Text from a speech given by Sue Jacobson, March 1984, at the opening of the exhibit
"A cabinet for sages built/Which kings might envy" is a line from Wordsworth that appeared on one of the windows of the Furness Library Building. Though meant in an entirely different context, it seems a wonderful description for a card catalogue, and an appropriate quote with which to inaugurate the celebration of the 100th anniversary of the University of Pennsylvania Library card catalogue.
The "cabinet" today at the University of Pennsylvania numbers some 7,800 drawers; it contains approximately 8 million cards, representing over 3 million volumes in the University libraries. This is a fax cry from the earliest catalogues of the library collection. The 1829 Catalogue of Books Belonging to the Library at the University of Pennsylvania listed some 1,700 titles, and the first card catalogue, begun in Spring 1884, had, one year later, 18,356 cards, representing 6,232 volumes.
While many people think of the catalogue as static, reading through the annual reports, minutes of meetings and procedural memoranda of the past 100 years presents an entirely different picture. The card catalogue has, in fact, been constantly evolving.
Obvious to any user of the card catalogue are the variations in the appearance of the cards, a result both of the technological means of producing cards and of the different cataloguing standards being followed.
The earliest cards were handwritten, in the finest library script. Melvil Dewey, in his "Brief Rules for Library Handwriting" instructs the librarian to write all "letters upright with as little, slant as possible ... [to] take great pains to have all writing uniform in size, blackness of lines, slant, spacing and forms of letters ... [and to] follow the library hand forms of all letters, avoiding any ornament, flourish, or lines not necessary to the letter." (1)
In 1919, the Cataloguing Department was moved to rooms in the basement of the library to allow for the expansion of the periodical room. Out of earshot of the public, cataloguers could now begin to use typewriters without disturbing library readers. The annual report of that year expressed the hope that this would mean more legibility and uniformity in cards and the production of more work per cataloguer. Unfortunately, the production of cards was curtailed during the War years, when all the typewriters but one were stolen from the library, and were not replaced until after the War. The introduction of various duplicating techniques in the '50's and '60's made the process of card production more efficient. A cataloguer could type a single unit card, which could then be duplicated and headings added by a typist. In the 1970's, the library began to receive cards generated by computer. While this meant an enormous decrease in the work of typing, the quality of these early cards left much to be desired. The cards received today at the University Library are laser-printed. They include not only description, call number, headings and tracings, but also department library location symbol, LC card number, cataloguer's initials, dates of cataloguing and of printing, and the designation "replacement" if necessary. All this is predetermined by the RLIN card production profile; the cards are received in alphabetical order, with headings printed, and ready to be filed in the catalogue.
The card catalogue has, alas, lost some colour over the years. The Handbook of the Library issued in 1927, advised that while "most of the cards are white, it may assist readers to know that all cards with the top edge coloured red and with the first line in red ink are subject cards." Green, blue and yellow cards were used as subject cards for books of Biography, Bibliography and Criticism respectively. Though the use of coloured cards was soon abandoned, the typing of subject headings in red continued. However, on computer-generated cards, subjects appear in upper-case letters, while other headings are printed in a combination of upper and lower case.
One of the most variable aspects of the catalogue is the amount of information on each card, and the diverse ways of displaying it. On the earliest handwritten cards, for example, the imprint appeared in the lower right of the card; later handwritten cards moved this information to follow the collation area. On both handwritten and typed cards, the library followed a number of complicated local practices, such as dashed-on entries for variant editions, double analytic entries, and ambiguous tracings indicated by underscoring words on the face of the card. Bibliographic description became simpler and more standardized with the adoption of the Rules for Descriptive Cataloguing in 1949, and the Anglo-American Cataloguing Rules in 1967. In 1974 the University Library along with the rest of the library community implemented the International Standard Bibliographic Description (ISBD) and the catalogue card became a plethora of dashes, slashes, colons and semi-colons. I will avoid here the argument about whether this is confusing or helpful to the library patron. But what I do want to point out is that these changes in card format reflect a shift from reliance on local practices to adherence to nationally and internationally accepted standards of bibliographic description. This allows the University Library to participate in the developing national bibliographic databases, able to accept other libraries' cataloguing and to contribute our own bibliographic records.
Over the years, the catalogue has constantly been reexamined with the view to more efficiently serve the needs of the library user. One of the issues that has constantly been debated is the question of the divided versus the single sequence catalogue. The first catalogue of 1884 was a divided catalogues with separate sequences for author and title and for subjects. In 1898, a single dictionary catalogue was substituted in the place of the double subject and author catalogues. This was not, however, the end of the debate. The question was raised in the 1940's and again in 1958-59, prior to the impending move to the new Van Pelt Building. There was strong conceptual support for the divided catalogue in each instance, the enormity of the project and the costs involved led to the decision to retain existing dictionary catalogue.
In connection with the creation of the single dictionary catalogue in 1898, a complete revision of the card catalogue was undertaken. A description of this project, taken from the Report of the Librarian to the Provost in 1899, tells us that "a beginning was made during the summer of 1898, when, with the assistance of some additional help, the subject and author catalogues were thrown together and, pending the completion of the revisions, guide cards were introduced in large numbers. A work force of sixty cataloguers was hired, as well as thirty-two persons occupied with more mechanical labour, and the work of revision was completed by the beginning of the semester."
Other catalogue revisions were less drastic. During the years 1907-08, work was begun on preparing a proper subject index for the catalogue, in connection with which a thorough revision of subject headings was made "in order to give them a uniform character." In 1927, Library of Congress cards were introduced into the catalogue along with the "Library of Congress method of cataloguing." This entailed the acceptance of Library of Congress form of headings, and the result was chaos.
Headings used by the Library of Congress were very different from the system in use by the University Library, and could not be interfiled in the card catalogue. An attempt was made to force the new LC cards into the old pattern, or, as time went on, the old cards into the the Library of Congress pattern, by systems of crossing out and numbering parts of headings on the catalogue cards. But the inconsistencies remained; filing became a nightmare, and the existing staff was unable to cope with the situation.
A second large-scale revision of the card catalogue was finally undertaken in 1947. This Project involved not only refiling and corrections, but a good deal of recataloguing as well. Among the tasks of the project were the addition of filing titles to voluminous authors, the replacement of old and practically unreadable entries, and the reorganization of large sections of analytic entries. Perhaps the major achievement of this revision was the development of a code of filing rules for the University of Pennsylvania Library. While this initially slowed the project, it meant that sections of the catalogue could be refiled concurrently with the adoption of the code and in conformity with the new rules. By 1957, this revision project as a major effort had basically ended. The Report of the Preparation Division advises that "improvements henceforth would come by replacement of substandard cards, unified rules for filing, and ad hoc correction of errors." It is significant to note that the position of "catalogue revisor" became that of "catalogue editor" -- recognition, perhaps, that the maintenance of the card catalogue is an ongoing process, and not just a special project.
The catalogue has been systematically gone through from beginning to end at least four times in recent history. Prior to the adoption of the LC classification scheme in 1967, the letter designations for departmental libraries on all cards were changed to location stamps, to avoid confusion between these letters and the letters of the LC call number. In 1974, and again in 1980, the catalogue was expanded by the addition of more drawers and cases, to accommodate the increasing size of the library collection and consequent number of cards. Also, in 1975, a major weeding project was undertaken, to remove cards showing out-of-date or inaccessible locations. These were not, per se, revisions of the catalogue. However, the scrutiny of almost any section of the catalogue is bound to reveal errors and inconsistencies, which then can be resolved.
In the 1980's the major problem of the card catalogue has again been the need to reconcile new forms of headings with those already filed in the catalogue. The entry changes mandated by AACR2 and the long-overdue modernization of many LC subject headings have given rise to innumerable guide cards and cross-references, and made the maintenance of the card catalogue an increasingly complex task.
While this exhibition focuses on the card catalogue and on the tasks involved in creating the catalogue, it is a fitting time, as well, to pay tribute to those who have accomplished those tasks, namely the cataloguing staff.
In his report to the Provost in 1885, Librarian James G. Barnwell describes the attributes of a cataloguer: "A cataloguer should possess a very high order of natural ability, a wide range of general knowledge, familiarity with books as books apart from the literature contained in them, and the ability to write in a very legible hand. A knowledge of one or more foreign languages is a valuable adjunct to these qualifications." (He goes on to complain of the difficulty in finding suitable people, and suggests it may be a "result of the opening of many new channels for female labour of the higher order and the consequent diminishing unemployed supply of that grade.")
Barnwell's appreciation of cataloguers' skills was not, I'm afraid, shared by all library directors. A report of the '40's states that "much of cataloguing work is so monotonous as to have a serious effect on a good many workers, producing the unimaginative and defeatist attitude which is sometimes known as 'the cataloguers' mind'." Among the librarians I know "cataloguers' mind" would seem to be a compliment and not a defect. As Rudolf Hirsh, then Director of Technical-Processing stated: "Only those who have actually worked in a cataloguing department know the amount of detail involved in the preparation of even the simplest record." Today's cataloguers must not only possess the skills described by Barnwell, but must comprehend a wide range of cataloguing rules and their interpretations, must be adept in the use of computers, must be familiar with the MARC formats for encoding bibliographic data into machine-readable form, and must be alert and informed of the constant changes in these areas.
We recognize that the age of the card catalogue at the University of Pennsylvania Library is drawing to a close. Within the next five years, the University Library will move toward implementing an online catalogue, providing users with greater flexibility and ease of access to bibliographic data than ever before possible. This 100th anniversary celebration is a moment of "Hale and Farewell" as we move into a new era.
(1) Library Notes, c.1, no.1, March 1887, p. 281.