Pino Buizza, Thesaurus and Heading Lists: Equivalences and Divegences in:

International Society for Knowledge Organziation (ISKO), Marianne Lykke, Tanja Svarre, Mette Skov, Daniel Martínez-Ávila (Ed.)

Knowledge Organization at the Interface, page 59 - 68

Proceedings of the Sixteenth International ISKO Conference, 2020 Aalborg, Denmark

1. Edition 2020, ISBN print: 978-3-95650-775-5, ISBN online: 978-3-95650-776-2,

Series: Advances in Knowledge Organization, vol. 17

Bibliographic information
Pino Buizza – Università degli studi di Firenze (Florence), Italy Thesaurus and Heading Lists Equivalences and Divegences Abstract The variety of indexing systems needs interoperability to satisfy global information. In mapping projects and related literature, the focus is predominantly on equivalence relationships between terms. Starting from examples of terms mapped between the Thesaurus of Nuovo Soggettario and LCSH and Rameau, the semantic relationships of terms are explored to verify correspondences and divergences in related terms. The different structure of the examined vocabularies leads to semantic networks that are not parallel. Specific and general remarks follow, in light of ISO 25964:2011-2013 and recent revisions. 1.0 Aim and background The great variety of indexing systems, together with the use of different languages, meets the specific needs of their patrons. However, today access to information on a global scale requires interoperability, a theme widely studied, and mapping is the way to reach data coming from different sources. Many mapping tests have been carried out and a rich literature is available on the matter.1 However, the focus is predominantly, if not exclusively, on equivalence relationships between concepts/terms. This attitude creates precise maps among single nuclei. However, it ignores the semantic relationships that are traditionally and diffusely represented in controlled vocabularies, hierarchical and associative relationships. Moreover, it loses sight of the overall correspondence of the envisaged systems. What happens if we explore their semantic networks? Do we find parallel or diverging nets? Do the correspondences go on step by step or stop at the starting points? This paper addresses this issue starting from mappings of the Thesaurus of Nuovo soggettario 2 (ThNS, source vocabulary) to Library of Congress Subject Headings3 (LCSH) and to Repertoire d’autorité-matière encyclopédique et alphabétique unifié4 (Rameau), as target vocabularies; all of them are characterized for general scope and produced by national bibliographic agencies. Some typical examples of mapping are reported, showing the equivalences and the most frequent divergences between the semantic networks and giving the starting point for justifying the divergences on the basis of the different features of the indexing languages or of their application criteria. The paper discusses some overall remarks on mappings and interoperability between different indexing languages and suggests possible alternative solutions drawn from ISO 25964:2011-2013, which is a sound reference to recognize and represent equivalences and their degree, to state meaningful and useful links where there is no equivalence, and to manage non-parallel systems. 1 To mention only a few papers: Hudon 1997, Doerr 2001, Riesthuis 2003, Zeng and Chan 2004, Jacobs, Mengel, and Müller 2010, Binding and Tudhope 2016, Balakrishnan, Voß, and Sorgel 2018, Kempf 2018, Zeng 2019. 2 3 4 60 Nuovo soggettario is an analytic-synthetic indexing language produced by the National Central Library in Florence and adopted by Italian libraries (Biblioteca nazionale centrale di Firenze 2006). Here, we are interested in the component Thesaurus, which is devoted to the control of vocabulary and semantic relationships between concepts. It is clearly distinct from the rules for concept analysis and strings construction, and from the authority file of assigned strings. In this separation of syntax from semantics, ThNS conforms to stipulations of ISO 25964-1:2011, and can function autonomously in postcoordinated indexing. The terms are not distinct in headings and subdivisions and there are no complex subjects. The terms included are only topic concepts, and terms that are valuable as genre/form are also included without distinction. Geographic terms, proper names and titles are not included. The structure of ThNS is not by discipline, but is founded on four macrocategories (Agents, Actions, Things, Time) divided into semantic categories, in number of thirteen (Organisms, Organizations, Persons and groups; Activities, Disciplines, Processes; Matter, Objects, Space, Tools, Forms; Structures; Time). Each of these categories is organized in facets and sub-facets, also adopting node labels. Within these categories, hierarchical relationships between concepts (BT and NT) and equivalence relationships between terms (USE, UF) are arranged without deviations. The connections between concepts belonging to different categories are recorded as associative relationships (RT). The use of polyhierarchy is limited to few controlled conditions. These features make ThNS different from other widespread indexing systems like LCSH and Rameau, but do not prevent from mapping correspondences useful for searching through different systems. ThNS mapping has reached almost 14000 equivalences with LCSH and 12000 with Rameau and it is still in progress. The equivalence mappings from ThNS are all recorded in skos as ‘closeMatch’, even when the equivalence could be an ‘exactMatch’. Compound equivalence, as foreseen in ISO 25964- 2:2013, 8.3, is not adopted: neither intersecting (EQ+) nor cumulative (EQ|). Mappings on broader or narrower levels (BM or NM) are not adopted, so each term admits only one mapping with one term of the same semantic category in the target vocabulary, and there are no double mappings from or to one term. The equivalences recorded in LCSH as ‘closely matching concepts’ with ThNS are the reciprocal of the mappings made in Florence, unfortunately incomplete, as they have not been added lately. In Rameau, Italian terms do not yet appear, but in they are now included among the resources ‘sur le Web: notice correspondante dans Le Nuovo Soggettario’ and also among the variants (autres formes du thème) with the qualification ‘italien’. Regardless of the following technical remarks, future reciprocal agreements are essential to the effectiveness of these works: for users’ searching functions and to manage the updating through the vocabularies. Mapping activity has shown cases of no correspondence (a concept represented in a vocabulary does not appear in the other) and of inexact correspondence (for instance, due to a different level of specificity). These mismatches can be due to language differences, different context and literary warrant, or different structure and application criteria of the considered indexing languages (for instance, the inclusion of complex subjects, requiring a double mapping from two terms of ThNS, or a different granularity). Some representative cases of difficult correspondence between single concepts have been previously reported (Buizza 2019). 61 2.0 Mapping concepts and exploring relationships This study focuses on six concepts represented by equivalent terms in the three vocabularies. Their hierarchical chains and related terms have been explored, in order to verify correspondences and differences in their semantic networks. The sample of examples tries to cover the most typical and interesting cases, without any presumption to be complete. No quantitive survey was carried out, to know the incidence of unsatisfactory results. The first example is a product globally widespread, the smartphone. ThNS records it as Smartphone (not translated), and its equivalents are Smartphones in LCSH and Smartphones in Rameau. Looking at the semantic relationships in the three vocabularies we can see immediately some differences, beginning from hierarchical ones (see Table 1). In ThNS, for the monohierarchical choice, we find only one BT Telefoni cellulari, even if the object is both a telephone and a computer. The term representing the second hierarchy, Palmari, is linked with an associative relationship, RT. In LCSH polyhierarchies are applied and we find two BTs: Cell phones and Pocket computers. In ThNS we have no NT due to the general choice of not recording proper names in the thesaurus, whether brand names or names of products. In LCSH some models from different brands are recorded as NT (for instance Samsung Galaxy S (Smartphone)). Therefore, only one equivalence can be set at the upper level (Telefoni cellulari EQ Cell phones) and no equivalence can be set at the lower level. Table 1. The case for Smartphone (simplified); double hierarchy is split. The superscripts mark recorded closeMatches. ThNS LCSH Rameau ··· BT Telefoni5 RT Radiotelefonia4 ·· BT Radiotelefoni · BT Telefoni cellulari2 Smartphone1 -------------------------- BT Computer portatili8 RT Palmari7 Smartphone1 ·· BT Telephone5 UF Telephone service, Telephones · BT Cell phones2,3 Smartphones1 · NT Samsung Galaxy S (Smartphone)6 -------------------------------------- ·· BT Portable computers8 · BT Poket computer7 Smartphones1 ·· BT Téléphone5, Radiotéléphonie4, Radiocommunications mobiles, Systèmes de communication sans fil · BT Téléphonie mobile3 EP Téléphones cellulaires Smartphones1 · NT Samsung Galaxy (Smartphones) ·· NT Samsung Galaxy S (Smartphone)6 ---------------------------------------- ·· BT Ordinateurs portatifs8 · BT Ordinateurs de poche7 Smartphones1 The comparison with Rameau gives similar results. In Rameau, two BTs are recorded, Téléphonie mobile and Ordinateurs de poche, and some NTs for specific brands, with relationships parallel to those of LCSH. However, there are two differences. The BT Téléphonie mobile has shifted the semantic category from objects to an activity – or rather, concepts of activity and of objects have been merged under one term, as shown by the French equivalents in Rameau (e.g. EP Téléphones cellulaires). Based on the category distinction, ThNS cannot set an equivalence with this term as it would cover only the component ‘object’. The parallelism is lost also with LCSH and Téléphonie mobile has correspondence with both Cell phones and Cell phone systems, with two distinct relationships. The second difference is how the two heading lists link the proper names of the smartphones to the noun term. LCSH links them directly as NT. Rameau 62 links them in two steps, through an intermediate term for the brand: Smartphones NT Samsung Galaxy (Smartphones) NT Samsung Galaxy S (Smartphone). The last term has as equivalent the homograph recorded in LCSH, while the intermediate term has no equivalent in LCSH. The ways hierarchical chains go to upper levels are now explored. In ThNS Telefoni cellulari has BT Radiotelefoni, which has BT Telefoni. In LCSH Cell phones has BT Telephone (in the singular, with variants: Telephone service and Telephones, with double category value) and also the complex subject Radio—Transmitters-receivers. ThNS sets equivalence between Telefoni and Telephone, even though the Italian word does not include the telephone service. No equivalence for Radiotelefoni, but the RT Radiotelefonia has equivalent Radiotelephone, that includes service and devices (its NT Walkie-talkies has a homograph in ThNS recorded as NT of Radiotelefoni). In Rameau, Téléphonie mobile has four BTs: Téléphone, Radiotéléphonie, Radiocommunications mobiles and Systèmes de communication sans fil, but the last term has the second and third as NT, besides Téléphonie mobile, and the second term is also NT of the first; an entry to be updated, probably – there are some in every vocabulary. The equivalences in ThNS go from Telefoni and Radiotelefonia to Téléphone and Radiotél- éphonie – equally from the English terms to the French. The correspondences between the three vocabularies are restored, but the paths from smartphones to this point are different and the exact correspondence between smartphones is not kept between the reference to one category in ThNS and the meaning merging object and functions in LCSH and Rameau. According to the choice of monohierarchy, ThNS establishes an associative relationship to the second hierarchy: Smartphone RT Palmari. The equivalents of Palmari are Pocket computers and Ordinateurs de poche. Their respective BTs, Computer portatili, Portable computers, and Ordinateurs portatifs, are equivalent. At upper levels we find the equivalence of the three languages. But in LCSH and Rameau coming back to smartphones is inside one hierarchical chain by NT relationships, while in ThNS it is necessary to pass through the associative relationship Palmari RT Smartphone. The complexity and the divergencies of these semantic networks are due, at least in part, to the double function of the considered object. Another example is the simple concept of ‘date’, the fruit, seen from different points of view. In ThNS the term Datteri has BT Frutta, whose BT is Alimenti, and RT Frutti, a term that has BT the node label [Organi e parti di piante], that has BT Piante. According to the meaning of the Italian words, Frutti has a scope note assigning it the botanic meaning, while for the works about fruits as food Frutta is used. In LCSH, the equivalent Dates (Fruit) has two BTs: Date palm products (with hierarchy: BT Palm products BT Plant products) for the economic side, and BT Fruit, for food, agricultural and botanic meanings, as is shown by its five BTs: Food, Food crops, Horticultural crops, Horticultural products, and Plants. In Rameau, the equivalent Dattes has BT Fruit, whose scope note states both botanic and food meanings. These are shown together with the economic meaning, in its three BTs: Diaspores (botanique), Plantes comestibles, and Produits horticoles. In LCSH and in Rameau, we find also Cooking (Dates) and Cuisine (dattes), with the activity of cooking qualified by the specific food, a method collecting a long sequence of terms for each type of food under the same first word. These complex terms cannot have correspondence in Nuovo soggettario, where strings are provided syntactically out of the thesaurus. 63 In this case the given concept is represented by the term for the food and a form term meaning the aim of the document: Datteri-Ricette. This example confirms that: adopting polyhierarchy or not is an important factor; scope notes may be crucial for stating an equivalence; the presence of complex subjects makes a difference between the vocabularies that is difficult to manage. Even if there are not different meanings, nor different contexts for a concept, the equivalence of the terms for a given concept does not grant the persistence of the equivalence in the other steps of the semantic networks. Looking at concepts representing a set of individuals, for instance deities, and considering a particular historical expression, say Greek deities, we find the equivalent terms Divinità greche, Gods, Greek, and Dieux grecs. In ThNS the subordinate terms are two subsets, Muse and Ninfe, while in LCSH we find some subsets of different kind (e.g. Gods, Minoan, referring to specific areas) and a number of proper names (e.g. Chaos (Greek deity)) but not the well-known gods of Olympus (recorded in Library of Congress Name Authority File). In Rameau we find some subsets (e.g. Dieux minoens), the female collective term (Deesses greques) and an exhaustive list of proper names of deities (e.g. Aphrodite (divinité grecque)). The choice of Nuovo soggettario to manage the proper names of individuals through the guidelines of the Manuale, causes this lonely position of ThNS. However, the other vocabularies do not agree with each other either: the distribution of subordinate sets is not equal and LCSH places some gods among names of persons. There are misalignments in superordinate relationships too. The Italian term has BT Divinità, the English one lacks BT, the French one has two study areas as BT Mythologie grecque and Religion grecque. The equivalent terms for Divinità are Gods and Dieux. Divinità has NTs for deities of various peoples and Divinità femminili, Divinità marine, Divinità salutari, etc. Gods has NTs for deities of nature (e.g. Water gods) and of specific religions (e.g, Hindu gods, with NT for some proper names), except for classical religions, that are orphan, and has BT Mythology, Classical. Dieux has NTs for the deities of phenomena (e.g. Dieux des vents), that can have NTs for proper names, like Vayu (divinité hindoue), while the terms qualified by people or religion have BT for the appropriate mythology and/or religion (e.g. Dieux hindous BT Hindouisme, Mythologie hindoue). Other differences in semantic relationships can be found in the category of persons. Donatori di sangue has equivalents Blood donors and Donneurs de sang (see Table 2). The term in ThNS has BT the node label [Persone secondo il comportamento], that has BT Persone. In LCSH the immediate BT is Persons. This is not a small difference: Persone has three node labels (according to activity and to conditions, in addition to behaviour) and each of them has NTs for other node labels in a widely faceted articulation. Under Persons a long list of terms represents specific types of persons (e.g, Saints, Slaves, Travelers), arranged alphabetically. In a different way, in Rameau Donneurs de sang has BT Donneurs d’organes, that has no BT, the same as the equivalent terms of the above examples: Saints, Esclaves, Voyageurs. As a result, the terms for categories of persons are fragmented and neither systematized in a pyramid as in ThNS, nor collected under a comprehensive top term as in LCSH. There are some groupings, for instance under Catégories socio-professionnelles (with hierarchies like: NT Commerçants NT 64 Libraires), and also terms with a BT. But some BTs might have moved to another semantic category, for instance, Personnes remariées has BT Remariage. Table 2. The case for Blood donors and Persons. The superscripts mark recorded closeMatches. ThNS LCSH Rameau Persone2 · NT [Secondo l’attività] · NT [Secondo la condizione] ·· NT [Secondo la condizione sociale] ··· NT Schiavi3 · NT [Secondo il comportamento] ·· NT Donatori di organi6 ·· NT Donatori di sangue1 ·· NT Viaggiatori4 ·· NT [Secondo la fede e le convinzioni religiose] ··· NT Santi5 Persons2 · NT Slaves3 · NT Organ donors6 · NT Blood donors1 · NT Travelers4 · NT Saints5 -- Esclaves3 Donneurs d’organes6 · NT Donneurs de sang1 Voyageurs4 Saints5 A term for a discipline, Dermatologia, has equivalents Dermatology and Dermatologie (see Table 3). In ThNS the broader term Medicina is reached through the interposed node label [Medicina applicata a specifici organi, apparati, sistemi, funzioni]. This node label is paired on the same array by two other labels: [Medicina applicata a categorie di persone] and [Medicina applicata a specifiche attività], which are all recorded as NTs under Medicina specialistica. The result is a faceted distribution of the branches of medicine. In LCSH and Rameau there is a direct relationship to the general discipline, BT Medicine and BT Médicine respectively. In the subordinate hierarchy, in ThNS there are no NTs and the term Dermatologia veterinaria is recorded as associated (RT). In a different interpretation, its equivalents Veterinary dermatology and Dermatologie vétérinaire are recorded as NT. Among other NTs for specialities there are also terms that are neither members nor parts of the superordinate concept, such as, in LCSH, agents: Dermatologists, and a technique: Radioisotopes in dermatology, or, in Rameau, a védette construite for an activity: Peau-Maladies-Soins infirmiers. Table 3. The case for Dermatology. The superscripts mark recorded closeMatches. ThNS LCSH Rameau ··· BT Medicina2 ·· BT Medicina specialistica · BT [Medicina applicata a organi…] Dermatologia1 RT Dermatologia veterinaria3 · BT Medicine2 Dermatology1 · NT Veterinary dermatology3 · NT Pediatric dermatology4 · NT Dermatologists · NT Radioisotopes in dermatology · BT Mèdicine2 Dermatologie1 · NT Dermatologie vétérinaire3 · NT Dermatologie pédiatrique4 · NT Peau-Maladies-Soins infirmiers Another disciplinary term, Diritto internazionale, and its equivalents International law and Droit international, present the same kind of difference at superordinate levels: direct relationships to Law and Droit, while in ThNS a node label is interposed: BT [Diritto secondo la materia] BT Diritto. The subordinate Italian hierarchy has specific branches (e.g. Diritto comunitario, Diritto internazionale marittimo), and associative relationships link objects, tools and activities of the discipline. In LCSH and Rameau 65 there are many NTs, including any kind of concept that can be considered in the semantic area of the discipline, including complex or syntactically constructed terms (e.g. Islands- Law and legislation or Iles-Droit) and parenthetic terms (e.g. Missing persons (International law) and Personnes disparues (droit international)). The sets of the two languages do not fully overlap, some terms have no equivalent in the other language (e.g. Women (International law) or Animaux (droit international)). In ThNS a parenthetic term is allowed only for disambiguation of homographs (normally not by the discipline). Concepts expressed with parenthetic terms in LCSH and Rameau are seen in ThNS as syntactical relationships between distinct concepts, to be treated according to the rules for string construction (if they are not pleonastic). The term Intervento (Diritto internazionale), with the same meaning of Intervention (International law) and of Intervention (droit international), is a non-preferred term to be expressed by a combination of terms: Intervento (Diritto internazionale) USE+ Diritto internazionale, Intervento militare. No equivalence is possible from the Italian terms to those of LCSH and Rameau. According to ISO 25964-2:2013, 8.3.2, an intersecting compound equivalence could be provided in the other direction: Intervention (International law) EQ Intervento militare + Diritto internazionale, and Intervention (droit international) EQ Intervento militare + Diritto internazionale. Obviously, this method is not adopted, at present. In ThNS Discipline is a semantic category and a top term. Its hierarchy comprises only concepts of discipline. Other concepts strictly related to a discipline (not simply falling into a discipline) are linked with an associative relationship, and remain in their semantic category (agents, activities, tools, etc.). Thus, in ThNS we do not find the hierarchies typical of classifications based on disciplines, where a discipline is the systematic container of everything it concerns. The free structure of LCSH and Rameau is quite different: categories are not considered and the relationships with narrower terms may be similar to an alphabetic index collecting all the topics belonging or attributed to that discipline, including complex and parenthetical terms. 3.0 Overall remarks There are several differences between the semantic networks of the three vocabularies, even though we can often start from equivalent terms for the same concept and we find other full equivalences along the paths. The differences occur mainly between ThNS and the two heading lists, due to the different structure of ThNS explained above. To summarize, some of the reasons are: firstly the lack of subdivisions, complex subjects, and proper names. Secondly, an architecture based on categories without disciplinary groupings, strictly categorial equivalences and hierarchies, faceting, and limited polyhierarchy, compared with more free relationships aiming to link what is inside a discourse rather than the accurate identities of concepts. Téléphonie mobile is not the same entity as Téléphones cellulaires, but when talking about the former one talks also about the latter. There are important differences even between the strongly connected LCSH and Rameau, due to different granularity or definition, including jumping hierarchical levels, merging concepts under one term or distinguishing them under different terms, different criteria of subdivision, non-identical choices in recording complex subjects. This means 66 that some equivalences are missing or couples of terms referred to the same concepts need different paths to be connected. The great extension of the vocabularies allows for the presence of most concepts in the three languages and their mapping, so the divergences of the networks may appear not very important: surfing among terms seems always possible. What are the drawbacks of these divergences, in regard to the primary value of the equivalence between terms representing the same concept? Terms have no magic power. They can neither fully represent the content and informative potential of the works entered under them, nor fully and exactly retrieve the searched information. They live as strictly connected elements of indexing and of natural language, controlled in indexing and free in users’ access. Terms literally say what their definitions or scope notes say, but they represent wider themes and thoughts, multiple and articulated connections to other concepts and themes. Semantic relationships just try to represent these articulated meanings in systematic, understandable and viable order, guiding indexers in choosing terms and users in searching and exploring resources. Different configurations of the semantic frames are quite legitimate and sometimes required. However, for interoperability and combined use, the largest overlapping of the systems, or, at least, knowing the different paths and the missing links is desirable. Going up and down hierarchies consistently makes the choice of broadening or narrowing the field of research easier, faster and more reliable, and avoids possible confusions. The same usefulness works in order to see the sibling terms, passing through the broader term. In indexing work, only the adopted system is involved and the problems of semantic networks are less relevant than the deficiencies in mapping. Some observations about the mentioned cases reduce their seriousness. The lack of proper names in ThNS simply means that a name is assigned following the Manuale and is recorded and searchable among the strings, not in the thesaurus. Complex subjects are ready to use, while in ThNS the strings are combined when needed, according to syntactic rules, or drawn from strings already recorded. The equivalence is restored at this level. Missing terms for a concept, probably due to lack of literary warrant, could be added and validated (e.g. Divinità minoiche in ThNS, or Samsung Galaxy smartphones in LCSH), unless an equivalence relationship with a term for a slightly different concept is preferred in the vocabulary policies. Only when two concepts are represented by the same term we can find different solutions in indexing. Obviously other differences in indexing results come from different policies and syntax (e.g. summarization vs depth indexing, or one coextensive string for one work vs more complementary strings). The results of searching through diverging semantic networks vary greatly depending upon the way indexes are presented and searches are carried out: through the vocabulary or directly on strings, by surfing the vocabulary or the strings or the records, with or without automatic extension to target vocabularies. ‘Exploding’ a search, that is, adopting a technique able to retrieve also the resources linked to the terms subordinate to the searched term, gives different results in systems with or without polyhierarchies, and adopting hierarchies with or without extra semantic categories. The same happens ‘expanding’ a search, that is, retrieving the resources linked to both subordinate and associated terms (ISO 25964-1:2011, 10.2.1). In general terms, simple searches are not problematic. When surfing, users must follow the different features of each indexing system. 67 Automatic exploring gives variously inaccurate results (with topics and resources that are not pertinent), insufficient (some relevant resources are not found) or redundant (relevant resources are retrieved together with off-topic ones, good for serendipitous diversions), depending on finding equivalent, lacking or overabundant relationships. Nowadays the connections between the equivalent terms of the different vocabularies allow users to search in the catalogues of the three libraries in few steps: any improvement of interoperability will be beneficial to users from the three areas. Structural differences and soundness of the systems seem unfavourable to future convergences. However, some suggestions are an obvious consequence of this survey. Adding other forms and levels of mapping is recommendable to clarify and enrich connections, even though it acts almost only at the level of single correspondences, without weakening divergencies between the semantic networks. Any convergences towards international standards, especially ISO 25964, would lead to positive effects, even though subject headings lists are too far from thesauri in their structure. At present, some elements offer suitable conditions to interoperability. Linked open data already support the three systems and work on data of the entities in a way that prefers modular systems, instead of systems fixing pre-coordinated entities in their vocabularies. Systems where each element is treated per se, in order to be combined with other elements, either steadily according to semantic paradigms or occasionally according to syntactic needs, are more suitable for linked data. In this perspective, ‘Réformer Rameau’ is a very encouraging venture launched in France5. In five years, among other things, it foresees to abolish the distinction between tête de vedette and subdivision, to split up the vedettes construites, and to delete special instructions for disciplines. This treatment of terms is similar to thesaural treatment. It is neutral as regards the destination to pre-coordination, and reduces the disciplinary bonds, that functionally collect any object of interest in a given field of research, but constrain concepts into limits more restricted than those where concepts normally move. As for ThNS, the second edition of the Guida to Nuovo soggettario, now in the final phase, confirms the adherence to standard ISO 25964. It does not change the features of the thesaurus and its criteria for mapping other vocabularies, even after a complete rewriting. We cannot exclude a future opening to varied mappings, which would be particularly helpful if accompanied by collaboration and agreed criteria between national libraries. What matters here, in this description of coherence, limits and problems of network mapping, is to point out that good mappings between single terms are not sufficient. This paper recommends, particularly to people involved in designing interoperability, going beyond mappings between single terms, towards clearly branched and connected maps, which would be useful to cross unknown or less well-known areas of knowledge. References Balakrishnan, Uma, Jakob Voß, and Dagobert Soergel. 2018. “Towards Integrated Systems for KOS Management, Mapping, and Access: Coli-Conc and its Collaborative Computer-Assisted KOS Mapping Tool Cocoda.” In: Challenges and Opportunities for Knowledge Organization in the Digital Age: Proceedings of the Fifteenth International ISKO Conference 9-11 5 68 July 2018 Porto, Portugal, edited by Fernanda Ribeiro and Maria Elisa Cerveira. Advances in knowledge organization 16. Baden-Baden: Ergon, 693-701. Biblioteca nazionale centrale di Firenze. 2006. Nuovo soggettario. Guida al sistema italiano di indicizzazione per soggetto. Prototipo del Thesaurus. Milano: Editrice bibliografica. Binding, Ceri, and Douglas Tudhope. 2016. “Improving Interoperability Using Vocabulary Linked Data.” Journal of Digital Information 17: 5-21. Buizza, Pino. 2019. “Indicare quasi la stessa cosa. Appunti di indicizzazione interlinguistica.” In Viaggi a bordo di una parola. Scritti di indicizzazione semantica in onore di Alberto Cheti, a cura di Anna Lucarelli, Alberto Petrucciani, Elisabetta Viti. Roma: Associazione italiana biblioteche, 33-49. Doerr, Martin. 2001. “Semantic Problems of Thesaurus Mapping.” Journal of Digital Information 1, no. 8. Hudon, Michèle. 1997. “Multilingual Thesaurus Construction. Integrating the Views of Different Cultures in One Gateway to Knowledge and Concepts.” Knowledge Organization 24: 84-91. ISO 25964-1-2:2011-2013. Information and Documentation. Thesauri and Interoperability with Other Vocabularies. Geneva: ISO. Jacobs, Jan-Helge, Tina Mengel, and Katrin Müller. 2010. “Benefits of the Crisscross Project for Conceptual Interoperability and Retrieval.” In Paradigms and Conceptual Systems in Knowledge Organization: Proceedings of the Eleventh International ISKO Conference 23-26 February 2010, Rome, Italy, edited by Claudio Gnoli and Fulvio Mazzocchi. Advances in knowledge organization 12. Würzburg: Ergon Verlag, 236-241. Kempf, Andreas Oskar. 2018. “The Need to Interoperate: Structural Comparison of and Methodological Guidance on Mapping Discipline-Specific Subject Authority Data to Wikidata.” In Challenges and Opportunities for Knowledge Organization in the Digital Age: Proceedings of the Fifteenth International ISKO Conference 9-11 July 2018 Porto, Portugal, edited by Fernanda Ribeiro and Maria Elisa Cerveira. Advances in knowledge organization 16. Baden- Baden: Ergon, 644-652. Riesthuis, Gerhard J.A. 2003. “Information Languages and Multilingual Subject Access.” In Subject Retrieval in a Networked Environment, edited by I.C. McIlwaine. München: De Gruyter Saur, 11-18. Zeng, Marcia Lei. 2019. “Interoperability.” Knowledge Organization 46: 122-146.. Zeng, Marcia Lei and Lois Mai Chan. 2004. “Trends and Issues in Establishing Interoperability Among Knowledge Organization Systems.” Journal of the American Society for Information Science and Technology 55, no.5: 377-395.

Chapter Preview



The proceedings explore knowledge organization systems and their role in knowledge organization, knowledge sharing, and information searching.

The papers cover a wide range of topics related to knowledge transfer, representation, concepts and conceptualization, social tagging, domain analysis, music classification, fiction genres, museum organization. The papers discuss theoretical issues related to knowledge organization and the design, development and implementation of knowledge organizing systems as well as practical considerations and solutions in the application of knowledge organization theory. Covered is a range of knowledge organization systems from classification systems, thesauri, metadata schemas to ontologies and taxonomies.


Der Tagungsband untersucht Wissensorganisationssysteme und ihre Rolle bei der Wissensorganisation, dem Wissensaustausch und der Informationssuche. Die Beiträge decken ein breites Spektrum von Themen ab, die mit Wissenstransfer, Repräsentation, Konzeptualisierung, Social Tagging, Domänenanalyse, Musikklassifizierung, Fiktionsgenres und Museumsorganisation zu tun haben. In den Beiträgen werden theoretische Fragen der Wissensorganisation und des Designs, der Entwicklung und Implementierung von Systemen zur Wissensorganisation sowie praktische Überlegungen und Lösungen bei der Anwendung der Theorie der Wissensorganisation diskutiert. Es wird eine Reihe von Wissensorganisationssystemen behandelt, von Klassifikationssystemen, Thesauri, Metadatenschemata bis hin zu Ontologien und Taxonomien.