guidelines for the construction, format, and …...ansi/niso z39.19 - 1993 issn: 1041-5653 (revision...

ANSI/NISO Z39.19 - 1993 ISSN: 1041-5653(Revision of Z39.19-1980)

Guidelines for the Construction,Format, and Management of

Monolingual Thesauri

Abstract: A thesaurus is a controlled vocabulary arranged in a knownorder and structured so that equivalence, homographic, hierarchical, andassociative relationships among terms are displayed clearly and identifiedby standardized relationship indicators that are employed reciprocally.The primary purposes of a thesaurus are (a) to facilitate retrieval ofdocuments and (b) to achieve consistency in the indexing of written orotherwise recorded documents and other items, mainly for postcoordinateinformation storage and retrieval systems. This standard provides guide-lines for constructing monolingual thesauri: formulating the descriptors,establishing relationships among terms, and effectively presenting theinformation in print and on a screen. It also includes thesaurus mainte-nance procedures and recommended features of thesaurus managementsystems.

An American National StandardDeveloped by the

National Information Standards OrganizationApproved August 18, 1993 by the

American National Standards Institute

Bethesda, Maryland, U.S.A.

Published byNISO Press4733 Bethesda AvenueSuite 300Bethesda, Maryland 20814

Copyright ©1994 by National Information Standards OrganizationAll rights reserved under International and Pan-American Copyright Conventions. No part of thisbook may be reproduced or transmitted in any form or by any means, electronic or mechanical,including photocopy, recording, or any information storage or retrieval system, without priorpermission in writing from the publisher. All inquiries should be addressed to NISO Press, 4733Bethesda Avenue, Suite 300, Bethesda, Maryland 20814.

Printed in the United States of America

ISSN: 1041-5653 National Information Standards seriesISBN: 1-880124-04-1

This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).

Library of Congress Cataloging-in-Publication Data

National Information Standards Organization (U.S.) Guidelines for the construction, format, and management of monolingualthesauri / developed by the National Information Standards Organization : ap-proved August 30, 1993, by the American National Standards Institute.

p. cm.—(National information standards series, ISSN 1041-5653;ANSI/NISO Z39.19-1993) Rev. ed. of: American national standard guidelines for thesaurus structure,construction, and use. New York: ANSI, c1980. Includes bibliographical references (p. ) and index. ISBN 1-880124-04-1 : $49.00 1. Subject headings—Standards—United States. I. American National Stan-dards Institute. II. American National Standards Institute. Subcommittee 25 onThesaurus Rules and Conventions. American national standard guidelines forthesaurus structure, construction, and use. III. Title. IV. Series.Z695.N36 1994 93-42543025.4'9—dc20 CIP

Contents

PageForeward ...................................................................................................................................................................viiiAbbreviations, Codes, and Conventions Used in the Standard ........................................................................ xii1. Introduction ............................................................................................................................................................ 1

1.1 Purpose and Structure.................................................................................................................................... 11.1.1 Purpose ................................................................................................................................................ 11.1.2 Structure .............................................................................................................................................. 1

1.2 Vocabulary Control ........................................................................................................................................ 1

2. Scope of the Standard ........................................................................................................................................... 12.1 Applications of the Standard ........................................................................................................................ 22.2 Monolingual Thesauri .................................................................................................................................... 22.3 Types of Terms ................................................................................................................................................ 2

3. Scope, Form, and Choice of Descriptors ............................................................................................................ 23.1 Single-word vs. Multiword Descriptors ...................................................................................................... 23.2 Scope of Descriptors ....................................................................................................................................... 2

3.2.1 Parenthetical Qualifiers ..................................................................................................................... 23.2.2 Scope Notes ......................................................................................................................................... 3

3.2.2.1 Reciprocal Scope Notes ............................................................................................................. 33.3 Types of Concept ............................................................................................................................................ 3

3.3.1 Unique Entities ................................................................................................................................... 43.4 Grammatical Forms of Descriptors .............................................................................................................. 4

3.4.1 Nouns and Noun Phrases ................................................................................................................. 43.4.1.1 Verbal Nouns .............................................................................................................................. 43.4.1.2 Noun Phrases .............................................................................................................................. 4

3.4.1.2.1 Adjectival Noun Phrases ................................................................................................... 43.4.1.2.2 Prepositional Noun Phrases .............................................................................................. 5

3.4.2 Adjectives ............................................................................................................................................ 53.4.2.1 Limiting the Number of Compound Descriptors .................................................................. 53.4.2.2 Economy of Cross-References .................................................................................................. 5

3.4.3 Adverbs ............................................................................................................................................... 53.4.4 Initial Articles ...................................................................................................................................... 5

3.4.4.1 Deletion ........................................................................................................................................ 53.4.4.2 Retention ...................................................................................................................................... 5

3.5 Singular and Plural Forms ............................................................................................................................. 63.5.1 Count Nouns ....................................................................................................................................... 6

3.5.1.1 Exception to Plural Count Nouns ............................................................................................ 63.5.2 Noncount (Mass) Nouns ................................................................................................................... 6

3.5.2.1 Abstract Concepts ...................................................................................................................... 63.5.2.2 Unique Entities ........................................................................................................................... 6

3.5.3 Coexistence of Singular and Plural Forms ..................................................................................... 63.6 Selection of Preferred Form ........................................................................................................................... 7

3.6.1 Usage .................................................................................................................................................... 73.6.1.1 Literary Warrant ......................................................................................................................... 73.6.1.2 Published Authorities ................................................................................................................ 73.6.1.3 Other Indexing Languages ........................................................................................................ 73.6.1.4 Personal Authorities .................................................................................................................. 7

3.6.2 Spelling ................................................................................................................................................ 73.6.2.1 Literary Warrant ......................................................................................................................... 7

(continued)

Contents

Page iv

(3. Scope, Form, and Choice of Descriptors continued)3.6.2.2 Authorities ................................................................................................................................... 7

3.6.3 Abbreviations, Initialisms, and Acronyms ..................................................................................... 73.6.3.1 Preference for Abbreviation ...................................................................................................... 73.6.3.2 Preference for Full Form ............................................................................................................ 7

3.6.3.2.1 General Use ......................................................................................................................... 73.6.3.2.2 Ambiguity ............................................................................................................................ 7

3.6.4 Neologisms, Slang, and Jargon ........................................................................................................ 83.6.4.1 Provisional Terms....................................................................................................................... 83.6.4.2 Cross-References from Slang Terms ........................................................................................ 8

3.6.5 Common Names and Trade Names ................................................................................................ 83.6.5.1 Preference for Common Name ................................................................................................. 83.6.5.2 Trademarks ................................................................................................................................. 83.6.5.3 Loss of Protected Status ............................................................................................................. 8

3.6.6 Popular and Scientific Names .......................................................................................................... 83.6.7 Loanwords, Translations of Loanwords, and Foreign-Language Equivalents ......................... 8

3.6.7.1 Loanwords ................................................................................................................................... 83.6.7.2 Translations of Loanwords ....................................................................................................... 83.6.7.3 Foreign-Language Equivalents ................................................................................................ 9

3.6.8 Proper Names ..................................................................................................................................... 93.6.8.1 Place Names ................................................................................................................................ 9

3.7 Capitalization, Punctuation, and Nonalphabetic Characters ................................................................... 93.7.1 Capitalization ...................................................................................................................................... 93.7.2 Nonalphabetic Characters ............................................................................................................... 10

3.7.2.1 Parentheses ................................................................................................................................ 103.7.2.2 Hyphens ..................................................................................................................................... 103.7.2.3 Apostrophes .............................................................................................................................. 10

3.7.2.3.1 Possessive Case ................................................................................................................. 103.7.2.3.2 Proper Names .................................................................................................................... 10

3.7.2.4 Diacritical Marks ...................................................................................................................... 103.7.2.5 Other Symbols and Punctuation Marks ................................................................................ 10

3.7.3 Romanization .................................................................................................................................... 10

4. Compound Terms ................................................................................................................................................ 104.1 General ........................................................................................................................................................... 10

4.1.1 Purpose of Guidelines on Compound Terms .............................................................................. 114.1.2 Precoordination and Postcoordination ......................................................................................... 114.1.3 Factors to Be Considered ................................................................................................................. 114.1.4 Elements of Compound Terms ...................................................................................................... 11

4.2 Criteria for Establishing Compound Terms as Descriptors ................................................................... 124.3 Criteria for Determining when Compound Terms Should Be Split ...................................................... 12

4.3.1 Factors to Be Considered ................................................................................................................. 124.3.2 Hierarchical Structure...................................................................................................................... 12

4.3.2.1 Focus and Difference ............................................................................................................... 124.3.2.2 Transitive Action ...................................................................................................................... 124.3.2.3 Intransitive Action .................................................................................................................... 13

4.4 Compound Descriptors vs. Node Labels .................................................................................................. 134.5 Order of Words in Compound Terms ....................................................................................................... 13

4.5.1 Cross-References from Inversions ................................................................................................. 13

5. Relationships ........................................................................................................................................................ 135.1 General ........................................................................................................................................................... 13

5.1.1 Semantic Linking .............................................................................................................................. 13(continued)

Contents

Page v

(5. Relationships continued)5.1.2 Kinds of Relationship ...................................................................................................................... 13

5.2 The Equivalence Relationship ..................................................................................................................... 155.2.1 General ............................................................................................................................................... 155.2.2 Synonyms .......................................................................................................................................... 155.2.3 Lexical Variants ................................................................................................................................ 165.2.4 Quasi-Synonyms .............................................................................................................................. 16

5.2.4.1 Generic Posting ......................................................................................................................... 165.2.5 Cross-References to Elements of Compound Terms................................................................... 16

5.3 The Hierarchial Relationship ........................................................................................................................... 165.3.1 The Generic Relationship ................................................................................................................ 17

5.3.1.1 Codes for the Generic Relationship ....................................................................................... 175.3.2 The Whole-Part Relationship ......................................................................................................... 17

5.3.2.1 Codes for the Whole-Part Relationship ................................................................................. 185.3.2.2 Parts of Multiple Wholes ......................................................................................................... 18

5.3.3 The Instance Relationship ............................................................................................................... 185.3.3.1 Codes for the Instance Relationship ...................................................................................... 18

5.3.4 Polyhierarchical Relationships ....................................................................................................... 185.3.5 Node Labels in Hierarchies ............................................................................................................ 18

5.4 The Associative Relationship ...................................................................................................................... 195.4.1 Relationships Between Descriptors Belonging to the Same Hierarchy .................................... 19

5.4.1.1 Relationships Between Overlapping Sibling Terms ............................................................ 195.4.1.2 Relationships Between Mutually Exclusive Sibling Terms ................................................ 195.4.1.3 Derivational Relationships ...................................................................................................... 19

5.4.2 Relationships Between Descriptors Belonging to Different Hierarchies ................................. 205.4.3 Node Labels for Related Terms ...................................................................................................... 205.4.4 Specifying Types of Related Term References ............................................................................. 20

6. Print Display ......................................................................................................................................................... 216.1 The Equivalence Relationship ..................................................................................................................... 21

6.1.1 Inverted USE References and Permuted Lists ............................................................................. 216.1.2 Inverted USE References and Hierarchical Display .................................................................... 216.1.3 Juxtaposition of Terms .................................................................................................................... 21

6.2 Hierarchical Relationships .......................................................................................................................... 226.2.1 Conventional Flat Thesaurus Structure ........................................................................................ 226.2.2 Multilevel Hierarchy Within Alphabetic Display ....................................................................... 22

6.2.2.1 Multilevel Broader and Narrower Terms ............................................................................. 226.2.2.2 Generic Structure ...................................................................................................................... 22

6.2.3 Tree Structure ................................................................................................................................... 226.2.3.1 Notation ..................................................................................................................................... 23

6.2.4 Graphic Displays .............................................................................................................................. 236.2.5 The Top Term Structure .................................................................................................................. 236.2.6 Two-Way Hierarchical Structure ................................................................................................... 236.2.7 Broad Categories .............................................................................................................................. 23

6.3 General Format Considerations .................................................................................................................. 236.3.1 Components of a Printed Thesaurus ............................................................................................. 236.3.2 Introduction ...................................................................................................................................... 236.3.3 Typography ....................................................................................................................................... 246.3.4 Capitals/Lowercase Letters ............................................................................................................ 246.3.5 Filing .................................................................................................................................................. 24

6.3.5.1 Alphabetic Characters ............................................................................................................. 246.3.5.2 Numerals ................................................................................................................................... 246.3.5.3 Non-Alphanumeric Characters .............................................................................................. 24

(continued)

Contents

Page vi

(6. Print Display continued)6.3.6 Running Heads ................................................................................................................................. 25

6.4 Multiple Versions of Printed Thesauri ...................................................................................................... 25

7. Screen Display ...................................................................................................................................................... 257.1 User Categories ............................................................................................................................................. 257.2 Factors to Be Considered ............................................................................................................................. 257.3 Types of Display ........................................................................................................................................... 26

7.3.1 Alphabetical Listing ......................................................................................................................... 267.3.2 Permuted Displays ........................................................................................................................... 267.3.3 Hierarchies and Classifications ...................................................................................................... 267.3.4 Term Detail ........................................................................................................................................ 267.3.5 Graphic Displays .............................................................................................................................. 26

7.4 Documentation .............................................................................................................................................. 267.5 Relation of Thesaurus Character Set to Search Commands ................................................................... 26

8. Thesaurus Construction...................................................................................................................................... 278.1 Avoidance of Duplicate Work .................................................................................................................... 278.2 Determination of Structure and Display Format ..................................................................................... 278.3 Methods .......................................................................................................................................................... 27

8.3.1 The Committee Approach............................................................................................................... 278.3.2 The Empirical Approach ................................................................................................................. 278.3.3 Combination of Methods ................................................................................................................ 27

8.4 Machine Assistance ...................................................................................................................................... 278.5 Term Records ................................................................................................................................................. 288.6 Term Verification .......................................................................................................................................... 28

8.6.1 Candidate Terms .............................................................................................................................. 288.6.2 Provisional Terms ............................................................................................................................ 28

8.7 Level of Specificity ........................................................................................................................................ 288.8 Unassigned Descriptors ............................................................................................................................... 298.9 Announcement and Deposit of Published Thesauri ............................................................................... 29

8.9.1 Notification of Intent ....................................................................................................................... 298.9.2 Pilot Thesaurus ................................................................................................................................. 298.9.3 Deposit with a Clearinghouse ........................................................................................................ 29

9. Thesaurus Maintenance ...................................................................................................................................... 299.1 Updating the Vocabulary ............................................................................................................................ 29

9.1.1 Addition of Descriptors ................................................................................................................... 299.1.2 Modification of Existing Descriptors ............................................................................................. 299.1.3 Deletion of Terms ............................................................................................................................. 30

9.2 Vocabulary Updates and Database Records ............................................................................................. 309.3 History of Changes ....................................................................................................................................... 30

10. Thesaurus Management Systems .................................................................................................................... 3110.1 Typography and Sorting ............................................................................................................................ 3110.2 Term Records and Display Format .......................................................................................................... 3110.3 System Constraints ..................................................................................................................................... 3110.4 Searching ...................................................................................................................................................... 3110.5 Editing .......................................................................................................................................................... 3110.6 Error Checking ............................................................................................................................................ 3110.7 Cross-References ......................................................................................................................................... 3210.8 Term Deletion .............................................................................................................................................. 3210.9 Candidate Terms ......................................................................................................................................... 32

(continued)

Contents

Page vii

(10. Thesaurus Management Systems continued)10.10 Field Definition ......................................................................................................................................... 3210.11 Reports ........................................................................................................................................................ 3310.12 Subsets ........................................................................................................................................................ 33

10.12.1 Copying Function ......................................................................................................................... 3310.12.2 Microthesauri ................................................................................................................................ 33

References ................................................................................................................................................................. 34

Glossary of Terms .................................................................................................................................................... 35

AppendixesA. Supplementary Figures .................................................................................................................................... 39B. Minithesaurus of Thesaurus Terms ................................................................................................................. 61C. User Interfaces .................................................................................................................................................... 62D. Thesaurus Clearinghouses ............................................................................................................................... 63E. Bibliography of Manuals on Thesaurus Construction.................................................................................. 64

Index ............................................................................................................................................................................

List of FiguresFigure 1. Semantic Relationships in a Thesaurus ................................................................................................ 14

Appendix A: Supplementary Figures

A1. The Relationship Indicators USE AND and USED FOR+ ....................................................................... 39A2. KWIC Index to a Thesaurus ......................................................................................................................... 40A3. KWOC Index to a Thesaurus ....................................................................................................................... 41A4. Flat Thesaurus Structure ............................................................................................................................... 42A5. Multilevel Broader and Narrower Terms .................................................................................................. 43A6. Generic Structure ........................................................................................................................................... 44A7. Tree Structure ................................................................................................................................................. 45A8. Hierarchical Thesaurus Display Featuring Node Labels and Line Numbers ....................................... 46A9. Graphic Display ............................................................................................................................................. 47A10. Top Terms in an Alphabetic Thesaurus ..................................................................................................... 48A11. Hierarchical Display of Top Terms ............................................................................................................. 40A12. Two-Way Hierarchical Display ................................................................................................................... 50A13. Category List .................................................................................................................................................. 51A14. Printed Alphabetic Thesaurus Designed for End-Users .......................................................................... 52A15. Printed Alphabetic Thesaurus Designed for Indexers and Searchers ................................................... 53A16. KWIC Screen Display of Thesaurus Terms ................................................................................................ 54A17. KWOC Screen Display of Thesaurus Terms .............................................................................................. 55A18. Screen Display of Term Record ................................................................................................................... 56A19. Thesaurus Term Review Form..................................................................................................................... 57A20. Menu of a Thesaurus Management System ............................................................................................... 58A21. Screens for Modification and Deletion of Terms....................................................................................... 59A22. Screen Display of Term Record Featuring Specialized Codes in a Company Thesaurus ................... 60

Page viii

Foreword

(This foreword is not part of the American National Standard Guidelines for the Construction, Format,and Management of Monolingual Thesauri, ANSI/NISO Z39.19 - 1993. It is included for information only.)

The first edition of this standard, published in1974, was prepared by Subcommittee 25 on The-saurus Rules and Conventions of American Na-tional Standards Committee Z39 on Standardiza-tion in the Field of Library Work, Documentation,and Related Publishing Practices (later known asLibrary and Information Sciences and Related Pub-lishing Practices) [1]. The subcommittee drewheavily on standards of practice developed by theEngineers Joint Council, the Committee on Scien-tific and Technical Information of the Federal Coun-cil for Science and Technology, and UNESCO [2,3,4].In 1980 ANSI published a revised edition of thethesaurus standard. In 1985, the National Informa-tion Standards Organization, the successor ofAmerican National Standards Committee Z39, bal-loted to reaffirm the standard. Two“no” votes ledto a default ballot in 1987, which resulted in addi-tional recommendations for revision.

In 1988 Standards Committee PP was appointedto draft a revised standard. The committee mem-bers, chosen for their expertise in both the theoryand the practice of the development and use ofthesauri in information systems, represented a widerange of subject interests (aerospace, business,medicine, social sciences, information science, andthe arts).

The current revision borrows heavily from thecorresponding international and British standards[5,6]. It differs substantially in structure and word-

ing from the 1980 American edition, but supportsmost of its recommendations. An important changeis in the handling of the whole-part relationship,which is now treated as a special case of the hierar-chical relationship rather than as an associativerelationship.

Another notable change is the deletion of theword use from the title because this standard doesnot prescribe how thesauri are to be used. Thephrase monolingual thesauri has been added to thetitle to distinguish this standard from a separatestandard for multilingual thesauri [7].

This edition continues to emphasize the role ofthesauri in information storage and retrieval sys-tems, but recognizes their applicability to otherfields, such as knowledge engineering. New sec-tions deal with screen display and thesaurus man-agement systems.

Unlike the previous edition, the revised stan-dard does not take a monolithic approach tothesaurus structure and display, but instead pointsout the advantages and disadvantages of variousformats. Many of these formats are illustrated in anappendix of supplementary figures.

Suggestions for improving of this standard arewelcome. They should be sent to the National In-formation Standards Organization , P.O. Box 1056,Bethesda, MD 20827 (301) 975-2814.

At the time it approved this standard, NISO hadthe voting members:

American Association of Law LibrariesGary J. Bravy

American Chemical SocietyRobert S. Tannehill, Jr.Leon R. Blauvelt (Alt)

American Library AssociationMyron ChaceGlenn Patton (Alt)

American Psychological AssociationMaurine F. Jackson

American Society for Information ScienceNolan Pope

American Society of IndexersPatricia S. Kuhr

American Theological Library AssociationMyron Chace

Apple Computer, Inc.Karen Higginbottom

Art Libraries Society of North AmericaPatricia J. BarnettPamela J. Parry (Alt)

Association of American University PressesMary Lou Menches

Association of Information and DisseminationCenters

Bruce H. Kiesel

NISO Voting Members

ANSI/NISO Z39.19-1993

Page ix


Association for Information and ImageManagement

Marilyn CourtotAssociation of Jewish Libraries

Bella Hass WeinbergPearl Berger (Alt)

The Association for Recorded Sound CollectionsDonald McCormickMartin Silver (Alt)

Association of Research LibrariesDuane Webster

AT&T Bell LabsM. E. Brennan

Baker & Taylor BooksChristian K. LarewStephanie Lanzalotto (Alt)

Library Binding InstituteSally Grauer

Library of CongressWinston TabbSally H. McCallum (Alt)

Mead Data CentralPeter RyallDave Withers (Alt)

Medical Library AssociationRick B. ForsmanRaymond A. Palmer (Alt)

MINITEXAnita Anker BraninWilliam DeJohn (Alt)

Music Library AssociationLenore CoralGeraldine Ostrove (Alt)

National Agricultural LibraryJoseph H. HowardGary K. McCone (Alt)

National Archives and Records AdministrationAlan Calmes

National Federation of Abstracting andInformation Services

Sarah Syen

National Institute of Standards and Technology,Office of Information Services

Jeff HarrisonMarietta Nelson (Alt)

National Library of MedicineLois Ann Colaianni

OCLC, Inc.Kate NevinsDon Muccino (Alt)

OHIONETGreg Pronevitz

Optical Publishing AssociationJohn NairnRichard A. Bowers (Alt)

PALINETJames E. Rush

Pittsburgh Regional Library CenterMary Lynn Kingston

Readmore Academic ServicesSandra J. GurshmanDan Tonkery (Alt)

The Research Libraries Group, Inc.Wayne DavisonKathy Bales (Alt)

The Boeing CompanySteven C. HillMichael Crandall (Alt)

Book Manufacturers’ InstituteStephen P. Snyder

CLSI, Inc.Andy Lukes

Colorado Alliance of Research LibrariesWard Shaw

Council of National Libraries and InformationAssociation

Thomas J. KempData Research Associates, Inc.

Michael J. MellingerJames Michael (Alt)

Digital Equipment CorporationJoan Blair

DynixRick Wilson

EBSCO Industries, Inc.Mary Beth Vanderpoorten

Engineering Information, Inc.Eric JohnsonMary Berger (Alt)

The Faxon Co., Inc.Fritz SchwartzJoe Santosuosso (Alt)

Gaylord Information SystemsRobert RileyBradley McLean (Alt)

IBM CorporationPeggy Federhart

Indiana Cooperative Library Services AuthorityBarbara Evans MarkusonJanice Cox (Alt)

Page x


U.S. Department of Defense Joint MedicalLibrary

Diane ZehnpfennigU.S. Department of Energy, Office of Scientific

and Technical InformationMary HallNancy Hardin (Alt)

U.S. ISBN Maintenance AgencyEmery Koltay

U.S.National Commission on Libraries andInformation Science

Peter YoungSandra N. Milevski (Alt)

VTLSVinod Chachra

H.W. Wilson CompanyGeorge I. LewickyAnn Case (Alt)

Society of American ArchivistsChristine WardVictoria Irons Walch (Alt)

Software AG of North America, Inc.James J. KoppJames E. Emerson (Alt)Special Libraries AssociationAudrey N. Grosch

SUNY/OCLC NetworkGlyn T. EvansDavid Forsythe (Alt)

UMIDon WillisJohn Brooks (Alt)

Unisys CorporationBill Payne

U.S. Department of Commerce Printing andPublishing Division

William S. LofquistU.S. Department of Defense Defense Technical

Information CenterGretchen SchlagClaire Tozier (Alt)

Standards Committee PP, which developed this standard, had the following members:

Bella Hass Weinberg, ChairSt. John’s University

Ronald L. BuchanNASA Scientific and Technical InformationProgram

Lois LymanDigital Equipment Corporation

Toni PetersenArt and Architecture Thesaurus

Caroline Reyes NewsBank/Readex

Peri SchuylerNational Library of Medicine

Hans H. WellischEmeritus, University of Maryland

Page xi


NISO Board of Directors

At the time NISO approved this standard, the following individuals served on its Board of Directors:

James E. Rush, ChairpersonPALINET

Michael J. Mellinger, Vice Chair/Chair-electData Research Associates

Paul Evan Peters, Immediate Past ChairpersonCoalition for Networked Information

Heike Kordish, TreasurerNew York Public Library

Patricia R. Harris, Executive DirectorNational Information Standards Organization

Directors Representing LibrariesSusan Vita

Library of CongressShirley Kistler Baker

Washington UniversityNolan Pope

University of Wisconsin Madison

Directors Representing Information ServicesMichael J. McGill

University of MichiganWilhelm Bartenbach

Engineering InformationBonnie Lawlor

Institute for Scientific Information, Inc.

Directors Representing PublishingJoel H. Baron

The Faxon Company, Inc.Constance U. Greaser

American HondaMarjorie Hlava

Access Innovations, Inc.

References

1. American National Standard Guidelines for The-saurus Structure, Construction, and Use. NewYork: American National Standards Institute,1974. (ANSI Z39.19-1974).

2. Thesaurus Rules and Conventions. New York:Engineers Joint Council, November 1966.

3. Guidelines for the Development of InformationRetrieval Thesauri. U.S. Federal Council for Sci-ence and Technology, Committee on Scientificand Technical Information (COSATI). Washing-ton, D.C.: Government Printing Office, 1967.

4. Guidelines for the Establishment and Develop-ment of Monolingual Scientific and TechnicalThesauri for Information Retrieval. Paris: Unit-ed Nations Educational, Scientific and CulturalOrganization, July 1970.

5. Documentation—Guidelines for the Establish-ment and Development of Monolingual Thesau-ri, 2nd ed. Geneva: International Organizationfor Standardization, 1986. (ISO 2788-1986).

6. Guidelines for the Establishment and Devel-opment of Monolingual Thesauri. London:British Standards Institution, 1979. (BS5723:1979). Superseded by: British StandardGuide to Establishment and Development ofMonolingual Thesauri. (BS 5723:1987=ISO2788-1986).

7. Documentation—Guidelines for the Establish-ment and Development of MultilingualThesauri. Geneva: International Organizationfor Standardization, 1985. (ISO 5964-1985).

Page xii


Abbreviations, Codes, and Conventions Usedin this Standard

Abbreviations (Thesaurus Codes) and Relationship Indicators

BT = broader termBTG = broader term (generic)BTI = broader term (instance)BTP = broader term (partitive)GS = generic structureHN = history noteNT = narrower termNTG = narrower term (generic)NTI = narrower term (instance)NTP = narrower term (partitive)RT = related termSEE = equivalent to U (USE)SN = scope noteTT = top termU = useUF = used forUF+ = used for . . . and . . .USE+ = use . . . and . . .X = see from (equivalent to UF); reciprocal of see

Conventions:Descriptors are in boldface.

Words discussed in the text are enclosed in ‘singlequotation marks’.

Underlined terms are defined in the glossary.


Page 1

1.1 Purpose and StructureA thesaurus may be considered in terms of both

its purpose and its structure.

1.1.1 PurposeFour principal purposes are served by a the-

saurus:a) Translation. To provide a means for translating

the natural language of authors, indexers, andusers into a controlled vocabulary used for in-dexing and retrieval.

b) Consistency. To promote consistency in the as-signment of index terms.

c) Indication of Relationships. To indicate semanticrelationships among terms.

d) Retrieval. To serve as a searching aid in retrievalof documents.

1.1.2 StructureA thesaurus displays through its structure the

equivalence (synonymous), hierarchical, and asso-ciative relationships among terms. These relation-ships are defined in section 5. Print and screendisplay formats used to show these relationshipsare recommended in sections 6 and 7, respectively.

1.2 Vocabulary ControlVocabulary control in a thesaurus is achieved

through three principal means: (a) the delineationof the scope, or meaning, of descriptors (see section3.2); (b) the linking of synonymous and nearlysynonymous terms through the equivalence rela-tionship (see section 5.2); and (c) the disambigua-tion of homographs (see section 3.2.1).

2. Scope of the StandardThis standard presents guidelines and conven-

tions for the contents, display, methods of con-struction, and maintenance of a monolingual the-saurus. It also notes the recommended features of athesaurus management system.

The standard deals with some aspects of termselection, in that it contains recommended proce-dures for vocabulary control, but it is particularlyconcerned with means for establishing and dis-playing certain kinds of relationships betweenterms.

This standard should be regarded as a set ofrecommendations based on preferred techniques

1. IntroductionA thesaurus*, for the purposes of this standard,

is a controlled vocabulary of terms in natural lan-guage that are designed for postcoordination. Theneed to control the formation and use of termsstems mainly from two basic features of naturallanguage, namely synonyms (different terms rep-resenting the same concept) and polysemes orhomographs (terms with the same spelling rep-resenting different concepts). The controlled vo-cabulary is established by information special-ists or lexicographers and is generally employedin indexing.

Indexing is the process by which subject termsor classification symbols are assigned to con-cepts dealt with in documents. It includes anysystem in which the selection and organizationof indexing terms call for human intellectualdecisions, regardless of whether computer assis-tance is also used to store and manipulate theseterms or to identify documents to which certainterms or combinations of terms have been as-signed. The effectiveness of indexing as a meansfor identifying and retrieving documents dependsupon a well-constructed indexing language.

Research in the field of information sciencehas shown that controlled vocabularies improveboth precision and recall in searching. For ex-ample, they improve precision by defining thescope of terms; they increase recall by retrievingdocuments that employ different terms for thesame concept. (Some of the manuals of thesaurusconstruction listed in Appendix E cite studiesthat have demonstrated these advantages of con-trolled vocabularies.)

Thesauri may be used in computer-assisted in-dexing, but this standard is primarily oriented to-ward thesauri employed for assignment indexingby humans (as opposed to derivative indexing bymachines).

It is important to note, however, that thesauri arenot limited to systems that apply vocabulary con-trol at the indexing stage. Thesauri may also behelpful to the user searching unindexed, i.e., natu-ral language, databases. The value of thesauri iscurrently recognized as well by researchers inknowledge engineering and hypertext. The manyuses to which thesauri may be put are recognized inthe standard, but not explicitly discussed.

* Underlined terms are defined in the glossary.

Guidelines for the Construction, Format, and Management ofMonolingual Thesauri


Page 2

and procedures. Optional procedures are, how-ever, sometimes described, e.g., for the display ofinter-term relationships in a thesaurus.

The conventions used in this standard to indi-cate the force of recommendations are: shall (re-quired for meeting the standard), should (recom-mended), and may (optional).

2.1 Applications of the StandardThe standard assumes that thesauri will be

applied mainly in postcoordinate informationstorage and retrieval systems. Guidelines for thedevelopment of lists of subheadings designed forprecoordination with descriptors are thereforenot included in this standard, but this does notpreclude the use of thesauri in precoordinateindexing systems.

The standard also does not provide guidelinesfor human or computer-aided indexing based onthesauri. For indexing procedures and practices,see ANSI Z39.4.

2.2 Monolingual ThesauriThe recommendations contained in this stan-

dard relate specifically to monolingual thesauri,with particular reference to English. No consid-eration has been given to the special require-ments of other languages, nor to those of multi-lingual thesauri in which terms are listed inparallel sequences showing conceptual equiva-lences in more than one language.

2.3 Types of TermsThis standard deals only with the formulation,

organization , and display of terms that form acontrolled subset of natural language. It does notsuggest procedures for organizing and displayingheadings consisting of mathematical and chemicalformulas, nor for establishing proper names. (Forinformation on these, refer to appropriate stan-dards, such as Anglo-American Cataloguing Rules,2nd ed. [AACR2].) Such headings may be includedin thesauri, however, and relationships among themmay be indicated through the devices presented inthis standard.

3. Scope, Form, and Choice of DescriptorsThis section deals with the principles of descrip-

tor selection and the determination of the correctform. Literary warrant is the guiding principle forthe selection of the preferred form of a descriptor.Section 3.2.1 treats the form of cross-referencesfrom nonpreferred terms to preferred terms. In thissection, the phrase ‘rather than’ is used to contrast

preferred and nonpreferred terms. Where twosynonyms are juxtaposed without indication of apreference, they are separated by a slash.

The development of an initial set of descriptorsis dealt with in section 8.

3.1 Single-word vs. Multiword DescriptorsEach descriptor included in a thesaurus should

represent a single concept (or unit of thought). Aconcept may be expressed by a single-word term orby a multiword term. Because of the difficulty ofdefining ‘single concept’ in the latter case, objectivecriteria for dealing with compound terms are pro-vided in section 4.

3.2 Scope of DescriptorsThe scope of descriptors is restricted to selected

meanings within the domain of the thesaurus. Eachdescriptor should be formulated in such a way thatit conveys the intended scope to any user of thethesaurus. Terms whose meanings overlap in gen-eral usage, and homographs, i.e., terms with iden-tical spellings but different meanings, should beavoided as far as possible in the selection of de-scriptors. Literary warrant, i.e., the occurrence ofterms in documents, and other criteria may, how-ever, justify the use of such terms as descriptors.Section 3.2.1 deals with methods of disambiguat-ing homographs. Section 3.2.2 explains the use ofscope notes for terms whose meanings overlap.

3.2.1 Parenthetical QualifiersThe use of homographic terms as descriptors re-

quires clarification of their meaning through a quali-fier (‘gloss’ in linguistic terminology). The qualifier,which is enclosed in parentheses, is part of the de-scriptor. The qualifier itself may be a descriptor, oftena broader term than the one being qualified. It shouldbe as brief as possible, ideally consisting of one word,but should not be a homograph. Qualifiers should bestandardized within a given thesaurus to the extentpossible, e.g., ‘biology’ and ‘bioscience’ should notboth be used as qualifiers.

A qualifier is not a scope note; however, a quali-fied descriptor may have a scope note appended toit. For guidelines on scope notes, see section 3.2.2.

Qualifiers should also be added to entry termswhen their meaning is ambiguous.

Examples:cranes (birds)cranes (lifting equipment)

mercury (metal)Mercury (planet)Mercury (Roman deity)

SCOPE OF THE STANDARD


Page 3

seals (animals)seals (law)seals (numismatics)

socialization (economics)socialization (social psychology)

a) A qualifier should be added to each homographicterm, even when one is used in the primary senseof the domain and the second in a differentsense. For example, cranes (lifting equipment)should be the descriptor in an engineering the-saurus that also includes cranes (birds).

b) A parenthetical qualifier should be appended tohomographic terms even when a descriptor isused in only one of its meanings within a thesau-rus, e.g., shells (structures) in an engineeringthesaurus. This facilitates cross-database search-ing and mapping of descriptors in disparatedomains.Examples:developing (photography)organism (philosophy)translation (genetics)

c) Parenthetical qualifiers should not be used torepresent compound concepts, e.g., ‘cookery(zucchini)’ or ‘pipes (plastic).’ ‘Plastic’ is used inthe latter example to indicate a type of piperather than to disambiguate the word ‘pipe.’Appropriate uses of qualifiers with the term‘pipes’ would be: pipes (musical instruments)and pipes (smoking implements). (See section 4for guidelines on formulating compound de-scriptors.)

d) Selection of a compound term as a descriptor ispreferred to use of a parenthetical qualifier witha single-word term, if usage permits, i.e., if thecompound occurs in natural language.Examples:phonograph records rather than records

(phonograph)religious tolerance rather than tolerance

(religion)Both of these compound terms are natural ex-

pressions that may be selected for thesauri in thefields of music and religion, respectively.

3.2.2 Scope NotesA scope note is used to restrict or expand the

application of a descriptor, to distinguish betweendescriptors that have overlapping meanings innatural language, or to provide other advice onterm usage to either the indexer or the searcher. Ascope note should state the chosen meaning of a

descriptor; it may also indicate other meanings thatare recognized in natural language, but which havebeen deliberately excluded from the controlledvocabulary. A scope note, unlike a parentheticalqualifier, is not part of a descriptor. While qualifiersare generally added only to homographs, a scopenote (SN) may be supplied for any descriptor.

Example:illuminations SN Includes both the ornamental decoration

and the illustrations in manuscripts as wellas in some early printed books, if done byhand.

The scope of descriptors is indicated addition-ally through the semantic relationships repre-sented in the thesaurus (see section 5.2). Changesin the scope of descriptors should be recorded inhistory notes (see section 9.3).

3.2.2.1 Reciprocal Scope NotesWhen reference is made to other descriptors in a

scope note, a reciprocal scope note should gener-ally be provided for each descriptor mentioned.

Examples:dogtrots SN Passages sharing a roof common with the

rest of a building, connecting two parts ofa log house of the American folk art tradi-tion. Distinguished from breezeways byits folk architecture tradition and log housecontext and its common roof.

breezeways SN Roofed passages connecting two parts of a

house or a house and a garage; commonafter 1930. Distinct from dogtrots, whichoccur in folk architecture log houses.

Even where the scope of only one of the descriptorsrequires clarification, it is useful to note in the termrecord for the second descriptor that it has been citedin a scope note of a different descriptor, e.g.:

information science X SN library science

The X indicates that there is a reference from thescope note of library science to information sci-ence. This reciprocal reference will ensure thatwhen a change is made to one of the descriptors, orit is deleted, the effect on the other descriptor willbe considered.

3.3 Types of ConceptThe concepts represented by descriptors can

be grouped into general types. The compiler ofa thesaurus needs to be aware of these types,

SCOPE, FORM, AND CHOICE OF DESCRIPTORS


Page 4

because the type of concept may affect some ofthe procedures used in thesaurus construction;for example, the choice of singular or plural form(see section 3.5) and applying a test for the valid-ity of a hierarchy (see section 5.3.1). The listedtypes are not exhaustive.a) things and their physical parts

Examples:birdscarburetorsmicroformsmountainsoil paintingsteddy bears

b) materialsExamples:adhesivesmustard gasoxygenpaintswater

c) activities or processesExamples:acidificationminiature golfpaintingparentingscuba divingsewing

d) events or occurrencesExamples:birthdayscivil warsholidaysrevolutions

e) properties or states of persons, things, materials,or actionsExamples:consciousnesselasticitypersonalityspeedtexture

f) disciplines or subject fieldsExamples:anthropologyinformation scienceorganic chemistrytheology

g) units of measurementExamples:hertzkilometers

3.3.1 Unique EntitiesUnique entities, or “classes-of-one,” are usually

expressed as proper nouns.Examples:AristotleEarthFourth of JulyLibrary of CongressMagna CartaWorld Health OrganizationZimbabwe

3.4 Grammatical Forms of DescriptorsThe following sections indicate the preferred

grammatical forms of descriptors. Entry termsfrom nonpreferred grammatical forms may be pro-vided.

3.4.1 Nouns and Noun PhrasesThe grammatical form of a descriptor should be

a noun or noun phrase. For an exception to thisrecommendation, see section 3.4.2.

Examples:drawingsliverspace shuttles

3.4.1.1 Verbal NounsVerbs expressed as infinitives (without ‘to’) or

participles should not be used alone as descriptors.Activities should be represented by nouns or gerunds.

Examples:catalysis rather than catalyzecookery rather than cookfreezing rather than frozenreading rather than readdistillation rather than distill

3.4.1.2 Noun PhrasesNoun phrases are compound terms that may be

established as descriptors if they represent a singleconcept (see section 4). Noun phrases occur in twoforms: premodified, or adjectival; and postmodified,or prepositional.

3.4.1.2.1 Adjectival Noun PhrasesAdjectival noun phrases, which are generally

premodified, are the preferred form.Examples:basal metabolismcold fusionhistorical dramapublic televisionrapid transit



Page 5

3.4.1.2.2 Prepositional Noun PhrasesPrepositional noun phrases are generally

postmodified. When possible, noun phrases shouldexclude prepositions; for example, use ‘carbohy-drate metabolism’ rather than ‘metabolism of car-bohydrates,’ and ‘children’s hospitals’ rather than‘hospitals for children.’

Descriptors in the form of prepositional nounphrases should be restricted to concepts that can-not be expressed in any other way, or that havebecome idiomatic.

Examples:accessories after the factburden of proofcream of tartarcoats of armsplaster of Parisprisoners of warsergeants-at-armsstrength of materials

3.4.2 AdjectivesAdjectives and adjectival phrases used alone

may be established as descriptors in thesauri in thespecial circumstances considered below.

3.4.2.1 Limiting the Number of CompoundDescriptors

As an alternative to the creation of multiplecompound descriptors, adjectives may appear asseparate descriptors when designed to beprecoordinated in indexing or postcoordinated insearching. They should generally not be assignedas indexing terms in isolation. Given the possibilityof false coordination in searching (e.g., the linkingof an adjective with the wrong noun), adjectivaldescriptors should be used sparingly.

Examples:Adjectives used Precoordinatedas descriptors indexing termsairborne airborne troopsmobile mobile homesoffshore offshore drillingportable portable heatersprestressed prestressed concreteCertain noun phrases can be used to modify other

nouns, e.g., ‘high frequency’ can modify the noun‘waves.’ The guidelines for adjectives may be appliedto such noun phrases.

3.4.2.2 Economy of Cross-referencesAdjectives may be used alone in general cross-

references to direct the user to or from a group ofdescriptors beginning with a corresponding noun,

e.g., “cardiac . . . see also the descriptors begin-ning with heart.” An example of a reference inthe opposite direction (noun to adjective) is:“France see also the descriptors beginning withFrench (French art, French language, French lit-erature, French wines).”

This guideline applies especially when theadjective and the noun to which it is related differwidely in their derivation and spelling, for ex-ample, ‘sea/marine,’ ‘ear/aural.’ It works well inprinted thesauri with a limited number of de-scriptors beginning with one of the terms, butexplicit references from all compound forms (e.g.,cardiac diseases USE heart diseases) may be nec-essary in thesauri displayed on a screen, or inprinted tools in which the compound terms be-ginning with a particular adjective span severalcolumns or pages, and the user may not noticethe general reference.

3.4.3 AdverbsAdverbs such as ‘very’ or ‘highly’ should not be

used alone as descriptors. A phrase beginning withsuch an adverb may be accepted as a descriptoronly when it has acquired a specialized meaningwithin a domain.

Examples:very high frequencyvery large scale integrationvery low density lipoproteins

3.4.4 Initial ArticlesThe use of initial articles in descriptors should be

avoided.

3.4.4.1 DeletionDelete the initial article when the descriptor is

clear without it. If not, use a parenthetical quali-fier.

Examples:arts rather than the artsstate rather than the state (political entity)

3.4.4.2 RetentionIf the initial article is an integral part of a

proper name, and should be searchable, it shouldbe included in the descriptor in direct order.Otherwise, invert the article. Whether or not theinitial article is included as an integral part of thename is dependent upon the language. In thefollowing examples, English is the assumedlanguage of the thesaurus.



Page 6

Examples:A Big Company [initialism: ABC]Computer Place, TheEl NiñoEl SalvadorLos AngelesNarrows, TheIn cases where the article is included in direct

order, a cross-reference from the element followingthe article should be provided, e.g., Salvador USEEl Salvador.

3.5 Singular and Plural FormsTerms can be divided into two categories: count

nouns (see section 3.5.1) and noncount (mass)nouns (see section 3.5.2). The guidelines for sin-gular and plural based on these categories applyto the formulation of both descriptors and entryterms.

3.5.1 Count NounsCount nouns are names of objects or concepts

that are subject to the question ‘How many?’ butnot ‘How much?’. These should normally be ex-pressed as plurals.

Examples:bookschemical reactionspenguinspolitical partiessingersvertebrateswindows

3.5.1.1 Exception to Plural Count NounsIf in the domain of the thesaurus there is liter-

ary or user warrant for the expression of countnouns in the singular, establishment of descrip-tors in that form is acceptable. For example, in thefield of biomedicine, the names of parts of thebody are generally formulated in the singular.

Examples:digestive systemearlungnosestomachIn a museum catalog, objects are typically treated

as unique items, and descriptors are given in thesingular.

Examples:chairoil paintingtapestry

3.5.2 Noncount (Mass) NounsNoncount nouns are names of materials or sub-

stances that are subject to the question ‘How much?’but not ‘How many?’. These should be expressedin the singular.

Examples:copperpaint (but cf. paints, below)snowwaterIf the community of users served by a thesaurus

regards a given substance or material as a class withmore than one member, the class should be ex-pressed in the plural.

Examples:plastics (i.e., various types of plastic)paints

3.5.2.1 Abstract ConceptsThe names of abstract concepts, e.g., systems of

belief, activities, emotions, properties, and disciplines,shall also be expressed in the singular. Some of theseterms are subject to the question ‘How much?’.

Examples:beliefs Judaism; Taoismactivities digestion; distribution; or processes migration; welding,

writingemotions anger; envy; love; pityproperties conductivity; silence; or states worth;disciplines architecture; business;

musicology; metallurgy

3.5.2.2 Unique EntitiesThe names of unique entities, whether concrete

or abstract, shall be expressed in the singular.Examples:Shangri-LaStatue of Liberty

3.5.3 Coexistence of Singular and Plural FormsWhere the singular and plural forms of a term

represent different concepts, separate descriptors foreach should be entered in the thesaurus as appropri-ate. The distinction should be indicated by a qualifier.

Examples:bridge (game)bridges (dentistry)bridges (structures)

damage (injury)damages (law)

wood (material)woods (forested areas)



Page 7

3.6 Selection of Preferred FormThe authority for the form selected should be

recorded in the term record (see section 8.5). Thevariant forms should be included as entry terms.The source(s) of variant forms may also be re-corded.

3.6.1 UsageDescriptors should reflect as much as possible

the usage of people familiar with the domain of thethesaurus. Neutral rather than pejorative or objec-tionable terms should be selected, e.g., developingnations rather than underdeveloped countries. Seealso section 3.6.4.2.

3.6.1.1 Literary WarrantWords and phrases drawn from the literature

of the field should determine the formulation ofdescriptors. When two or more variants haveliterary warrant, the most frequently used termshould be selected as the descriptor.

3.6.1.2 Published AuthoritiesGuidance in the selection of descriptors may

be found in reference works of the domain, suchas dictionaries, glossaries, encyclopedias, andauthoritative treatises of the field.

3.6.1.3 Other Indexing LanguagesOther thesauri, abstracting and indexing ser-

vices, and subject heading lists may be useful assources for selection of preferred forms.

3.6.1.4 Personal AuthoritiesOpinions of subject experts regarding the pre-

ferred form of descriptors may be sought.

3.6.2 Spelling

3.6.2.1 Literary WarrantThe most widely accepted spelling of words

should be adopted. If variant spellings exist and arecommonly recognized, each should be entered inthe thesaurus, and a cross-reference should be madefrom the nonpreferred to the preferred form.

Examples:Romania rather than Roumaniatheater rather than theatre

3.6.2.2 AuthoritiesSpelling should follow the practice of well-

established dictionaries or glossaries. If a choicebetween spellings is made for dialectal reasons (forexample, between American and British English),

the choice should be adhered to consistentlythroughout the thesaurus; exceptions may be madefor proper names.

Examples:catalogs rather than cataloguescolor rather than colourlabor rather than labour but Labour Party

3.6.3 Abbreviations, Initialisms, andAcronyms

This section deals with the choice of full vs.abbreviated forms of descriptors. Capitalizationis dealt with in section 3.7.1.

3.6.3.1 Preference for AbbreviationAbbreviations and acronyms should be selected

as descriptors only when they have become so wellestablished that the full form of the term or propername is rarely used. Cross-references should bemade from the full forms.

Examples:AIDS rather than acquired immune(disease) deficiency syndromeDNA rather than deoxyribonucleic acidlasers rather than light amplification by

stimulated emission of radiation

UNESCOrather than United Nations Educational, Scientific, and Cultural Organization

VTOL rather than vertical takeoff and aircraft landing aircraft

3.6.3.2 Preference for Full Form

3.6.3.2.1 General UseThe full form of terms should be selected as

descriptors when the abbreviated form is notwidely used and generally understood. Cross-references should be made from the abbreviatedforms.

Examples:automated teller machines rather than ATMsdriving while intoxicated rather than DWIprisoners of war rather than POWs

3.6.3.2.2 AmbiguityMany acronyms and abbreviations stand for

more than one word or phrase; the full form ofthe term should therefore be selected as the de-scriptor in preference to the abbreviated form,even when the abbreviation has only one value in



Page 8

the domain of the thesaurus.Examples:artificial insemination rather than AIartificial intelligence rather than AI

3.6.4 Neologisms, Slang, and JargonNeologisms, slang, and jargon terms often cover

new concepts originating within a particular spe-cialty, subculture, or social group. Such terms aregenerally not included in standard dictionaries.When no widely accepted alternative exists, theneologism, slang, or jargon term should be ac-cepted as a descriptor.

Examples:kludgeyuppies

3.6.4.1 Provisional TermsA neologism, slang, or jargon term may be

labeled provisional (see section 8.6.2), and maybe elevated to full descriptor status as the termbecomes accepted into the language.

Example:burnout

3.6.4.2 Cross-references from Slang TermsA slang term that is an alternative to an existing

and well-established term should be admitted asan entry term to the descriptor.

Example:psychiatrists rather than shrinks[descriptor] [cross-reference]

3.6.5 Common Names and Trade Names

3.6.5.1 Preference for Common NameA product is frequently known by a widely

recognized trade name. Where an equivalent com-mon name exists, the common name should beused in preference to the trade name, because thegeneric name is the more accurate representation ofthe concept.

Example:diazepam rather than ValiumBoth common and trade names may be descrip-

tors in thesauri that cover generic products as wellas specific brands, with the latter treated as nar-rower terms (see section 5.3.1).

3.6.5.2 TrademarksBecause trademarks are recognized by law as

proprietary, they should be identified as such in athesaurus. The identification should be an integralpart of the term. One of the following designations

may be used, as appropriate: (Trademark), (TM),(Reg), or (R).

Examples:Current Contents (R)Scotch Tape (TM)

3.6.5.3 Loss of Protected StatusA trademark may be used as a descriptor with-

out qualification only when its legal protection hasbeen lost.

Examples:aspirinescalatorthermos

3.6.6 Popular and Scientific NamesIf a popular and a scientific name refers to the

same concept, the form most likely to be sought bythe users of the thesaurus should be chosen as thedescriptor, and a cross-reference provided from thenonpreferred term. For example, penguins might bechosen as the descriptor in a nontechnical thesau-rus, but the scientific equivalent, Sphenisciformes,may be selected for a zoological thesaurus.

3.6.7 Loanwords, Translations ofLoanwords, and Foreign-LanguageEquivalents

3.6.7.1 LoanwordsLoanwords are terms borrowed from other lan-

guages that have become naturalized in the borrow-ing language. If such terms are well established, theyshould be admitted into the thesaurus. Diacriticsshould be included if required by the orthographicauthorities for the borrowing language.

Examples:coup d’étatgestalthabeas corpuspas de deuxweltanschauung

3.6.7.2 Translations of LoanwordsOccasionally a loanword and a commonly

accepted translation coexist in the language ofthe thesaurus. The loanword should be preferredif it is more widely accepted in the domain of thethesaurus.

Example:amicus curiae rather than friend of the court

[in a legal thesaurus]If the translation becomes well established,

however, it should be selected as the descriptor.



Page 9

In all cases where a concept can be expressed byboth a loanword and a translated equivalent, across-reference should be made from thenonpreferred term.

Examples:braking radiation rather than bremsstrahlungpen name rather than nom de plume

3.6.7.3 Foreign-Language EquivalentsForeign-language terms, i.e., terms which have

not become naturalized in the language of thethesaurus, should be linked to terms in the pre-ferred language in cases where the foreign termsare likely to be sought by users. The languagechosen for the descriptor should be that whichthe user would likely expect, with a cross-reference from the equivalent term in the otherlanguage. (For guidelines on multilingual the-sauri, see ISO 5964.)

Example:Shavuot / Pentecost[Hebrew] [English]

3.6.8 Proper NamesVariant forms of proper names of persons, insti-

tutions, organizations, and places, as well as titles,if uncontrolled, create problems for searchers. Fre-quency of need for proper name access points gen-erally determines which of the following practicesis adopted.

Proper names of persons, institutions, organiza-tions, places, and titles:a) may be controlled by inclusion in a thesaurus of

subject (topical) descriptors;b) may be controlled through a separate name

authority file;c) may be left uncontrolled.

In options b) or c), proper names may be as-signed to documents as identifiers, differentiatedfrom topical descriptors.

Different methods may be adopted for control-ling various types of names. For example, geo-graphic names may be included in a thesaurus oftopical descriptors, and personal names establishedin an authority file.

Reasons for merging the files of name headingsand topical descriptors are (a) that the borderlinebetween them is not sharp, and (b) that it is oftendesirable to link a class with its instances hierarchi-cally, e.g., museums and Louvre (see section 5.3.3).

When proper names are included in a thesaurus,the form of the name should be selected in accordancewith a recognized code of cataloging practice, such asthe Anglo-American Cataloguing Rules, 2nd ed.

3.6.8.1 Place NamesNames of countries and geographic regions

frequently vary from language to language. Vari-ant terms referring to the same place also occurwithin a single language community, for reasonssuch as the following:a) An official and a popular name are both in

common use:Example:Netherlands / Holland

b) Anglicized and vernacular forms coexist:Examples:Cambodia / KampucheaIsrael / YisraelThe form most familiar to the users of the thesau-

rus should be designated as the descriptor, andcross-references should be provided from the vari-ants. Other things being equal, preference shouldbe given to the official rather than the popularname. The short form of the official name should bepreferred. Standard authorities, such as the publi-cations of the U.S. Board on Geographic Names,should be consulted for the official forms.

Examples:Arabian Peninsula rather than ArabiaNetherlands rather than HollandPhilippines rather than Republic of

the Philippines

3.7 Capitalization, Punctuation, and Non-alphabetic Characters

3.7.1 CapitalizationIt is recommended that predominantly lower-

case characters be used for descriptors in the-sauri. (The display of descriptors in indexes isoutside the scope of this standard.) Capitalsshould be used only for the initial letter(s) ofproper names, tradenames, and for those compo-nents of taxonomic names, such as genus, whichare conventionally capitalized; for all the lettersof initialisms; or where featured in unusual posi-tions in product or corporate names. Becauselowercase letters may also occur in unusual posi-tions in proper names, using a combination ofcapitals and lowercase letters in thesauri indi-cates to the user the correct orthography of a termin natural language and serves to distinguishcommon nouns from similar proper names.

Examples:dBASE IV (Trademark)DNAinformation systemsInformation Systems Corp.NewsBank



Page 10

3.7.2 Nonalphabetic CharactersTo simplify filing and searching, the use of

symbols and punctuation marks in descriptorsand entry terms should be minimized. Estab-lished orthographic authorities determine whensuch characters are essential.

3.7.2.1 ParenthesesParentheses should be used only to enclose quali-

fiers (see section 3.2.1) and trademark indicators(see section 3.6.5.2), or when they constitute part ofa term.

Example:benzo(a)pyrene

3.7.2.2 HyphensTo obviate problems in filing and searching,

hyphens should be avoided in thesauri wheneverpossible. Retain hyphens in topical terms only whendropping them would lead to ambiguity.

Examples:high temperature testingn-body problemnonfictionHyphens should be retained where they occur as

part of abbreviations, trademarks, chemical names,or proper nouns.

Examples:MS-DOSNewton-John, Oliviap-benzoquinoneSaur, Karl-OttoStratford-upon-Avon

3.7.2.3 Apostrophes

3.7.2.3.1 Possessive CaseApostrophes indicating the possessive case,

whether in the singular or plural, must be retained.(For the choice of singular or plural descriptors, seesection 3.5.) The singular form is most frequentlyfound in eponyms. Standard orthographic au-thorities should be consulted for the position ofthe apostrophe in common nouns.

Examples:singular plural

Boyle’s law artists’ marksDown’s syndrome workers’ compensationPascal’s triangle

3.7.2.3.2 Proper NamesApostrophes that are part of proper names must be

retained.Example: O’Dwyer’s tubes

3.7.2.4 Diacritical MarksDiacritical marks may be used if they are re-

quired for proper names or by the accepted stan-dards of a discipline. (See section 3.6.7.1 for the useof diacritics in loanwords and section 3.7.3 for theuse of diacritics in romanization.)

Examples:El NiñoGuillain-Barré syndromeMössbauer effectrésumés

3.7.2.5 Other Symbols and PunctuationMarks

Symbols, such as the ampersand, and punctua-tion marks should be used only in trademarks andproper names because they create filing and search-ing problems.

Examples:A & W Food ServicesPL/C

3.7.3 RomanizationCommonly accepted spellings for terms or proper

names from languages written in non-roman scripts,as found in authoritative reference sources, arepreferable to systematic romanization, i.e., apply-ing a table to convert the characters in a non-romanscript into roman ones. A cross-reference should beprovided from the systematic romanization to theestablished spelling.

Examples:Chebyshev rather than Tchebyshev polynomials polynomials

or Cebysev polynomialsKoran rather than Qur´anWhen romanizing (transliterating or transcrib-

ing), the procedure described in a relevant stan-dard or a nationally accepted library scheme shouldbe followed as far as possible. (Library schemesused in the U.S. have been collected in ALA-LCRomanization Tables, 1991.) If a choice exists, aromanization system that uses few or no diacriticalmarks should be selected.

4. Compound Terms

4.1 GeneralCompound or multiword terms in natural

language are considered lexemes, i.e., words boundtogether as lexical units. Dictionaries differ in theirpolicies regarding the inclusion of variouscategories of compounds, and so they are notalways authorities in this regard for thesaurus

COMPOUND TERMS

ˆˆ


Page 11

designers. For this reason, guidelines on thehandling of compound terms in thesauri have beendeveloped.

To be acceptable as a descriptor, a compoundterm should express a single concept or unit ofthought, capable of being arranged in a genus-species relationship within a hierarchy or treestructure. The compound phrase ‘children andtelevision,’ for example, cannot be accommodatedeither in a strict genus-species hierarchy on ‘chil-dren’ or one on ‘television,’ while the compoundterm ‘adopted children’ represents a type of chil-dren, and the compound term ‘educational tele-vision’ represents a type of ‘television.’ (A com-pound phrase such as ‘children and television’could be subordinated to a hierarchy on ‘childrenand media,’ but this would not represent a singlegenus-species relationship.)

4.1.1 Purpose of Guidelines on Compound Terms

The establishment of procedures for dealingconsistently with compound terms is one of themost difficult areas in the fields of thesaurusconstruction and indexing. The guidelines belowhave the following purposes:a) to aid in achieving intra-indexer, inter-indexer,

and inter-organizational consistency;b) to avoid overcomplexity of the indexing lan-

guage;c) to achieve a thesaurus whose structure is based

upon principles that are amenable to logicalexplanation;

d) to enhance the ease and precision of searching.

4.1.2 Precoordination and PostcoordinationDescriptors should represent single concepts,

expressed by a single word or by a multiword term.To express complex subjects, descriptors may bepostcoordinated in search statements using Bool-ean or other operators, e.g., liver AND anatomy, orprecoordinated into semantically linked, heading-subheading combinations, e.g., liver—anatomy.

4.1.3 Factors to Be ConsideredThe factors enumerated below may be consid-

ered in deciding which multiword terms should besplit into separate descriptors and which should beretained in compound form. It should be empha-sized that these procedures need not be appliedrigidly. Administrative policies or system con-straints may govern which of these practices isadopted by a given organization.

a) Literary warrant. A compound term may be em-ployed so frequently within the literature of thedomain covered by the thesaurus that splittingthe term into its components would be unac-ceptable to users who consider it a lexeme.

b) Regulating the number of terms in the thesaurus.Compound terms increase the number of de-scriptors in a thesaurus.

c) Printed versus computerized retrieval systems. If athesaurus is used for printed as well as electronicpublications, the requirements of users of bothformats should be taken into account. Printedindexes may require precoordinated headingsfor compound concepts, so that the number ofentries per heading can readily be scanned by asearcher, whereas in computerized retrieval,access is often enhanced by the indexing of com-pound concepts through separate descriptors,which can subsequently be postcoordinated bythe searcher. Rules may be developed for com-bining descriptors to form precoordinate index-ing terms; while this practice reduces the size ofa thesaurus, it should be kept in mind that theinclusion of compound descriptors may sim-plify the indexing process.

d) Avoidance of false drops in retrieval. Compounddescriptors often obviate certain kinds of falsedrops (unwanted combinations of terms) inretrieval when the same words in a differentsequence have a different meaning. For ex-ample, if the concept ‘library science’ is repre-sented by the separate descriptors ‘library’and ‘science,’ a search on this combinationwill also retrieve ‘science library.’

e) The nature of terminology in a given field. Thismay call for special criteria to regulate thetreatment of compound terms.

4.1.4 Elements of Compound TermsElements of a compound term are distinguished

by their logical roles or functions, which are rel-evant to the criteria for determining whether acompound term should be split (see section 4.2).The majority of compound terms, including bothpre- and postmodified noun phrases, can be ana-lyzed into the following components:a) the focus (also known as the ‘head noun’): the

noun component that identifies the broader classof things or events to which the term as a wholerefers.Examples:concrete in the compound term reinforced

concreteglass in the compound term stained glass

COMPOUND TERMS


Page 12

b) the difference (also known as the ‘modifier’):the part of a compound term that refers to acharacteristic, or logical difference, which nar-rows the denotation of the focus by specifyinga subclass of the broader concept representedby the focus.Examples:reinforced, which specifies a subclass, i.e., a

type, of concrete in reinforced concretestained, which specifies a subclass of glass in

stained glass

4.2 Criteria for Establishing CompoundTerms as Descriptors

Compound or multiword terms should be estab-lished as descriptors in the following circumstances:a) Splitting the parts would lead to ambiguity or a

loss of meaning.Examples:plant foodrose windows

b) One component of the term is not relevant to thescope of the thesaurus or is too vague to exist asan independent term.Examples (nondistinctive element underlined):composite drawingsfirst aidstone cutters

c) The modifier in the term has lost its originalmeaning, so the meaning of the compound termas a whole is not the sum of the meaning of itsparts.Examples:deck chairstrade winds

d) The modifier suggests a resemblance, as in ametaphor, to an unrelated thing or event.Examples:butterfly valvestree structures

e) The term contains an adjective that does notdefine a subclass of the focus, and is not actuallya member of that class at all.Examples:artificial limbspaper flowersrubber duckstin soldiers

f) The term is a proper name, or includes propernouns or proper adjectives.Examples:Freudian slipsHudson’s Bay CompanyUnited Nations

g) The compound term has become so familiar incommon use, or in the field covered by thethesaurus, that it is considered for practical pur-poses to represent a single concept.Examples:data processinggross domestic product

4.3 Criteria for Determining whenCompound Terms Should Be Split

4.3.1 Factors to Be ConsideredIt is more difficult to specify exact criteria for

recognizing the kinds of compound terms thatshould be split into separate components, each ofwhich is then entered as a noun or noun phrase inthe thesaurus. The following recommendations arebased upon general criteria, but they should not beregarded as mandatory instructions to be appliedrigidly in all circumstances. As noted above, suchdecisions are often based on the nature of terminol-ogy in a given field, which may call for specialtreatment of compound terms, or on the necessityfor controlling the size of the thesaurus.

4.3.2 Hierarchical StructureThe following recommendations are offered as a

means of achieving consistent treatment of com-pound terms that should be split. They also ensurea correct hierarchical structure in which each termfits into a logical conceptual framework. Each ofthese recommendations is accompanied by a con-verse condition that can be used to identify classesof terms that can usually not be split without risk ofambiguity.

4.3.2.1 Focus and Differencea) A compound term should be split when its focus

refers to a property or part, and its differencerepresents the whole or possessor of that prop-erty or part.Examples:aircraft engines = aircraft + engineshospital personnel = hospitals + personnelsoil acidity = soils + acidity

b) Conversely, a compound term should not besplit when the focus term refers to a whole andthe difference is a term for its part or property.Examples:acid soilsskilled personnel

4.3.2.2 Transitive Actiona) A compound term should be split if it consists of

a term representing a transitive action modified

COMPOUND TERMS


Page 13

by a term for the object on which the action isperformed.Examples:office management = offices [object] + man-

agement [action]textile printing = textiles [object] + printing

[action]b) Conversely, a compound term should not be

split when its component term for a thing ormaterial is modified by a term for the actioncarried out upon it.Examples:cast steel (cast [action] + steel [object])printed textiles (printed [action] + textiles

[object]

4.3.2.3 Intransitive Actiona) A compound term should be split if it consists of

a term for an intransitive action modified by aterm for the performer (agent) of the action.Examples:bird migration = birds [agent] + migration

[action]metal deterioration = metal [agent] + deterio-

ration [action]b) Conversely, a compound term should not be

split when the term for an object is modified bya term for the intransitive action in which theobject is involved.Example:dancing shoes (dancing [action] + shoes

[object])

4.4 Compound Descriptors vs. Node LabelsA compound descriptor should not be created

solely for the reason that it forms a logical level ina hierarchy and would serve to group a set ofnarrower terms. For this purpose node labelsshould be used (see section 5.3.5).

4.5 Order of Words in Compound TermsNoun phrases should be entered in a thesaurus

in natural language order, i.e., direct order. Forfurther guidance on the form of premodified andpostmodified noun phrases, see section 3.4.1.2.

Examples:birds of preyoral surgerysecondary schools

4.5.1 Cross-references from InversionsThe inverted form of a noun phrase may be

included as a cross-reference to the preferred termin natural language order.

Examples:surgery, oral USE oral surgeryschools, secondary USE secondary schoolsSee also the discussion of permuted lists in sec-

tion 6.1.1.

5. Relationships

5.1 General

5.1.1 Semantic LinkingOne of the main differences between thesauri

and other vocabulary lists is that thesauri displayand clearly distinguish the basic semantic relation-ships that link the terms they contain by means ofrelationship indicators. Other types of term lists donot provide such relationship indicators, or do notprovide them systematically. Ideally, a thesaurusshould not include any orphan terms, descriptorsthat are not related to any other descriptors.

A variety of formats for displaying semanticrelationships are described in section 6; this sec-tion is concerned with defining and illustratingthe relationships themselves.

5.1.2 Kinds of RelationshipRelationships of three kinds should be included

in thesauri:a) the equivalence relationship (see section 5.2);b) the hierarchical relationship (see section 5.3);c) the associative relationship (see section 5.4).

These relationships are illustrated in Figure 1,p.14.

Each relationship possesses the property of reci-procity, i.e., every relationship indicated betweenterm A and term B has a corresponding relation-ship from term B to term A.

The conventional abbreviations for relationshipindicators are used in the examples below. Addi-tional abbreviations for specialized purposes arefound in the following sections. (A complete list ofabbreviations used in this standard appears onpage xii. Typographic aspects of relationship indi-cators are discussed in section 6.3.3.)

Relationship Relationship Indicator Abbreviation

Equivalence USE none or U

(Synonymy) USED FOR UF

Hierarchy BROADER TERM BT

NARROWER TERM NT

Association RELATED TERM RT

RELATIONSHIPS


Page 14

Figure 1. Semantic Relationships in a Thesaurus

The circles in the diagrams below represent the scope of the terms.

The diagram impliesequivalent sets. Circle Aand Circle B overlap.

Figure 1a. The equivalence relationship, e.g., teenagers USE adolescents. (A) (B)

The diagram implies classinclusion.

Figure 1b. The hierarchical relationship, e.g., mammals NT cows. (A) (B)

The diagram impliessemantic overlap, i.e., thatthere is an element ofmeaning common to bothterms.

Figure 1c. The associative relationship, e.g., gold RT money. (A) (B)

A = B

A B

A B


Page 15

The relationship indicators are binary (or paired)operators. Some indicators are symmetric, and someare asymmetric.

RT is symmetric: if A RT B, then B RT A.USE and UF are asymmetric: if A USE B, then B UF A.Similarly, for BT and NT: if A BT B, then B NT A.

5.2 The Equivalence Relationship

5.2.1 GeneralWhen the same concept can be expressed by two

or more terms, one of these is selected as the pre-ferred term, i.e., the descriptor. The relationshipbetween preferred and nonpreferred terms is anequivalence relationship in which each term isregarded as referring to the same concept. Thedescriptor in effect substitutes for other terms ex-pressing equivalent or nearly equivalent concepts.A cross-reference to the descriptor should be madefrom any synonym or quasi-synonym that mayfunction as an entry term for the user.

Reciprocity of the equivalence relationship isexpressed by the following conventions:

U or USE, which leads from anonpreferred (entry)term to the descriptor,and

UF or USED FOR, the reciprocal, whichrecords entry termsleading to the descriptor.

Examples:Aves USE birds outline USE shapebirds UF Aves shape UF outlineThese relationship indicators are the equiva-

lents of see and x (see from), respectively, in manytraditional subject heading lists. This equivalencerelationship covers three basic types of term:a) synonyms (see section 5.2.2);b) lexical variants (see section 5.2.3); andc) quasi-synonyms (see section 5.2.4).

5.2.2 SynonymsSynonyms are terms whose meanings are re-

garded as the same or nearly the same in a widerange of contexts. True synonyms are rare innatural language. Although the terms may beinterchangeable in many circumstances, usagemay vary as a result of such factors as level offormality, professional vs. lay context, or

pejorative vs. neutral vs. complimentary conno-tation. The examples listed below are illustrativeof the various classes of synonyms that may beencountered in practice. A slash is used to sepa-rate the synonyms, indicating that the preferredterm has not been selected. (For this reason, theterms are not in boldface.)a) terms of different linguistic origin

Examples:cats / felinesfreedom / libertysodium / natriumsweat / perspiration

b) popular terms and scientific namesExamples:aspirin / acetylsalicylic acidgulls / Laridaesalt / sodium chloride

c) generic nouns and trade namesExamples:petroleum jelly / Vaselinephotocopies / Xeroxesrefrigerators / Frigidairestissues / Kleenex

(See section 3.6.5.2 for the designation of trade-marks.)d) variant names for emergent concepts

Example:hovercraft / air cushion vehicles

e) current or favored terms versus outdated ordeprecated termsExamples:poliomyelitis / infantile paralysisdeveloping countries / underdeveloped coun-tries

f) common nouns and slang or jargon termsExamples:helicopters / whirlybirdspsychiatrists / shrinks

g) dialectal variantsExamples:elevators / liftssubways / undergroundsIn these and other cases, preferred terms should

be selected to serve the needs of the majority ofusers, bearing in mind the criteria enumerated insections 3 and 4. For the sake of predictability,these criteria should be applied consistentlythroughout the thesaurus. If, for example, Ameri-can spelling is preferred to British spelling, orscientific names are preferred to popular names,this decision should be noted in the introduction(see section 6.3.2) and should be applied consis-tently in the formulation of descriptors.

RELATIONSHIPS


Page 16

5.2.3 Lexical VariantsLexical variants differ from synonyms in that

synonyms are different terms for the same concept,while lexical variants are different word forms forthe same expression. These forms may derive fromspelling or grammatical variation or from abbrevi-ated formats.a) lexical variants, including direct versus

inverted order, orthographic variants, stemvariants, and irregular pluralsExamples:radar antennas / antennas, radarRomania / Rumania / Roumaniaground water / ground-water / groundwateronline / on-linepediatrics / paediatricsmice / mouse

b) full names and abbreviationsExamples:International Federation for Documentation /

FIDpi mesons / pionspolyvinyl chloride / PVC

5.2.4 Quasi-SynonymsQuasi-synonyms are terms whose meanings are

generally regarded as different, but which aretreated as equivalents for the purposes of athesaurus, e.g., ‘REM sleep’ and ‘rapid eyemovements.’ The extent to which terms are treatedas quasi-synonyms depends in large measure uponthe domain covered by the thesaurus and its size.

Quasi-synonyms are frequently antonyms orrepresent points on a continuum:

Examples:wetness / drynesssmoothness / roughnessAn indexing agency may decide to designate

either ‘wetness’ or ‘dryness’ as the preferred termon the understanding that the specialist studyingone of these properties should be able to retrieve allinformation on the other.

As a general rule, terms should be treated asquasi-synonyms only in subject areas that are pe-ripheral to the domain of the thesaurus. Whenconcepts can be distinguished in the thesaurusdomain with sufficient precision to justify theirrepresentation as separate terms in a thesaurus,they should be individually defined and retainedas descriptors. If two concepts cannot be consis-tently and reliably differentiated from one another,however, a term for one concept should be selectedas the descriptor and a USE reference made fromthe other.

5.2.4.1 Generic PostingGeneric posting is a technique in which the

name of a class and the names of its members aretreated as quasi-synonyms, with the broader classname functioning as the preferred term.

Examples:waxes plant waxes UF plant waxes USE waxesfurniture UF beds beds USE furniture UF chairs chairs USE furniture UF desks desks USE furniture UF tables tables USE furnitureIf employed, this technique should be limited to

the peripheral areas of a subject field, or used whenthe number of documents on the members of a classdoes not warrant its being split into subclasses.This practice places limits on the specificity of thethesaurus, and should be used with discretion. Thenarrower terms are useful entry terms.

5.2.5 Cross-references to Elements ofCompound Terms

A USE....AND (or USE+) reference may be madefrom a compound term to its components in caseswhere the term is split and both components mustbe used in indexing or searching.

Examples:coal mining USE coal AND (or+) miningferromagnetic films USE ferromagnetic materials

AND filmsUF+ (USED FOR . . . AND . . .) is a code that may

be used for the reciprocal of the reference from thecompound term, as UF alone may suggest that oneof the two components may be used alone in search-ing for the concept.

Example:coal UF+ coal miningmining UF+ coal mining

This type of cross-reference is illustrated in Figure A1.

5.3 The Hierarchical RelationshipThis basic relationship is the primary feature that

distinguishes a systematic thesaurus from an un-structured list of terms, such as a glossary. It is basedon degrees or levels of superordination and subordi-nation, where the superordinate descriptor repre-sents a class or a whole, and subordinate descriptorsrefer to its members or parts. Reciprocity may beexpressed by the following relationship indicators:BT (Broader Term), a label for the superordinate

descriptorNT (Narrower Term), a label for the subordinate

descriptor

RELATIONSHIPS


Page 17

Example:mammals vertebrates BT vertebrates NT mammalsIn the flat format thesaurus (see section 6.2.1)

BT and NT indicate one level broader and one levelnarrower, respectively. There are other types ofalphabetical display that indicate multiple levelsof hierarchy (see section 6.2.2).

Hierarchical relationships may also be indicatedby systematic presentations such as tree structures(see section 6.2.3) or graphic displays (see section6.2.4).

The hierarchical relationship covers three logi-cally different and mutually exclusive situations:a) the generic relationship (see section 5.3.1);b) the whole-part relationship (see section 5.3.2);c) the instance relationship (see section 5.3.3).

All these types of relationship may apply to asingle descriptor, and special codes may be used todistinguish them. Each of the relationships leads tohierarchies that are amenable to a logical test (seebelow) through reference to the basic types of con-cept represented by the terms, for example, thoselisted in section 3.3. Every subordinate descriptorshould refer to the same basic kind of concept as itssuperordinate descriptor, that is, both the broaderand the narrower term should represent a thing, anaction, a property, etc.

Example:a) anatomy (a discipline) and central nervous

system (a body part that may be an object ofstudy of that discipline) represent differentkinds of concept; therefore, these terms can-not be related hierarchically;

b) central nervous system and brain both repre-sent body parts; these terms can therefore berelated hierarchically.

5.3.1 The Generic RelationshipThis relationship identifies the link between a

class and its members or species. A simple way toapply the test for validity described above is toformulate the statement “[narrower term] is a[broader term].” This relationship is also amenableto a logical ‘all-and-some’ test as shown in thefollowing diagram:

succulent plants

cacti

SOME A

The diagram indicates that some members of theclass ‘succulent plants’ are known as ‘cacti’ and

that all ‘cacti,’ by definition and regardless of con-text, are ‘succulent plants.’ This test would usuallyensure that a descriptor such as ‘cacti’ is not subor-dinated to a class such as ‘desert plants,’ since cactiare not desert plants by definition. This particularrelational link can be represented by the followingdiagram:

SOME SOM

desert plants

cacti

The diagram illustrates that some members ofthe class ‘desert plants’ are known as ‘cacti,’ andthat some, but not all, cacti are desert plants. Thesedescriptors should therefore be assigned to differ-ent hierarchies in the thesaurus, and both descrip-tors should be assigned to the same documentwhen indexing a work on “cacti as desert plants.”

5.3.1.1 Codes for the Generic RelationshipThe generic nature of a relationship may be

identified by the BT/NT coding described above,or if more refined coding is desired for the varioustypes of hierarchical relationship, by the followingabbreviations:BTG = Broader term (generic)NTG = Narrower term (generic)

Example:rats rodents BTG rodents NTG rats

5.3.2 The Whole-Part RelationshipThis relationship covers situations in which one

concept is inherently included in another, regard-less of context, so that the descriptors can be orga-nized into logical hierarchies, with the whole treatedas a broader term. This relationship can be appliedto several types of term; the four types enumeratedbelow are not intended to be exhaustive. In thefollowing examples, parts are indicated throughindention.a) systems and organs of the body

Example:nervous system

central nervous systembrainspinal cord

b) geographic locationsExample:Canada

OntarioOttawaToronto

RELATIONSHIPS

SOME

ALL


Page 18

c) disciplines or fields of discourseExample:science

biologybotanyzoology

d) hierarchical organizational, corporate, social, orpolitical structuresExamples:countries

states/provincescities

armiesdivisions (military)

battalionsregiments

5.3.2.1 Codes for the Whole-PartRelationship

The hierarchical whole-part relationship may beindicated specifically by the following abbrevia-tions:BTP = Broader term (partitive)NTP = Narrower term (partitive)

Example:central nervous system nervous system BTP nervous system NTP central

nervous system

5.3.2.2 Parts of Multiple WholesWhen a whole-part relationship is not exclusive

to a pair of terms, i.e., the part can belong to mul-tiple wholes, the name of the whole and its part(s)should be linked associatively rather than hierar-chically in the thesaurus (see section 5.4). Carbure-tors, for example, are parts of machines other thancars. Therefore, ‘cars’ and ‘carburetors’ should notbe linked in a whole-part relationship in a thesau-rus. See section 4.3.2.1 for further guidance on thispoint.

5.3.3 The Instance RelationshipThis relationship identifies the link between a

general category of things or events, expressed bya common noun, and an individual instance of thatcategory, often a proper name.

mountain regions —class— state capitals Alps —instances— Albany Himalayas TrentonIn the first example, the Alps and the

Himalayas are assigned to subordinate positionsin a hierarchy, yet they are neither kinds norparts of mountain regions, but represent specificexamples or instances.

5.3.3.1 Codes for the Instance RelationshipThe hierarchical instance relationship may be

indicated specifically by the following abbrevia-tions:BTI = Broader term (instance)NTI = Narrower term (instance)

Example:fairy tales

NTI CinderellaRumpelstiltskin

5.3.4 Polyhierarchical RelationshipsSome concepts belong, on logical grounds, to

more than one category. They are then said topossess polyhierarchical relationships.

Example:

stringed instruments percussion instruments

musical instruments

pianos

In this example, the term pianos is assigned tosubordinate positions on the basis of its genericrelationship to two broader terms, in other words,pianos would be an NT to both stringed instru-ments and percussion instruments.

In other cases, polyhierarchical links may bebased upon whole-part relationships.

Example:

biology chemistry

biochemistry

In some cases, polyhierarchical links can be basedon logically different relationships.

Example:

bones head

skull

In this example, the link between bones andskull is based upon the generic relationship (theskull is a kind of bone), whereas the link betweenhead and skull is based on the hierarchical whole-part relationship (the skull is part of the head).These relationships may be represented as

skull BTG bones BTP head.

5.3.5 Node Labels in HierarchiesWhen descriptors are arranged in hierarchies

(such as tree structures) in a thesaurus (see section

RELATIONSHIPS


Page 19

6.2), node labels may be inserted into arrays toshow the principles of division among a set ofsibling terms (descriptors that share a broader term).Although their function is similar to that of BTs,node labels are not descriptors, and must not be usedas indexing terms. They are often typographicallydistinguished from descriptors, e.g., through theuse of italics and/or enclosure in angle brackets.

Example:cars by motive power diesel cars electric cars by purpose racing cars sports carsAdditional examples of node labels are in Figure

A8 and Appendix B. For the use of node labels inthe associative relationship, see section 5.4.3.

5.4 The Associative RelationshipThis relationship covers associations between

descriptors that are neither equivalent nor hier-archical; yet the terms are semantically or con-ceptually associated to such an extent that thelink between them should be made explicit in thethesaurus, on the grounds that it may suggestadditional descriptors for use in indexing or re-trieval. The relationship is symmetrical, and isgenerally indicated by the abbreviation RT (re-lated term). (Cf. section 5.4.4.)

Example:cells cytology RT cytology RT cellsThe associative relationship is the most difficult

one to define, yet it is important to make explicit thenature of the relationship between descriptorslinked in this way and to avoid subjective judg-ments as much as possible; otherwise, RT refer-ences may be established inconsistently.

As a general guideline, whenever one term isused, the other should always be implied withinthe common frames of reference shared by theusers of the thesaurus. Moreover, one of theterms is often a necessary component in anyexplanation or definition of the other; the term‘cells,’ for example, forms a necessary part of thedefinition of ‘cytology.’

Either of the following types of terms can belinked by the associative relationship:a) those belonging to the same hierarchy (see sec-

tion 5.4.1);b) Those belonging to different hierarchies (see

section 5.4.2).

5.4.1 Relationships Between DescriptorsBelonging to the Same Hierarchy

5.4.1.1 Relationships Between OverlappingSibling Terms

RT references are required for sibling terms withoverlapping meanings, such as ‘ships’ and ‘boats,’where each of the terms can be precisely defined (sothey do not form an equivalence set), yet they aresometimes used loosely and almost interchange-ably; the user interested in one should therefore bereminded of the other.

This link does not need to be made explicit in thesystematic section of a thesaurus containing orga-nized hierarchies (see section 6.2.3), since thismethod of display automatically brings such termstogether. The link should be indicated, however, inan alphabetical thesaurus, and in the alphabeticsection of a hierarchical thesaurus.

Example:boats ships BT vehicles BT vehicles RT ships RT boats

5.4.1.2 Relationships Between MutuallyExclusive Sibling Terms

It is not necessary to interrelate all sibling terms.For example, there is no need to associate termssuch as ‘roses’ and ‘daffodils,’ which share thebroader term ‘flowers,’ because the meaning of theterms does not overlap, i.e., they are mutuallyexclusive.

5.4.1.3 Derivational RelationshipsConcepts linked by a familial or derivational

relationship (i.e., one of the concepts is derivedfrom the other) also require RT references. Thisguideline applies to the relationship betweensuch terms as ‘parents’ and ‘children’ or to therelationship of ‘mules’ to ‘donkeys’ and ‘horses.’In the graphic display that follows, donkeys,horses, and mules are all subclasses of equines.In the alphabetic display, RT references are pro-vided between mules and donkeys as well asmules and horses. Horses and donkeys are notlinked by RTs because they do not share a deriva-tional relationship.

Example (Graphic display):

donkeys horses

equines

mules

RELATIONSHIPS


Page 20

Example (Alphabetic display):donkeys horses BT equines BT equines RT mules RT mulesequines mules NT donkeys BT equines NT horses RT donkeys NT mules RT horses

5.4.2 Relationships Between DescriptorsBelonging to Different Hierarchies

It is possible to establish many grounds forassociating descriptors belonging to differenthierarchies. Related term references are often madebetween etymologically related descriptors, i.e.,that contain the same root, but which do not repre-sent the same kind of thing. The following are somerepresentative examples of typical relational situ-ations. For guidance on coding the precise natureof a relationship, see section 5.4.4.a) Disciplines or fields of study and the objects or

phenomena studied, or the discipline’s prac-titionersExamples:mathematics mathematicians RT mathematicians RT mathematicsneurology nervous system RT nervous system RT neurologybotany plants RT plants RT botany

b) Operations or processes and their agents orinstrumentsExamples:temperature control thermostats RT thermostats RT temperature

controlhunters hunting RT hunting RT hunters

c) Objects or processes and their counteragentsExamples:plants herbicides RT herbicides RT plantsinflammation anti-inflammatory

agents RT anti-inflammatory RT inflammation agents

d) Actions and their productsExamples:weaving cloth RT cloth RT weavinglacrimation tears RT tears RT lacrimation

e) Actions and their targets

Examples:harvesting crops RT crops RT harvesting

binding books RT books RT binding

f) Objects or substances and their unique propertiesExamples:poisons toxicity RT toxicity RT poisons

liquids surface tension RT surface tension RT liquids

g) Concepts linked by causal dependenceExamples:death bereavement RT bereavement RT death

infections pathogens RT pathogens RT infections

h) Concepts and their units or mechanisms of mea-surementExamples:electric current amperes RT amperes RT electric current

temperature thermometers RT thermometers RT temperature

i) Phrases in which the noun is not a true broaderterm (see section 4.2e). For example, a ‘rubberduck’ is not a ‘duck.’Examples:ducks rubber ducks RT rubber ducks RT ducksfishes fossil fishes RT fossil fishes RT fishes

5.4.3 Node Labels for Related TermsIn order to bring closely related concepts together

in the alphabetical array under a given descriptor,related terms may be divided into categories that donot form part of a logical hierarchy. These relatedterms should then be identified by a node label.

Example:books RT

<operations>bindingprinting

In this example, operations functions as a nodelabel.

For the use of node labels to indicate principlesof division for logical hierarchies, see section 5.3.5.

5.4.4 Specifying Types of Related TermReferences

In certain thesauri, it may be considered desir-

RELATIONSHIPS


Page 21

able to refine related term references in order tomake the nature of the relationships explicit. Codesfor such relationship indicators and their recipro-cals may be developed locally. (See, for example,Figure A22.) These local codes should be clearlyexplained and illustrated in the introduction (seesection 6.3.2) or documentation (see section 7.4) tothe published or machine-readable thesaurus.

6. Print DisplayThe three basic types of format in a printed

thesaurus are: (a) alphabetical, showing all theimmediate relationships of each term; (b) permutedor rotated, giving access to every word in eachdescriptor and entry term; and (c) hierarchical,showing a display of all levels of hierarchies. Vari-ous versions of each format are illustrated by samplepages from thesauri in Appendix A.

A major principle in the design of the format ofa printed thesaurus is the minimization of doublelookup, i.e., the need to consult more than onesequence.

The most commonly used methods of present-ing equivalence relationships and term hierarchiesare described and evaluated in sections 6.1 and 6.2,respectively. A discussion of general format con-siderations follows in section 6.3.

6.1 The Equivalence RelationshipUSE references from nonpreferred terms should

be incorporated into the main alphabetical listingof a thesaurus rather than being relegated to anauxiliary “access vocabulary” or separate list ofentry terms.

6.1.1 Inverted USE References andPermuted Lists

Some thesauri include an auxiliary permuted orrotated list in KWIC or KWOC format that givesaccess to every word in descriptors and USE refer-ences (see Figures A2 and A3). The permuted listshould not be used as a substitute for the inclusionof useful inversions as USE references in the mainalphabetic list.

For example, it is conceivable that in a thesaurusin the domain of library and information science, auser will search the term ‘automation, library.’ Ifthe preferred term is library automation, a USEreference is essential. A reference from the invertedform of the term information science is, in contrast,not essential and would, moreover, be misleadingas it might be interpreted as ‘science information.’An automatically generated permuted display will,however, display the term information scienceunder science as well as under information.

6.1.2 Inverted USE References andHierarchical Display

The choice of format for the display of hierarchi-cal relationships (see section 6.2) affects decisionson the necessity for USE references. For example, ifnarrower terms are included in the alphabetic dis-play in a case such as

libraries NT academic libraries

public librariesspecial libraries

the following USE references may be unnecessary:libraries, academiclibraries, publiclibraries, specialIf, however, academic libraries has the NTs

college libraries and university libraries, in theflat format (see section 6.2.1), these would not bevisible under libraries; therefore, USE referencesfrom inversions of all the compound terms shouldbe made. Moreover, where a descriptor has numer-ous BTs, NTs, and RTs, an inverted USE referencemay be easier to find than the desired narrow term.

If hierarchical information is given only by a treestructure (see section 6.2.3), USE references frominversions would be desirable within the mainalphabetic sequence of the thesaurus.

6.1.3 Juxtaposition of TermsJuxtaposition of terms also plays a role in the

creation of USE references for a printed thesau-rus. For example, an entry term that would im-mediately precede or follow the descriptor towhich it leads may be suppressed. (This guide-line would not apply to thesauri for databasessearched by computer — see section 7.3.1.) Somesingular and plural forms of terms may be widelyseparated in a printed alphabetic list, however,and a USE reference from one to the other wouldthen be warranted, even if the policy regardingpreference for singular or plural is stated in theintroduction to the thesaurus (see section 6.3.2).In the following example, several terms occurbetween the singular and plural forms of ‘cat,’and so a reference from one to the other is desir-able, especially because the singular form is thefirst word of a compound descriptor.

Example:cat USE catscat fleascatalogscatharsiscats

PRINT DISPLAY


Page 22

The addition of descriptors that file betweenthe position of a potential entry term and itscorresponding descriptor, or the interfiling ofdescriptors with uncontrolled terms in an indexmay create a need for additional USE references.

Similar considerations apply to the establishmentof related term references in a printed thesaurus.

Most printed thesauri are produced from ma-chine-readable files that also generate a screendisplay. Because a screen displays less informa-tion than a printed page, a more generous entryvocabulary is required on a screen than on apage. The reader’s eye takes in the larger pan-orama of the printed page, and it is easier tolocate descriptors that are variants of the searchterm. A possible approach is to include all poten-tially useful references in the machine-readablefile and suppress unnecessary ones in the printedproduct (see section 7.3.1).

6.2 Hierarchical RelationshipsThe hierarchy of descriptors may be displayed

in a variety of ways. This section explains theadvantages and disadvantages of the most com-monly used formats for displaying hierarchicalrelationships among descriptors in print. Inden-tion should be used as a visual cue for hierarchicallevels in all the formats.

6.2.1 Conventional Flat ThesaurusStructure

The most commonly used thesaurus format ar-ranges all descriptors in alphabetical order andgives for each some or all of the following: scopenotes, used for (UF) references, terms that are onelevel broader (BT), terms that are one level nar-rower (NT), and related terms (RT). This is therecommended order of relationships. Narrowerterms are generally arranged alphabetically, butnode labels may be used to group both narrowerand related terms in categories (see section 5.3.5and Appendix B).

Indention should be used as an additional visualcue for narrower terms.

Example:vertebrates

NT mammalsreptiles

An example of the flat format is found in theERIC thesaurus (see Figure A4).

6.2.2 Multilevel Hierarchy WithinAlphabetic Display

This type of display format saves the time of

the thesaurus user in that the broadest and mostspecific descriptors relating to the sought termare immediately identified, whereas in the flatformat, the user must proceed one step at a time,going from descriptor to descriptor, to examinethe full hierarchy. In both formats described inthe following sections, the sibling terms of thesought descriptor can only be identified by goingto the broader term. Because the hierarchy isrepeated over and over again in these formats,they are far more space-consuming than the flatformat and hence less desirable for a printedthesaurus. If the hierarchies are not very deep,e.g., only three or four levels, the extra spaceoccupied may not be excessive.

6.2.2.1 Multilevel Broader and NarrowerTerms

This format differs from the flat format in that itgives all levels of broader and narrower termswithin the alphabetic display. It employs specialnotation such as BT1, BT2 (one level broader, twolevels broader) and NT1, NT2 (one level narrower,two levels narrower). A model for this format is theInternational Energy Subject Thesaurus (see Figure A5).

6.2.2.2 Generic StructureMultiple levels of hierarchy may be indicated

without BT/NT notation, by using the abbrevia-tion GS (generic structure) with indention andpunctuation marks such as period and colon ascues to the levels of hierarchy. A model for thelatter system is found in the NASA Thesaurus (seeFigure A6).

6.2.3 Tree StructureIn this format, each descriptor is assigned a

classification notation or line number, and thisleads the user from the alphabetic display to the fullhierarchical display (sometimes called ‘systematicdisplay’ or ‘classified display’) in a separate se-quence. Medical Subject Headings (MeSH) serves asa model for the format featuring classification no-tation (see Figure A7). The Art and ArchitectureThesaurus provides a model of a hierarchical dis-play with line numbers (see Figure A8), althoughthe system also employs classification notation.

There is no redundancy in this method of the-saurus display, and it is economical of space. Thepanorama of the complete hierarchy is a distinctadvantage to the user.

Another advantage of the tree structure is thatthe complete array of classes and subclasses servesas a check on the logical consistency of the

PRINT DISPLAY


Page 23

hierarchies. In the flat format or generic structure,hierarchical relationships may be built from thebottom up; in a tree structure, they are often builtfrom the top down.

The power of a tree structure may be fully ex-ploited in online searching (see section 7.3.3). It is,however, possible for a computer to build a com-plete hierarchy from a thesaurus employing a flatformat or generic structure.

6.2.3.1 NotationIf a hierarchical classification scheme is ap-

plied to a tree structure, its notation must becarefully developed so that it will be hospitableto interpolation at any level. Computer-gener-ated or humanly-assigned line numbers may beeasily revised when descriptors are added, butthe notation is not expressive, i.e., it does notreflect the levels of hierarchy.

Pure notation, consisting of either letters ornumbers, has a smaller base than mixed or alpha-numeric notation and is therefore less hospitableto new descriptors. Mixed notation may, how-ever, create problems for novice users, who mayhave difficulty locating the codes. This suggestsan advantage of the generic structure: all levelsof broader and narrower terms are displayed tothe user within the alphabetic list, and no secondlookup is required. Sibling terms are not dis-played in the generic structure, however, whilethey are an important feature of the tree struc-ture.

6.2.4 Graphic DisplaysGraphic displays are logically equivalent to a

tree structure, but may not have a notation (seeFigure A9). They are difficult to produce and main-tain, as well as being wasteful of space. A simplegraphic display may make it easier to take in therelationships of a descriptor at a glance. If too manyrelated terms and levels of hierarchy are included,however, the display becomes difficult to compre-hend.

6.2.5 The Top Term StructureIn this format, the alphabetic display includes all

the relationships found in the flat format, with theaddition of a relationship indicator for the top term(TT) of the hierarchy. The latter is equivalent to thebroadest term in a generic structure. The top termleads the user to a separate sequence of the thesau-rus in which top terms are arranged alphabetically,with each top term followed by all its narrowerterms, arranged hierarchically on various levels.

Since it does not include a classification notation, atop term structure does not provide guidance to thelocation of a descriptor within the array under itstop term. The Inspec Thesaurus provides a model forthis format (see Figures A10 and A11).

This format is more economical in terms of spacethan the generic structure, but like the tree struc-ture, requires a second lookup.

6.2.6 Two-Way Hierarchical StructureThis is generally appended to a flat format the-

saurus. Each descriptor is an access point, and alllevels of broader and narrower terms are displayed,generally without notation and with indention as acue to hierarchy. A model of this is found in theERIC thesaurus (see Figure A12). This format islogically equivalent to the generic structure, i.e., ina single display one can view all of the broader andnarrower terms of a descriptor. In combinationwith the flat format, the two-way hierarchical struc-ture requires more space than the top term struc-ture.

6.2.7 Broad CategoriesThe alphabetic arrays of some thesauri include

numbers that identify the broad category to whicheach descriptor belongs. The notation for the cat-egory may be searchable online or used for pur-poses of selective dissemination of information.Such thesauri generally feature a separate sectionthat displays the descriptors for each numberedcategory or subcategory in an alphabetic sequence,undifferentiated in terms of hierarchical levels. Anexample of such category numbers may be found inthe TEST thesaurus (see Figure A13).

6.3 General Format Considerations

6.3.1 Components of a Printed ThesaurusGiven the variety of displays available, it is not

possible to stipulate a standard form of layout for aprinted thesaurus. Minimally, every thesaurusshould include an alphabetic display.

Every published thesaurus should include a titlepage, a table of contents identifying all the sections,and an introduction (see section 6.3.2).

6.3.2 IntroductionAll printed thesauri should contain a compre-

hensive introduction that clearly statesa) the purpose of the thesaurus;b) the scope, i.e., the subject field(s) covered, with

core and fringe areas separately identified;c) the meaning of all conventions, abbreviations, and

PRINT DISPLAY


Page 24

any punctuation marks used in nonstandard ways;d) the rules and authorities adopted in selecting the

preferred forms of descriptors, and in establish-ing their relationships;

e) whether the thesaurus complies with a nationalor international standard for thesaurusconstruction;

f) the filing rules employed, citing an appropriatenational standard when used;

g) the total number of terms, with separate totalsfor descriptors and entry terms (the number ofdescriptors indicates the size of the indexing/searching vocabulary; the ratio of entry terms todescriptors indicates the accessibility of the the-saurus to end-users);

h) the date on which the computer file for thethesaurus was put into final form before print-ing, which may be the date of the last update;

i) a statement on the updating policy, and thename and address of the responsible organiza-tion to which comments and suggestions shouldbe sent.

6.3.3 TypographyDescriptors, nonpreferred terms, relationship

indicators, and textual notes should be typographi-cally distinguished. Suggested typographic speci-fications are: lightface or italics for nonpreferredterms, all capitals for relationship indicators suchas USE, and boldface for descriptors.

Example:teenagers USE adolescents

6.3.4 Capitals / Lowercase LettersThe thesaurus should serve as an orthographic

authority in addition to noting preferred termi-nology. A combination of capitals and lowercaseletters should therefore be used in thesaurusterms. Topical descriptors and entry terms shouldbe in lowercase; proper names should be capital-ized as in standard usage. Initialisms are gener-ally all uppercase; acronyms may have an initialcapital followed by lowercase letters.

6.3.5 FilingThis section provides general rules for the filing

of alphabetic characters and numerals in thesauri.Guidance on specific filing problems may be found inALA Filing Rules (1980), Library of Congress Filing Rules(1980), and British Standard Alphabetical Arrangementand the Filing Order of Numerals and Symbols (BS1749:1985). These filing codes are, however, not com-patible with each other in certain details, and only oneof them should be chosen as an authority.

6.3.5.1 Alphabetic CharactersTerms consisting of letters should preferably

be filed word-by-word rather than letter-by-let-ter. In word-by-word filing, a space is significant;this filing principle is called “nothing beforesomething.” It juxtaposes terms that begin withthe same word.

Examples:Word-by-word filing Letter-by-letter filinggas USE gases gas USE gasesgas coolants gas coolantsgas welding gasesgases gasketsgaskets gas weldingAnother advantage of word-by-word filing is

that the space is a significant character in computerprograms that may be used to sort terms and toprint the thesaurus.

A disadvantage of word-by-word filing is that itseparates agglutinated compounds from those con-sisting of two words, e.g., ‘folk song, folk tune,folklore, folktale.’ USE references from the variantspellings may compensate for this problem.

6.3.5.2 NumeralsSome operating systems file numerals at the

beginning of the alphabet and some at the end.Either position is acceptable, provided that thefiling position of numerals is explained in the intro-duction to the thesaurus.

Applying the principles of ascending order andword-by-word filing to numerals, a computer willfile 19 before 2, and 2001 before 30. A special sortroutine is required to get the computer to file onarithmetical value, e.g., 1, 1.5, 10, 151, 1000. Specialsorting instructions must be inserted by humans ifnumbers are to be ignored, e.g., in chemical terms,or filed as spelled out, e.g., 21 [twenty-one] gunsalute.

6.3.5.3 Non-Alphanumeric CharactersPunctuation marks should generally be ignored

in filing entry terms and descriptors. In the follow-ing example, the comma in the inverted USE refer-ence is ignored:

cat breedscat, Siamese USE Siamese catscatalogsParentheses within terms should be ignored in

filing, e.g., benzo(a)pyrene should be treated as if itwere spelled ‘benzoapyrene.’

Parentheses around qualifiers require specialtreatment. It is desirable for the user to see homo-

PRINT DISPLAY


Page 25

graphic terms juxtaposed to assist in the selec-tion of the desired term. Ignoring the parenthe-ses around qualifiers would lead to the dispersalof homographic terms among compound terms.The desired sequence is illustrated in the follow-ing example:

Earth (planet)earth (soil) USE soilearth science

6.3.6 Running HeadsWhenever a printed thesaurus includes mul-

tiple sequences, e.g., an alphabetic list, a rotatedlist, and a hierarchical display, each page shouldfeature a running head to identify the sequence.Running heads are useful even if the sequences arein separate volumes.

6.4 Multiple Versions of Printed ThesauriThesauri may be produced in various editions:

a basic list of terms, references, and relationshipsdesigned for the end-user or occasional searcher,and a more complete version designed for theindexer and the expert searcher, which mayinclude detailed scope notes, indexing instruc-tions, information on term history, and postingsdata.

A model for multiple print formats may befound in the two editions of Medical Subject Head-ings: MeSH as found in Index Medicus (also knownas “Black and White” MeSH), which is designedfor end-users, and the Annotated Alphabetic List,which is designed for indexers and searchers (seeFigures A14 and A15).

7. Screen Display

7.1 User CategoriesThe design of screen displays should take into

account the needs of each anticipated class of user:a) thesaurus maintainers who have expertise in

indexing and thesaurus construction, and arelikely to be experts in the subject domain of thethesaurus. They must have access to all viewsof a thesaurus and complete information abouteach term, with the ability to edit and manipu-late term records, cross-references, classifica-tion notation, and hierarchies. They require“housekeeping” displays not needed by end-users of a thesaurus. (See section 10 for speci-fications of thesaurus management software.)

b) expert searchers and indexers who have exper-tise in indexing, online information retrieval,

use of thesauri, or all of these. Indexers arelikely to have expertise in the subject domainof the thesaurus; expert searchers may or maynot have such expertise. These sophisticatedusers require the ability to search and viewcross-references, definitions, and notes fordescriptors as well as various levels of theclassification or hierarchies. Postings data areespecially important for searchers (see section7.3.1). Sophisticated thesaurus displays andterminology are appropriate for these users.

c) end-users who are not likely to be experiencedin the jargon and complexities of online infor-mation retrieval, but who can be expected tohave expertise in the subject field, and to un-derstand its terminology. Abbreviationsshould be avoided, or explained in a windowor at the bottom of the screen on which theyappear. The types of displays available to ex-pert searchers can be useful to end-users aswell, when designed with their needs in mind.End-users may benefit from on-screen instruc-tions in addition to any printed documenta-tion that may exist. It is important to take theneeds of each anticipated class of user intoaccount in designing screen displays.

7.2 Factors to Be ConsideredA thesaurus is generally used in the context of

an information storage and retrieval system, andthe choice of screen display formats will dependupon the domain, the searching behavior ofusers in that domain, and the overall design ofthe information system of which the thesaurus isa component.

Viewing information on a screen differs fromviewing printed information: with a screen, it isharder to browse and remember one’s context;the screen is more difficult and tiring to viewthan printed media; and the available screen“page” size can make it difficult to graspinformation that is perfectly comprehensible inprinted form.

A great deal of research has been done and isongoing concerning user search behavior andcognitive learning styles in the context of interac-tive search systems, library catalogs, videotex,and hypermedia. In addition to existing stan-dards in computer-human interaction, workinggroups are producing draft standards for userinterfaces, screen layouts, menus, windows, andother interface formats. These standards shouldbe applied rigorously to all user interface designs.See Appendix C for sources of standards and

SCREEN DISPLAY


Page 26

references to authoritative texts on interfacedesign.

7.3 Types of DisplayThe following types of display are recommended

for consideration (see also section 6). The selectionwill depend on the requirements of a given systemand its users.

7.3.1 Alphabetical ListingWhen a user enters a search term—whether it is

a descriptor, entry term, or term not in the thesau-rus—the system should display an alphabeticalsequence of terms preceding and following thesearch term. This list should contain both descrip-tors and entry terms with USE references to de-scriptors.

Spelling variations, common misspellings, andcommon typographical errors should be stored asUSE references in the machine-readable thesaurusto enhance retrieval, but need not be displayed tousers (see section 10.2).

Thesauri linked to databases may also displaythe number of postings for assigned descriptors.The postings data in a thesaurus display should notinclude free-text occurrences of single-word de-scriptors.

7.3.2 Permuted DisplaysRotated or permuted listings (e.g., KWIC and

KWOC displays) of multiword terms are usefulto all classes of users (see Figures A16 and A17).

7.3.3 Hierarchies and ClassificationsIt should be possible to display a hierarchy at

various levels. For example, if the hierarchy of afull thesaurus has five levels, it should be pos-sible to display an outline of the first three levels.Indention or other cues to the hierarchy are es-sential.

Users should be able to broaden or narrowtheir searches by selecting descriptors at variouslevels in the hierarchy, and to add some or allbroader or narrower terms for a descriptor. Forexample, beginning with the descriptor verte-brates, the user may request that the search beexpanded to include all of its narrower terms—atone or more levels—or only mammals and, op-tionally, its narrower terms. A tree structure withnotation facilitates such search expansion, al-though the same effect may be achieved withother structures. Alternatively, the system couldassign to each descriptor in a hierarchical displaya line number, which the user may refer to (rather

than rekeying terms) in expanding a search. High-lighting is another option to obviate rekeying.

Node labels in hierarchies, which are not to beused in indexing or searching, should be distin-guished from descriptors, e.g., by placing themin angle brackets. Node labels may also be as-signed line numbers.

7.3.4 Term DetailA user should be able to choose a term from

any listing and see an expanded view of thedetail for that term, either fully or in part. Usersshould have the option of calling up a descriptor’shistory, scope note, or definition, as well as allterm relationships — equivalence, hierarchical,and associative—plus any specialized relation-ships created for the thesaurus. These relation-ships should be appropriately labeled, avoidingabbreviations in the case of end-user displays(see Figure A18). Windowing techniques may beemployed effectively in term-detail screens.

7.3.5 Graphic DisplaysResearch has shown that graphic displays may

communicate relationships among concepts moreeffectively than linear displays to some users.Graphic displays of thesauri should be consid-ered, taking into account the domain and searchhabits of users. (Cf. section 6.2.4.)

7.4 DocumentationFor publicly available databases, thorough de-

scriptions of the thesaurus and its use should beavailable in the database documentation. Thebasic information sheets provided by databasevendors should explain the structure of the the-saurus and the commands for searching it. The-saurus software packages should provide clearlywritten, well-organized and indexed documen-tation for thesaurus maintainers, as well as forindexers and searchers where the thesaurus ispart of a complete information retrieval package.

7.5 Relation of Thesaurus Character Set toSearch Commands

Nonalphabetic characters used in a machine-readable thesaurus should not conflict withspecial characters used in search commands.For example, parentheses used in qualifiersshould not be interpreted as nesting indicatorsin a search statement. If the potential for ambi-guity exists, a different character should besubstituted, e.g., square brackets may enclosequalifiers.

SCREEN DISPLAY


Page 27

8. Thesaurus Construction

8.1 Avoidance of Duplicate WorkThe compiler should ascertain, through refer-

ence to one of the clearinghouses listed in Appen-dix D, whether an existing thesaurus covers thesame or an overlapping domain of knowledge.Complete duplication of subject coverage is rare,but access to one or more thesauri in related fieldscan frequently serve as a useful starting point. Forexample, a thesaurus developer in the field of mal-practice law should begin by consulting existingthesauri in medicine and law, which will surelycontain many of the terms and relationships re-quired for the more specialized, interdisciplinarythesaurus.

8.2 Determination of Structure andDisplay Format

If possible, the form of the thesaurus (e.g., flat,generic structure, hierarchical display, or graphicdisplay) should be decided before terms are col-lected and considered as candidates for inclusion.As demonstrated in section 6, the display format ofthe thesaurus affects the types of cross-referencesand relationship indicators that are provided. Forexample, if a tree structure with classification nota-tion is adopted, the thesaurus may not include BT/NT references. The format should therefore be se-lected before relationships among terms are con-structed. (Although it is possible to convert onethesaurus display format to another, one must be-gin with consistent coding of relationships in asingle system.)

8.3 MethodsThree initial approaches to thesaurus construc-

tion are recommended.

8.3.1 The Committee ApproachExperts in the subject domain of the thesaurus

draw up a list of the key terms in the field andindicate the relationships among them, with assis-tance from experts in thesaurus design.

8.3.2 The Empirical Approacha) The deductive method. Terms are extracted from

documents (by humans or computers; see sec-tion 8.4 for a discussion of machine assistance),optionally during a preliminary stage of index-ing, but no attempt is made to control the vo-cabulary, nor to determine relationships betweenterms, until a sufficient number of terms hasbeen collected. All terms are then reviewed by a

group of experts, preferably consisting of bothinformation and subject specialists. They shouldfirst identify terms that represent the broadestclasses, and then allocate remaining terms tothese classes on the basis of their logical relation-ships, so that the hierarchies tend to be estab-lished on a broader-to-narrower basis. Vocabu-lary control should be applied at the stage wherehierarchies are established, following the prin-ciples described in sections 3 and 4.

b) The inductive method. New terms are selected forpotential inclusion in the thesaurus as they areencountered in documents. Vocabulary controlis applied from the outset. Each term, as it isadmitted, is designated as a member of one ormore broader classes that are constructed on anad hoc basis at an early stage. The thesaurus istherefore established on a narrower-to-broaderterm basis. Thesaurus construction is regardedfrom the outset as a continuous operation. Assis-tance from subject experts is strongly recom-mended; these experts may serve as members ofa formal editorial board or committee.

8.3.3 Combination of MethodsIn practice, more than one of these approaches

is likely to be employed at one stage or anotherduring the construction of a thesaurus. For ex-ample, hierarchies of terms that were first estab-lished inductively may later be examined from adeductive viewpoint. Both techniques are essen-tially empirical, and it should be accepted fromthe outset that some decisions regarding the termsand their interrelationships that were made dur-ing the early stages of compilation may have tobe revised as further experience is gained.

The compilers should check terms and hierar-chies frequently to ensure consistent application ofprinciples in such procedures as the construction ofinter-term relationships and the splitting of com-pound terms.

8.4 Machine AssistanceIt is assumed throughout this standard that the-

saurus construction calls for intellectual decisions.Machine assistance can be employed, however, forterm identification tasks such as the following:a) Identification of candidate terms. Candidate terms

may be identified automatically from machine-readable text, e.g., titles and/or abstracts. The num-ber of potential terms should be reduced by useof a stop list (a list of function words—articles,conjunctions, and prepositions such as ‘a,’ ‘and,’and ‘for’—plus other words considered to be of

THESAURUS CONSTRUCTION


Page 28

no value for retrieval). Remaining terms shouldbe matched against those already recorded inthe thesaurus. Unrecorded terms may be consid-ered as candidates for inclusion.

b) Registering frequency of descriptor assignment. Incomputerized indexing systems, the frequencywith which a descriptor has been used in index-ing can be registered automatically. Descriptorswith exceptionally high or low scores can beconsidered as candidates for modification ordeletion.

c) Recording terms in user queries. Terms found inuser queries that do not match one or moredescriptors or entry terms may also be consid-ered for inclusion, especially when a given termoccurs in multiple queries. To ensure privacy,users should not be identified.

8.5 Term RecordsAn individual record should be created for ev-

ery descriptor, and optionally for every entry term,as soon as it is admitted into a thesaurus. Recordsfor entry terms may include source notes as well asthe date of admission into the thesaurus. For de-scriptors, the record may contain any or all of thefollowing elements:

• descriptor;• source(s) consulted for descriptors and entry

terms. This field is especially important forneologisms or unfamiliar terms; it may in-clude citations to published sources or thenames of personal authorities consulted;

• scope note;• synonyms, i.e., Used For references;• nondisplayable variations, e.g., common spell-

ing errors (see section 10.2);• broader terms;• narrower terms;• related terms;• locally established relationships (see, for ex-

ample, Figure A22);• category or classification number;• history note (generally coded HN), includ-

ing minimally the date added, as well as therecord of changes, if any (see section 9.3). Ifappropriate, the history note may includethe date discontinued, the term that suc-ceeded the descriptor, and/or the term thatpreceded it.

An example of a term record is in Figure A18.Section 10.10 discusses field definition in the-

saurus management systems.

8.6 Term VerificationBefore a descriptor is admitted into a thesaurus,

it should be validated according to the rules insection 3 concerning scope, form, and choice ofdescriptors. The compiler should also review rela-tionships between each new descriptor and otherdescriptors in the hierarchy to which it is assigned.The following types of authority within the subjectdomain should be checked before candidate termsare accepted for inclusion:a) technical dictionaries, glossaries, scholarly mono-

graphs, reference texts, and encyclopedias;b) existing thesauri;c) classification schemes.

In problematic cases, expert advice should besought on the selection of a descriptor from variantforms of terms.

8.6.1 Candidate TermsCandidate terms are proposed descriptors that

have not gone through all acceptance procedures.These terms should be marked by a special symbolor phrase in the term record. As soon as a candidateterm is approved as a descriptor, the symbol orphrase should be deleted.

In an online system in which the thesaurus islinked to a single database, candidate terms aregenerally not displayed to users; in thesauri thatare not linked in this way, candidate terms may bedisplayed.

8.6.2 Provisional TermsWhere no published authority exists, descrip-

tors for new concepts may be labeled provisional.Provisional terms are checked periodically to de-termine whether there is literary warrant for estab-lishing them as descriptors. When provisional termsare upgraded to descriptors, the label “provisional”is removed.

8.7 Level of SpecificityThe addition of highly specific descriptors is usu-

ally restricted to the core area of the subject fieldcovered by a thesaurus, because the proliferation ofsuch descriptors in fringe areas is likely to lead to athesaurus that is difficult to manage (see section5.2.4.1). Although the cost of computer storage of avery large thesaurus may be insignificant, the humancost of establishing relationships among numerousdescriptors at the periphery of the domain is high.

In an organization that deals with documentscovering more than one domain of knowledge, itmay be necessary to develop a number of special-ized thesauri, each linked to, and compatible with,

THESAURUS CONSTRUCTION


Page 29

a general thesaurus that has a lower level of speci-ficity, produced by the same organization. A modelfor this may be found in the Macrothesaurus of theUnited Nations. (See also section 10.12.2.)

8.8 Unassigned DescriptorsWhen hierarchies are being established in a the-

saurus, descriptors that have not yet been used inindexing are frequently admitted into the thesau-rus on the grounds that they are necessary tocomplete a hierarchy (for example, as broaderterms), and that they have potential value as index-ing terms in their own right, where the thesaurus islinked to a particular database. These descriptorsshould be marked by a special symbol in the hier-archical display and in the term record, or by aphrase such as “not yet used.” (The phrase wouldnot be necessary in cases where all thesaurus dis-plays and term records indicate the number ofpostings.) As soon as the descriptor is used inindexing, the symbol or phrase should be deleted.

8.9 Announcement and Deposit ofPublished Thesauri

8.9.1 Notification of IntentWhen an organization has decided to compile

and publish a new thesaurus, it should submit anotice of intent to an appropriate professional jour-nal, to minimize the possibility of two organiza-tions simultaneously developing thesauri in thesame domain.

8.9.2 Pilot ThesaurusA new thesaurus should be tested by means of

a series of trial runs prior to publication in printedor electronic form. A pilot version should bedistributed to a select group of users, and sugges-tions for changes of terms and their relationshipsshould be considered by the editors. The reviewof established thesauri is treated in section 9.1.

8.9.3 Deposit with a ClearinghouseA copy of the first and any subsequent edition of

a printed thesaurus or a printout of a thesaurusmaintained only in electronic form should be de-posited with the appropriate clearinghouse (seeAppendix D).

9. Thesaurus MaintenanceThis section suggests procedures for adding,

modifying, and deleting terms. The term recordshould note the date of each change and identifythe individual responsible for it.

9.1 Updating the VocabularyThesauri are reflections of language, and they

are therefore dynamic instruments. Policies andprocedures should be established for periodic re-view of terminology, establishment of new de-scriptors, and replacement of obsolete descriptors,especially in fields where the terminology changesrapidly. Thesaurus editors should update the the-saurus at intervals that will be determined by thefrequency and volume of changes made, and by themethod of distribution. Printed revisions will nec-essarily be produced less frequently than revisionsto a computerized system, in which updates can beaccessible more quickly than is possible with print— instantaneously in some systems.

9.1.1 Addition of DescriptorsWhenever an appropriate descriptor or combi-

nation of descriptors cannot be found in a thesau-rus, the indexer or searcher should nominate a newdescriptor as a candidate term (see section 8.6.1).The nomination, including an adequate definitionof the term, may be submitted electronically, possi-bly via an on-screen form, or by means of a printedform. A sample form is in Figure A19. New termsand those nominated for reconsideration should bereviewed by the editor of the thesaurus, and pref-erably by an editorial board that functions in thesame manner as the group responsible for the origi-nal edition. The following decisions should be madefor each term:a) Should this term be added to the thesaurus, or is

it already covered by an existing descriptor orcombination of descriptors? If it is covered,should an additional USE reference be added?

b) Are scope notes or other notes needed? (Seesection 3.2.2.)

c) If the term is to be added, what is the correctform, considering syntax, number, and spell-ing? (See section 3.)

d) How should the new term be interrelated withexisting descriptors on a hierarchical and re-lated-term basis? (See section 5.)

9.1.2 Modification of Existing DescriptorsIndexers and searchers should also be able to

propose modifications to existing descriptors ortheir relationships, explaining the rationale andsupplying supporting documentation for the pro-posed changes. Like candidate term nominations,such proposals may be communicated electroni-cally or via printed forms. (Figure A19, ThesaurusTerm Review Form, accommodates both candidateterms and modifications to existing descriptors.)

THESAURUS MAINTENANCE


Page 30

Such proposed changes should be considered bythe thesaurus editor and board, using the criteriafor descriptor selection in sections 3 and 4. If adescriptor is modified, the date of the changeshould be recorded in the history note (see sec-tion 8.5), and a USE reference should be madefrom the old form to the new form.

If the thesaurus is used in an indexing system,the date on which an old descriptor was last as-signed should be included in the history note. If therelationships are modified, a record of the old onesshould be maintained in the history note as well.

9.1.3 Deletion of TermsOverused descriptors and descriptors that have

been assigned very infrequently in indexingshould be considered candidates for modifica-tion or deletion, as both kinds of term are gener-ally ineffective in retrieval. In some cases, anoverused descriptor can be replaced by two ormore descriptors of greater specificity, or thesecan be added to it as narrower terms.

9.2 Vocabulary Updates and DatabaseRecords

Whenever a descriptor is modified or deletedfrom a thesaurus, the impact of the change on theability to search previously indexed databaserecords must be considered, unless the modified ordeleted descriptor has never been assigned. Theproblem of searching modified or deleted descrip-tors can be handled in several ways. The choice ofmethod depends upon the size of the database,available human and financial resources, the na-ture and number of changes to be made, and whetherautomated procedures exist.a) A deleted or modified descriptor may be re-

tained in the thesaurus for retrieval purposesonly. It should be marked, e.g., “for retrievalpurposes only,” and the date of its change instatus should be recorded in the history noteand displayed to users. If it is superseded by anewer descriptor, a cross-reference to the newdescriptor should be provided. The supersededdescriptor should also be listed in the historynote of the newer descriptor, with the dates forwhich the former is valid.

b) One-for-one descriptor changes, where one termsimply replaces another, can be implemented inthe database automatically if resources permit.In very large systems, the cost of such file main-tenance may be prohibitive. The history note inthe term record then serves as a guide for retro-spective searching.

c) More complicated term changes require humanintervention, e.g., where two or more specificdescriptors are added to replace a more generalterm or become narrower terms to it.Deleting a descriptor in the database without

replacing it with another descriptor may, over time,seriously reduce access to documents if no replace-ment indexing is carried out. If it is discovered,however, that a descriptor has been establishedand assigned in error, once the descriptor has beendeleted from database records, its record may bedeleted from the thesaurus as well.

9.3 History of ChangesEach descriptor requires a history note record-

ing its date of entry (see section 8.5). The noteshould also track the history of modifications (seesection 9.1.2), recording previous forms with rea-sons for the change. Where obsolete descriptors areretained in the thesaurus but are not assigned inindexing, the date of replacement and reasons for itshould be given. Where terms are deleted com-pletely, it may be desirable to maintain a separatefile of deleted descriptors, retaining a record oftheir scope and history in order to track suchdecisions in case a term from this group is laterreconsidered.

Examples of history notes of various types andwith varying degrees of detail are found in the figuresin Appendix A:• Figure A1: The Code DA notes date of entry into

the thesaurus.• Figure A4: The entry:

Action Programs (Community) (1966 1980) USE COMMUNITY ACTION

indicates the dates for which the former descriptorwas valid.• Figure A5 illustrates the opposite pattern in the

last entry:FEDERAL REGION III

(Prior to June 1982 this concept was in-dexed to CENTRAL REGION.)

• Figure A14 documents a change in an entryterm:

FERTILIZATION IN VITROApr 79; TEST-TUBE FERTILIZATIONwas TEST TUBE BABIESsee FERTILIZATION IN VITRO Apr-Dec 1979X TEST-TUBE FERTILIZATION

• Figure A14 also provides an example of therecord of a change in status of a descriptor:

FERULA was see under PLANTS, TOXIC 1986-90

This history note means that, for the years

THESAURUS MAINTENANCE


Page 31

indicated, the descriptor was not authorized foruse in the printed index, only in the online data-base. The see under reference led the indexer orsearcher to a broader descriptor that was autho-rized for use in the print index.

10. Thesaurus Management SystemsBecause thesauri are generally used in the con-

text of information storage and retrieval systems,the design or choice of a thesaurus managementsystem is necessarily dependent on and should bean integral part of the overall design of the informa-tion system of which it is a component.

Maintainers of thesauri should have access tosystems that can deal with the special needs of athesaurus. It should not be necessary to adapt thethesaurus to an inadequate system. The systemshould be flexible enough to allow thesaurusmanagers to take advantage of emerging tech-nologies.

The following are general recommendations forfeatures of thesaurus management software to beused by thesaurus maintainers. See also the recom-mendations in section 7 regarding screen displayfor users.

10.1 Typography and SortingThe system should allow printing or screen dis-

play in both capitals and lowercase letters, and itshould be able to handle any special charactersrequired by the subject field or language of thethesaurus. (See section 3.7, Capitalization, Punc-tuation, and Nonalphabetic Characters.) The sortsequence should treat the capital and lowercaseversions of a letter as equivalent and should com-ply with the filing guidelines presented for printedthesauri in section 6.3.5.

10.2 Term Records and Display FormatAt minimum, term records (see section 8.5)

should contain the data and specify the relation-ships recommended in this standard; but an indi-vidual thesaurus may require additional data, suchas term weights for retrieval, special relationships,or codes from standards such as the USMARCAuthority Format. The system should be capable ofhandling hierarchical thesaurus displays, whererequired.

The system should also be able to distinguishbetween the nonpreferred terms (USE references)that should be displayed to thesaurus users andthose that should not be displayed (e.g., closespelling variations and common typographicalerrors entered to enhance computerized retrieval).

10.3 System ConstraintsThere should be no limits on the number of

characters in a term, the number of levels in ahierarchy, or the number of relationships(broader, narrower, associative, or equivalence)that can be established for a given descriptor.

10.4 SearchingThesaurus maintainers, indexers, and interme-

diaries should be able to search for and retrievecomplete or truncated terms, includingnonpreferred terms. System sensitivity to capitalsand lowercase letters may lead to retrieval failure;a search term in capitals should therefore match alowercase descriptor or entry term and vice versa.

Browsing displays should be available andshould include hierarchical, alphabetical, and (op-tionally) KWIC or KWOC displays. (The need forthe last type of display may be obviated by a text-word search capability.) From any of these dis-plays, it should be possible to view a term in thecontext of its relationships as well as a completeterm record.

Another desirable feature, particularly in largesystems, is to allow for retrieval of descriptors andentry terms on the basis of quantitative characteris-tics, such as the number of characters or words in aterm. Such data are useful for the determination ofrequirements for field length in term records, forthe calculation of system capacity, as well as forprojections of growth of the thesaurus.

10.5 EditingMaintainers should be able to create and edit

terms and to navigate editing screens easily usingtemplates or prompts. When new relationships arebeing entered, or old ones deleted, the existingrelationships should be visible on the screen. Itshould be possible to edit a descriptor’s position ina hierarchy, and to simultaneously display mul-tiple hierarchies for a descriptor where these exist.

Editing capabilities should include standardword processing features, i.e., the ability to add,modify, and delete characters, words, and lineswithout rekeying an entire field, and to navigateeasily backward and forward between fields andrelationship types.

A sample menu from a thesaurus managementsystem is shown in Figure A20.

10.6 Error CheckingThe system must be able to detect duplicate

terms, whether entry terms or descriptors.Duplicates detected by the system should be

THESAURUS MANAGEMENT SYSTEMS


Page 32

displayed to the thesaurus editor, who will judgewhether they are true duplicates or homographsrequiring qualification. The system must be able toidentify duplicates with typographic variations(capitals and lowercase letters) for qualification,e.g., mercury (metal) and Mercury (planet). Itshould be able to check for inappropriate codingand content, as well as errors in term relationshipsand inconsistencies in the modification or deletionof terms, as described in the following sections.

10.7 Cross-ReferencesWhen cross-reference links are created between

terms, the system should:a) check for the existence of a term chosen as a

broader, narrower, related, or USE term, andcheck for the validity of references, so that:

• no cross-references are made to nonexistentterms;

• the only type of reference that leads from anonpreferred to a preferred term is a USEreference. Nonpreferred terms have no BT,NT, or RT references;

• a term is not related to itself;• no conflicting references are made for the

same descriptor pairs (e.g., both a BT/NT andRT/RT reference);

• duplicate cross-references are eliminated.

b) create reciprocal references for BT/NT (orBTG/NTG and BTP/NTP), RT/RT, and USE/UF pairs. If the reciprocal descriptor for a BT,NT, or RT reference does not exist, the systemshould inquire whether a new descriptorshould be created, but should not create itautomatically. However, if the reciprocal termof a UF reference does not exist, the systemmay create that term. See section 5.1.2 forlogical formulations of the relationships.Examples:editor creates passenger cars

UF carssystem creates cars

USE passenger carseditor creates motor vehicles

NT passenger carssystem creates passenger cars

BT motor vehiclesc) maintain these references reciprocally, so that

when a term is modified or deleted, the refer-ences to it are changed accordingly in the recordsfor all the terms related to it. For example, if the

editor changes passenger cars to automobiles inthe above example, the system should make thesame change throughout the thesaurus, e.g.,motor vehicles NT automobiles.

d) When a thesaurus is integrated with a hyper-media system, the appropriate standards forchecking and maintaining hypermedia linksapply.

10.8 Term DeletionWhen a term is deleted, the system should first

display the term and then prompt for verificationof the deletion. The system must make deletionsreciprocally so that all links to and from deletedterms are deleted with the term (see Figure A21).The system should inform the thesaurus editorwhen the deletion creates an orphan term (adescriptor without associative or hierarchical re-lationships).

10.9 Candidate TermsThe system should include a mechanism for the

entry of candidate terms, designated as such, andcandidate relationships, which are not admitted tothe thesaurus until reviewed and approved bythesaurus editors. It should be possible to retrieve,display, and print out these candidate terms andrelationships. Candidate relationships should bereciprocated automatically. Whoever enters theproposed term should be able to record its sourceand to enter suggested relationships and a pro-posed scope note, because that person may belooking at relevant explanatory material at the timeof entry. The structure of candidate term recordsshould be the same as that of descriptor records(see section 8.5), because the former may be up-graded to the latter.

It is also useful for the system to be able tocapture terms used by database searchers, as thesemay be useful entry terms for the thesaurus. Thesystem may also provide a mechanism for search-ers to propose candidate terms directly.

A merge or transfer function that allows for thecomparison and merger of two similar terms can beuseful in choosing among candidate terms enteredby different indexers, or where revisions of a the-saurus are necessary in a quickly changing envi-ronment. The preferred term receives the relation-ships and notes from the nonpreferred term, andthe nonpreferred term becomes a USE reference.

10.10 Field DefinitionThe system should provide for the definition of

fields other than those enumerated for term records



Page 33

(section 8.5), for purposes such as assigning codesfrom other systems to descriptors, e.g., StandardIndustrial Classification (SIC) codes to the namesof industries; or providing term definitions as wellas scope notes. (Scope notes often differentiatebetween two descriptors for indexing purposes;definitions give the dictionary meaning.)

It may also be useful to be able to define andname specialized relationship types which aremore specific than USE/UF, and BT/NT or RT/RT, e.g., designating abbreviations, or express-ing the exact nature of various types of narrowerterms, such as the subsidiary of a company. Seesection 5 for a discussion of term relationships,and Figure A22 for an example of a screen dis-play featuring specialized codes.

10.11 ReportsIt should be possible to obtain printouts and

screen displays in whatever format the userrequirements for an individual thesaurus dictate.(See section 6 on print displays and section 7 onscreen displays.) In addition to reports designedfor both users and maintainers, thesaurusmanagers need specific “housekeeping” reports.These reports may be displayed on a screen orprinted, but in either case, they should be clearand legible and should use space and paper aseconomically as possible. Section 6 discusses thespace implications of various printed thesaurusdisplay formats. Screen displays should followthe guidelines in section 7. Thesaurus reportsmay include, but are not limited to, the following:a) term listings, including:

• terms added or modified since a certain date;• descriptors that do or do not have a particular

type of note or relationship (SN, UF, BT, NT,or RT);

• orphan terms, with no hierarchical or associa-tive relationships;

• top terms; b) displays of term relationships, including:

• all the relationships for an individual term;• hierarchical displays;

• listings at various levels of a hierarchy;• reports of special links created for a given

thesaurus;• the entire hierarchical context for an indi-

vidual term;c) statistical reports on terms and their charac-

teristics, including:• the number of descriptors and of all terms in

the thesaurus;• the number of terms featuring a certain char-

acteristic, for example, those that consist ofthree words or those that end in a certainstring of characters, e.g., itis;

• the number of terms added, modified, de-leted, or merged since a certain date;

• the number of BTs, NTs, UFs, RTs, SNs, orother value in the thesaurus;

• the average number of characters per termand the size of the longest term;

• postings reports for thesauri linked to data-bases.

10.12 Subsets

10.12.1 Copying FunctionIt is useful to be able to copy a subset of terms

from one thesaurus to another where more thanone thesaurus is being maintained on a system.

10.12.2 MicrothesauriOmnibus indexing services (organizations that

produce multiple specialized indexes through asingle indexing operation) or documentation cen-ters may maintain a thesaurus of broad scope,subsets of which, known as microthesauri, areused in specific indexing products. When suchsubsets are extracted, only relationships to de-scriptors in the microthesaurus should be in-cluded. A microthesaurus must be able to standalone.

Microthesauri may contain additional highlyspecific descriptors required for their application,but such descriptors should map to the hierarchicalstructure of the broad thesaurus.



Page 34

ReferencesThis list includes sources cited in the text as well as in the Glossary, which follows. References to olderthesaurus standards that are not cited in the text are found at the end of the Foreword. A bibliography ofmanuals of thesaurus construction is in Appendix E.

[ISO 5964] Documentation—Guidelines for the Es-tablishment and Development of Multilingual The-sauri. Geneva: International Organization for Stan-dardization, 1985.

Library of Congress Filing Rules. Prepared by JohnC. Rather and Susan C. Biebel. Washington, D.C.:The Library, 1980.

Macrothesaurus for Information Processing in theField of Economic and Social Development. 3rded., prepared by Jean Viet. New York: United Na-tions, 1985.

[MESH, 1992a] Medical Subject Headings: Anno-tated Alphabetic List, 1992. Bethesda, MD: Na-tional Library of Medicine, 1991. (NTIS: PB 92-100031).

[MESH, 1992b] Medical Subject Headings: Supple-ment to Index Medicus vol. 33, 1992. Bethesda, MD:National Library of Medicine, 1992. (NIH Publica-tion No. 92-265).

[MESH, 1992c] Medical Subject Headings: TreeStructures, 1994. Bethesda, MD: National Libraryof Medicine, 1993. (NTIS: PB 94- 100054).

NASA Thesaurus. NASA, Scientific and TechnicalInformation Branch. Springfield, VA: NTIS, 1988. 3vols. Vol. 1: Hierarchical Listing. Vol. 2: AccessVocabulary. Vol. 3: Definitions.

[TEST] Thesaurus of Engineering and ScientificTerms. New York: Engineers Joint Council, 1967.

USMARC Format for Authority Data, IncludingGuidelines for Content Designation. Network De-velopment and MARC Standards Office. Washing-ton, D.C.: Cataloging Distribution Service, Libraryof Congress, 1987.

[AACR2] Anglo-American Cataloguing Rules, 2nded. Rev. Michael Gorman and Paul W. Winkler,Eds. Chicago: American Library Association, 1988.

ALA Filing Rules. Chicago: American Library As-sociation, 1980.

ALA-LC Romanization Tables. Washington, DC:Library of Congress, 1991.

[ANSI Z39.4] American National Standard for Li-brary and Information Sciences and Related Pub-lishing Practices - Basic Criteria for Indexes. NewYork: American National Standards Institute, 1984.

Art and Architecture Thesaurus. Second edition.Toni Petersen, Director. New York: Oxford Univer-sity Press, l994. 5 vols.

British Standard Alphabetical Arrangement andthe Filing Order of Numerals and Symbols. Lon-don: British Standards Institution, 1985. (BS1749:1985).

[ERIC] Thesaurus of ERIC Descriptors. 12th edi-tion. James E. Houston, Ed. Phoenix, AZ: OryxPress, 1990.

Inspec Thesaurus, 1993. London: Institution of Elec-trical Engineers, 1993.

International Energy Subject Thesaurus. Oak Ridge,TN: U.S. Department of Energy, Office of Scientificand Technical Information, 1990. (ETDE/PUB-2;DE90008750).

[ISO 2788] Documentation—Guidelines for the Es-tablishment and Development of MonolingualThesauri, 2nd ed. Geneva: International Organiza-tion for Standardization, 1986.

[ISO 5127/1] Information and Documentation -Vocabulary - Part 1. Basic Concepts. Geneva: Inter-national Organization for Standardization, 1983.

[ISO 5127/6] Information and Documentation -Vocabulary - Part 6. Documentary Languages.Geneva: International Organization for Standard-ization, 1983.

REFERENCES


Page 35

Glossary of TermsThe following definitions are derived from several related standards and specialized dictionaries, but allhave been modified for the purposes of this standard. References to other standards appear in brackets.Underlined terms within definitions are defined elsewhere in this glossary.

assignment indexing. An indexing method bywhich descriptors or subject headings from a con-trolled vocabulary are selected by a human orcomputer to represent the concepts or items in adocument. The terms may or may not occur in thedocument. Cf. derivative indexing.

authority file. A set of records of establishedheadings and the cross-references to be made toand from each heading, often citing the authorityfor the preferred form or variants. Types of author-ity files include name authority files, subject au-thority files, and thesauri.

autoposting. See up-posting.bound term. A descriptor consisting of a

compound term. (The term was originated byMortimer Taube in his Studies in Coordinate Index-ing, vol. 1, 1953, p. 43.)

broader term. A descriptor to which anotherdescriptor or multiple descriptors are subordinatein a hierarchy. The relationship indicator for thistype of term is BT.

candidate term. A term considered for admis-sion into a thesaurus because of its potentialusefulness as a descriptor. Cf. provisional term.

category. A grouping of descriptors that aresemantically or statistically associated, but whichdo not constitute a strict hierarchy based on genus-species or part-whole relationships.

classified display. See tree structure.compound term. A term consisting of more than

one word. A compound term may be selected as adescriptor.

concept. A unit of thought, formed by mentallycombining some or all of the characteristics of aconcrete or abstract, real or imaginary object. Con-cepts exist in the mind as abstract entities indepen-dent of terms used to express them.

controlled vocabulary. A subset of the lexicon ofa natural language. A list of preferred andnonpreferred terms produced by the process ofvocabulary control. Types of controlled vocabu-lary include subject heading lists and thesauri.

cross-reference. A direction from one term toanother. Cross-references are of three types: (a)equivalence relationship, beginning with the word seeor USE, which leads to one or more descriptors thatare to be used instead of the term from which thecross-reference is made; (b) associative relationship,

beginning with the words see also, or related term(RT), which leads from one descriptor to otherdescriptors that are related to or associated withit in the context of a thesaurus; (c) hierarchical,beginning with the words broader term (BT), ornarrower term (NT), which represent generic andspecific relationships, respectively. (It is possibleto display hierarchical relationships without theuse of cross-references.)

derivative indexing. An indexing method bywhich words occurring in the title or text of adocument are extracted by a human or computerto serve as indexing terms.

descriptor. A type of heading that is a termchosen as the preferred expression of a concept ina thesaurus.

difference. In a compound term, one or morecomponents that serve to narrow the extension ofa focus and specify one of its subclasses. Alsoknown as ‘modifier.’

document. Any item, printed or otherwise,that is amenable to cataloging and indexing. Theterm applies not only to written and printedmaterials in paper or microform versions (e.g.,books, journals, maps, diagrams), but also tononprint media (e.g., machine-readable records,transparencies, audiotapes, videotapes) and, byextension, to three-dimensional objects or realia(e.g., museum objects and specimens).

entry term. The nonpreferred term in a cross-reference that leads to a descriptor in a thesaurus.Also known as ‘lead-in term.’ The relationshipindicator for this type of term is USE; its reciprocalis UF (used for).

entry vocabulary. The set of nonpreferred terms(USE references) which lead to descriptors in athesaurus. (This term is used by some thesaurusdesigners to represent the preferred as well as thenonpreferred terms in a thesaurus.)

eponym. A term incorporating the name of areal or mythical person, generally the discovererof a phenomenon or inventor of an object, e.g.,Herculean labor, Parkinson’s disease, pasteuriza-tion.

facet indicator. See node label.false drop. An irrelevant reference retrieved

when natural language terms or descriptors arepostcoordinated.

GLOSSARY


Page 36

flat format. An alphabetical display format ofthesauri in which only one level of broader termsand one level of narrower terms are shown for eachdescriptor.

focus. In a compound term, the noun compo-nent that identifies the class of concepts to whichthe term as a whole refers. See also difference.

free text. Antonym of controlled vocabulary.Natural language terms appearing in documents,which may complement descriptors in an informa-tion storage and retrieval system. In free text search-ing, descriptors may also be retrieved. Cf. key-word.

generic posting. (1) In thesauri, the treatment ofnarrower terms as quasi-synonyms, e.g., furnitureUF beds; UF sofas. (2) In indexing and subjectcataloging, the assignment of a broader term in-stead of a specific term, e.g., furniture to a docu-ment on sofas. Cf. up-posting.

generic structure. A thesaurus format that indi-cates all hierarchical levels of descriptors within analphabetic display by means of codes, indention,and/or punctuation marks.

gloss. See qualifier.heading. A preferred name or term. Types of

heading include proper name headings (which maybe called identifiers), subject headings, and de-scriptors. A heading may include a qualifier.

hierarchy. Generic (broader)-specific (narrower)or whole-part relationships, which are generallyindicated in a thesaurus through codes or inden-tion. See also broader term; narrower term.

history note. A note in a term record in athesaurus that provides the date of entry of adescriptor as well as the history of modificationsto its scope, relationships, etc.

homograph. One of two or more words thathave the same spelling, but different meanings andorigins. In thesauri, homographs are generally dis-tinguished by qualifiers.

identifier. (1) A proper name (or its abbreviationor acronym) of an institution, person, place, object,or process, optionally treated as a category of head-ing distinct from descriptor. Identifiers may beheld in a separate file (cf. authority file), and theirform may be controlled (e.g., the name of an inter-national organization having different names invarious languages, only one of which is selected).(2) In some systems, a provisional term that may beupgraded to descriptor status, or a highly specificterm that is not eligible for descriptor status, butwhich is considered useful for retrieval and is as-signed to one or more documents without vocabu-lary control.

indexing. An operation intended to representthe results of the content analysis of a document bymeans of a controlled indexing language or bynatural language. [ISO 5127/1]

indexing language. A controlled vocabulary orclassification system and the rules for its applica-tion. An indexing language is used for the repre-sentation of concepts dealt with in documents andfor the retrieval of such documents from an infor-mation storage and retrieval system. [ISO 5127/1]

indexing term. The representation of a conceptin an indexing language, generally in the form of anoun or noun phrase. Descriptors, subject head-ings, and heading-subheading combinations areexamples of indexing terms.

information storage and retrieval system. Aset of operations and the associated equipment,software, and documentation by which docu-ments are indexed and the records are stored, sothat selected records can be retrieved in responseto requests employing commands that can behandled by the system.

keyword. A word occurring in the natural lan-guage of a document that is considered significantfor indexing and retrieval. Cf. free text.

KWIC (Key Word In Context) index. A type ofindex, arranged alphabetically, in which each sig-nificant word in a string of text serves as an accesspoint, by being graphically emphasized and sur-rounded by the rest of the string. The keyword isgenerally in a centered column and is followed onthe right by the continuation of the string, whichprovides the context. The balance of the string, ifany, is positioned to the left of the keyword. AKWIC index of descriptors and entry vocabularymay serve as an auxiliary alphabetic display in athesaurus.

KWOC (Key Word Out Of Context) index. Atype of index, arranged alphabetically, in whicheach significant word in a string of text serves asan access point, usually positioned in the left-hand column of a page, followed by the completestring. The keyword may therefore not be in theimmediate context of the words that surround it.A KWOC index of descriptors and entry vocabu-lary may serve as an auxiliary alphabetic displayin a thesaurus.

lead-in term. See entry term.lexeme. A lexical unit in a language; a vocabu-

lary item, which is not necessarily limited to asingle word.

literary warrant. Justification for the representa-tion of a concept in an indexing language or for theselection of a preferred term because of its frequent

GLOSSARY


Page 37

occurrence in the literature. Cf. user warrant.microthesaurus. A subset of a thesaurus, cover-

ing a limited range of topics within the domain ofthe thesaurus. A microthesaurus may contain highlyspecialized descriptors that are not in the broadthesaurus. Such descriptors should map to the hier-archical structure of the broad thesaurus. Amicrothesaurus is internally consistent with re-spect to relationships among terms.

modifier. See difference.narrower term. A descriptor that is subordinate

to another descriptor or to multiple descriptors in ahierarchy. The relationship indicator for this typeof term is NT.

natural language. A language used by humanbeings for verbal communication. Words extractedfrom natural language texts for indexing purposeswithout vocabulary control are often called key-words.

near-synonym. See quasi-synonym.node label. A “dummy” term, often a phrase,

that is not assigned to documents when indexing,but which is inserted into the hierarchical section ofsome thesauri to indicate the logical basis on whicha class has been divided; also known as a ‘facetindicator’ [ISO 2788]. Node labels may also be usedto group categories of related terms in the alpha-betic section of a thesaurus.

nonpreferred term. One of two or moresynonyms or lexical variants that serves as anentry term.

orphan term. A descriptor that has no associa-tive or hierarchical relationship to any other de-scriptor in a thesaurus.

polyseme. A word with multiple meanings. Inspoken language, polysemes are called homonyms;in written language they are called homographs.Only the latter are relevant to thesauri designed fortextual information.

postcoordination. The combining of descriptorsat the searching stage rather than at the subjectheading list construction stage or indexing stage.

postings. The number of documents to which adescriptor is assigned.

precoordination . The formulation of amultiword subject heading or the linking of aheading with a subheading at the subject head-ing list construction stage or at the indexing stageto express a compound concept, e.g., catalogingof serials, cataloging—serials, or serials cata-loging. Precoordination differs from the estab-lishment of compound terms as descriptors.

preferred term. One of two or more synonymsor lexical variants selected as a descriptor.

provisional term. A descriptor with temporarystatus in a thesaurus; it often represents a newconcept in a field in which the terminology has notyet been standardized. Cf. candidate term; identi-fier (definition 2).

qualifier. A defining term, also known as ‘gloss,’used in a controlled vocabulary to distinguish ho-mographs. A qualifier is considered part of a de-scriptor, subject heading, or entry term, but is sepa-rated from it by punctuation (the qualifier isgenerally enclosed in parentheses).

quasi-synonym. A term whose meaning is notexactly synonymous with that of another term, yetwhich may nevertheless be treated as its equivalentin a thesaurus.

related term. A descriptor that is associativelybut not hierarchically linked to another descrip-tor in a thesaurus. The relationship indicator forthis type of descriptor is RT.

relationship indicator. A word, phrase, abbre-viation, or symbol identifying a semantic relation-ship between terms. Examples of relationship indi-cators are USE, UF (used for), and RT (relatedterm).

romanization. The conversion of a non-romanscript by means of transcription or transliterationor a combination of the two methods.

scope note. A note following a descriptor ex-plaining its coverage, specialized usage, or rulesfor assigning it.

sibling. A descriptor that shares the samebroader term (one level higher) as otherdescriptors.

stop list. A list of words considered to be of novalue for retrieval. It consists primarily of functionwords—articles, conjunctions, and prepositions—but may also include words that occur very fre-quently in the literature of a domain.

subheading. A term appended to a heading inorder to modify or delimit the heading by indicat-ing a particular aspect or relationship pertaining toit. A term with a subheading may be subject tofurther modification. Cf. precoordination.

subject heading. A word or phrase, or anycombination of words, phrases, and modifiersused to indicate the overall content of a docu-ment or part of a document. Precoordination ofterms for multiple and related concepts is a char-acteristic of subject headings that distinguishesthem from thesaurus descriptors.

subject heading list. An alphabetical list ofsubject headings with cross-references fromnonpreferred terms and links to related terms.These lists often include separate sequences of

GLOSSARY


Page 38

standardized subheadings that may be combinedwith all or only some subject headings. Rules forapplying subheadings usually accompany suchlists.

synonym. A word or term having exactly orvery nearly the same meaning as another word orterm. [Cf. ISO 5127/6]

systematic display. See tree structure.term. One or more words designating a concept.

See also compound term; descriptor; entry term.term record. A set of fields of information about

a descriptor in a thesaurus, including the history ofthe descriptor, its relationships to other terms, and,optionally, authorities for the descriptor.

thesaurus (pl. thesauri, also thesauruses). Acontrolled vocabulary arranged in a knownorder in which equivalence, homographic,hierarchical, and associative relationships amongterms are clearly displayed and identified bystandardized relationship indicators, which mustbe employed reciprocally. Its purposes are topromote consistency in the indexing ofdocuments, predominantly for postcoordinatedinformation storage and retrieval systems, andto facilitate searching by linking entry terms withdescriptors. Thesauri may also facilitate the re-trieval of documents in free text searching.

Note: The term “thesaurus” is the Latin form ofthe Greek word thesauros, originally meaning “trea-sure store.” In the 16th century, it began to be usedas a synonym for “dictionary” (a treasure store ofwords), but later it fell into disuse. Peter MarkRoget resurrected the term in 1852 for the title of hisdictionary of synonyms. The purpose of that workis to give the user a choice among similar termswhen the one first thought of does not quite seem tofit.

A hundred years later, in the early 1950s, theword “thesaurus” began to be employed again asthe name for a word list, but one with the exactlyopposite aim: to prescribe the use of only one term

(a “descriptor”) for a concept that may have syn-onyms. A similarity between Roget’s Thesaurus andthesauri for indexing and information retrieval isthat both list terms that are related hierarchically orassociatively to descriptors, in addition to syn-onyms.

top term. The broadest descriptor in a thesaurushierarchy, sometimes indicated by the abbrevia-tion TT.

transcription. The process of recording the pho-nological and/or morphological elements of a lan-guage in terms of a specific writing system.

transliteration. The process of recording thegraphic symbols of one writing system in terms ofthe corresponding graphic symbols of another writ-ing system.

tree structure. A thesaurus display format inwhich the complete hierarchy of descriptors isshown. Each descriptor is assigned a tree numberor line number which leads from the alphabeticaldisplay to the hierarchical one; the latter is some-times called ‘systematic display’ or ‘classified dis-play.’

up-posting. The automatic assignment ofbroader terms in addition to the specific descriptorby which a document is indexed. Also known as‘autoposting.’ Cf. generic posting.

user warrant. Justification for the representationof a concept in an indexing language or for theselection of a preferred term because of frequentrequests for information on the concept or free-textsearches on the term by users of an informationstorage and retrieval system.

vocabulary control. The process of organizinga list of terms (a) to indicate which of two or moresynonymous terms is authorized for use; (b) todistinguish between homographs; and (c) to in-dicate hierarchical and associative relationshipsamong terms in the context of a thesaurus orsubject heading list. See also controlledvocabulary.

GLOSSARY


Appendix A(This Appendix is not a part of the American National Standard Guidelines for the Construction, Format,and Management of Monolingual Thesauri, ANSI/NISO Z39.19-1993. It is included for information only.)

Supplementary Figures

The following samples of thesaurus displays and forms are included to illustrate specific points made inthe standard, primarily in sections 6 and 7 (print and screen display). Because certain illustrations may notcomply with all recommendations of the standard, they have been placed in an Appendix rather thanembedded in the text.

Figure A1. The Relationship Indicators USE AND and USED FOR+. These are applied, respectively, to refer from compoundterms to two descriptors that represent them, and to record the reciprocal of the cross-reference under each of the descriptors. Thesecodes are discussed in section 5.2.5.

Source: International Energy Subject Thesaurus. 1990.

ANSI/NISO Z39.19-1993APPENDIX A

Figure A2. KWIC (Keyword in Context) Index to a Thesaurus. Every word in descriptors and entry terms is an access point. Thisdisplay is discussed in section 6.1.1.

Source: Thesaurus of ERIC Descriptors, 12th ed., 1990, edited by James E. Houston, with permission from The Oryx Press, 4041 N.Central at Indian School Rd., Phoenix AZ 85012, 602-265-2651.

ANSI/NISO Z39.19-1993 APPENDIX A

Figure A3. KWOC (Keyword Out of Context) Index to a Thesaurus. Words are extracted from descriptors and entry terms. Thisdisplay is discussed in section 6.1.1.

Source: Thesaurus of Engineering and Scientific Terms, 1967.


Figure A4. Flat Thesaurus Structure. For each descriptor, hierarchical information (BT, NT) is limited to one broader level andone narrower level. This display is discussed in section 6.2.1.

Source: Thesaurus of ERIC Descriptors, 12th ed., 1990, edited by James E. Houston, with permission from The Oryx Press, 4041 N.Central at Indian School Rd., Phoenix AZ 85012, 602-265-2651.


Figure A5. Multilevel Broader and Narrower Terms. All the broader and narrower terms of a given descriptor are given withinthe alphabetic display. This display is discussed in section 6.2.2.1.

Source: International Energy Subject Thesaurus. 1990.


Figure A6. Generic Structure. A full hierarchical display, excluding sibling terms, is given for each descriptor within thealphabetic sequence of the thesaurus. This display is discussed in section 6.2.2.2.

Source: NASA Thesaurus. Vol. 1. Hierarchical Listing, 1988, p. 409.


Figure A7. Tree Structure. Indention and the length of the tree number (center column) indicate hierarchical level. The treenumbers on the right (in smaller type) indicate additional hierarchies in which the descriptor is found. This display is discussed insection 6.2.3.

Source: Medical Subject Headings. Tree Structures. 1992.


Figure A8. Hierarchical Thesaurus Display Featuring Node Labels (italicized, in brackets) and Line Numbers. Indentionindicates hierarchical level. This display is discussed in section 6.2.3.

Source: Art and Architecture Thesaurus, 2nd ed. 5 vols. New York: Oxford University, 1994. Vol. 2, p. 486.


Figure A9. Graphic Display. Synonyms and narrower terms are contained within the boxes. Other relationships are indicated bylines connecting the boxes. This display is discussed in section 6.2.4.

Source: Euratom Thesaurus. Part II. Terminology Charts, 2nd ed., 1967.


Figure A10. Top Terms (TT) in an Alphabetic Thesaurus. These represent the highest level in a hierarchy and may be the sameas the BT (broader term). This display is discussed in section 6.2.5.

Source: 1993 Inspec Thesaurus, with permission from INSPEC, the Institution of Electrical Engineers, Michael Faraday House, SixHills Way, Stevenage Herts. SG1 2AY UK.

APPENDIX A


Figure A11. Hierarchical Display of Top Terms. Periods represent hierarchical level. Whereas the alphabetic display containsonly a single level of narrower term, this one displays the complete hierarchy. This display is discussed in section 6.2.5.

Source: 1993 Inspec Thesaurus, with permission from INSPEC, the Institution of Electrical Engineers, Michael Faraday House, SixHills Way, Stevenage Herts. SG1 2AY UK.

APPENDIX A


Figure A12. Two-way Hierarchical Display. For every descriptor in the thesaurus, complete hierarchical information is given.Colons precede broader terms; periods indicate narrower terms. This display is discussed in section 6.2.6.

Source: Thesaurus of ERIC Descriptors. 12th ed., 1990.

APPENDIX A


Figure A13. Category List. A numerically arranged list of broad categories. Each descriptor is assigned to one category orsubcategory. This display is discussed in section 6.2.7.

Source: Thesaurus of Engineering and Scientific Terms (TEST). 1967, p. 569.

APPENDIX A


Figure A14. Printed Alphabetic Thesaurus Designed for End-Users. The term records are not as detailed as they are in FigureA15. This display is discussed in section 6.4.

Source: Medical Subject Headings. Supplement to Index Medicus, vol. 33, 1992, p. 232.

APPENDIX A


Figure A15. Printed Alphabetic Thesaurus Designed for Indexers and Searchers. The term records include far more detail thanthose in the corresponding end-user thesaurus illustrated in Figure A14. This display is discussed in section 6.4.

Source: Medical Subject Headings: Annotated Alphabetic List. 1992, p. 345.


Figure A16. KWIC Screen Display of Thesaurus Terms. Like the corresponding print display illustrated in Figure A2, this oneprovides access to every word in descriptors and entry terms. This display is discussed in section 7.3.2.

Source: Project Management Enterprises, Inc. Bethesda, MD.

APPENDIX A


Figure A17. KWOC Screen Display of Thesaurus Terms. Like the corresponding print display illustrated in Figure A3, alldescriptors and entry terms containing the keyword — in initial, medial, or final position — are identified. This display is discussedin section 7.3.2.


APPENDIX A


Figure A18. Screen Display of Term Record. The record includes unabbreviated relationship indicators and a history note. Thisdisplay is discussed in section 7.3.4.


^

><

>

Switches Copy Print Format

Application Edit Print Refresh Options Exit Help

LEXICO/2 - TGM - II

TermRecord TypeApproval StatusApproval Status DatePublic Note

Cataloger's NoteUsed For

Broader TermRelated Term

TGM-I CompatibilityDate CreatedLatest Update

TermRecord TypeApproval StatusApproval Status DatePublic Note

Cataloger's NoteBroader Term

BBiibbllee ccaarrddss

Main Termy26-Aug-93Cards with a scriptural picture or quotation, or both. Sometimes issued in sets and often used as rewards of merit in Sunday schools.MARC filed 655.Scripture cardsSunday school cardsDevotional imagesCollecting cardsRewards of meriti10-Nov-9226-Aug-93

Contents of Record by: TERM

BBii ll llbbooaarrdd ppoosstteerrss

Main Termy20-Aug-93Large multi-sheet posters; intended for posting on billboards, fences or similar surfaces.MARC filed 655Posters

APPENDIX A


Figure A19. Thesaurus Term Review Form. The form may be used to add, modify, or delete descriptors and entry terms. Additionof descriptors is discussed in section 9.1.1.

Source: NASA Center for AeroSpace Information.


Figure A20. Menu of a Thesaurus Management System. This menu displays and defines commands. Editing capabilities ofthesaurus management systems are discussed in section 10.5.

Source: BASISplus Thesaurus Manager, Information Dimensions, Inc., Dublin, Ohio.

APPENDIX A


Figure A21. Screens for Modification and Deletion of Terms. The system requests confirmation of a command to delete adescriptor. Term deletion is discussed in section 10.8.

Source: BASISplus Thesaurus Manager, Information Dimensions, Inc., Dublin, Ohio.

APPENDIX A


CO = Company descriptor, in this case a company nameDIV = Division a type of narrower termJVW = Joint Venture with a type of related termPAR = Parent a type of broader termSUB = Subsidiary a type of narrower termUF = Used For UF reference, in this case a former name

Figure A22. Screen Display of Term Record Featuring Specialized Codes in a Company Thesaurus.

^

><

>

Switches Copy Print Format

Application Edit Print Refresh Options Exit Help

LEXICO/2 - TGM - II

CO

UFSBO

DIVDIV

SUBSUBSUB

JVW

SNOMORE Group [USA]

Snow More CorporationNew England Technologies [USA]

SNOMORE Flood Control Systems [USA]SNOMORE Icemore Division [USA]

Allweather Energy Systems Ltd. [UK]SNOMORE Financial Services [USA]SNOMORE GmBH [GER]

Mariette Motors, Plowmore, Div. [USA]

Contents of Record by: TERM

APPENDIX A


Appendix B(This Appendix is not a part of the American National Standard Guidelines for the Construction, Format,and Management of Monolingual Thesauri, ANSI/NISO Z39.19-1993. It is included for information only.)

Minithesaurus of Thesaurus TermsThe following minithesaurus incorporates several optional features of thesaurus display that are

described in the standard.

Generic Structure Example

thesauriGS indexing languages

. classification schemes

. controlled vocabularies

. . subject heading lists

. . thesauri

. . . alphabetical thesauri

. . . descriptors

. . . entry terms

. . . faceted thesauri

. . . macrothesauri

. . . minithesauri

. . . monolingual thesauri

. . . multilingual thesauri

. . . relationship indicators

. . . systematic thesauri

Legend:<> node labelBT Broader Term(s)BTP Broader Term PartitiveGS Generic StructureNT Narrower Term(s)NTP Narrower Term PartitiveRT Related Term(s)SN Scope NoteUF Used For

In the flat display format, optional node labels areused to organize narrower terms that are alphabeti-cally arranged in the generic structure. Node labelscould be used in the generic structure as well. Thecodes BTP and NTP indicate the whole-part rela-tionship.

Related Terms, Used For references, and ScopeNotes are the same for both systems. These examples donot provide a full set of semantic relationships.

Flat Display Example

controlled vocabulariesBT indexing languages NT subject heading lists

thesauri

descriptorsUF thesaurus termsBTP thesauri

entry termsSN the lead-in vocabulary of a thesaurusUF use referencesBTP thesauri

thesauriBT controlled vocabularies

NT<by display format> alphabetical thesauri faceted thesauri hierarchical thesauri<by size> macrothesauri minithesauri<by number of languages> monolingual thesauri multilingual thesauri

NTP descriptorsentry termsrelationship indicators

RT authority filescomputer-aided indexingdictionariesglossariesnomenclature

thesaurus termsUSE descriptors

use referencesUSE entry terms

APPENDIX B


Appendix C(This Appendix is not a part of the American National Standard Guidelines for the Construction, Format,and Management of Monolingual Thesauri, ANSI/NISO Z39.19-1993. It is included for information only.)

User Interfaces

I. StandardsHuman Factors Society. Human Computer Inter-action (HCI) Standards Committee. This commit-tee serves both as the ANSI accredited standards-developing body for HCI standards and as the U.S.Technical Advisory Group (TAG) to ISO TC159/SC4/WG5 (its international counterpart). Specificareas of standards development include: dialoguetechniques (including menus, commands, forms,and direct manipulation); information presenta-tion (including text, graphics, sound, and voice);user guidance (including system messages, statusinformation, feedback, prompts, error handling,and help); window-based interaction; and the de-sign process for developing successful HCI. Ad-dress: Human Factors Society, Inc., P.O. Box 1369,Santa Monica, CA 90406.

ISO 9241 (parts 1-n). Ergonomic Requirements forOffice Work with Visual Display Terminals (draftsand works-in-progress). Part 14: Menus.

II. Authoritative TextsGalitz, Wilbert O. Handbook of Screen FormatDesign, 3rd ed. Wellesley, MA: QED InformationSciences, Inc., 1989.

Mayhew, Deborah. Principles and Guidelines toSoftware User Interface Design. Englewood Cliffs:Prentice Hall, 1992.

Shneiderman, Ben. Designing the User Interface:Strategies for Effective Human-Computer Inter-action, 2nd ed. Reading, MA: Addison-Wesley,1992.

APPENDIX C


Appendix D(This Appendix is not a part of the American National Standard Guidelines for the Construction, Format,and Management of Monolingual Thesauri, ANSI/NISO Z39.19-1993. It is included for information only.)

Thesaurus Clearinghouses

Thesauri should be deposited with one of the following international clearinghouses:

b)Thesauri in other languages:

Instytut Informacji Naukowej,Technicznej i EkonomicznejClearinghouseul. Zurawia 3/500-926 WARSZAWAPoland

a) Thesauri in the English language, including mul-tilingual thesauri containing English-languagesections:

Thesaurus ClearinghouseThe LibraryFaculty of Library ServiceUniversity of Toronto140 St. George StreetTORONTOOntario M5S 1A1Canada

APPENDIX D


Appendix E(This Appendix is not a part of the American National Standard Guidelines for the Construction, Format,and Management of Monolingual Thesauri, ANSI/NISO Z39.19-1993. It is included for information only.)

Bibliography of Manuals on Thesaurus Construction

Details on some of the structures or procedures suggested in this standard may be found in the followingEnglish-language manuals on thesaurus construction. Inclusion of a publication on this list does not indicateapproval of its contents by the committee developing this standard.

Aitchison, Jean; Alan Gilchrist. Thesaurus Con-struction: A Practical Manual. 2nd ed. London:Aslib, 1987. 173 pp.

Lancaster, F.W. Vocabulary Control for Informa-tion Retrieval. 2nd ed. Arlington, VA: Informa-tion Resources Press, 1986. 270 pp. (First edition,1972, contains material omitted from the 2nd ed.)

Orna, Elizabeth. Build Yourself a Thesaurus: AStep-By-Step Guide. Norwich: Running Angel,1983. 32 pp.

Soergel, Dagobert. Indexing Languages and The-sauri: Construction and Maintenance. Los Ange-les: Melville, 1974. 632 pp.

Townley, Helen M.; Ralph D. Gee. Thesaurus-Making: Grow Your Own Word-Stock. London:Andre Deutsch, 1980. 206 pp.

APPENDIX E

ANSI/NISO Z39.19-1993 INDEX

Page 65

Index

Abbreviations, 7-8, 13, 16used in standard, xii

Abstract concepts, 6Access vocabulary, 2, 21, 22

See also Cross-references, USE referencesAcronyms, 7-8Activities, 4, 6Addition of new descriptors, 29Adjectival noun phrases, 4Adjectives, 5Adverbs, 5ALA Filing Rules, 24ALA-LC Romanization Tables, 10Alphabetic character filing, 24Alphabetical displays, 21, 22, 26American National Standard for Library and Information Sciences

and Related Publishing Practices - Basic Criteria for Indexes,(ANSI Z39.4), 2

Ampersand (&), 10Angle brackets used for node labels, 19, 26Anglo-American Cataloguing Rules (AACR2), 2, 9Annotated Alphabetic List (MeSH) (fig. A15), 25Announcement of thesaurus, 29Antonyms, 16Apostrophes, 10Art and Architecture Thesaurus (fig. A8)

Classification and line numbers, 22Node labels, 19

Articles, 5-6Assignment indexing, 1, 35Associative relationships, 14,19-21

in printed thesauri, 22Authorities, 7, 10, 11, 24, 28Authority files, 9, 35

BASISPlus Thesaurus Manager (Information Dimensions Inc.)menu (fig. A20), 31term deletion /modification screen (fig. A21), 32

Bound terms, 10, 35British Standard Alphabetical Arrangement and the Filing Order of

Numerals and Symbols, 24Broader term (generic), 17Broader term (instance), 18Broader term (partitive), 18Broader terms, 13, 16-19, 35

as qualifiers, 2false, 20

BT indicator, xii, 13, 16BTG indicator, xii, 17BTI indicator, xii, 18BTP indicator, xii, 18

Candidate terms, 27, 28, 32, 35See also Provisional terms

Capitalization, 9, 24Categories of descriptors, 23, 35

in term record, 28Categories of users, 25Changes to descriptors, 29-31Classification notation, 22-23, 26

in term record, 28

Classified display, 22See also Tree structure

Clearinghouses, 27, 29Codes used in standard, xii, 13, 16

See also Relationship indicatorsCoexistence of singular and plural forms, 6Committee approach to thesaurus construction, 27Common names vs. trade names, 8, 15Common use, 6, 7, 9, 12, 15Compound concepts, 3Compound descriptors, 11-13

preferred to qualifiers, 3Compound terms (Section 4), 2, 3, 10-13, 35

cross-references to, 13, 16, 21Computer-assisted indexing, 1Computer-assisted thesaurus construction, 27-28Computerized filing, 24Concepts, 35

abstract, 6compound, 3single concept, 2,11types of, 3-4

Consistency of hierarchies, 12, 22-23Consistency of indexing, 1, 11Consistency of thesaurus construction, 27Controlled vocabulary, 1, 35Conventions in a printed thesaurus, 23Conventions in standard, xii, 2Count nouns, 6Counteragents of processes as related terms, 20Cross-references, 2, 15-16, 19-21, 35

See also USE references, Related termsadjectives used alone as, 5between descriptors in different hierarchies, 20between full name and abbreviation, 7,16between sibling terms, 19economy of, 5error checking of, 32to both elements of a split compound term, 16

Database documentation, 26Date of thesaurus, 24Dates for descriptors, 30Deductive method of construction, 27Definitions of terms vs. scope notes, 33Deletion of descriptors, 30, 32Deposit of published thesaurus, 29Derivational relationships, 19-20Derivative indexing, 1, 35Descriptors (Section 3), 2-10

adding to thesaurus, 29choice of, 7-9compound, 11-13dates of, 30definition of, 35deletion of, 30, 32division into categories, 23grammatical forms of, 4-6modification of existing, 29-30,qualifiers for, 2-3scope of, 2-3specificity of, 28-29

INDEX ANSI/NISO Z39.19-1993

Page 66

Descriptors (continued)types of, 3-4unassigned, 29verification of, 28

Diacritical marks, 8, 10Dialectal variations, 7, 15Dictionaries, 7, 10, 28Difference in compound terms, 12, 35Discontinued descriptors, 30Display formats (Section 6), 21-25

deciding in advance, 27screen display (Section 7), 25-26, 31

Documentation, 25, 26Documentation - Guidelines for the Establishment and Development

of Multilingual Thesauri (ISO 5964), 9Documents, 1, 35Double lookup, 21, 23Duplicate term checking, 31-32Duplicate work avoidance, 27

Editing in thesaurus management systems, 31Editorial boards, 3, 27, 29, 30Emergent concepts, 8, 15Emotions, 6Empirical approach to construction, 27Encyclopedias, 7,28End users, 25Entry terms, 2, 35

narrower terms as, 16print vs. screen displays, 22

Entry vocabulary, 21, 22, 35Eponyms, 10, 35Equivalence relationship, 14, 19-21

See also Entry Terms, USE referencesin print displays, 21-22

ERIC ThesaurusSee Thesaurus of ERIC Descriptors

Error checking, 31-32EURATOM Thesaurus, part 11, graphic display (fig. A9), 23Events, 4Expert searchers, 26Experts, 7, 27

False coordination of search terms, 5False drops, 11, 35Field definition in thesaurus management systems, 32-33Filing rules, 24Flat format, 17, 22, 36Focus in compound terms, 11-12, 36Foreign-language terms, 9Form of descriptors, 4-5, 7-9Formats for printed thesauri, 21-25Free text, 36Frequency of descriptor assignment, 28

See also Postings data, Unassigned descriptorsFull form of term vs. abbreviation, 7-8, 16

Generic nouns vs. trade names, 8, 15Generic posting, 16, 36Generic relationship, 17Generic structure, 22, 23, 36Genus-species relationship, 11, 17Geographical names, 9Gerunds, 4Glossaries as authorities, 7, 28

Graphic displays, 17, 23, 26GS code, xii, 22

Head noun in compound terms, 11-12Headings, 2, 36Hierarchical displays, 21, 22-23, 33

use of node labels in, 18-19Hierarchical relationships, 14, 16-19

print display, 22-23used in online searching, 26

Hierarchical structureconsistency of, 12, 22, 23testing validity of, 4, 17two-way, 23

Hierarchy, definition, 36History notes, 28, 30-31, 36

recording changes of scope in, 3Homographs, 1, 2- 3, 36

juxtaposed in filing, 25singular and plural forms, 6

Housekeeping reports, 33Hyphens, 10

Identifiers, 9, 36Idiomatic phrases, 5Indention, 22, 23, 26Index Medicus (fig. A14), 25Indexers, 25Indexes, printed, 11Indexing, 1, 36Indexing consistency, 1, 11Indexing language, 1, 36Indexing services, 28, 33Indexing terms, 1, 36Inductive method of construction, 27Information storage and retrieval systems, 2, 25, 31, 36Initial articles, 5-6Initialisms, 7-8INSPEC Thesaurus, top terms (fig. A10, All), 23Instance relationship, 18International Energy Subject Thesaurus

history note (fig. A1, A5), 30Multilevel BT and NT (fig. A5), 22UF+ cross-reference, (fig. A1), 16

Intransitive action in compound terms, 13Introduction to a thesaurus, 23-24Inversions of terms as cross-references, 5, 13, 21Italics, 19, 24

Jargon, 8, 15of information retrieval, 25

Juxtaposition of termsin filing, 24-25in print displays, 21-22

Key word, definition of, 36KWIC and KWOC indexes, 21, 26, 31

Letter-by-letter filing, 24Lexemes, 10, 11, 36Lexical variants as synonyms, 16Lexico System (Project Management, Inc.)

KWIC and KWOC displays (fig. A16, A17), 26Term record (fig. A18), 28

Library of Congress Filing Rules, 24


Page 67

Limiting number of terms 5, 11Line numbers, 22, 26Literary warrant, 2, 6, 7, 11, 28, 36-37Loanwords, 8-9Locally-defined fields, 32- 33Locally-defined relationships, 20-21, 28, 33Lowercase letters, 9, 24

Machine assistance in thesaurus construction, 27Mass nouns, 6Materials, 4, 6Medical Subject Headings (MeSH)

history note (fig. A14), 30-33multiple print formats (figs. A14, A15), 25tree structure (fig. A7), 22

Merge function, 32Microthesauri, 33, 37Modification of descriptors, 29-30Modifier in compound terms, 12Monolingual thesauri, 2, 9Multilevel hierarchical displays, 17, 22Multilingual thesauri, 2Multiple versions of print thesauri, 25Multiple wholes, parts of, 18Mutually exclusive sibling terms, 19

Narrower term (generic), 17Narrower term (instance), 18Narrower term (partitive), 18Narrower terms, 16-19, 22, 37NASA Thesaurus

generic structure (fig. A6), 22term review form (fig. A19), 29

Natural language, 1, 37Natural language order, 13Neologisms, 8,15Node labels, 37

in flat thesaurus displays, 22in hierarchies, 13, 18-19not used as descriptors, 19screen display, 26to group related terms, 20typography, 19, 26

Nonalphabetic characters, 10, 24, 31impact on searching, 26

Noncount nouns, 6Nonpreferred terms, 2, 15, 22, 36

See also Cross-references, USE referencesNotation, 22-23, 26

in term record, 28Notification of intent to publish, 29Nouns and noun phrases, 4-5NT indicator, xii, 13, 16NTG indicator, xii, 17NTI indicator, xii, 18NTP indicator, xii, 18Numerals in filing, 24

Obsolete descriptors, 30Omnibus indexing services, 33On-screen instructions, 25Online display, 22, 25-26Online searching

See SearchingOrder of words in compound terms, 13

Orphan terms, 13,33,37

Parentheses, 2, 10, 26in filing, 24-25

Parenthetical qualifiersSee Qualifiers

Part-whole relationship, 17-18Pejorative terms, 7Peripheral subject domains, 16Permuted lists of terms, 21, 26Personal authorities, 7, 27Pilot version of thesaurus, 29Place names, 9Polyhierarchical relationships, 18Polysemes, 1, 37Popular vs. scientific names, 8, 15Possessive case, 10Postcoordination, 1, 11, 37Postings data, 25, 26, 28, 29, 33, 37Postmodified noun phrases, 4, 5, 13Precision in searching, 1, 11Precoordination, 2, 11, 37Preferred terms, 2, 7-9, 15, 16, 37

See also DescriptorsPremodified noun phrases, 4, 5,13Prepositional noun phrases, 5Print displays, 21-25

vs. screen displays, 22, 25Printed indexes, 11Printed thesauri, 23-25Processes, 4, 6, 20Proper names, 2, 9, 10, 18Properties (characteristics), 4, 6, 20Provisional terms, 8, 28, 37

See also Candidate termsPublished authorities, 7, 10, 11, 24, 28Published thesauri, 29Punctuation marks, 10, 24-25, 31

impact on searching, 26Purpose of thesauri, 1Purpose of thesaurus stated in introduction, 23

Qualifiers, 2-3, 6, 37effect on searching, 26in filing, 25

Quantitative data reports, 31, 33Quasi-synonyms, 15,16,37

Recall in searching, 1Reciprocal scope notes, 3Reciprocity of relationships, 13, 16, 32Registered trademark designation, 8Related terms, 19-21, 37

in printed thesauri, 22Relationship indicators, xii, 13, 16-19, 37

locally defined, 20-21Reports in thesaurus management systems, 33Romanization, 10, 37RT indicator, xii, 13, 19Running heads, 25

Scientific vs. popular names, 8, 15Scope notes, 2, 3, 33, 37

in term record, 28Scope of a thesaurus, 23

INDEX ANSI/NISO Z39.19-1993

Page 68

¯

Scope of descriptors, 2, 3Scope of standard, 1Screen display (Section 7), 22, 25-26

vs. print display, 22, 25Search expansion, 26Searchers, 25Searching

false coordination in, 5false drops in, 11impact of nonalphabetic characters, 26in thesaurus management systems, 31precision and recall, 1, 11use of hierarchies, 26use of tree structure, 23

See From reference, 3, 15Semantic relationships, 13, 14Sibling terms, 19, 22, 23, 37Single concept, 2, 11Single-word vs. multiword descriptors, 2Singular vs. plural form, 4, 6

juxtaposed in displays, 21Slang, 8, 15Software

See thesaurus management systemsSorting instructions, 24Specificity of descriptors, 28-29Spelling, 7Spelling variations, 15, 16

not displayed, 26, 31Splitting compound terms, 12-13Standard

applications of, 2conventions in, xii, 2scope of, 1

Standard compliance, 24States (of being), 4, 6Statistical reports, 33Stop list, 27-28, 37Subheadings, 2, 37Subject experts, 7, 27Subject heading lists, 7, 37Subject headings, 7, 37Subordination and superordination, 16Symbols, 10Symmetrical relationships, 15, 19

See also Reciprocity of relationshipsSynonyms, 1, 15-16, 38

in term records, 28System constraints, 31Systematic display, 22

See also Tree structure

Term detail in screen displays, 26Term records, 3, 28, 29, 31, 38Term reports, 33Terms, 38

types of, 2verification of, 28

TEST thesaurusSee Thesaurus of Engineering and Scientific Terms

Thesauriconsistency of, 12, 22-23, 27construction of (Section 8), 27-29definition of, 1, 38

Thesauri (continued)display of, 21-26documentation, 25, 26format considerations, 21, 23-26level of specificity, 28-29maintenance of (Section 9), 29-31publishing, 29purpose and use, 1reports and listings, 33

Thesaurus clearinghouses, 27, 29Thesaurus maintainers, 25Thesaurus management systems (Section 10), 31-33

documentation, 26Thesaurus of Engineering and Scientific Terms

categories of terms (fig. A1 3) , 23history note (fig. AQ - history note, 30KWOC display (fig. A3), 21

Thesaurus of ERIC Descriptorsflat format display (fig. AQ, 22KWIC display (fig A2), 21two-way hierarchies (fig. A12), 23

Thesaurus relationships (Section 5), 13-21See also Relationship indicatorsassociative, 19-21equivalence, 15-16hierarchical, 16-19

Thesaurus updates, 29-30Top terms, 23, 33, 38Trademarks, 8, 10, 15Transcription, 10, 38Transfer (merge) function, 32Transitive action in compound terms, 12-13Translations of loanwords, 8-9Transliteration, 10, 38Tree structure, 21, 22-23, 38

compound terms in, 11to indicate hierarchies, 17use in searching, 23, 26

Truncated term searches in thesaurus managementsystems, 31

TT indicator, 23Two-way hierarchical structure, 23Typography, 24, 31

for node labels, 19, 26

U/UF indicators, xii, 13, 15UF+ indicator, xii, 16Unassigned descriptors, 29

See also Postings dataUnique entities, 4, 6Units of measurement, 4, 20Up-posting, 38Updating the vocabulary, 29-30Uppercase letters, 9, 24Usage, 6, 7, 9, 12, 15USE indicator, xii, 13, 15USE references, 15-16

in alphabetical listings, 21in screen displays, 26not displayed, 26, 31

USE+ indicator, xii, 16USE....AND.... relationship, 16USED FOR....AND.... relationship, 16User categories, 25


Page 69

User query terms, 28, 32User warrant, 6, 38

See also UsageUSMARC Authority Format, 31

Validation of hierarchy, 4, 17, 22Verbal nouns, 4Verbs not used as descriptors, 4Verification of terms, 28Vocabulary control, 1,27, 38

Whole-part relationship, 17-18Word-by-word filing, 24

X (See From) indicator, xii, 3,15