Knowledge taxonomies: what's the role for information professionals?

Information Outlook, June, 2005 by Jo Anne Cote

Taxonomy--the classification of items within subject domains--is especially effective in helping with today's information access difficulties. These structures are particularly good at representing open systems and are useful in the visual world of the Web. They are helpful in portraying abstract concepts and reflecting the various forms knowledge can take, whether tangible or intangible. Taxonomies do not necessarily use pre-existing classification schemes and are often based on a synthesis derived from user need and language.

[ILLUSTRATION OMITTED]

Like traditional library and information science classification arrangements, they can be hierarchical systems that also use controlled vocabularies and thesauri, but they are not always based on the accepted standards used in a traditional library setting. The multifaceted subject headings of a taxonomy have the added bonus of reflecting connections and processes in a semantic structure, or ontology, allowing for potentially wider consultation and application of knowledge objects. Taxonomies leave room for growth and constant revision, depending on external or internal factors.

In a presentation to the American Society of Indexers in May 2004, Denise Bedford, thesaurus manager and senior information officer for the World Bank Group, explained taxonomy architecture (see figures 1-4). Her overview of this architecture is helpful in understanding the taxonomical approach and how library and information professionals can integrate the approach with their existing skills and competencies. Taxonomies encompass many different aspects of information; for example, content, metadata and metadata repositories, navigation architectures, search architectures, and portal architectures. These structures can be presented explicitly (through data structures or interface features) or implicitly (by being designed into applications or content).

[ILLUSTRATION OMITTED]

Traditional LIS Skills and Competencies

Taxonomies can be constructed by building on the many practices typically considered part of the librarian's skill set, such as analysis, abstracting, research skills, cataloguing and classification systems, synthesis, subject headings, controlled vocabularies, indexing and thesauri (see table 1).

It is clear that many of the traditional skills of LIS professionals would be transferable to taxonomy creation, perhaps modified to apply to wider, more open organizational structures. For instance, the lists of household items, countries, or controlled vocabularies given as examples of a flat taxonomy structure (figure 1) could easily be the bibliographies or lists of books that librarians traditionally create using research skills, analysis, and abstracting abilities.

In the cataloguing process, librarians are already familiar with the process of creating MARC records, in which aspects of a document are analyzed and entered into the record as metadata. These skills are applicable to facet taxonomies (figure 2).

Hierarchical taxonomy architecture (figure 3) directly reflects the broad-term and narrow-term relationships and the notion of authority control apparent in traditional subject headings. Further, the process of synthesis can be applied in both instances to more accurately render concepts or contents.

Network taxonomies (figure 4) are more complex, existing as a combination of hierarchy and star architectures and expressing very rich knowledge objects and their relationships to one another. This semantic structure is another form of controlled vocabulary or authority control. Traditional LIS skills such as analysis, thesaurus creation, and indexing would definitely be applicable to network taxonomy.

In addition to information management skills, library and information professionals are traditionally taught basic interview and listening skills, and are expected to be able to communicate effectively in both written and verbal formats. These capabilities are useful in the construction of knowledge taxonomies, especially in ascertaining a common user language and appropriate subject categories.

Another factor to consider in assessing the applicability of traditional library and information professional skills to taxonomy is the common notion that a library is a physical structure--a school, university, or public library--dealing in physical manifestations of knowledge, usually in print format. This one-dimensional vision of a library is no longer accurate. Special libraries--for example, corporate or digital libraries--often do not fall into this category and do not necessarily share the same educational or public-orientated goals. The needs of these libraries are not usually reflected in the traditional LIS curriculum. In figure 5, I list the traditional LIS skills that are relevant to knowledge taxonomies.

[FIGURE 1 OMITTED]

New LIS Skills and Competencies

While existing LIS skills provide an adequate base on which to build an understanding of today's information reality, certain additional competencies and mind-sets would be useful, especially with regard to taxonomy development. In particular, the ability to see the big picture, think conceptually, and deal with ambiguity are important to synthesizing a useful taxonomy. Lateral thinking and general contextual awareness are also important abilities when visualizing multifaceted applications of subject headings that reflect as much as possible the relationships and value of knowledge objects to their user groups. Also, the traditional book-library focus must be broadened to include other kinds of libraries, information formats, and users.