[b-hebrew] Two X letters

Yitzhak Sapir yitzhaksapir at gmail.com
Mon Oct 29 06:35:35 EDT 2007

On 10/29/07, Isaac Fried wrote:

> OK, OK, you are right. If only for the sake of peace and harmony on
> this distinguished list I will climb down the rhetorical tree and
> settle upon the solid ground of stolid and dignified discourse.
> The claim made [correct me please if I am wrong] is that once upon a
> time in the very distant and foggy past the ancient Hebrews, or
> ancient Canaanites [the mythical proto-Semites], had two close, yet
> semantically distinct, speech-sounds that merged into one sound that
> was eventually recorded by the inventors of the alphabet as the
> single letter X.

This is not exactly correct.  The claim is made that in the distant
past, there was a community of speakers who spoke Proto-Semitic.
This language had additional phonemes not found in the Tiberian
vocalization of Hebrew.  Some of these did survive as independent
phonemes in Classical Arabic.  The inventors of the alphabet
originally had different letters for them, but a radically simplified
version of the alphabet had only 22 letters, suggesting that already
then some sounds had begun to coalesce.  In Hebrew, some of the
sounds remained distinct even through the Biblical period.  On this
you can read Steiner's article here:

> This claim of the twin X seeks to draw its approval from its
> existence in Arabic, fortified by the assertion that the Arabs,
> conservative in their ways, have preserved the sounds of the ancient
> Semites. It is a plausible argument but like many other apparently
> logical arguments it may well be a dud.

The claim of the multiple X is not only based on its presence in
Arabic.  It is based on its presence in several languages, ancient
and modern, including Classical Arabic and Ugaritic, but also
other languages in Semitic and the even larger family called
Afroasiatic.  More importantly is the fact that cognate words can
be identified in these languages.  Let us take a hypothetical root
ABC, where any one of the consonants ABC may be X1 or X2.
What we find is that where Arabic has, say, A-X2-C, and Ugaritic
has a similar root A-Xn-C, in a close meaning, then Ugaritic will
have X2 just like Arabic.  Where Arabic has X1, Ugaritic will have
X1.  The general relationship of these words, and there are many
such words, therefore seems clear.  There are several different
ways of explaining this:
1) X1/X2 in Arabic and X1/X2 in Ugaritic are reflexes of two similar
phonemes in a parent language from which these two evolved.
The most immediate parent language of Arabic and Ugaritic is
Central Semitic.  So, option #1 says that in Central Semitic
there were two different consonants X1 and X2 and these were
maintained in Classical Arabic and Ugaritic but lost some time
until Tiberian Hebrew was recorded.
2) X1/X2 in Arabic and X1/X2 in Ugaritic are two parallel
developments of a split in some consonant that was preserved
whole in Tiberian Hebrew.  This option is very unlikely since
we have a great many words in both languages that show
this cognate relationship.  We have to explain this "split"
somehow.  That is, to find a plausible reason for X to split
into X1 and X2.  The problem is that in modern languages,
sound changes like that take place irrespective of meaning
but instead, are conditioned by the phonetic environment
surrounding the consonant.  The cognates show a variable
phonetic environment but consistent correlation in meaning,
making a split highly unlikely.

> The Arabic alphabet in its
> present form is a [relatively] modern committee-made product
> introduced only after the revelation of the Koran and motivated by
> the desire to immediately have a permanent written record thereof.

There is also the Old South Arabic alphabet as well as the Ugaritic
alphabet to take into consideration.

> The over-dot was possibly added to accommodate various dialectical
> differences [thus they are signs for "phones" not "phonemes"],

The only interest here is Classical Arabic's dialectical representation at
that time.  It is recognized that Classical Arabic dates from approximately
the same time as Tiberian Hebrew.  This is not the case with older Arabic
dialects and Ugaritic.

> or indeed for the purpose of semantic cleavage judged desirable due
> to the proliferation of Arabic roots. In other words, the process may
> have well advanced in the opposite direction: in the beginning there
> was only one X, then later it branched into a pair of Xs.

See the above discussion.

> This is
> what might have happened in Hebrew, a single letter spawning, as the
> language grew and demand for finer semantic distinctions increased,
> the five equivalents G, X, H, K, Q now coexisting in it.

Linguists assume as a principle (called the Uniformitarian Principle) that
linguistic processes observed today (most importantly, sound change and
analogy) were active in the past as well.  Linguistic processes not
observed today should not be assumed for the past as well.  Thus,
your five equivalents, in order to stand up to rigorous linguistic scrutiny,
must be explainable on the basis of either sound change or analogy.
Furthermore, there must be evidence, either from internal reconstruction
or from an attested language, that the five equivalents were originally
one.  Lastly, this issue of five equivalents must also be explainable in
terms of development of languages.  That is, developments in Hebrew
should first be reconstructed for Proto-Canaanite, which included at
least Hebrew and Phoenician and maybe also Ugaritic (this is a point
where linguists disagree).  Changes traced to Proto-Canaanite should
be then explained in terms of Phoenician, since Phoenician also
developed from Proto-Canaanite.  Only then, should changes be traced
to Proto-Northwest Semitic.  Changes that are traced to Proto-NWS
should then be explained in terms of Aramaic, since Aramaic also
developed from Proto-NWS.  Then to Central Semitic, with changes
explained in terms of Arabic.  Further stages include West Semitic,
Semitic, probably North Afroasiatic (Egyptian and Semitic), and
lastly Afroasiatic.  Changes that happened before Afroasiatic can
only be reconstructed using internal reconstruction of Afroasiatic
since we don't know of any cognate language to Proto-Afroasiatic
with which we can compare Proto-Afroasiatic.  In general, internal
reconstruction is somewhat limited in the amount of information it
can produce.  Having mapped out all the changes as above would
still not prove that these changes took place, but it would provide us
not only "they might have..." but "this is how it happened."  Just
having to explain the "how" means that all kinds of side-effects
might be theorized and those side-effects can then be checked.
In other words, a prediction of the theory that can be verified.
An example of such a theory is the Laryngeal Hypothesis for
Indo-European, which was confirmed by Hittite.

> Pronunciation is a fickle thing --- spoken Hebrew, for instance, is
> rapidly losing at this very moment many of its H sounds.

Modern spoken Hebrew is a hybrid language that can hardly be used
to reconstruct any ancient form of Hebrew.  So the developmental
methodology described above stops very early and can't even
progress beyond "Common Hebrew" much less to Proto-Canaanite
or Proto-Semitic or Proto-Afroasiatic.

> The personal
> pronouns are invariably reduced now to a mere U [pronounced oo as in
> 'foot'] for HU), 'he', and a mere I [pronounced ee as in 'feet'] for
> HI), 'she'. Educated Hebrew speaking people will routinely and
> nonchalantly say TA-BAYIT, 'the house', instead of the full ET HA-
> BAYIT. We are all for the short and the easy.

These processes are understood well within the contexts of
sound changes and analogy.

> Deep throated ayin is also out, and Hebrew is not poorer for it.
> The way foreigners record Hebrew names containing sounds strange to
> their ear is little proof as to how these names are pronounced by the
> Hebrews. Look at what they did to the name Isaac. All I am ready lo
> learn from AZA being said GAZA is that the city name was pronounced
> by some indigenous people with a baffling, non-reproducible, deep
> throated ayin.

Again, the issue is the ability to trace such Ayin-Ghayin words to
similar cognate words in Ugaritic and Classical Arabic, among others.

> Modern Hebrew linguists are espousing the dual-X-theory with such
> enthusiasm  since by dint of this hypothesis they grant themselves
> the license to consider, practically at will, any X as being either
> one of two "different" letters. It makes their life easier.

No.  They can only work on the basis of recognized cognates in other
languages.  Easy is one thing it is not.

> Any Hebraist is constantly being called to explain, to himself, or to
> fended off pressing students, this strange phenomenon of Hebrew
> having two [or more] equiconsonantal words, say NAXAL, 'brook', and
> NAXAL, 'inherit', of apparently remote meaning. One easy way of
> extricating oneself from this dilemma is to make the dogmatic claim
> that these are two "different" X letters,

Without cognates, one can only suggest that these may be originally
two different letters, not that they definitely are.

> and hence the two words are
> not the "same". They look the same but are actually genetically
> distinct. One is then in an instant in the free --- nothing left to
> explain.

Again, see the above for the amount of explanation required of both

> I don't buy, not for a moment, this sneaky, ad-hoc, deus ex machina,
> argumentative device, which in my opinion is no more than an
> instrument of deception, self delusion, and slippery dodging.

Up until now, it was rather a civilized representation, but don't you think
it would be appropriate to first read up on what linguists really say, before
you go off contesting what they say?

Some good articles are:
Ringe, Don. "Reconstructed Ancient Languages" in Cambridge Encyclopedia
of the World's Ancient Languages

Faber, Alice. Genetic Subgrouping of the Semitic Languages. in
The Semitic Languages (ed. Robert Hetzron; London: Routledge, 1997),

Huehnergard, J., "Historical Phonology and the Hebrew Piel," in Linguistics and
Biblical Hebrew (ed. W. R. Bodine; Winona. Lake, IN: Eisenbrauns, 1992) 209-

Huehnergard, J.  1990.   "Remarks on the Classification of the
Northwest Semitic Languages," in The Balaam Text from Deir Alla
Re-evaluated: proceedings of the international symposium held at
Leiden, 21-24 August 1989.  Pp.282-93.

> Moreover, it provides only the briefest of reprieves since it fails
> to apply to similar situations with other letters. The Hebrew words
> KPAR, 'village', KPOR, 'frost', KPYIR, 'lion cub', KOPER, 'pitch',
> KOPER, 'ransom',  KAPORET, 'cover' and KAPARAH, 'expiation' are all
> derivatives of the root KPR, yet no resort is made here to an
> argument of multiple kaps, probably because Arabic is devoid of such.

Reconstructing multiple phonemes in cases that they have coalesced
in all descendant languages in the exact same way is very hard as it
can only be done by internal reconstruction.

> Being hardly a science [it is certainly not an exact science],
> Linguistics thrives on consensus.

Why is it not a science?

> Without this, much of it would have
> fallen apart and dissolved under the weight of internecine bickering
> and dissenting opinions of horrifying heretics questioning its
> fundamental premises. This is why Linguistics is so apprehensive of
> the fringe and holds so dear the "mainstream", making it largely the
> scholastics of the quote, "she points out in her paper, he says in
> his book, they write in their dictionary". It tends also, as in other
> sciences, to harden into universal truths obvious, self-serving,
> platitudes and fallacies, stifling thereby any attempt at innovative
> thinking.

Let's make believe you didn't say this last paragraph, ok?

> It is the duty of every free thinking, Hebrew loving, man to decry
> them as such as soon as he sees one; in the words of Psalms 137:7

Free thinking is definitely encouraged.  I myself have various theories
that contest the mainstream view.  I generally don't advocate them on the
lists that I am signed up, however, and when I do present a position I
try to differentiate my (as yet unpublished and/or not reviewed) theory
from the general consensus of scholarship.

But free thinking cannot be a way to let in through the back door wild
guesses that have no basis in any methodology or any way of verification,
and maybe even have no dependence on previous studies in the field.
Free thinking cannot be a reason to reinvent the wheel.  Scholars have
debated the issues you are raising.  They considered them from all
aspects.  Before you go off saying that they just don't want to listen,
first listen to all they have to say and especially try to find if they
considered your ideas in the past and if so, what were their objections.

Yitzhak Sapir

More information about the b-hebrew mailing list