Organizing Knowledge. An Introduction to Managing Access to Information

The first edition of this book was published in 1987, followed by the second edition in 1992 and the third edition in 2000. The present edition is restructured and organized this way:

Content outline:

1. 1.

   List of figures.

2. 2.

   Introduction.

3. 3.

   List of acronyms and abbreviations:

   • •

     Part I Structuring and Describing– Knowledge, information and their organization– Formatting and structuring knowledge– Describing documents

   • •

     Part II Access– Users and user behaviour– Subjects as access points– Classification and order– Further concepts and tools for subject access– Access through author names and titles

   • •

     Part III Systems– Organizing knowledge in the digital environment– The evaluation and design of information retrieval systems– Organizing knowledge without IT– Management of knowledge systems

4. 4.

   Index.

This is a standard textbook in the field, which has been rather widely used. It has many fine qualities, but it also reflects the crisis in the field, primarily related to the lack of theoretical foundation in relevant fields such as semantics and epistemology. The book presents a lot of systems constructed by and/or used by information professionals. It seems to be strongest in the approach to Knowledge Organization (KO) known as facet‐analysis. It fails, however, in the reviewer's opinion, to present the general theoretical foundations of KO, its different theoretical assumptions and its basis in interdisciplinary research.

To illustrate this point, let us start by considering what this book writes – and does not write – about “disciplines”. On p. 180 is written: “All current classifications base their main classes on disciplines. Disciplines are ways of looking at the world. The narrowest definition of disciplines postulates a small number – perhaps six or seven – of fundamental disciplines, and contrast them with phenomena, the objects of the world, any of which can be studied from more than one disciplinary viewpoint… ”.

The reader is not told who has proposed this “narrow definition” of disciplines and the possible six or seven candidates are not mentioned. Neither is this “narrow definition” is confronted with any criticism or alternatives. Most important is that such a postulate of six or seven fundamental disciplines suggests a certain understanding of disciplines which is in direct conflict with the understanding of disciplines as social divisions of intellectual labour reflected in, for example, the organization of higher education (which is the reviewer's understanding of disciplines). This last understanding opens the way to the empirical study of disciplines and thus to an empirical base of the part of Knowledge Organization working with this concept. What is written about disciplines by Rowley & Hartley is mainly common sense considerations, not based on examination of the research literature on the issue. A very important explanation and critical analysis of the developments of the system of disciplines in the social sciences were made by Wallerstein and reviewed by Hjørland (2000) for the community of knowledge organization. There are other relevant contributions to the study of disciplines, including Dogan (2001). The failure of our community of KO to incorporate such contributions in our own knowledge and use them in developing our field is, in my opinion, a sign of an intellectual crisis. The concept of discipline is not, unfortunately, the only example. Such omissions represent the rule rather than the exception. This criticism is directed towards our community as a whole, not just to our textbook‐writers, who rely on the input from researchers. When this input is lacking, it becomes almost impossible to write high‐level textbooks.

Perhaps the readers of this review (as well as the authors of the book being reviewed) will find my above formulated criticism unjustified. After all, the first author has written many papers including papers defining core concepts such as information and the data‐information‐knowledge‐wisdom (DIKW) hierarchy. However, the understanding that KO is fundamentally about concepts and their semantic relations, about semantic tools and knowledge organizing processes, that different views of knowledge, language, concepts and disciplines have implications for theories of KO etc. seems to be lacking. It is this reviewer's claim that as long as this is lacking the field of KO will not be able to develop in a deeper sense but will mainly be about the use of technology produced by other professions and will produce rather fragmented and ad hoc textbooks.

There is one important exception from the criticism formulated above. This edition has maintained a view of indexing from the former edition: “In order to achieve good consistent indexing, the indexer must have a thorough appreciation of the structure of the subject and the nature of the contribution that the document is making to the advancement of knowledge within a particular discipline” (Rowley and Hartley, 2008, p. 109). I find this formulation really insightful and extremely important. I also find it in accordance with the view expressed by myself: “The subjects of a document are its informative potentials” (Hjørland, 1992, 1997). This view may be hard to accept for our profession because it implies that the core knowledge needed to perform KO is not something that can be mastered without deep subject‐specific knowledge. However, if it is true – as I believe it is – the profession cannot help itself by avoiding facing reality. As long as the profession had a kind of de facto monopoly on organizing knowledge in physical libraries, it was able to survive without taking the full consequence of this insight. Today it is urgent for our survival that we develop a more realistic understanding of KO. In my opinion this should be based on the theory of knowledge and the sociology of knowledge. These are difficult but necessary fields to relate to.

I will now discuss a few selected parts of the book. Chapter 10: “The evaluation and design of information retrieval systems” is a well written text‐book chapter. It covers the basics and it is fine that it presents Keen's research which the authors claim are under‐reported. Their doing so also proves that they are well read in the literature of information science. However, the chapter remains with a “traditional” understanding. It does not, for example, consider the formerly discussed principle that “the indexer must have a thorough appreciation of […] the contribution that the document is making to the advancement of knowledge”. If they had done so, they should have related to how such indexing might be evaluated. A next step would possible be to recognize that experts may disagree, that there exist different “paradigms” and that each “paradigm” implies its own criteria of relevance.

The important concept of literary warrant – although being mentioned in the book (e.g. p. 177) – is not in the index, nor is there any deeper introduction or discussion of it. Also the important principle of request oriented indexing is not in the index, and it is also clear from the book that this concept is either unknown or not recognized by the authors. For example p. 198: “class at the discipline which receives the greater emphasis [in the book to be classified]. This is clearly an expression of content oriented indexing, i.e. the opposite of request oriented indexing.”

Although Carl Linné is mentioned (p. 175‐6) there is in reality no discussion of scientific classification, its philosophy, methodologies, and implication for our field. It is stated (p. 138): “Semantic relations between terms are, as their name implies, built into the meanings of the terms. They are permanent, in that they do not change according to whatever document is being indexed or searched. Semantic relations are stable … ” I think this is simply wrong, and it is not related to any recognized theories about semantics. Meanings are not built into the words, but are changing with contexts (see Wittgenstein's famous concept of language games). Semantic relations are also theory‐dependent. Whether a drug is a tranquilizer or not, for example, depends on medical research, and different experiments may provide different conclusions. Pluto used to be classified as a “planet”, but in August of 2006 The International Astronomical Union redefined it along with some asteroids as a dwarf planet. Whether or not birds are considered reptiles depends on the taxonomic theory. According to the phenetic school in biological taxonomy they are not, but from the point of view of cladism they are considered kinds of reptiles. These are examples demonstrating that semantic relations are theory dependent and changes from one document to another. They demonstrate the importance of not ignoring scientific classification and the importance of basing a theory of semantic relations on proper studies.

I will end this review by arguing – as I did in my review of: Chowdhury and Chowdhury (Hjørland, 2008) – that it is critical that the book does not relate to bibliometrics as a way of organizing knowledge – although it does introduce the concept of a “citation”. There are two main arguments. First: as documented by, for example, Schneider (2004) it is possible to use bibliometric methods to identify potentially relevant descriptors for thesaurus construction, thus bibliometrics have been demonstrated to be a possible method for KO. Second: let us for the third time return to the quote: “the indexer must have a thorough appreciation of the… nature of the contribution that the document is making to the advancement of knowledge”. Only the most qualified researchers at the research front are probably qualified to do so, and they may do it by referring to the relevant literature in their papers. Bibliometric maps may thus be highly dynamic systems of organizing knowledge based on experts' evaluation of “the nature of the contribution that the document is making to the advancement of knowledge”.