Hebraica Home

Non-Roman Script Issues



New: keeping or adding original script 6XX fields.

Appendix C.5. Oriental numbers
Appendix C.3. Arabic Numerals vs. Numbers Expressed as Words
Justification in AACR2 for Retaining Original Script
[PCC Core Guidelines for Multiple Character Sets]
Conser Editing Guide: Appendix O (re CJK serials, but makes sense to apply to all JACKPHY, and for all formats)

Unicode bidirectional script formatting (from RLIN21 "Tech notes")

Formatting of original script 260 field subfield $c (Old RLIN)

Coding for original script parallel fields for established headings

Proposal 5JSC/LC/2

AACR2R Appendix C.3. Arabic Numerals vs. Numbers Expressed as Words
C.3B1:Substitute arabic numerals for numbers expressed as words in the following areas and elements of the bibliographic description:

a) in an edition statement
b) in the material (or type of publication) specific details area unless a particular rule directs otherwise (see 3.3B2)
c) in the date of publication, distribution, etc., element
d) in the physical description area
e) in the series numbering.

AACR2R Appendix C.5. Oriental Numbers
C.5A. In cataloguing Arabic alphabet, Far Eastern, Greek, Hebrew, Indic, etc., materials, substitute roman numerals or Western-style arabic numerals for numerals in the vernacular as instructed in the following rules.

C.5B. Use roman numerals in romanized headings for persons identified by numerals (e.g., rulers).

C.5C. Use Western-style arabic numerals in romanized headings for corporate bodies and in uniform titles.

Thawrat 25 Māyū, 1969
Lajnah al- ‘Ulyā li-Iḥtifālāt 14 Tammūz.

C.5D. Use Western-style arabic numerals in the following areas and elements of the bibliographic description:

1) in an edition statement
2) in the material (or type of publication) specific details area unless a particular rule directs otherwise (see 3.3B2)
3) in the date of publication, distribution, etc., element
4) in the physical description area
5) in the series numbering

C.5E. Consider inclusive dates and other numbers to be a single unit in languages that are read from right to left.

1960-1965 not 1965-1960

Add punctuation to the left of the inclusive dates or numbers.

CONSER Editing Guide Appendix O

CONSER Editing Guide
Appendix O has the following general guidelines for CJK records which
probably can be applied to other non-Roman records.
4.   Participants who opt to create CJK records must supply non-Roman
script data for all elements listed below (when applicable):

Main entry/Uniform title:   1XX, 240
Descriptive areas: 1-4 and 6   245, 250, 260, 362, 440, 490
Non-topical subject headings:   600-630, 651
Added entries:  246 (when first indicator is set to "1" or "3"),
700-730, 800-830
Linking entries:   580, 765-787 (when available...)

5. Descriptive fields 245, 250, 260, 362, 440, and 490 are considered
"core" fields; parallel non-Roman script data is given only for those
elements in these fields that appear in non-Roman script and which are
transcribed from the item itself.

6. Non-Roman script data should never be given for note fields (such
as 310, 321, and 5XX) other than 580. These fields are always given in
English, since the language of CONSER cataloging records is English.

Parallel fields for est. headings

Original script representation of established English/Romanized headings: Hebraica Team supplies original script parallel fields for 100, 700, 440, 830, 600, headings. Even though the original script headings are not (yet) under bibliographic control, Hebraica Team will provide the same tagging and indicators as are given for the established Roman-script counterparts.

For example:


$aMusṭerṿerḳ fun der Yidisher liṭeraṭur ;$v30sṭer. bd.


‏‏מוסטערװערק פון דער ײדישער ליטעראטור ; 30סטע בד.

Non-Roman 6XX fields

When copy cataloging, if 6XX headings are already matched by corresponding original script headings, let them remain. However, do not add new 6XX original script records unless there is reason to believe that in this particular case the added access point makes a significant difference (e.g., adding רמב״ם).


Justification for Retaining Original Script

AACR2 states a preference for exact transcription over systematic romanization. For example, 1.0E says that for title and statement of responsibility, edition, publication, distribution [...], and series data, one should give information transcribed from the item itself in the language and script (wherever practicable) in which it appears there. In the age of Unicode, original script transcription is definitely practicable, and therefore (according to AACR2) should be preferred. With respect to access points, however, the rules insist that headings and references must be represented in roman characters. This is stated in 22.3C2, e.g., where the cataloger is instructed to romanize personal names according to approved conversion tables. Similar instructions are given for corporate bodies (24.1B1) and uniform titles (25.2D).

Now, while AACR2 states that all references need to be romanized, it doesn't say the these references couldn't *also* be represented in their original scripts. Assuming these rules are retained in the forthcoming Resource Description and Access, could this loophole could be exploited to get original script access points back into our catalogs?

Descussion of RDA Proposals


RDA A1.4E. Date.

(a) For published resources, transcribe the date in which the resource was published as it appears on the resource, usually only as year(s).

[i] Date not in Western-syle Arabic numerals. Add the year(s) in such numerals in square brackets as needed for comprehensibility.

This would seem to support a new practice of transcribing exactly what cataloger finds on the source, and represent a change, I think, from current AACR2 practice, which stipulates:

"1.4F1. For published items, give the date (i.e., year) of publication, distribution , etc., of the edition revision, etc., named in the edition area. If there is no edition statement, give the date of the first publication of the edition to which the item belongs. Give dates in Western-style arabic numerals . If the date found in the item is not of the the Gregorian or Julian calendar , give the date as fo8und and follow it with the year(s) of the Gregorian or Julian calendar ..." [emphasis mine]

As Keiko points out, "current AACR2 [...] says to use Western-style Arabic numerals" not just in Appendix C but even clearly spelled out in the descriptive rules. Contrariwise, the RDA proposal says to transcribe data exactly as it appears on source. Thus, in example provided by Keiko (taken from AACR2 1.4F1, and used in lieu of oriental numerals which defy non-Unicode software), the new proposal would support the following transcription:

anno XVIII [anno 18, 1939]

rather than current rules, which would prescribe:

anno 18 [1939]

On the other hand, if Appendix C.5D3 (covering representation of Oriental numerals in date transcription) carries over into RDA, the loop-hole would be closed again, wouldn't it. I'm not sure whether RDA is moving toward a single publication data element, or whether it will continue to support the (MARC, etc.) parallel fields technique for original vs. romanized transcription. If the former is the case, then I think the current JSC proposal (assuming a similar revision of Appendix C) might be a welcome change for us. As of July 2005, CCAAM (Committee on Cataloging Asian and African Materials) and MELA (Middle East Librarians Associaiton) are reviewing this proposal. Perhaps AJL needs to be involved as well. [cf. 7/9/05 email message to Heidi]

Back to top

This file last modified 05/04/06