Authority and Vocabulary Control in Image Collections

by Karen Spencer


Due to the introduction and implementation of on-line image catalogs and databases containing entire image collections there is a need for some sort of universal language for cataloging image collections. It is now possible to access on-line various image collections, but some sort of universal language is needed to catalog certain aspects of images, to make possible to search for specific images. Yet, there are many problems related to creating universal cataloging standards for visual collections, and I will only deal with the vocabular and authority control needed to describe art, architecture, and museum images.

Authority and Vocabulary Control

It is very difficult to imagine that a universal descriptive language can be created and used proficiently. Image collections have not developed a universal standard for organization because most collections have extremely diverse information needs. For example, art historians tend to use iconographic descriptors when cataloging images, while museum curators use straightforward information taken directly from the physical aspects of an art work. If an image collection is serving a small group of scholars, vocabulary to describe an image can usually be chosen fairly easily, because the cataloger understands the needs of the patron. However, if the audience becomes the entire on-line world, can a single, universal, descriptive tool be used to satisfy all of the users?

Art historians argue that it is not possible to restrict the language used to catalog an image. One image can contain an inexhaustible range of meanings, and it can take more than one thousand words to describe it. There is not an adequate list of terms available today to describe all of the aspects of images contained in image collections. In fact, few reliable descriptors exist for images: "It is that pictures need words. Words for artists. Words for style. Words for dates. Words for subjects. Words for things. In other words, many, many words."

It is conceivable to expect catalogers to use some type of universal authority control and controlled vocabulary in specific fields such as the artist/architect name and date field and the title field, and to maintain the remaining fields that they feel necessary for their users in the same manner as they have in the past. The use of controlled vocabulary in such basic fields as artist/architect name and title fields will enable cooperation between institutions while maintaining the necessary unique fields for each individual image collection. Cooperation between institutions will not necessarily call for complete standardization in descriptions, but will call for compatible software systems. Nevertheless, questions remain about which control tools are the most adequate for describing art and architecture images.

Controlled vocabulary and authority control have developed independently in individual art, architecture, and museum image collections . Endless methods for cataloging images have been created due to the diverse needs of the users. The types of indexes that have been created also vary enormously. In addition, the nature of the image collections vary greatly, even within the same discipline. For example, in the field of art history, some collections may be based on one very narrow subject such as Italian Renaissance architecture and may use a very limited controlled vocabulary, while others may contain thousands of general art and architecture images and use no controlled vocabulary at all. The only way some of the images can be located is by actually searching through the physical location where the images are stored.

Image collection developers have, in the past, used whatever tools they felt necessary to organize their specific collections. These tools include geographical dictionaries, the Art and Architecture Thesaurus (AAT), the ICONCLASS, the subject list developed by Elisabeth Betz Parker (of the Library of Congress Prints and Photographs Division), and various naming protocols. The librarians and curators that I spoke with commented that in the future they planned to consult the AAT and the Union List of Artists Names developed by the Getty Art History Information Program (ULAN) in the future because these tools are available electronically and will be easier to access when cataloging. The image collection organizers admitted that in the past they have not been consistent in their use of controlled vocabulary and authority lists. Although many of the collections did have existing "home-made" authority and vocabulary controls available to them, the lists were not consulted consistently. Cataloging for these image collections is based on the "historical knowledge" of the collection, which is usually acquired during their training (from their predecessor). Because of this, many of the collections have not handled information consistently from one generation to the next. The second-generation and third- generation catalogers generally disagree with various aspects of the previous catalogers' methodology, so they change methods to suit their own needs. This was never a problem until image collections began to create computerized index systems.

Many image collections converted their records of old paper indexes to computerized databases. During the conversion it became blatantly apparent that there was a problem with inconsistency in the information being used to catalog images. Database structure demands consistency of information for search capabilities. No matter how good the database software is, "technology will not dramatically change how an item is indexed without the cataloger or indexer first changing how he/she goes about the process of documentation."

Use of Thesaurus vs. ICONCLASS

An image does not tell us what it is about. Images contain information that is not only useful to art historians but also to social historians, historians of music, medicine, engineers, indeed anyone who wants to learn about the appearance of historical people and objects: "...A set of photographs of a busy street scene a century ago might be useful to historians wanting a 'snapshot' of the times, to architects looking at buildings, to urban planners looking at traffic patterns or building shadows, to cultural historians looking at changes in fashion, to medical researchers looking at female smoking habits...." Who is to judge what terms should be used to describe an image?

The two most common sets of terms consulted by image collection organizers are the Art and Architecture Thesaurus and ICONCLASS. The AAT was created by the Art History Information Program (AHIP) and is trademark of the J. Paul Getty Trust. The AAT contains extensive art and architecture terms (although it does "skimp" on the terms used to describe images in the fine arts). It is organized in a hierarchical way, lists words with definitions, and contains notes on usage. The AAT provides a flexible base for "defining" an image. Most image collections that use the AAT use it as a supplement of some sort in conjunction with other vocabulary lists they have created for their individual collections. ICONCLASS, on the other hand, is not as flexible: "[ICONCLASS] takes everything, absolutely everything, you can think of. Then divides it into one hierarchical and interlocking set of broad categories with examples." ICONCLASS was developed by Professor H. Van der Waal to deal with the traditional iconography of art history. It contains codes that represent various iconographic elements used in cataloging descriptions. The codes are difficult to remember (you have to look them up in the nine volume set), and because of the interlocking nature of the system, additional codes are difficult to incorporate. Catalogers can not always say what they want to say using these codes.


It is clear that there is a need for a universal vocabulary and authority control in image collections. Many computerized projects have experienced problems in trying to define what information should be contained in specific fields (e. g. What do you put in the location field? Do you use the location of the object depicted in the image, or the location of the holding institution, or maybe the location of the stored surrogate image?). The problem of what type of vocabulary to use in the description of art or architectural images has not really been addressed in the on-line image systems I have seen. Controlled vocabulary and authority have been more of an issue for individual image collections. Controlling the vocabulary used is a very important aspect of future organization of advanced image collections. The realization that standards were needed not only for consistency within one's own image collection but also for future collaborations with other institutions' image collections resulted in the creation of the AAT and ULAN. Additional projects are also in progress, such as the Visual Resources Association project to develop standards for descriptive elements. Conforming to the vocabulary used in ULAN and the AAT are both good beginnings. However, if AAT is to be used for future cataloging, the terms will need to be expanded. Still, I feel the AAT is the most useful tool for describing art, architecture, and museum objects, and for controlling the consistency of vocabulary in database indexes.


1. Roddy, Kevin. "Subject Access to Visual Resources: What the 90s Might Portend." Library HI Tech, Vol. 9, No. 1 (1991), pp. 45-60.

2. Barnett, Patricia J. "An Art Information System: From Integration to Interpretation." Library Trends, Vol. 37, No. 2 (Fall 1988), pp. 194-205.

3. Besser, Howard. "Visual Access to Visual Images: The UC Berkeley Image Database Project." Library Trends, Vol. 38, No. 4 (Spring 1990), pp. 787-98.

4. Roddy, Kevin. "Subject Access to Visual Resources: What the 90s Might Portend." Library HI Tech, Vol. 9, No. 1 (1991), pp. 45-60. l


Barnett, Patricia J. "An Art Information System: From Integration to Interpretation." Library Trends, Vol. 37, No. 2 (Fall 1988), pp. 194-205.

Bearman, David. "Considerations in the Design of Art Scholarly Database." Library Trends, Vol. 37, No. 2 (Fall 1988), pp. 206-219.

Besser, Howard. "Getting the Picture on Image Databases -- The Basics." Database, Vol. 18, (April/May 1995), pp. 12-16+.

Besser, Howard. "Visual Access to Visual Images: The UC Berkeley Image Database Project." Library Trends, Vol. 38, No. 4 (Spring 1990), pp. 787-98.

Harping, Patricia. "The Architectural Subject Authority of the Foundation for Documents of Architecture." Visual Resources, Vol. 7, No. 1 (1990), pp. 55-63.

Leung, C.H.C., D. Hibler, and N. Mwara. "Picture Retrieval by Content Description." Journal of Information Science, Vol. 18, No. 2 (1992), pp. 111-119.

Reed, Patricia Ann, and Jane Slegdg. "Thing About Museum Information." Library Trends, Vol. 37, No. 2 (Fall 1988), pp. 220-231.

Roddy, Kevin. "Subject Access to Visual Resources: What the 90s Might Portend." Library HI Tech, Vol. 9, No. 1 (1991), pp. 45-60.

Stam, Deirdre. "Pondering Pixeled Pictures: Research Directions in the Digital Imaging of Art Objects." Visual Resources, Vol. 10, No. 1 (1994), pp. 25-39.

Sutherland, John. "Image Collections: Librarians, Users and Their Needs." Art Libraries Journal, Vol. 7, No. 2 (Summer 1982), pp. 41-60.