Biblical Hebrew Forum

["Yitzhak Sapir" <yitzhaksapir AT gmail.com>]

On 3/10/06, Herman Meester wrote:
> I would like to ask the interested members' opinion on the origin of
> the word שתים shtayim.
> Remarkably, this word has a shwa with the shin, but a dagesh in the taw.
> A so-called shwa na` (mobile) is supposed not to be followed by a
> sharp begedkepet consonant sign: it should have a soft
> "veghedhkhefeth" variant, if the shwa we are dealing with had once
> been a vowel - i.e. if it really is a shwa na`.

I think there is a simple answer to your question: the schwa in
$tym is a vocal schwa, and the t has a daggesh xazaq.  No Massoretic
rules are broken, and it makes the most sense (to me, at least) from
a comparative linguistic point of view.  That is $tym is to be read as
"$:ttaym."

The dagesh in taw was not, at least not originally, a dagesh qal, which
is the one that maintains the bgd-kpt principle.  It is the "feminine"
form of "$naim".  The "n" before the "t" drops out leaving a dagesh
xazaq in the "t".  The suffix "aim" is the same in both because it is the
dual suffix and it doesn't change between feminine and masculine words.
Compare qarnaym (dual) but qarnot (plural), yadaym (dual) but yadot
(plural), and also the Biblical word "xomotaym" (two walls).  In his article
on ProtoSemitic, Huehnergard reconstructs a final "-m" and "-na" in a
non-construct form of singulars/plurals and duals (respectively).  While
he suggests the final -im plural endings resulted from something else,
for me it appeared that they were vestiges of this original ending.  This
would explain forms like xomotaym: "xom" (root), "o" (from "a:", perhaps
from some kind of plural marker in Semitic), "t" (feminine ending),
"ay" (from Semitic genitive-accusative dual marker "ay"), and "m" (the
normal non-construct ending of words, although "na" is reconstructed
for the dual).

Taking all this together, it seems that $taym is roughly equivalent to
xomotayim.  The t is "hard" because it was doubled, assimilating the
original "n".  This suggests: $ttaym < $ntaym.  Now, if we suggest that
$ in the beginning of this word was non-vocal, not only do we break the
Massoretic rule "words begin with a vocal schwa" but also a Massoretic
rule of having a vocal schwa before a doubled consonant.  Rather than
break two rules like that, I think that this form "$ttaym" suggests the
Massoretic "$:ttaym" (ie, vocal schwa na) and if it is to be read as
"$taym" as Ken Penner quotes Lambdin that is probably a secondary
development of losing the schwa na and the doubling as a result.
Perhaps Lambdin did not realize that the dagesh in the t is probably
a hard one and did not want to read "$:taym" where Massoretic
bgdkpt rules suggest "$:thaym".

David Testen has suggested that the original form of certain words had
no vowel, with a sonorant *l, *m, or *n taking the place of the vowel in
the syllable.  "Examples of this include the ancient substantives *bn,-
'son',  and *sm,- 'name', the pronominal forms *sm, 'they (masc.)' and
*sn, 'they (fem)', and the proclitic asseverative particle *l,-." 
(Huehnergard,
in Afro-Asiatic, Cambridge Encyclopedia of the World's Ancient
Languages, which I quoted above)  He later notes that also the second
person pronouns ("-tm," "-tn,") as well as the number "two" ("*thn,-").  He
also comments, when dealing with syllabic structure, that "Clearly, the
generalizations just enunciated would need to be modified if such forms
as *bn + case ending -- that is, *[bn,um] 'son' (CCVC?, CVVC?) -- are to
be considered valid in Proto-Semitic."  He references David Testen, 1998,
"Parallels in Semitic Languages: The Development of Arabic la- and
Related Semitic Particles." Studies in Semitic Languages and Linguistics
26. Leiden: Brill.  So Testen's theory may explain quite a few features of
various Semitic languages, but it reconstructs form in Proto-Semitic that
don't follow the normally observed consonant patterns.

In an earlier (1985) article on the subject, David Testen writes "[Arabic]
)ithnata:ni calls for [a proto-form] *thn-at-.  Corresponding to the forms
with a vowelless suffix, on the other hand, are such words as ...
from *thn-, Akkadian $ittu,:, Arabic thinta:ni, Aramic tarte:n, and
evidently, Hebrew $ta'yim."  That is, he suggests that the original Proto-
Semitic form of the root was a *thnt (*thn for the root, and -t for the
feminine ending).  Together with Huehnergard's earlier comments on PS
endings, I guess this means that "two" was in PS "thnta:na" (nom.) or
"thntayna" (gen./acc.).  This article is to be found in "The Significance
of Aramaic r<*n" JNES 44, 143-146.

David Testen replied to a related question of mine on this subject here:
https://listhost.uchicago.edu/pipermail/ane/2004-May/013572.html

Yitzhak Sapir