Mei Mei Wu
Professor and Chairperson, Graduate Institute of Library and Information Studies, National Taiwan Normal University, Taiwan (R.O.C.)
E-mail: meiwu@ntnu.edu.tw
Keywords: Classification Theory; Knowledge Organization; Linked Data; Ontology;
Semantic Web
【Abstract】
This article argues that the biggest difference between the knowledge organization method and the traditional information organization in the Internet age is that the boundaries of information objects in the Internet age no longer exists, the interconnected characteristics of information and knowledge, and the concept and method of knowledge organization are fundamentally changed. From the island‐like knowledge organization for the single information object as the knowledge organization unit, evolving to the multi‐faceted knowledge organization method, of which the major functions are to identify and search for information. The Semantic Web approach of knowledge organization allows remapping and reconstructing new knowledge in terms of interpretation and thus offering new insights of knowledge.
This article addresses the related theories and cases of such knowledge organization to highlight the new development of dialectical knowledge organization.
【Long Abstract】
Introduction
This paper proposes that there has been a paradigm shift in the way knowledge is organized (knowledge organization, KO); the traditional reasons for organizing information and the methods used have changed with the development of the Internet. It first argues that in the Internet era, the term “knowledge organization” is more appropriate than the term “information organization” as in our digital world, information can no longer be confined to a specific physical location as in the past. Lee (2015, p.100) supports the use of the term “knowledge organization” rather than
DOI: 10.6245/JLIS.2017.431/725
“information organization” because embedded within the term are concepts, terminology, meaning, relationships, order, structure, approaches, perspectives, context, etc. These are not included in an understanding of “information”. By applying the new methods of knowledge organization, such as ontologies, Linked Data, and triple RDF, knowledge will have enhanced meaning, structure, and order. Furthermore, in "knowledge representation" there will be the opportunity for reinterpretation and finding new knowledge and meaning. This paper also discusses the work of Dao Feng, the 10th century Chinese officer who made the printed Chinese classic sutra available to highlight that the bias of knowledge affects the scope of knowledge and knowledge organization.
Observation and thoughts
The evolution of information technology in the Internet age has encouraged the shift in the KO model.
The Semantic Web approach of KO allows the remapping and reconstructing of knowledge in terms of interpretation, thereby offering new insights. The following eight issues, which highlight the movement of knowledge organization in the digital environment, are considered:
1. Characteristics of media affect KO methods: The methods used to organize knowledge in the Internet era differ from traditional methods in that there are no longer boundaries to the single knowledge object. This shift is the result of the interconnection between information and the characteristics of knowledge.
2. New KO methods expand KO functions: The major function of traditional information organization was to manage the information in order to be able to rapidly retrieve it. KO in the Semantic Web environment allows new interpretations, new representation. The reconstruction of knowledge is thus a new function.
3. New KO methods can be explained by Classification theory: Sparck Jones (1970, 2005) proposed three basic elements of classification: class, object/entity, and attributes. By using these three basic elements, eight classification models are developed. It is usually assumed that academic classification, the first category, is rigorous, but there are always still some grey areas that lead to differences of opinion as well as some difficulties. For example, the problem of tomatoes and avocadoes: there is always some controversy about whether they should be classified as vegetables or fruit as they could be either.
These sorts of problems can be solved in Sparck Jones’ classification model as it allows objects to be represented in multiple categories with compound attributes.
4. Classification is difficult: Jorge Luis Borges wrote “John Wilkins's Analytical Language” (Borges,
1993) to criticize a set of classification symbols developed by John Wilkins, the 17th century British mathematician and linguist. Borges made up 14 categories of animals from a fictional Chinese encyclopedia called the “Celestial Empire of benevolent Knowledge.” The story inspired Foucault and was documented in the foreword of The Order of Things (1994). The 14 categories have become a famous classification exercise to show LIS students that classification is difficult and that a perfect classification system may not exist.
5. Current classification in practice: There are different levels of complexity in classifying natural objects and artifacts in regard to scientific knowledge and the classification of documents and digital objects in the Internet. KO tools such as ontologies, topic maps, Linked Data, etc., are now available for innovative knowledge representation.
6. Evolution of KO: The traditional way of organizing knowledge is basically an island-like type for a single object of information. The development of the multi-oriented KO method allows multi-faceted classification; however, its function is still identifying and searching for knowledge. The triple RDF Linked Data concepts establish the innovative knowledge representation functions to weave events, people, time, space, characters, and facts together.
7. Developing all types of ontologies: To reconstruct knowledge and obtain in-depth understanding, case studies and modeling are required for developing more usable ontologies so that knowledge can be identified and translated from human to machine.
8. Transforming to Linked Data: Those who work in libraries are most concerned about the latest development of KO methods. The Library of Congress released 54 ontologies and Linked Data, such as subject headings, classification schemes, machine-readable catalogs, international language standards, metadata, ontologies, etc., to support and improve KO.
Learning from the individual cases
Chen (2016) and researchers at Academia Sinica Center for Digital Cultures’ Linked Data project employing a case study, used the works of modern painter, Chen Chengbo, to link collections, people, organizations, and events. They did this by using the relevant standard vocabulary, such as Schema.org, DBpedia, SKOS, GettyVocabularies, DCMI Metadata Terms, RDF Schema, etc. In the study, 30 categories and 57 attributes of the knowledge ontology of art history were identified. The significant contribution of this project is that a new KO can be represented by the “discoverability" through the link between the information and the knowledge ontology.
The Linked Jazz Oral History project Led by Pattuelli at Pratt Institute in New York is another
example of a successful project. Based on the Rutgers University Jazz Oral History Project, the Linked Jazz project transferred the traditional flat data to RDF triples by applying the Resource Description Framework. The triad describes 19,197 people and their relevant information, including musical instruments, dates of birth/death, locations, and occupations. It uses the RDF triples to describe 154 jazz events, including the names of performers and orchestra, dates, venues (such as the main hall), and the main titles of the concert programs.
The China Biographical Database Project (CBDB) project, which includes 36 million items of biographical information, is another case generated by the new KO system. The main function of CBDB is not to provide the full text of historical biography; rather, through the establishment of knowledge ontology and link information, its function is to provide indications of social circles, explore social relationships, and present travel routes. Then it can offer a re-understanding of the history and culture, that is, a historical portrayal of its multi-oriented life course, through kinship and non-kinship associations, social status (status, entry into Government offices / postings, places, writings), and so on.
This differs from traditional database searches which have only limited access points and superficial information query functions.
The fourth case is Shanghai Library’s Genealogy project which uses the Linked Open Data (LOD) technology to develop a database of genealogies, including genealogy records, ancient books, archives and periodicals, to improve and standardize the format of traditional digital collections. The Shanghai Library has provided the Open Link Information Service with four established basic knowledge bases, namely, historical figures / celebrity knowledge base, historical geographic knowledge base, historical chronology knowledge base, and historical event knowledge base. Among many, "Sheng Xuan-huai Archives Knowledge Base" and "Genealogical Knowledge Base" are two of the applications. The links provide information on the people and researchers related to the history or genealogy study. Combined with visualization technology, the retrieval of traditional literature databases with the possibility of innovation and knowledge can be presented.
Conclusions
Based on the observation on the development of information technology and the cases learned, three professional areas of knowledge and skills in KO are suggested for people who will be working in the field of KO in the future:
1. Developing domain knowledge with a broad understanding of the related information resources 2. Developing insights and in-depth analytical skills as well as good interpretative ability
3. Becoming familiar with machine languages, as well as platform and query languages to be able to operate Semantic Web KO
【Romanization of Chinese references is offered in the paper.】