“Isms” in information science: constructivism, collectivism and constructionism Sanna Talja The University of Tampere Centre for Advanced Study (UTACAS), The Research Institute for Social Sciences, University of Tampere, Tampere, Finland Kimmo Tuominen Library of Parliament, Helsinki, Finland, and Reijo Savolainen Department of Information Studies, University of Tampere, Tampere, Finland Abstract Purpose – Describes the basic premises of three metatheories that represent important or emerging perspectives on information seeking, retrieval and knowledge formation in information science: constructivism, collectivism, and constructionism. Design/methodology/approach – Presents a literature-based conceptual analysis. Pinpoints the differences between the positions in their conceptions of language and the nature and origin of knowledge. Findings – Each of the three metatheories addresses and solves specific types of research questions and design problems. The metatheories thus complement one another. Each of the three metatheories encourages and constitutes a distinctive type of research and learning. Originality/value – Outlines each metatheory’s specific fields of application. Keywords Philosophy, Information science, Libraries Paper type Conceptual paper Introduction Why bother with metatheories[1]? Gorman (2001, p. 24), for instance, takes the view that “we cannot spend a great deal of time and effort on speculative enquiry” but should seek to resolve the very serious practical problems that confront libraries, librarians, and library users today. Solutions to practical questions are, however, always developed on the basis of theoretical and epistemological assumptions. As stated by Hjørland (2003a, p. 805), researchers and practitioners “cannot choose between using a specific philosophical framework and not using any philosophical framework”. Even the most rudimentary metadata solutions and information retrieval algorithms are based on metatheoretical assumptions (Hjørland, 1998, p. 606). The Emerald Research Register for this journal is available at The current issue and full text archive of this journal is available at www.emeraldinsight.com/researchregister www.emeraldinsight.com/0022-0418.htm This article is an extended and substantially modified version of a paper presented at the CoLIS 4 Conference (Tuominen et al., 2002). The authors wish to thank Marcia Bates, Jenna Hartel, Birger Hjørland, and the anonymous referees for comments that greatly helped to specify the arguments presented in the article. “Isms” in information science 79 Accepted 27 August 2004 Journal of Documentation Vol. 61 No. 1, 2005 pp. 79-101 q Emerald Group Publishing Limited 0022-0418 DOI 10.1108/00220410510578023