The Library of Congress >> Librarians, Archivists >> Standards

Report on the RDA/MARC Working Group Discussion of Identifying Entity Types in MARC Records

The RDA/MARC Working Group had discussions of the many points that were brought out in the MARBI discussion of 2011-DP03, Identifying Work, Expression, and Manifestation records in the MARC 21 Bibliographic, Authority, and Holdings formats.  They found that while there was some sentiment for defining such a field, there was disagreement in the group and in the community about how the values in the field should be defined.  It was pointed out in the discussion that there were various reasons for establishing this marker:  to be able to validate that a record for a certain level contained the appropriate fields for that level; to correctly organize displays; and for applications that do not yet exist.

Strict Entity Type Definition

One approach is to make the definition of an entity type value require strict adherence to record content having only specified data elements that apply to that entity type, which in the present case would be FRBR level. With this approach a work, expression, manifestation, or item record could be validated to assure that it had only the allowed data elements.  This would require a listing of the data elements that are valid for each entity type.  Such lists have been worked out in several initiatives:  LC had the MARC format analyzed by Tom Delsey in 2001 to determine the FRBR level of each MARC element, VTLS has done the same for their FRBR implementation, and OCLC had carried out work in this area.  There are probably other efforts as well.  A definitive listing has not yet been established, however.

As analyzed and described in Using MARC 21 with FRBR: Record Configurations, one problem with strict use of the entity markers is that the format does not support the inclusion of some entity level elements at this time.  For example the Authority format does not support including subject headings related to works as they are contained in Bibliographic records.  Thus no Bibliographic records could be marked as manifestation but include work-related subjects and, theoretically, no work record in the Authority format could be marked as work because of the missing elements. 

Another issue however is the realization that any given data element may well be needed on more than one FRBR level depending on the history or type of a particular work. What is seen as a hard-and-fast manifestation element for one type of material may be a bit more slippery and not quite so hard-and-fast for one or more other types.   The assumption that each MARC element could be assigned to one and only one FRBR entity level is a flawed assumption.  For example, Authority "work" records for manuscript material could include an array of other elements that are considered manifestation for the conventional book.  This would both be a problem of the type exemplified for subjects and also would dilute the validation possibilities -- or require validation to be dependent on the type of material and other factors. 

Another problem is that RDA is not fully worked out yet in this area and the tying of elements to entity levels will probably go through an adjustment in the next few years.  This could require frequent adjustment of the "tables" and teaching of catalogers.

And finally it was noted that FRBR is intended as a Conceptual Model rather than a practical application manual.

Loose Entity Type Definition

The other approach is to go with loose definition of the entities, so that essentially the "heading" for each record would be the part of the record to which the entity identification applied:  the 1XX in an Authority record and the 1XX/245 or the 1XX/240 in Bibliographic records. (Note if the work and expression records were encoded as Bibliographic records, then the 1XX/240 would be the applicable "heading" as there would not be a 245 in those records).  This was the initial idea that came from the cataloging directorate at LC as they especially wanted to be able to determine a work Authority record from an expression Authority record. 

However, it has not been fully explored whether the title or name/title authority record for a unique or other "no-book-like" material would be a work record with elements that usually would apply to a manifestation or a manifestation record that exemplified a work.  The loose approach would be less dependent on having a strict list of elements valid for each entity type (although probably not without disagreement and controversy until the RDA entity levels for headings are definitively settled).  However, this approach would have limited helpfulness for validation of records for the correct elements associated with the entity type designation.  It would be more supportive of rapid identification by catalogers of the entity type of a record heading.

Conclusion and Ways Forward

Given these two differing views, each of which could be argued, the group could not come to a decision on a proposal.  The strict definition approach needs a great deal of mapping work before it could be implemented and the loose definition approach worried some and had its usefulness questioned.  One cannot yet produce guidelines because of the limited nature of the RDA test, which identified the issue but did not point to a clear solution.

A suggestion was made that in the next year, as more use is made of RDA, systems experiment with a "local" field, perhaps the one suggested in the Discussion Paper, field 883.  Provisional guidelines for use of the field could be published in the format.  This would enable experimentation with the concepts and also experimentation with the issues of these records in an exchange environment.  The issues could be reexamined and analyzed in conjunction with the experiments to enable more useful results.

MARC 21 HOME >> Report on the RDA/MARC Working Group Discussion of Identifying Entity Types in MARC Records