Domain Analysis for Knowledge Organization

Tools for Ontology Extraction

Richard Smiraglia

• Éditeur: Chandos Publishing
• Année: 2015
• ISBN:
  9780081001882
• Sujet: Business
• Sous-sujet: Gestione dell'informazione

Domain analysis is the process of studying the actions, knowledge production, knowledge dissemination, and knowledge-base of a community of commonality, such as an academic discipline or a professional community. The products of domain analysis range from controlled vocabularies and other knowledge organization systems, to scientific evidence about the growth and sharing of knowledge and the evolution of communities of discourse and practice.In the field of knowledge organization- both the science and the practice­ domain analysis is the basic research method for identifying the concepts that will be critical building blocks for knowledge organization systems. This book will survey the theoretical rationale for domain analysis, present tutorials in the specific methods of domain analysis, especially with regard to tools for visualizing knowledge domains.

• Focuses on the science and practice of organizing knowledge
• Includes step-by-step instructions to enable the book to be used as a textbook or a manual for researchers

1

Discourse domains and their role in knowledge production dissemination and organization

Abstract

Domain analysis for knowledge organization is the embrace of techniques for discovering the knowledge base of specific communities, for the purpose of informing the science of the order of knowledge and its application in knowledge organization systems. Domain analysis has become a core paradigm within the knowledge organization community in the postmodern environment. The papers that now are seen as catalytical in knowledge organization came from Hjørland and Albrechtsen (1995), which suggested domain analysis as a new approach to information science; a follow-on article Hjørland and Albrechtsen, 1999 focused directly toward the knowledge organization community; Hjørland (2002), which lays out 11 formal techniques; and, a special 2003 issue of Knowledge Organization on domain analysis edited by Hjørland and Hartel. Tennis (2003) defines two axes for the functioning of domain analysis as a methodological paradigm for the discovery of transferable analytical ontology. Just as domain analysis for knowledge organization has incorporated many theoretical perspectives, so has it been demonstrated to be a multimethod paradigm. There are many ways to elicit a terminological or thematic knowledge base by using empirical techniques for documentary analysis or qualitative techniques for ethnographic observation of a domain. Critical theory, semiotics, and discourse analysis provide social and cultural approaches to understanding domain epistemological perspectives. Domain analysis is a multitheoretical, multimethods core paradigm in the science of knowledge organization.

Keywords

Domain analysis

Knowledge organization

Ontology

Multimethods

Critical theory

1.1 Domain analysis for knowledge organization

This book is about domain analysis for knowledge organization. We take the term “knowledge organization,” often represented as KO in these pages, to mean the science of the order of knowledge and its application in knowledge organization systems (KOSs). Elsewhere (Smiraglia, 2014) I have defined knowledge as that which is known, and suggested that the science of knowledge organization is concerned not only with the metalevel multi- and interdisciplinary comprehension of knowledge but also with the heuristics for the conceptual ordering of that which is known. Research in knowledge organization takes place in many arenas, from the philosophical to the basics of every science. Because ordering knowledge is an essential aspect of the development of systems for information retrieval, often we find a focus on knowledge that is recorded in documents. But this need not be a formal criterion for the science of knowledge organization. We may study the heuristics for the ordering of concepts in documents, but we also may study the heuristics by which natural phenomena (such as biological phenomena) seem to be ordered, or to order themselves in reality. Thus, in KO, we are concerned with the ability to study the natural order of phenomena in every context—a frighteningly complex context for research.

However, the science of knowledge organization has emerged from centuries of practice—taxonomy and typology for certain, but also epistemology and ontology, and the evolution of controlled vocabularies (thesauri) and symbolic notational systems (classifications). In every instance, we require empirical understanding about the knowledge bases of contexts. In the science of knowledge organization as it has evolved from proposals by Dahlberg (2006), the search for atomic knowledge elements has focused on concepts. Concepts are themselves complex, (Smiraglia and van den Heuvel, 2013; Hjørland, 2009), and their constitutional aspects must also be comprehended. Toward the end of the twentieth century, the KO community turned to a postmodern view of knowledge (Mai 1994; Smiraglia 2012) in which domain-centric points of view and interoperability among them replaced the search for global (universal, catholic, unitary, etc.) systems. In this new reality domain analysis, or the study of the knowledge bases of specific, definable contexts, has become a core paradigm within the knowledge organization community. This book is about specific techniques, primarily those that are empirical, for discovering, documenting, and analyzing domains.

1.2 Catalysts for domain-analytical thought

From about the mid-1990s knowledge organization has been focused on efforts to interpret diverse domains and points of view (epistemologies) together, or in tandem, rather than continuing to seek a single universally applicable approach to the organization of all human knowledge. Mai (1999) was among the first to call for such a postmodern view of knowledge organization and its use as an approach to KOSs. Explaining the modern view that the role of knowledge organization was to represent (if, in fact, it could discover) the actual order of things in the universe of which humanity was a part, Mai suggested a distinction between a modern approach to knowledge organization based on discovery of ontology and a postmodern, existentialist approach to knowledge organization based on epistemology, or the points of view of diverse groups of people. Postmodern approaches acknowledge that a KOS is itself just one single point of view about a particular knowledge set, that other points of view can coexist and could be formed into KOSs, and finally that the role of postmodern knowledge organization as a science is to discover the multiplicity of points of view and search for techniques to bridge or crosswalk them, an approach that came to be known in part as interoperability. More recently, Mai (2010) has told the story of the modern period in knowledge organization in more detail, tracing the search for realism in applicability of universal structures for knowledge and its order. Mai also makes reference to an important distinction that arose in the postmodern era of knowledge organization between classifications that are nascent or “naïve” (Beghtol, 2004, 19), which are the product of emergent scholarly activity in its observational (and hence nascent, prehypothetical) stage, and those that are “professional information-retrieval” orderings designed to facilitate discovery and retrieval of previously recorded knowledge elements. The natural consequence of the shift toward postmodern epistemological approaches and the acknowledgment of the naïve-professional dichotomy is the need for domain analysis, a methodological paradigm for the discovery of a knowledge base in a given, specific community.

The papers that now are seen as catalytical in knowledge organization came from Hjørland and Albrechtsen (1995) and were originally oriented to the information science community. Their 1995 paper suggested domain analysis as a programmatic approach or “new horizon” for information science (p. 400), “to study the knowledge domains as thought or discourse communities, which are part of society’s division of labor.” They suggested that components of knowledge organization were “reflections of the objects of the work” of discourse communities, incorporating even the criteria of individual community members. Domain analysis for the determination of system requirements as an aspect of computer science had found its way into information science as a means of orienting systems for information retrieval based on the specific requirements or information needs of research communities. The subtle shift from system requirements to discovery, recording, and tracking of a discourse community’s ontology, or knowledge base was seen as a means of formalizing research for KOSs as a paradigmatic aspect of information science. Principle methods for domain analysis in information science, primarily informatics, were evolving from studies of communication among scientists into approaches to the mapping of disciplines and their evolution, particularly their paradigmatic evolution. A follow-on article (Hjørland and Albrechtsen, 1999) appeared in Knowledge Organization and was focused directly toward the KO community identifying domain analysis as a trend in classification research. In this paper, the pair wrote about the role of disciplines in modern KO research, the rise of inter- and multidisciplinarity, and the epistemological distinctions between scientific and bibliographic classifications (or, between taxonomy and typology of scientific disciplines, and classification of documents as it is known in information science). A consequence of pragmatic epistemology, they suggested, was the shift in knowledge organization (which they see as a component of information science) toward classification of knowledge in historical, social, and cultural contexts. Another consequence of pragmatic epistemology is to move away from the rationalist approaches to facet analysis that seeks to find universally applicable dimensions of knowledge, and to move toward non documentary-based or inter textuality-based approaches—empirical approaches—to the discovery of knowledge structures. In other words, although they do not say so explicitly in this paper, a shift is seen as evolving toward domain-analytical approaches to the discovery of domain-centric ontologies and to domain-ep...