Library Association focused on books. There were also separate rules for other types of materials, such as those published in 1919 and 1931 providing guidance for cataloging serial publications of societies and institutions, and rules in 1934 for cataloging government publications. By 1949, Rules for Descriptive Cataloging in the Library of Congress included separate rules for monographs, serials, maps, scores, facsimiles, and incunabula. Anglo-American Cataloging Rules in 1967 added rules for manuscripts, motion pictures and filmstrips, phonorecords, and pictures. AACR2 prescribed standards for description of two additional formats: computer files and three-dimensional artefacts/realia. All the codes provide general principles of description of a cataloged object based on the principle stated in Section 0.24 of AACR2: "The starting point for description is the physical form of the item in hand, not the original or any previous form in which the work has been published."

Libraries have never cataloged everything, not even everything in their collections (e.g., yes to monographs, no to vertical files). Moreover, libraries have set limits on the level of analysis of cataloged objects (e.g., yes to monographic series, no to periodical articles). In some cases, libraries as a group have ceded responsibility to other agencies for more detailed levels of description than they could provide; a case in point is the emergence of commercial periodical indexes. Catalogers of the day may have criticized the structure and consistency of the indexes and lamented their separation from the catalog, but for economic reasons libraries did not make the investment in this level of description needed by users. Libraries would not pay for the article level cataloging provided during the ALA experiment with publishers at the end of the last century. Consequently, the work was done by others, by other standards. In the age of manual cataloging and primarily physical resources, one moved among the periodical index volumes, the card catalog, and the stacks to dig up the treasures stored in bound journals.

In our age, the physical fact that a personal computer is now used for many formerly separate tools blurs the primacy and even the definition of the library catalog as catalogers have constructed it and the collection as selectors have built it. Some people can move among electronic periodical indexes, the online catalog, and the stacks. Somewhat fewer people can afford to have a service send the desired article and avoid that trip to the stacks altogether, or work in a journal-dependent field in which the literature itself is becoming available electronically. One-stop information retrieval from a personal computer may not be a reality for many, but it is at least available to some and a dream for many more.

In that portion of the bibliographic universe that catalog librarians have traditionally described, the two classic principles of data organization have been in various states of balance. The first principle is that things that are unlike in some important way require distinction or separation; such separation supports finding a particular object. The second principle is that things that are alike in some important way require gathering or collocation; such collocation supports finding all objects representing a particular work, a particular author, etc. Distinguishing unlike objects has traditionally resulted in a separate catalog record with more or less detailed description of distinguishing features. Physical carrier has always been treated as a distinguishing feature.

The electronic publishing field is producing an expanding quantity of electronic objects that are often simultaneously produced in hard copy, forcing a reexamination of definitions for like and unlike characteristics. Currently, if the carriers differ, the objects are deemed unlike. If the conditions for likeness, including physical carrier, are met, the same catalog record is used to describe it. If likeness exists but not all conditions are met, separate records are created and access points shared by other similar objects are assigned. These shared access points include names, uniform titles, subject terms, and classification numbers. (1)

Electronic objects not only call into question the factor of physical carrier as a distinguishing feature, they also prompt new uncertainty about what we are cataloging. In three important respects, electronic objects are like one another and unlike traditional (i.e., physical, immutable) objects. First, electronic resources available over the Internet have a less known and stable physical location. Stability of location may be an important variable when acquiring and providing access to electronic resources, but standards for stability and the mechanisms for handling instability (e.g., forwarding addresses, cross-references or pointers at the old site) are not yet settled.

A second similarity among electronic objects is the instability and uncertainty of their intellectual content. A given URL may lead to the same spot as last time, but the content of what is found is not necessarily what was found before. Electronic publishing is not the relatively tidy process to which we are accustomed with modern commercial publishing. The content is not fixed as a print run or film production is fixed; electronic data can be easily changed.

In another sense, comparing the introduction of digital systems and objects to the early years of the printing press and moveable type is particularly apt. Early printed books lacked title pages, for example, and often failed to provide accurate information on author, publisher, date of publication and so on. Electronic objects also fail to provide standard information about responsibility, publication, and so on.

Finally, electronic objects resemble each other and differ from traditional objects in the conditions of their maintenance and storage.(2) An electronic object may be stored on a local server, but the conditions of its storage and the predictability of its maintainers are much less under the control of the library than are those of any of the physical resources we have ever housed in library buildings. When an electronic object is stored more remotely, such control or influence can be assumed to be even less, in the absence of any binding understanding among the object's producer, storer, and users, whether in a library or a coffee shop. The implications of this malleability of content and instability of storage are staggering for cataloging and catalog maintenance, to say nothing of research. The challenge, then, is to describe the object when we are unsure of stability of location, maintenance, and content.

In other important respects that have not yet been fully explored, electronic resources are only apparently similar to one another. Many can be differentiated from each other using much the same categories catalogers use to distinguish individual physical objects. At this point, it is not clear to what extent using the known categories for physical objects will provide adequate ways of thinking about electronic objects. In her chapter on cataloging Internet resources in the soon-to-be-published The Cybrarian's Manual, Judy Myers calls for the cataloging community to undertake an effort to categorize or type electronic objects, rather than simply matching them with their likeliest traditional counterparts. This step, she argues, would usefully precede or at the least shape efforts to apply existing cataloging expertise to these objects.

Because access to all these objects is through personal computers and because the forms themselves are new, the way objects are represented and formatted may be more apparently similar to one another or like traditional forms than they are in fact. The Web page of a well-established research institute with a solid future may appear very similar to the Web page of a struggling new dental hygiene clinic that has paid a good consultant to develop one. Sites rich with local content appear similar to those composed of little but links to other sites. The seriality of electronic journals may be clear, but the description of their publication patterns may ultimately have little to do with that of physical journals. For example, the regular physical packaging of short articles may have had more to do with physical constraints (e.g., need to justify printing and binding costs) than with other aspects of serials. Electronic forms such as electronic pre-prints are most like traditional forms of information exchange among the invisible college. To say that access must be provided for electronic pre-prints is not to say that librarians, or catalogers in particular, will or should provide that access. The electronic environment may give new options for providing access to these materials.

It is easy to overstate the adequacy of our current codes and standards for cataloging the variety of materials that already come our way. In fact, even materials that we embrace as part of the traditional cataloging task have long strained the limits of the detailed codes and standards we have devised. Examples of this include materials created separately but assembled after the fact, loose-leaf publications, and materials created outside traditional publishing streams. (3) Much of the response to electronic resources has been to adapt current codes and standards to accommodate the new materials. We treat electronic resources, especially those resources available over the Internet, as a new form of computer file. The rules for cataloging computer files were designed when most cataloged computer files were data files contained on magnetic tape and loaded on mainframe computers. These rules, adapted from the original cataloging rules for printed books, have already been much reworked to fit newer computer file categories such as software for personal computers and CD-ROMs. The rules are now being further reworked and expanded to fit all the possible categories of electronic objects that library users might wish to find, to relate to and distinguish from similar objects, and to find again.

It is arguable that the current explosion of electronic resources in volume and in forms could precipitate a kind of crisis in library cataloging. The central questions are these: can we really adapt and expand each relevant chapter of the current rules to accommodate whatever portion of the electronic universe we thoughtfully and consciously accept into our cataloging mission? If a form should emerge that fits no existing precedent, would adding a new chapter to the "Anglo-American Cataloguing Rules" (chapter numbers 14 through 19 are, after all, theoretically available) be the best solution? Is there anything to be learned from the unifying impulses that have propelled format integration to fruition and continue to animate discussions of multiple versions? There are fewer occasions to have the physical object in hand for cataloging, rapidly proliferating numbers of electronic resources, and greater likelihood that works will be available in multiple formats. Has the time come to handle works and their manifestations so that the balance between distinguishing description and collocating descriptive access better fits the current publishing/distribution model and better supports the context in which modern information-seeking and research is done?

Footnote 1. Many cataloging rules guide the cataloger to create separate records for otherwise like objects. Identical information published in multiple physical formats currently requires the creation of separate records for each format. For example, microfilm newspapers and magazines, which requires the creation of separate records for each format. For example, microfilm newspapers and magazines, which many space-starved libraries archive instead of the paper subscriptions they buy for current use, are cataloged on records distinct from the records used for the current paper issues. If machine indexing defaults widely separate a title with a General Material Designation (e.g. [microform]) from a title without one (e.g. a print version), the hapless patron may never know that the library's holdings are divided among multiple records. Cataloging codes before AACR2 permitted adding variant copies to existing records in a liberal range of circumstances. Although the current cataloging code does not address this issue, a Library of Congress rule interpretation (to rule 1.0) prescribes that objects published in two separate places should be cataloged on separate records, even when those objects have been published by the same publisher (Cataloging Service Bulletin 49 (Summer):10).

Footnote 2. This statement is true even when the object is not on the Internet, and predates the rise of the Internet. Archivists lamented the loss of information about the writing process occasioned by the increasing use of the personal computer in literary writing. Yet even an early draft of a novel has a physical form on a floppy diskette or a hard drive; if this draft is maintained in a form that can be read by contemporary hardware, or refreshed, it can be used indefinitely.

Footnote 3. Two developments tend to unify aspects of cataloging that have been traditionally distinguished in cataloging practice. The first, MARC format integration, has recently been implemented. Format integration ends the practice of limiting machine-coded information for a given object to codes specified for the object's format. Machine coding schemes had somewhat independent object to codes specified for the object's format. Machine coding schemes had somewhat independent development streams and were thus not in perfect harmony, and were not universally applicable across all physical formats. Format integration has been implemented in the sense that newly created cataloging records will utilize a single set of machine codes and corresponding meanings for all formats of material. Of course, the vast body of existing records in databases local, regional, and national may or may not ever conform to the new standards. The work to bring all records up to current standards is great and the long-term benefits are not yet perfectly clear. The second development grows out of a recognition by the cataloging community that separate records for multiple versions of the same work or manifestation of that work may not always be in the best interest of the users of the catalog. Groups have been gathered to study what is now known as The Problem of Multiple Versions, but solutions acceptable to all affected parties have not yet been found. Review and revision of the structures through which modern cataloging is done involves a significant number of national and international groups; communication and consensus involving these groups is understandably challenging. Any revision to these structures also involves every online catalog vendor and every library with an online catalog; the reluctance to change in a way that brings discontinuity with catalogs as they exist currently is also understandable.