FAQ on romanization

What are the differences between romanization, transcription and transliteration?

Romanization refers to the general act of converting a text that does not use the Latin alphabet into the Latin alphabet; it can be done by transcription, transliteration or a combination of both.

Transcription is the representation of a phonetic unit (a sound) of a language by a written unit (a letter or syllable) of a writing system other than the one normally used for that language. It therefore relies on the phonetic aspect rather than the exact correspondence between characters to convert a text in one script to another. Transcription is the only method that can be used for the conversion of non-fully alphabetic languages and ideographic scripts.

Transliteration is the operation of converting a written unit (a character) of one writing system into a written unit of another system, independently of the pronunciation. The operation is carried out with the help of a conventional system that aims at the complete restitution of the original writing. This is called reversibility.

In the context of the PFAN, the aim is to romanise non-Latin scripts; this is done with the help of romanization tables, which use either transliteration, transcription or a mixture of both.

What is the purpose of romanization?

Romanization is the method prescribed by RDA in the basic instructions on names found in a non-preferred script, presumably because it supports the principle of representativeness by respecting the original language as much as possible. Romanization also allows for the circulation of documents in the library operations chain, from document processing to user retrieval to acquisition, even in the absence of linguistic expertise among library staff or users. In addition, some library systems may not support all non-Latin scripts. Romanization therefore facilitates the sharing and exchange of data while allowing a greater number of users to access a greater number of documents. A romanised authorised access point, for example, will allow resources associated with the entity to be processed by a cataloguer, referenced in the catalogue and retrieved by a user even if none of them are fluent in the original script of the author or the resource.

When a person's name in a Latin resource is derived from a non-Latin script, can I create the AAP using the form found in the resource?

It must first be determined whether instruction 9.2.2.5.3 applies in this case by looking at the related policy statement. In the case of textual works, if the text of that person's first work to be catalogued was originally written in a Latin script language, 9.2.2.5.3 does not apply and the CIP may be based on the form in the catalogued resource. If it later becomes apparent that most of the person's works were written in a non-Latin script, the provisions of the policy statement should be applied.

If 9.2.2.5.3 is applicable, cataloguers must follow the alternative of this instruction, as directed in the associated policy statement.

What do I do when I need to create the AAP for a person for whom I have the non-Latin name in the resource I have in my hands?

In this case, 9.2.2.5.3 is applicable and cataloguers must follow the alternative of this instruction, following the guidelines in the related policy statement.

For a name in Chinese, Japanese, Korean or Cyrillic script, I have consulted the three preferred sources (Encyclopédie Larousse, Le Petit Robert and Encyclopédie Universalis) in vain. I must now use the systematically romanised form of the name. What does this mean? How do I proceed?

This means determining the person's preferred name in non-Latin script by consulting the sources mentioned under 9.2.2.2 and then converting that preferred name to Latin script using the PFAN-approved romanization table for the language or script concerned. The PFAN-approved tables can be found on the PFAN wiki.

Should preferred titles and variant titles of works be transliterated in fields 1XX and 4XX of authority records?

The basic instruction under 6.2.2.7 says to transliterate a title of a work found in a script that differs from a script preferred by the agency creating the data according to the transliteration method adopted by the agency. RDA provides an alternative for selecting as the preferred title a title or form of title established by usage in the reference sources in a language preferred by the agency creating the data. According to the PFAN PS for 6.2.2.7, PFAN participants may apply the alternative only "for anonymous works created before 1501 that are not in Greek or Latin script." For these works, "choose as the preferred title a title established by usage in French if one exists." For works in Greek created before 1501, whether anonymous or not, refer to 6.2.2.5. In all other cases, the basic instruction should be applied.

I have located the authority of the person for whom I need to create the AAP in the LC/NAF file after checking that 9.2.2.5.3 applies. Can I take the preferred name form as is?

Not necessarily. Firstly, it may happen that the name or form of name registered as the preferred name by a PFAN cataloguer is not the same as that chosen by the LC/PCC cataloguer. Secondly, although LC and PCC follow the same criteria on the applicability of instruction 9.2.2.5.3 and also apply the alternative under 9.2.2.5.3, they prefer reference sources in English while PFAN prefers sources in French. According to the sources consulted, it could therefore be that the form chosen by the LC/PCC cataloguer is different from that found in the French reference works. It could also be that the form chosen for one file was found in a reference source in the language preferred by the program, while the form chosen for the other file had to be romanised following unsuccessful searches in the reference sources. Furthermore, the romanization tables used are not necessarily the same depending on the language, so that a romanised form found in the LC/NAF file would not necessarily be valid in Canadiana.

Where 9.2.2.5.3 is applicable (see related policy statement), the LC/NAF form can only be taken as is if both of the following conditions are met:

the preferred name used as the basis for the AAP is in a non-Latin script and is the same as that which would be used by the PFAN;
the ALA-LC romanization table used to convert the name into Latin script is the one approved by the PFAN for the language or script concerned.

Can I take the form found in the BnF authority file?

Not necessarily. Where 9.2.2.5.3 is applicable (see related policy statement), the form found in the BnF file can only be taken as is if both of the following conditions are met:

the preferred name used as the basis for the AAP is in a non-Latin script and is the same as that which would be used by the PFAN;
the romanization table used by the BnF to convert the name into Latin script is the one approved by the PFAN for the language or script concerned.

8. After establishing my AAP, I see that the LC/NAF file contains an authority record for this person with a multitude of access point variants, some of which are in non-Latin script. Is it good practice to copy and paste all the non-Latin variants into my record?

These variants can be copied as long as they are justified in a 670 field. It is up to the cataloguer to determine which variants are most useful in the context of the PFAN and should be added. It should be borne in mind that any non-Latin variant should normally also be included in a 400 field in romanized form according to the tables approved by the PFAN. If the LC/NAF record includes a 400 field with an equivalent romanized form according to the ALA-LC tables, be sure to also include a 400 field with a romanized form according to the PFAN-approved tables where this form differs.

Can I correct the AAP in an existing notice that contains the note " CETTE NOTICE VERSÉE EN LOT DOIT ÊTRE ÉVALUÉE AVANT D'ÊTRE UTILISÉE, POUR S'ASSURER QU'ELLE RÉPONDE AUX NORMES DU PFAN "?

Any AAP in a record with this note should be reassessed. Care should be taken to ensure that the AAP is consistent with the standards of the PFAN, and therefore with the application of the PS for 9.2.2.5.3.

Does the PFAN plan to use the MARC 880 (alternative script) field or use the 4XX fields instead?

Like NACO, the PFAN applies MARC 21 Model B for records with different writing systems. For this reason, the use of the 880 field is not currently permitted under the PFAN_-_Name_Authority_Manual_-_MARC_21_Supplement#880_Alternate_Graphic_Representation MARC21 Supplement to the Guide to Name Authorities.

Do I have to create authorities for names and titles that are not originally in Latin script?

According to the PFAN General Contribution Rules, " participating libraries are not obligated to establish an authority record for each access point included in a bibliographic record nor all the authority records related to a particular access point except in the following cases”. This includes access points for names and titles not originally in Latin script. For a list of these exceptions, see the contribution rules.

If a PFAN participant does not have the language skills to prepare a AAP using a systematically romanised form that conforms to PFAN practices where required, they should refrain from creating an authority record.

Last update: 2023-02-02