Birger Hjørland. Information seeking and subject representation: an activity-theoretical approach to information science. (New directions in information management. ISSN 0887-3844; 34). Westport, CT, USA: Greenwood Press, 1997. 213 p. Aus$?. Hard cover. 0 313 29893 9.

The title says it all – but probably requires further explanation. Hjørland has written extensively for more than a decade on knowledge organisation and information retrieval from a psychological perspective. Here, he amplifies his earlier work contending that activity theory has much to offer user studies and information retrieval research.

Activity theory derives from the philosophical pragmatism of John Dewey, who presaged the movement away from a rationalist, positivist approach in psychology, to one that attends more to the socioeconomic milieu. Hjørland, applying epistemology to information science, considers that users’ behaviour should be interpreted in the light of a scientific situation in a given area, or discourse community. In more practical terms, document representation with indexing languages or classification schemes, should be carried out with particular domains of knowledge in mind. The indexer should ask ‘Where is the user coming from?’ - or as Hjørland puts it: ‘different epistemic communities have different needs’ . This is akin to Soergel’s request-oriented rather than content-oriented indexing (1).

The thesis has much to commend it, but I think he could have illustrated it more convincingly. He uses library classification and subject headings that by their nature are not discipline-specific, to exemplify the difficulties of controlled vocabularies in addressing particular user domains. He gives examples of thesauri that are discipline-specific, but seems not to use these in support of his activity-based approach. He argues that if a given group of people begin to collect information systematically (example: florists), then they have in effect established an applied science, with particular information needs. Then he forsakes the individual by saying if there is no collective group with well-defined information needs, then there is no information system to optimise. He opines that knowledge representation rather than user-friendliness is the central theoretical issue in information retrieval, but at the same time contends that knowledge representation contributes to user friendliness. He queries the international standard (2) on subject analysis for being ‘document-centered’. It is supposed to be.

Hjørland sets down his ideas in chapters on a conceptual approach to subject retrieval, subject analysis including polyrepresentation, epistemological positions on the concept of subject, methodological consequences, a framework for information seeking, and information needs. In these he argues for a notion of subject analysis that identifies the informative potentials of documents for users who may have different levels of aspiration.

Although he states that there is no "correct" subject analysis of a document, curiously, there is no reference to Derrida’s theories on multiple interpretation of texts. I would also have welcomed some exploration of how a thesaurus (and its developers) may go beyond reflecting the paradigm of a discipline, to influencing a discipline’s evolution.

Despite this carping, I consider that the book provides a useful bridge between philosophy and classification, and a significant theoretical discourse. Its questioning of the technological solution to retrieval problems, makes it salutary for cataloguers and indexers seeking a fundamental basis for why they do what they do.

(1) Dagobert Soergel. Organizing information. London: Academic, 1985.

(2) International Standards Organizaton. Methods for examining documents, determining their subjects, and selecting index terms. (ISO5963). Geneva: ISO, 1985.