The Library of Congress >> Especially
for Librarians and Archivists >> StandardsThe Library of Congress >> Especially
for Librarians and Archivists >> Standards
HOME >> MARC Development >> Proposals List
DATE: May 22, 2014
REVISED:
NAME: Designating Relationships Between Subject Headings from Different Thesauri in the MARC 21 Authority Format
SOURCE: German National Library
SUMMARY: This paper proposes a way to designate relationships between entries of different thesauri in a MARC authority record.
KEYWORDS: Fields 7XX (AD), Relationship designators (AD); Thesauri (AD); Heading linking entry fields (AD); Subfield $i, in 7XX fields (AD); Relationship information (AD); Subfield $4, in 7XX fields; (AD); Relationship code (AD)
RELATED: 2014-DP02
STATUS/COMMENTS:
05/22/14 – Made available to the MARC community for discussion.
06/29/14 – Results of MARC Advisory Committee discussion: Approved as submitted.
07/24/14 - Results of MARC Steering Group review - Agreed with the MAC decision.
Mapping subject headings from different thesauri
In Germany there is a tradition of mapping different thesauri to each other. In the project MACS ("Multilingual Access to Subjects") as well as in the so called "Crosskonkordanzen" a detailed mapping from subject headings of the Schlagwort-Normdatei (SWD, Authority File for Subject Headings) to several other thesauri is done.
End users can benefit from these mapping results when they start searching and navigating based on a known thesaurus, and are guided by relationships to other thesauri, visibly or in the background.
Currently, the results of these intellectual efforts are kept in separate records. With the integration of the SWD into the "Gemeinsame Normdatei" (GND, Integrated Authority File) there is a need to develop a format solution. Because the exchange format for the GND is MARC Authority, and because the cataloging format of the GND is based on MARC Authority, we have to find a solution in the MARC Authority format.
The mappings are using the framework given by ISO 25964 "Information and documentation - Thesauri and interoperability with other vocabularies", part 2 "Interoperability with other vocabularies" (ISO 25964-2). There, table 1 "Additional abbreviations and symbols used in mappings" lists the following mapping types:
EQ
Equivalence; the term that follows the tag is the preferred term in a target vocabulary that is closest in meaning to the preferred term preceding the tag, from the source vocabulary.=
This symbol (an equals sign, Unicode character U+003D) should be used in conjunction with a mapping abbreviation to indicate that the mapping is exact. Specifically, "=EQ" signifies exact equivalence.~
This symbol (a tilde, Unicode character U+007E) should be used in conjunction with a mapping abbreviation to indicate that the mapping is inexact. Specifically, "~EQ" signifies inexact equivalence.BM
Broader mapping; the term that follows the tag represents a concept having a wider meaning.NM
Narrower mapping; the term that follows the tag refers to a concept with a more specific meaning.RM
Related mapping; the term that follows the tag labels an associated concept, but is not a synonym, a quasi-synonym, a broader term or a narrower term.|
This symbol, (a vertical bar or "pipe line", Unicode character U+007C), is shown between two or more preferred terms from a target vocabulary, the scopes of which, in sum, best cover the scope of a broader preferred term in the source vocabulary. Each of the target concepts represents part of the scope of the source concept. When converting index terms, all the target vocabulary preferred terms should be applied to a record indexed with the source vocabulary term. When converting search statements, the target vocabulary preferred terms should be combined with Boolean OR.Example:
Source vocabulary Target vocabulary Livestock sheep| cattle | pigs | poultry +
This symbol (a plus sign, Unicode character U+002B) is shown between two or more preferred terms from a target vocabulary that are used in conjunction to represent a compound concept in the source vocabulary. Each of the target concepts represents an aspect of the source concept. When converting index terms, all the target vocabulary preferred terms should be applied to a record indexed with the source vocabulary term. When converting search statements, the target vocabulary preferred terms should be combined with Boolean AND.Example:
Source vocabulary Target vocabulary women executives women + executives
In the Crosskonkordanzen file, many of the detailed mappings are of the type "equivalence" ("EQ"). There is no usage of the distinction between "EQ" (closest equivalence), "=EQ" (exact equivalence) and "~EQ" (inexact equivalence). In other cases, there is a hierarchical relationship between one subject heading from thesaurus A to a second subject heading from thesaurus B, either of the type "broader mapping" ("BM") or "narrower mapping" ("NM"). In addition, there are cases where the relationship is an associative one, called a "related mapping" ("RM").
Combined with this distinction, there is a majority of cases where there is a one-to-one-relationship. However there are as well relationships of the type one-to-many (one-to-two, one-to-three). Wherever there is a one-to-many relationship, the combination of the terms from the target thesaurus is meant in a way that ISO 25964-2 expresses with a "+" (plus sign), (Boolean AND). There are no combinations of the "|" (vertical bar) type (Boolean OR). There are a few relationships of the type many-to-one (two-to-one, three-to-one).
In the MARC Authority format, the 7XX fields contain the "Heading Linking Entries". According to the documentation at http://www.loc.gov/marc/authority/ad7xx.html:
"The 7XX Heading Linking Entry fields provide a machine link within a system between equivalent headings whether they are structured in the same or different form, are from the same or different authority files or printed thesauri, or whether they exist as separate authority records."
In our context of mappings between thesauri, we intend to use the 7XX fields to "provide a machine link within a system between equivalent headings [...] from [...] different authority files". The fields contain a subfield $0, so linking via an "Authority record control number or standard number" is feasible. The fields contain a subfield $2, so that the "Source of heading or term" can be expressed, as listed under the "Subject Heading and Term Source Codes". And the fields contain a subfield $8 "Field link and sequence number", so that grouping different fields in the same record is possible.
What is missing in the 7XX fields is the designation of a relationship beyond "equivalence". In cases of hierarchical (i.e. broader or narrower) relationships, or associative relationships, fields 7XX lack a way of expressing these types of relationships.
We analyzed subfield $w "Control subfield" of the fields 7XX, and of the fields 4XX and 5XX, respectively. Because $w and its positions are defined differently between 7XX and 4XX/5XX, there were concerns during the discussion of the preceding MARC Discussion Paper that adding a position /2 to 7XX $w might cause confusion, so we omitted this as an option.
We analyzed the 4XX and 5XX fields of the MARC Authority format, which contain subfields $i "Relationship information" (R) and $4 "Relationship code" (R). As the MARC Advisory Committee was in favor of defining subfields $i and $4 in the 7XX throughout the Authority format to enable the process of relating headings from different thesauri in both human readable and coded forms, we now propose the definition of subfields $i "Relationship information" (R) and $4 "Relationship code" (R) in the 7XX fields.
The content of subfield $i is textual, controlled by cataloging rules or other policy regulations. The content of subfield $4 is based on rules or other regulations, but controlled by a list of codes, maintained by LC's NDMSO, and documented in the MARC Code List for Relators.
There is one field in the 7XX range, field 788 "Complex Linking Entry Data", where relationships can be described "when the linking relationship between headings from different subject heading systems or thesauri cannot be adequately conveyed by fields 700-785". Field 788 is the only one of the 7XX range where a subfield $i "Explanatory text" (R) is defined. We don't intend to use 788, as it can transport textual information, and it doesn't tell by its tag which kind of entity it links to. In terms of consistency, there may be a need to consider the re-definition of 788 subfield $i as "Relationship information" (R), and the definition of 788 $4 "Relationship code" (R).
The cases of relationships of the type many-to-one (two-to-one, three-to-one) were discussed by the MARC Advisory Committee during its meeting in January 2014: If there is a mapping between two (or more) entities from Thesaurus A to one entity from Thesaurus B, there doesn't seem to be a MARC solution in a record contained in Thesaurus A (an Authority record with two fields in 1XX is not a valid option). Given the fact that these cases are very few ones, we can keep them out of scope here, and develop a local solution.
Notes on the examples:
The examples are real-life ones, taken from existing mapping records. They are describing German language topical terms and their mappings.
The proposed new subfields $i and $4 and their contents are underlined.
The contents of subfields $4 are the tag values in ISO 25964-2. Here, they are used in an illustrative way, and not meant to anticipate future decisions.
"EQ" - Equivalence
"BM" - Broader mapping
"NM" - Narrower mapping
"RM" - Related mapping
There are no examples for "=EQ" (exact equivalence) and "~EQ" (inexact equivalence).
The contents of $2 are:
"stw"
Standard-Thesaurus Wirtschaft = STW Thesaurus for Economics (Kiel: ZBW - Deutsche Zentralbibliothek fur Wirtschaftswissenschaften)
"thesoz"
Thesaurus for the Social Sciences (Germany: GESIS - Leibniz Institute for the Social Sciences)
Some 7XX fields are missing a $0, which is possible in an interim status of a record.
Example 1: Equivalence
LDR xxxxxnz#a22yyyyyn#4500
001 966658655
003 DE-101
035 ## $a (DE-588)4145068-1
150 ## $a Betriebsstörung
750 #7 $0 ([ISIL STW])19247-6 $i Äquivalent $a Betriebliche Störung $4 EQ $2 stw
Example 2: Broader mapping
LDR xxxxxnz#a22yyyyyn#4500
001 970531885
003 DE-101
035 ## $a (DE-588) 4142176-0
150 ## $a Ambulante Behandlung
750 #7 $i Oberbegriff $a Ambulante Versorgung $4 BM $2 thesoz
Example 3: Narrower mapping
LDR xxxxxnz#a22yyyyyn#4500
001 966660064
003 DE-101
035 ## $a (DE-588) 4631982-7
150 ## $a Energieversorgungsnetz
750 #7 $0 ([ISIL STW])18361-6 $i Unterbegriff $a Stromnetz $4 NM $2 stw
Example 4: Related mapping
LDR xxxxxnz#a22yyyyyn#4500
001 970532083
003 DE-101
035 ## $a (DE-588) 04225753-0
150 ## $a Bezugsquelle
750 #7 $i Verwandter Begriff $a Anbieter $4 RM $2 thesoz
Example 5: Equivalence, 1-to-2
5.1. Without subfields $8:
LDR xxxxxnz#a22yyyyyn#4500
001 966661230
003 DE-101
035 ## $a (DE-588) 4139099-4
150 ## $a Gastarif
750 #7 $0 ([ISIL STW])14603-0 $i Äquivalent $a Gas $4 EQ $2 stw
750 #7 $0 ([ISIL STW])10213-5 $i Äquivalent $a Preis $4 EQ $2 stw
5.2. With subfields $8:
LDR xxxxxnz#a22yyyyyn#4500
001 966661230
003 DE-101
035 ## $a (DE-588) 4139099-4
150 ## $a Gastarif
750 #7 $8 1 $0 ([ISIL STW])14603-0 $i Äquivalent $a Gas $4 EQ $2 stw
750 #7 $8 1 $0 ([ISIL STW])10213-5 $i Äquivalent $a Preis $4 EQ $2 stw
Example 6: Broader mapping, 1-to-2, with subfields $8:
LDR xxxxxnz#a22yyyyyn#4500
001 966663691
003 DE-101
035 ## $a (DE-588) 4627895-3
150 ## $a Kleinunternehmerin
750 #7 $8 1 $0 ([ISIL STW])13992-3 $i Oberbegriff $a Unternehmer $4 BM $2 stw
750 #7 $8 1 $0 ([ISIL STW])15920-4 $i Oberbegriff $a Frauen $4 BM $2 stw
Example 7: Narrower mapping, 1-to-2, with subfields $8:
LDR xxxxxnz#a22yyyyyn#4500
001 971391130
003 DE-101
035 ## $a (DE-588) 4120950-3
150 ## $a Ausbildungsnachweis
750 #7 $8 1 $i Unterbegriff $a Ausbildung $4 NM $2 thesoz
750 #7 $8 1 $i Unterbegriff $a Zeugnis $4 NM $2 thesoz
BIBFRAME currently has the capability to express these relationships to external thesauri in MADSRDF <http://www.loc.gov/standards/mads/rdf/v1> with the following properties: madsrdf:hasExactExternalAuthority (=EQ), madsrdf:hasBroaderExternalAuthority (BM), madsrdf:hasNarrowerExternalAuthority (NM), madsrdf:hasCloseExternalAuthority (EQ or ~EQ), madsrdf:hasReciprocalExternalAuthority.
In the MARC 21 Authority format:
5.1. Define a repeatable subfield $i "Relationship information" and a repeatable subfield $4 "Relationship code" in the fields 700, 710, 711, 730, 748, 750, 751, 755, 762, 780, 781, 782, and 785.
5.2. Define a list of relator codes and their terms for the six types of relationship "EQ", "=EQ", "~EQ", "BM", "NM", and "RM".
HOME >> MARC Development >> Proposals List
| The Library of Congress >> Especially for Librarians and Archivists >> Standards ( 08/11/2014 ) |
Legal | External Link Disclaimer | Contact Us |