The Music Thesaurus: A Faceted Approach to LCSH

As early as 1981 Pauline Cochrane, and perhaps even others of you who may be present here today, encouraged the development of subject-specific thesauri to act as supplements to LCSH. (1) Based on my own work over the past several years, I would say that the reciprocal is also a possiblity: LCSH can act as a supplement to and even a foundation for the development of subject-specific thesauri. LCSH offers substantial advantages for thesaurus building. It has a rich vocabulary which can be factored and enhanced. Its vocabulary contains not only phrases and words which describe various disciplines, but it also encapsulates an inventory of fundamental concepts related to those disciplines. While it maintains a structurally inconsistent architecture, many of the contexts represented can be exposed and strengthened. And, because LCSH enjoys worldwide acceptance, it has a preeminence which cannot be ignored.

Today's presentation recounts how analysis and evaluation of concepts represented in the LC music subject headings is leading to the organization of a formal music thesaurus a thesaurus which, in its early stages, is tied both conceptually and through a shared vocabulary directly to Library of Congress subject headings. LC headings are used as a point of departure, not a destination for the music thesaurus. It is planned that other subject lists, supplemented as necessary by terms from additional sources, will eventually be integrated into the music thesaurus. But for now, back to the point of departure.

Preliminary work on the music thesaurus began in 1991 with funding from the Council on Library Resources. (2) A computer database was created of about 12,000 LC music subject entries (including both preferred and nonpreferred headings) which were compiled into a thesaurus format. The subject headings could be viewed and searched in three standard projections in an alphabetical index in separate thesaurus records and in a hierarchical tree display.

While this projection improved access to the headings, it also brought to the surface their many structural weaknesses. One of the greatest impediments to LC's syndetic structure is that instead of single-concept terminology, LCSH contains multi-conceptual statements. Other examples of these statements are:

• CONCERTOS (ACCORDION WITH STRING ORCHESTRA)
• FOLKSONGS, BULGARIAN
• MUSIC AND ARCHITECTURE

Analysis of the LC music headings (both their strengths and their weaknesses) along with study of the music classification schemes of Dewey, British Catalog of Music and LCC, the RILM thesaurus and the Art and Architecture Thesaurus, as well as Ranganathan's theories (just to name a few sources consulted) helped me formulate a basic overall tentative hierarchical structure for the music thesaurus. It includes these facets:

TENTATIVE FACETS FOR THE MUSIC THESAURUS

• AGENTS
  • musicians
• EVENTS
  • ceremonies*
  • fasts/feasts/festivals*
  • holidays/seasons*
  • services*
• FORMS/GENRES
  • instrumental
  • vocal*
  • sacred*
  • secular*
• GEOCULTURAL ATTRIBUTES
  • languages*
  • locations
  • religions*
• SOUND DEVICES
  • instrumental
  • vocal*
• TEXTS*
  • sacred*
  • secular*
• OTHER TOPICS
  • terms/facets not yet categorized

*in progress

Once the tentative facets had been identified, the next step in the process was to undertake a thorough sorting and decoordination of the 12,000 entries. (3) I use a homegrown thesaurus construction package, ARIS (Anderson Rowley Information Systems), created by Fred Rowley and James D. Anderson. One of the software's components is nicknamed "ARISCoCo," for its co-collating capabilities. This component permits a variety of searching, sorting and collating routines which allowed groups of headings to be created and manipulated. To give you an idea of how the sorting and decoordinating process took place, let me demonstrate some sample routines that I used. Headings can be tagged and sent to a separate file, a file based on a conceptual theme, such as forms/genres or musical instruments.

1. they can be tagged individually
2. they can be tagged globally
3. or they can be searched hierarchically, culling headings with narrower, related or equivalent relationships as established by LCSH

By using these procedures, sorting became primarily a process of elimination. Once headings had been tagged and sent to a concept-based index, they could be removed from the original index. In this way, the original index got progressively smaller and smaller.

As long as the extracted headings exclusively belonged to one facet or designated index, the sorting process was easy. But many headings belonged to more than one category. I dealt with this problem in two ways.

1. As headings were being sorted, they could be sent to as many indexes as necessary. For example, the subject heading SUITES (RECORDER AND HARPSICHORD), was sent to two indexes: Forms/Genres and also Sound Devices.
2. The second way to resolve this problem was to decoordinate the headings. Headings were either manually decoordinated after they had been sorted into the concept-based indexes or they were automatically decoordinated once headings were collected into a resource file. As pointed out earlier, headings sent to resource files are saved in their original form and they are also factored into core vocabulary. The unique word list in the resource file could be quickly scanned, edited and validated by comparing the individual words to the heading strings. For example, what value do the words BAMBOO or BLOCK have to a music thesaurus? We can look down at the list of headings and find that they belong as possible boundterms: BAMBOO PIPE or BLOCK FLUTE. Only those terms relative to a particular facet were kept. The remaining vocabulary was sent to another index.

Decoordination made possible the reorganization of the many individual terms and modifiers that are embedded in the precoordinated LC subject strings. Once the headings had been sorted and decoordinated, it was possible to begin the real work of building a thesaurus. (Work which I've begun to realize may take a VERY long time.)

The outline onscreen shows the facets that have been partially developed for the music thesaurus to date. Terms within the facets have been assigned a temporary hierarchical arrangement. This design is a thumbnail sketch, based on the vocabulary represented in the LC music headings. It is expected that the outline and the hierarchical arrangements will change as terms are added and as more expertise is applied to the analysis of this fundamental work.

Having access to the two related collections of terms (the LC headings and the developing music thesaurus) in one database has several benefits. Both sets of vocabulary will be more accessible and may help complement each other. The varying contexts or combinations in which a word is used can be noted in either the Music Thesuarus or the LC music headings. Also, the user is able to trace the verbal and conceptual ties between the two lists. It is hoped that this faceted approach will assist users in identifying what to search and how to search it.

Going beyond the point of departure and looking toward the future, the goal of the music thesaurus is to track the vocabulary of the entire discipline of music. (4) By representing the most comprehensive vocabulary possible, by logically organizing that vocabulary and by making available an online product, whether stand-alone or interface, capable of sophisticated searching techniques, the music thesaurus offers many potentials to increasing access to music materials across a variety of databases.

Endnotes

1. Pauline Cochrane and Monika Kirtland, Critical Views of LCSH the Library of Congress Subject Headings: A Bibliographic and Bibliometric Essay and an Analysis of Vocabulary Control in the Library of Congress List of Subject Headings (Syracuse, NY:ERIC Clearinghouse on Information Resources, ED208 900, 1981).

2. Harriette Hemmasi, "ARIS Music Thesaurus: Another View of LCSH," Library Resources and Technical Services, vol. 36, no. 4, October 1992, p. 487503.

3. Harriette Hemmasi with Fred Rowley and James D. Anderson, "Isolating and Reorganizing Core Vocabulary from Library of Congress Music Headings for Use in the Music Thesaurus," Proceedings of the 4th ASIS SIG/CR Classification Research Workshop, October 1993, p. 89101.

4. Harriette Hemmasi, "The Music Thesaurus: Function and Foundation," NOTES, vol. 50, no. 3, March 1994, p. 875882.

Return to Proceedings Home Page

Advanced Search | Careers at OCLC | Feedback | Privacy Policy | ISO 9001 Certificate | ©2003 OCLC Online Computer Library Center, Inc.