[b-hebrew] T-SADE

Arnaud Fournet fournet.arnaud at wanadoo.fr
Mon Jul 26 01:41:47 EDT 2010

----- Original Message ----- 
From: "Will Parsons" <wbparsons at alum.mit.edu>
To: <fournet.arnaud at wanadoo.fr>
Cc: <jc.bhebrew at googlemail.com>; <b-hebrew at lists.ibiblio.org>
Sent: Sunday, July 25, 2010 5:38 PM
Subject: Re: [b-hebrew] T-SADE

> On Sun, 25 Jul 2010 05:30:05 +0200, "Arnaud Fournet" 
> <fournet.arnaud at wanadoo.fr> wrote:
>> Actually we have plenty of evidence from Armenian, Hurro-Urartian and
>> Hittite on how these "letters" were pronounced in Akkadian in particular.
>> It's quite clear that the propagation of Hebrew phonetics into cuneiform 
>> was
>> inadequate and unfortunately probably impossible to emend.  Cuneiform $
>> (=shin) is in fact -s- and Cuneiform s and z (tsamekh) is in fact -ts- 
>> and
>> -dz-.  Anytime you move out of the Semitic "bubble", then it becomes 
>> clear
>> that there is a problem with graphemes $ and s which do not have the 
>> correct
>> values.  It is also clear from Armenian that the emphatic s. in Akkadian 
>> was
>> an affricate. Place names and other words written with <s.> in Cuneiform
>> invariably pop up in Armenian with an affricate.
> There is a fundamental problem in logic here.  Transcribing X in language 
> A
> by Y in language B does not mean that X was pronounced as Y.  If Armenian
> (for example) transcribes an "emphatic" with an affricate, that does not
> mean the "emphatic" was in fact realized as an affricate, only that when
> faced with indicating a phonetic feature that was not present its own
> language, some accommodation must be done, and how that problem is 
> resolved
> may be difficult to predict ahead of time.
> This is the same situation with the so-called transcription of צ by Greek
> τσ/ts.

Armenian has a large set of inherited fricatives and affricates, so I tend 
to think that speakers are sensitive to the contrast between sibilant and 
In addition the transfer from cuneiform to historically attested Armenian is 
extremely consistent.
So it does teach us something that Armenian has or does not have affricates.
What we can see is that Akkadian had a phoneme /s/ unfortunately transcribed 
s^, three affricates written s, z and emphatic s. unfortunately not 
transcribed as affricates. It also most probably had a lateral fricative 
which is not distinguished from /s/ in spelling.
Hittite and Hurro-Urartean basically confirm that situation.
Late north-eastern Assyrian probably fused s and ts as there is a tendency 
to use sh- and s- signs indiscriminately.

The case in Greek is different. Greek did not have the large set of 
inherited fricatives and affricates that Armenian had, so its value is 


>> The "aesthetical" argument was referring to your last sentence: "In 
>> either
>> case, we have parallel series /s - S - z/ like /t -T - d/ like /k - K -
>> g/. An affricate like [ts] doesn't fit this model."
>> "fit that model".
>> /s - S - z/ is actually /ts TS dz/.
> Interesting idea.  I take it that /s/ would still be in the phonemic
> inventory corresponding to sin?  What about plain /z/, do think that was
> absent?
As far as I know, Semitic had :
*s (=shin) *s. (fused with tsade in Hebrew)
*ts (=tsamekh) *dz *ts. (=tsade)
It seems logical that *z also existed but I'm not aware that we can prove 
the existence of a contrast between *z and *dz somewhere within the 
perimeter of Semitic languages.
Modern Arabic only has /z/ and that situation must be old for a number of 
reasons but northern Semitic languages like Akkadian and Ugaritic had dz in 
my opinion. And from Egyptian transcriptions which use t_ for z like Gezer = 
q t_ r it can be inferred that early Hebrew and other related "Cananean" 
dialects had an affricate dz.
Needless to say that I consider the approach of Saenz-Badillos (A History of 
the Hebrew Language) to be complete nonsense and not far from 
antiscientific. There is an overwhelming body of data that shows that 
approach to be completely wrong.


>> If the original value of tsamekh were /s/, then it's a wonder that this
>> letter has *never* been used anywhere from Greece to Ethiopia to Etruria 
>> to
>> write the sound /s/.  As a matter of fact about everything has been used 
>> in
>> Greek to write /s/ except precisely that very letter.  For that reason,
>> tsamekh cannot have been /s/, it was /ts/.
> Thanks for the clarification.  And in case I have given a wrong 
> impression,
> I'm not actually opposed to the idea that samekh may have been pronounced
> [ts] at some point in its past, only that it's going too far to claim
> that Egyptian transcriptions can "prove" such a pronunciation.
It can at least prove that it was not /s/ as tsamekh is never transcribed 
with Egyptian grapheme for s.
Now I agree that the exact value of the grapheme t_ is unclear and 

> As far as Greek goes, the fact that shin became the basis of sigma I would
> take as indicating that of the variety of sibilant letters available in
> Phoenician at the time of the Greek adoption of the alphabet shin/sin was
> closest to the Greek sound.  This can be taken to mean that the Phoenician
> sound of shin was [s], but it may also be the case that Greek sigma was
> pronounced farther forward than the normal [s] as pronounced in English or
> French.  (It should be noticed that Modern Greek has such a pronunciation,
> which makes it sound somewhat like an English sh.)
> So, emphasizing again that I am not saying that samekh *wasn't* pronounced
> [ts], the Greek evidence cannot prove that it was.
The conclusion is mainly negative. Whatever it was, it certainly was not a 

> First of all, we *don't* have TS in the Codex Vaticanus in the text at 
> all.
> There are numerous occurrences of Hebrew name with sadhe transcribed with
> simple sigma in the Vaticanus and other manuscripts of the LXX.  The one
> claimed use of tau+sigma is in a heading to the acrostical portion of
> Lamentations, and that seems uncertain.
I do not see why it is uncertain.
That occurence is certain!? even if it may be isolated.

Arnaud Fournet

More information about the b-hebrew mailing list