I have to agree with Peter. I think something is amiss or just not
If the position over the base is not important, then why have both zarqa and
tsinnor in Unicode (0598, 05AE). The glyph shape is all that matters then.
Note that tsinnor is also used only in the 3 books, so it could also be
distinguished from zarqa (only in the 21 books) -- just as you point out
tsinnorit and zarqa could be distinguished. So, why two glyphs in Unicode?
In the 3 books, there are two functionally distinct accents, tsinnor and
tsinnorit, and they are placed differently over their resp. bases. Here, the
placement is graphemic; it does distinguish the two accents. There would be
no other way to distinguish them, unless one applies phonologically complex
parsing rules. Based on this observation, I think Unicode needs two glyphs,
namely to distinguish tsinnorit from tsinnor/zarqa. Tsinor and zarqa look
the same and are placed the same over the base (upper left). Tsinnorit is
the centered glyph.
That the accent applies to the word or syllable (is suprasegmental) isn't
relevant, because we are dealing with the graphemic problem that the accents
are placed over particular base letters, not the phonological level. Also,
we are already assuming that exact placement of the glyph depends on the
width of the base letter (a subgraphemic variation). Rather, we are arguing
that there is a graphemic distinction between tsinnor and tsinnorit, both
used in the 3 books. No graphemic distinction exists between tsinnor and
So, we do need two glyphs, but tsinnorit needs to be one. The other can be
for tsinnor and zarqa.
I believe 0598 = tsinnorit, not zarqa (unless you interpret "zarqa" as the
generic, place-it-where-you-will glyph shape -- ok); and 05AE =
Note on spelling: Unicode spells tsinnor "zinor". I'm now taking my
spellings from the Tabula Accentuum that comes with the Biblica Hebraica
Stuttgartensia, which provides exacts transliterations.
Sorry to belabor such a seemingly minor issue...
From: Jonathan Rosenne [mailto:firstname.lastname@example.org]
Sent: Friday, January 29, 1999 9:57 AM
To: Unicode List
Cc: Unicode List
Subject: Re: question re: Hebrew accents
There are indeed two characters, Zarqa and Tsinor, 0598 and 05AE.
Zarqa and Tsinorit look the same and have similar meaning, but cannot be
confused because Zarqa is used only with the 21 books of the Old Testament
and Tsinorit only with the 3 EMET books (Job, Proverbs, Psalms).
The placement of accents depends on the shape of the letter and the points
it carries. They actually belong to the word, not to a specific letter of
it, and are normally placed on the consonant of the syllable that has the
The placement of the accents depends on the
At 07:52 29/01/99 -0800, Peter_Constable@sil.org wrote:
> J>Zarqa is basically "above". The detailed positioning rules of
> Hebrew points and accents are too complex and imprecise and are
> of no interest to character standards. Unicode provides just a
> general indication of the placement.
> J>For example, the placement of Dagesh, basically a dot in the
> center of the letter, is affected by aesthetic considerations
> and conventions such that most font makes prefer to have the
> each letter with Dagesh a separate font. Or Sheva, basically
> "below", is moved to the right on some letters.
> I don't think a single character zarqa really is adequate.
> Consider dagesh: as you've indicated, aesthetic considerations
> can require a slightly different position for nearly every
> consonant that it occurs with. Yet, for a given consonant it
> would only ever occur in a single position. The is different
> from the situation with the zarqa/zinor/zinnorit glyph, which
> can appear over a given consonant in more than one position
> with the different positions having different significance. So,
> there are two distinct characters needed, whatever it seems
> best to call them. Indeed, in each case aesthetics require
> careful positioning on a consonant by consonant basis.
> Nevertheless, it is my understanding that two distinct
> characters are needed here.
> Peter Constable
> Non-Roman Script Initiative, SIL
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:44 EDT