[b-hebrew] Tidying up Unicode Hebrew: Please indicate your support

Peter Kirk peter.r.kirk at ntlworld.com
Mon Jul 21 16:49:53 EDT 2003

If you would like the problems with Unicode and biblical Hebrew to be 
fixed once and for all, please read on, or skip to "==================", 
read the letter that follows, and if you agree add your name and other 
details and return it to me.

I have had an interesting variety of responses in just a few days to my 
recent message "A threat to the integrity of the Hebrew language?" Most 
have been strongly opposed to the other proposal which I which I 
suggested was a threat. Only one person has given it qualified support. 
There has little comment on my alternative proposal, but then that is to 
be expected from an audience of people who mostly have limited 
understanding of the detailed technical issues. Further responses are of 
course welcome.

Peter Constable, the author of the proposal which I was responding to, 
has told me that in the light of criticism already received "this 
proposal has a very low likelihood of being pursued". On the other hand, 
it has apparently not been withdrawn or rejected. On that basis I will 
continue to collect reactions to it, but I will not distribute them 
further for the moment.

The most interesting response has been from some people within the 
Unicode Technical Committee. They have indicated that there is a real 
possibility that the problem could be fixed properly, by correcting the 
errors in the combining classes. I was told earlier that "UTC absolutely 
will not consider changing combining classes", but it seems that this 
position is not really quite so absolute. But if these classes are to be 
changed it will still require considerable evidence that scholars and 
other users of biblical Hebrew, software implementers and content 
providers are united in supporting the principle of the change.

And so, to give you an easy way to demonstrate your support, I am 
distributing the following letter, which is intended to be from the 
biblical Hebrew community to the Unicode Technical Committee, and also 
to the Internet Engineering Task Force because the UTC needs to convince 
the IETF to agree to this  change. Please read the letter, and if you 
agree with it, fill in your name and other details and return the whole 
letter to me. I will add your details to a master copy, and send it on 
at an appropriate time. If you wish to make more detailed or 
personalised comments, I can pass them on as well.

Note that this letter calls for acceptance in principle of changing the 
combining classes but does not specify the details of such a change. The 
list of "custom combining classes" in Appendix B of 
may be taken as a starting point for drafting a detailed proposal. See 
also the section "The normalisation issue" of this manual (a draft, I 
think). Another, essentially equivalent, starting point is the list of 
canonical combining classes for the proposed new biblical Hebrew only 
characters in 
section E.3. But, as I have written in the last paragraph of the letter 
below, it is important that there is wide consultation before specific 
proposals for change are considered.

Peter Kirk
peter.r.kirk at ntlworld.com


To: The Unicode Technical Committee and the Internet Engineering Task 
Force (IETF)
From: Experts on ancient Hebrew and Aramaic

We, the undersigned, comprise a group of:

Scholars of biblical and other ancient Hebrew and Aramaic;
Implementers of software and fonts for input, processing and display of 
these languages;
Content providers making available texts in these languages.

We understand that some errors were made in specifying the combining 
classes of some of the characters in the Hebrew Points and Punctuation 
section (U+05B0 to U+05C4) of the Hebrew block of the Unicode standard. 
These errors were introduced in earlier versions of the standard and are 
still present in version 4.0.

The errors relate to the Hebrew vowel points and some other combining 
characters. It was intended that these vowels and points should be used 
for all stages and varieties of the Hebrew language. However, the 
canonical combining classes defined for these characters are not 
appropriate for ancient Hebrew and Aramaic texts, including the Bible, 
but only for modern languages including modern Hebrew in which vowel 
sequences are not used. With these classes as defined, the Unicode 
normalization procedure reorders these vowels and points in such a way 
that semantically significant distinctions in these ancient texts are 
lost. Therefore these texts, which are in regular use by large 
communities, cannot be represented properly in Unicode.

The canonical combining classes as currently defined are also in 
conflict with the principle that "Characters have the same class if they 
interact typographically, and different classes if they do not" (The 
Unicode Standard, Version 3.0, section 3.10, p.51, 
http://www.unicode.org/book/ch03.pdf). These Hebrew vowels and points do 
interact typographically with each other and with other marks with which 
they co-occur, especially Hebrew Cantillation Marks (U+0591 to U+05AF). 
The errors are therefore an inconsistency within the Unicode standard.

We recognize that The Unicode Standard Stability Policy 
(http://www.unicode.org/standard/stability_policy.html) does not 
generally allow changes to be made to the canonical combining classes of 
existing characters. However, we note that this policy does allow for 
exceptions to be made "when there is a clear and evident mistake 
identified in the Unicode Character Database". We consider that "a clear 
and evident mistake" has been identified in these combining classes for 
Hebrew Points and Punctuation, in that the classes as currently defined 
conflict with the general principles of Unicode and make it impossible 
to use these characters for the purpose for which they were defined.

The disruption to existing texts which might result from this change 
will be minimal. Modern Hebrew texts will be unaffected as vowel 
sequences are not used. Only a small number of content providers have 
attempted to encode in Unicode the ancient texts which might be 
affected. In general they support changing these combining classes 
because this fixes a problem which they have identified. They have 
indicated that they will make minor adjustments to their content if 
necessary in line with adjustments to the Unicode standard.

We call on the Unicode Technical Committee and the Internet Engineering 
Task Force (IETF) to allow changes to be made to the canonical combining 
classes of the combining characters in the Hebrew Points and Punctuation 
section (U+05B0 to U+05C4) of the Hebrew block of the Unicode standard.

We also call on the Unicode Technical Committee and the IETF to consult 
broadly with the community of scholars, implementers and content 
providers for all stages of the Hebrew language, including with 
ourselves, before formally considering specific proposals for 
adjustments of these canonical combining classes or for other changes to 
the Hebrew block of the Unicode standard. There are a number of other 
small issues with Unicode support for ancient Hebrew, and it may be 
necessary to propose a small number of new characters or other changes. 
It is important that any changes made now are made correctly and in ways 
which are acceptable to all users of Hebrew texts.

Name: Peter Kirk
Position: Associate Editor, Key Terms in Biblical Hebrew project
Affiliation: Independent researcher working on a project led by SIL 
E-mail address: peter.r.kirk at ntlworld.com

E-mail address:

More information about the b-hebrew mailing list