Michèle Hudon, associate professor at l’École de bibliothéconomie et des sciences de l’information at l’Université de Montréal, spoke at the ISC conference about ISO25964, a new standard for thesauri.
ISO25964 will replace ISO2788 and ISO5964, out-of-date standards for monolingual and multilingual thesauri, respectively. These standards don’t address subject headings or taxonomies and are ill-suited to a networked environment of linked data, so in 2008, an international working group was struck to create the new standard based on the content of BSI8723, created by British information specialists. The working group, of which Dr. Hudon is a part, includes both practitioners and researchers.
Officially titled “Information and Documentation—Thesauri and interoperability with other vocabularies,” the ISO25964 project consists of two parts, the first of which, “Thesauri for information retrieval,” was published in August 2011 and covers general principles for developing and managing monolingual and multilingual thesauri. Part 2, “Interoperability with other vocabularies,” addresses crosswalks and mappings and is scheduled to be published at the end of this year.
“Interoperability”—a bit of a mouthful, as Hudon admits—refers to an ability to “act together coherently, effectively, and efficiently to achieve common objectives.” In the world of information science, it means the “capability of agents, services, systems and computer applications to exchange data, information and knowledge while preserving their integrity and full meaning.” (Zeng and Chan, 2009)
Thesauri are great tools for information retrieval for local users, but there may be multiple thesauri on the same topic that have different classification schemes and subject headings and thus can’t talk to one another. Having multilingual thesauri adds another layer of complexity.
In traditional information systems, thesauri allow a searcher to use the same search terms and strategies to search several databases and provide an efficient way to cross the language barrier. With the web, semantic interoperability becomes even more relevant. It allows for effective searching in several situations, including with the same language in different countries, two or more natural languages, a natural language and a language of specialty, a natural language and an indexing and retrieval language, and one or more indexing and retrieval languages (e.g., Library of Congress and Dewey).
Interoperability implies equivalence, but many would argue that absolute equivalence, particularly between distinct languages, doesn’t exist. Part 2 of the standard gives recommendations for establishing and maintaining mappings between multiple thesauri or between thesauri and other types of vocabularies. As Hudon said in her talk, “Because it necessarily exists in a particular cultural, social, professional and linguistic context, semantic and terminological interoperability of indexing languages depends on compromises to compensate for the lack of absolute equivalence between concepts and between terms.” She also emphasizes that semantic equivalence is dynamic and ever evolving.