Polyrepresentation

Polyrepresentation

“In short, the theory of polyrepresentation “. . . hypothesises that overlaps between different cognitive representations of both users’ information needs as well as documents can be exploited for reducing the uncertainties inherent in IR [information retrieval], and thereby improve the performance of IR systems” (Larsen and Ingwersen, 2002, p. 397). Ingwersen developed the theory of polyrepresentation through the 1990s. The theory is fully expanded in the Journal of Documentation article from 1996 (Ingwersen, 1996), which remains the main publication of the theory. Prior to that, the idea of polyrepresentation is mentioned throughout Ingwersen’s book Information Retrieval Interaction from 1992 as a high precision tool, and an early version was presented at SIGIR (Ingwersen, 1994). “(Larsen, 2004, p. 26).

The background for this concept is much older. Following the introduction of online retrieval systems in the 1960s much interest was directed towards the relative contributions of titles, abstracts, references and other subject access points in relation to recall and precision in IR. It was generally known among online searchers that different strategies could be used to increase recall and precision. In order to increase recall, one may, for example, use synonyms (e.g. natural language), search broader terms, limit the use of the Boolean “and” or search both "natural language fields” (e.g. titles and abstracts) and “controlled fields”. Similarly, precision might be improved, for example, by delimit the search to certain fields (e.g. titles, descriptors, and identifiers). It was also well known among experienced researchers that one can increase precision by demanding that a given word shall be represented in more than one field (e.g., search retrieval/ti and retrieval/de). This is what is later termed “polyrepresentation” and “cognitive overlaps” by Ingwersen and claimed to be a theory and part of “a cognitive view” in IR. Parts of this research on subject access points have been discussed in much greater detail by, for example, Hjørland & Kyllesbech Nielsen (2001).

Larsen (2004) maintains that polyrepresentation is closely associated with the cognitive view in information science. No attention is paid to the discussion concerning this view in the literature of information science or to the role played by alternative views such as semiotics or the domain-analytical view. The most distinctive feature of Larsen’s dissertation is the application of bibliographical references in information retrieval. In this respect the dissertation draws on, among others, Pao (1993) and Pao & Worthen (1989). This idea was not presented in Ingwersens early writings on polyrepresentation (e.g. Ingwersen, 1992) but was suggested by Hjørland (1993). This is at least one indication that what has been described as a theory of polyrepresentation connected to the cognitive view could as well (or perhaps better) be seen as a theory connected to the domain analytic view proposed by, among others, Hjørland (1993, 1997).

We may ask: What is the difference between a cognitive theory of polyrepresentation and a domain-oriented theory of polyrepresentation? Anderson (2003, p. 474) can be cited for a view that can be taken to represent a cognitive view related to polyrepresentation: “Research in psychology and information science has repeatedly demonstrated the enormous variability in the use of language by humans describing or seeking information (Collantes 1995). This is true for searchers as well as for indexers (Saracevic et al. 1988). Study after study indicates that indexers and searchers agree on terms about 23 per cent of the time. Most of the variability appears to be due to differences in choice for terms, and the remaining to differences in perception or conceptualizations of topics or features (Iivonen, 1994). Furnas et al. (1987) suggest that a good information retrieval system needs to provide as many as fifteen ways to express a topic or feature in order to accommodate up to 80 per cent of the search statements submitted by users”.

The view quoted above is a cognitive view because it is assumed that the human mind has a special tendency to produce variability, which may be uncovered by studies in cognitive psychology. The domain analytic theory, on the other hand, assumes that the variability in languages, e.g. the number of homonyms is related to sociological rather than psychological/cognitive factors. For example is the variation smaller in domains in which the researchers are disciplined by a precise special language.

Literature:

Anderson, J. D. (2003). Organization of knowledge. IN: International Encyclopedia of Information and Library Science. 2nd. ed. Ed. by John Feather & Paul Sturges. London: Routledge (pp. 471-490.

Brooks, T. A. (1993). All the right descriptors: a test of the strategy of unlimited aliasing. Journal of the American Society for Information Science, 44(3), 137-147.

Collantes, L. Y. (1995). Degree of agreement in naming objects and concepts for information retrieval. Journal of the American Society for Information Science, 46(2), 116-132.

Furnas, G. W.; Landauer, T. K.; Gomez, L. M. & Dumais, S. T. (1987). The vocabulary problem in human-system communication. Communications of the ACM, 30(11), 964-971.

Hjørland, B. (1993). Emnerepræsentation og informationssøgning. Bidrag til en teori på kundskabsteoretisk grundlag. Göteborg: Valfrid, Distributionsföreningen för inst Bibliotekshögskolan vid Högskolan i Borås och Centrum för biblioteks- och informationsvetenskap vid Göteborgs Universitet.

Hjørland, B. (1997): Information Seeking and Subject Representation. An Activity-theoretical approach to Information Science. Westport & London: Greenwood Press.

Hjørland, B. & Kyllesbech Nielsen, L. (2001). Subject access points in electronic retrieval. Annual review of information science and technology, 35, 249-298.

Iivonen, M. (1994). Consistency in the selection of search concepts and search terms. Information Processing & Management, 31(2), 173-190.

Ingwersen, P. (1992). Information Retrieval Interaction. London: Taylor Graham. http://www.db.dk/pi/iri/#the_whole_book

Ingwersen, P. (1994). Polyrepresentation for information needs and semantic entities: elements of a cognitive theory for information retrieval interaction. IN: Croft, W. B. & van Rijsbergen, C. J. (eds.). SIGIR ’94: Proceedings of the seventeenth annual international ACM-SIGIR conference on research and development in information retrieval, organised by Dublin City University, 3-6 July 1994. Dublin, Ireland. Lindon: Springer-Verlag, p. 101-110.

Ingwersen, P. (1996). Cognitive perspectives of information retrieval interaction: elements of a cognitive IR theory. Journal of Documentation, 52(1), pp. 3-50.

Larsen, B. (2004). References and citations in automatic indexing and retrieval systems - experiments with the boomerang effect. Copenhagen: Royal School of Library and Information Science, Denmark. (PhD thesis). http://www.db.dk/dbi/samling/phd/birgerlarsen-phd.pdf(Visited February 23, 2004).

Larsen, B. & Ingwersen, P. (2002). The boomerang effect: retrieving scientific documents via the network of references and citations. IN: Beaulieu, M., Baeza-Yates, R., Myaeng, S. H. and Järvelin, K. eds. Proceedings of SIGIR 2002 - the Twenty-Fifth Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, August 11-15, 2002, Tampere, Finland. New York: ACM Press, 2002, p. 397-398.(Poster paper). http://www.db.dk/blar/Files/Larsen_Ingwersen_SIGIR'02_poster.pdf

Pao, M. L. (1993). Term and citation retrieval: A field study. Information Processing & Management, 29(1), 95112.

Pao, M. L., & Worthen, D. B. (1989). Retrieval effectiveness by semantic and pragmatic relevance. Journal of the American Society for Information Science, 40(4), 226235.

Saracevic, T., Kantor. P., Chamis, A. Y., & Trivison, D. (1988). A study of information seeking and retrieving. I. Background and methodology. Journal of the American Society for Information Science, 39 (3), 161-176. http://www.scils.rutgers.edu/~tefko/JASIS1988part1.pdf