Questions for Rolf

Rolf Furuli furuli at
Mon Apr 30 10:53:50 EDT 2001

Dear Henry,

Let me try some observations and some questions in order to see if there is
some sense in the nonsense.

>>> From: Greg Doudna
>>> (2) is it not the case that the waw-conjunctive/ waw-conversive
>>> distinction is reflected in the MT vowel pointing?  If so, this
>>> means there was such a distinction understood at least by the
>>> Middle Ages.  In your theory did such a distinction enter
>>> post-Qumran era? Any ideas on when and how such a distinction arose
>>> secondarily, post-BH and post-QH?
>> Origen in his Hexapla does not distinguish between WE- and WA(Y) as
>> prefixes to YIQTOL; both are transcribed as OU-. Several manuscripts
>> with Palestinian pointing differ from the Masoretic text as to
>> pointing; WEYIQTOLs are pointed as WAYYIQTOLs vice versa.  I think
>> there is good reason to believe that while the Masoretes were
>> extremely faithful to what they heard in the synagogue, and would
>> not dream of inventing anything new, the difference between the
>> "consecutive" forms is graphically their invention.  There is
>> evidence that the default pronunciation of shewa in Masoretic times
>> was an "a"-sound, and I think that the Masoretes "semantically"
>> would use shewa and patah indiscrimately but their phonetic rules
>> distinguished between them.  My suggestion, therfore, is that the
>> difference between WAYIQTOL and WEYIQTOL was invented by the
>> Masoretes, not on semantic grounds (they were not grmmarians and did
>> not even know about the triradical nature of roots, and their
>> Massora is not grammatical), but partly because of their phonetic
>> rules, and partly because they saw little or no "semantic"
>> difference between patah and shewa.  However, from the tenth century
>> onward, in the infancy of hebrew grammar, the graphic and phonetic
>> differences in the text were interpreted as *semantic* differences,
>> and the four-component model of Hebrew conjugations was born.  This
>> model have survived to the present.  If there is an important
>> semantic difference between WEYIQTOL and WAYYIQTOL, why did not the
>> scribes at Qumran mark this,in order to avoid ambiguity, when they
>> else used so many plene vowels? In some Aramaic inscriptions, shewa
>> is written as YOD. Are you aware of any examples from Qumran where
>> patah is written plene?
>I'm sorry, Rolf, but this part is pretty much nonsense.  The Masoretes
>were not abstract "metaphysical" theorists, and did not
>self-consciously "invent" things to fit in with any grand preconceived
>schemes (certainly none of Kahle's accusations of "artificial"
>features in the Masoretic transcription -- as opposed to
>"conservative" features -- have really stuck).  And yet _wayyiqtol_
>and _w at yiqtol_ are not distributed randomly through the text in terms
>of overall contextual positioning, so that if anybody "invented" the
>distinction, it had to be the Masoretes, not the grammarians.  Also,
>the Masoretes were rather careful about recording fine distinctions in
>their pronunciation.

I appreciate your great expertice in connection with phonology; for my part
I studied the work of the Masoretes 8 years ago when I worked on my
thesis. I will return to this in the future, so some of my comments are
taken form memory.

We agree that the Masoretes did not invent any new grammatical traits, and
that their work is based on what they heard recited in the synagogue. What
I ask, is whether the difference between WAYYIQTOL and WEYIQTOL is
*semantic* in a grammatical sense, or whether its basis is something else.
I have suggested that the Masoretes could have introduced the mentioned
difference without having a semantic difference in mind.

As to empirical data, there are no traces of a difference between WAYYIQTOL
and WEYIQTOL in the DSS or the Hexapla, and Palestinian manuscripts differ
from the MT as to the difference between the two forms. So, there is no
graphical difference between the two before the 8th century C.E. I agree
with you that other phonemic traits neither are graphically visible, so a
lack of evidence does not prove anything, but it should be kept in mind.

The basic differences between WAYYIQTOL  and WEYIQTOL  are: 1) patah versus
shewa, 2) gemination versus simnple consonant,and 3) penultimate accent
versus ultimate accent. I do not count long form versus short form because
many WEYIQTOLs have the short form and some WAYYIQTOLs use the long form
when the short is available. I have a list of 147 WEYIQTOLs which are
apocopated (either consonant or vowel), and a list of 83 WEYIQTOLs which
could have been apocopated but which are not.

QUESTION 1. In those verbs where the stress is retracted, is it possible to
view the three differences as one and the same thing? I leave out
diachronic views, and just ask if this is a possible explanation.

QUESTION 2. Is there any semantic reason why the article HA and the
interrogative MA(H) causes gemination of the following consonant but the
interrogative HA: does not cause gemination? If there is no semantic
reason, why the difference?

If we use the particle MA(H) as an example, we find a doubling of the yod
in substantives with yod as first radical, but where the first consonant in
the following word is a laryngeal we find MF rather than MA(H) (both occur
in 2 Kings 4:2). This is the same as with WAYYIQTOL where the waw has a
qamets in 1. p. sing. verbs. Where a substantive with YE: as first syllable
follows MA(H)  there is no gemination as is the case with WAYYIQTOLs. But
what about apocopation? Because there are very few instances where a verb
with yod as its first letter follows MA(H) I am aware of only one example
where a test is possible, and that is Psalm 21:2. In Habakkuk 1:15 we find
a WEYIQTOL of the verb in pause with ultimate stress as we expect, and in
Psalm 16:9 we find a WAYYIQTOL with penultimate stress, which we also
expect. What is interesting, however, is that this WAYIQTOL is pointed by
segol as is the case in Psalm 21:2 where we also find penultimate stress.
This suggests, in my view, that the patah, gemination, and retraction of
stress in WAYYIQTOLs do not signal anything semantic, because the same can
occur with other prefixes or pre-words which cause gemination, as in the
case of MA(H).

  Though (as I discuss at length in my
>dissertation) they did transcribe some non-phonemic features of their
>pronunciation, while not transcribing one or two of the
>(non-suprasegmental) phonemic distinctions in their pronunciation
>(this precisely because they didn't really have any overarching
>linguistic theories), they were very consistent and systematic in
>transcribing those features of the pronunciation that they did choose
>to make part of their orthography, so that it's somewhat ridiculous to
>say that they somehow loosely sloppily erratically randomly confused
>WAYYIQTOL and W at YIQTOL, forms which have a two-fold difference in
>pronunciation (in non 1st. sg. forms) -- both the vowel following the
>consonant of the conjunction and also the gemination of the verb
>prefix consonant.  Reduced vowels (such as vocal _sh at wa_) and
>unreduced vowels (such as _pathah._) are two quite different kinds of
>entities in the Tiberian system, with different patterns of
>occurrences, and are only very rarely violate the distributional and
>other restrictions placed on each category (almost the only real case
>is the _qamees._ vowel in the initial syllable of _qodhaashiim_ and a
>few similar forms, where a _qames. vowel would be expected).
>It's certainly true that vocalic _sh at wa_ tended to coarticulate with
>neighboring sounds (rather than having a fixed definite mid-central
>quality), and that in absence of strong coarticulation it often tended
>to take on a lowish quality, at least in some pronunciations (as
>evidenced in the sporadic transcription of vocal _sh at wa_ with
>orthographic pathah._ when not following a guttural
>consonant, seen in various manuscripts), but this did not necessarily
>lead to any systematic confusion between vocalic _sh at wa_ and
>pathah._ -- much less to any confusion between prosodically reduced
>vocalic _sh at wa_ (which is never followed by a geminate consonant) and
>prosodically unreduced _pathah._ followed by a geminate consonant!
>So far you haven't presented any plausible scenario which explains in
>adequate phonological/morphological detail how the geminate consonant in
>WAYYIQTOL originated (much less the stress distinction between W at YIQTOL
>and WAYYIQTOL), nor how the Masoretes were motivated to artificially
>create a distinction which did not exist in previous stages of the
>language.  This is not one of those things (like sound changes) that
>simply arise out of the blue; it's a restructuring and reanalysis of the
>whole system that requires a highly specific explanation.  You can see
>my attempt to provide such an explanation for the origin of W at QATAL
>coming to have a separate morphological and phonological status (to the
>degree that it does) in section 4.4.2 of my dissertation; until you
>provide a comparable explanation for your proposed differentiation
>between W at YIQTOL and WAYYIQTOL, then your theory is too incomplete to be
>convincing to historical morphologists/phonologists.

I have no quarrel with you regarding your points above; I agree. I am not
saying that the Masoretes "somehow loosely sloppily erratically randomly
WAYYIQTOL and W at YIQTOL". To the contrary, I suggest that the Masoretes
consciously and systematically (though there are many "slips" which suggest
that the system was their "invention") chose shewa after waw with some
YIQTOLs and patah with most YIQTOLs.


First I will cite some evidence that the default pronunciation of shewa in
Masoretic times was an "a"-sound. K. Lewy did a thorough study of the
subject. He translated and commented upon a treatise called "Die Abhandlung
über das Schwa" which he ascribes to the time and tradition of Ben Asher.
Lewy´s translation shows that shewa was pronounced as patah (Levy Kurt,
"Zur Masoretischen Grammatik", Bonner Orientalistische Studien, Heft 15.

"Einer von ihnen <handelt vom Schwa> am anfang der Wörter; und das ist
stets bewegt /sewah mobile/, und seine Artikulation und Aussprache ist mit
Patah, wie (...) u.a. Sie alle werden beim Lesen mit Patah gesprochen. Und
dies ist ein Grundsatz, der niemals  durchbrocken oder aufgehoben wird, das
nähmlich jedes Schwa, welches am Anfang des Wortes steht, sich nur bewegt
findet /Only four words are exceptions/ Bei jedem mit Dagesch versehenen
Konsonanten in der Mitte des Wortes, unter dem ein Schwah steht, wird das
Schwa stets, ohne Ausnahme, mit Patah gelesen, und er gibt dabei überhaupt
keine Abweichung."

This situation is not very different from Qumran. In his discussion of the
Isaiah scroll, Kutcher also discusses phonological traits. His material
shows that shewa was probably pronounced either as the following vowel or
as an a-sound. Says he (Kutcher E Y. The Language and the Linguistic
Background of the Isaiah Scroll. Studies in the Texts of the Desert of
Judah, Vol 6, J van der Ploeg, Leiden. 1974:500-501):

"The transcriptions indicate that the shwa mobile is likely to have the
quality of the following vowel. This is the case in the Sept. Josephus, in
Jerome and also Origenes."

Because of an alef in 34:6 where MT has shewa he also concludes that "Shwa
mobile = a (or one of its reflexes) which also is confirmed by the LXX and

The Masoretes based their pointing on what they heard. If patah and shewa
were pronounced similarly, how could they distinguish between them? We do
not know. However, the pausal forms indicates that there was a king of
rythm in the reading, and vowels at the end of the clauses would get some
kind of stronger stress; this was marked by a lengthening of the vowel.
Generally speaking, the environments where WAYYIQTOL  and WEYIQTOL occur
are different. Of the WEYIQTOLs 65% are modal, 41,9% FOLLOW a YIQTOL, 25%
follow an imperative, and only 2% follow a QATAL. The WAYYIQTOLs for the
most part occur in narrative, and represent staccato consecutive action. I
therefore suggest that the prononciation of patah and shewa were similar,
but either 1) the strength/marking of the narrative sounds was different
from the strength/marking of those expressing modality,or 2) the
strength/marking was similar but the Mesoretes realized that there were two
completely different environments where the sounds occurred. On the basis
of either 1) or 2), or a combination of both, the Masoretes chose to use
shewa for one group of YIQTOLs+waw and patah for the other group.

It is true that shewa and patah function differently in their system, but
the choice of one of them need not be phonemic. If both represented an
"a"-sound, either could be used without a difference in meaning. To show
that this at least is a possibility I refer to Murtonen. He has done much
work on non-Tiberian material. Regarding Palestinian and Babylonian texts
he found tens of roots where /e/  occurred in free variation with /a/
without any difference in meaning and concludes that /e/ is used in a way
as if it was an allophone of the /a/ phoneme. In Samaritan manuscripts he
found that /ä/ usually and /e/ frequently interchanged with /a/ (His //).
(Murtonen A, 1986. "Hebrew in its West semitic setting, Part 1, A
comparative Lexicon", Leide:Brill)

I also quote Harviainen (Harviainen T, 1977. "On the vocalism of closed
unstressed syllables in Hebrew", Studia Oriantalia 48:1  Helsinki 1977 ):

"From the point of view of quality shewa  usually had the same sound as
pathah . As a rule §e§ was realized as a vowel resembling (a) in the Tib.
reading tradition, i.e. equal to §a§. Nevertheless there are two signs
(...) Besides these facts, scrutiny of the transcriptions of Jerome and
Pal. punctuations has brought me to the conclusion that the Tib §e§ is a
sign of compromise par exelence solving the great problem of graphical
simplicity from the point of view of punctuators, for the readers of Tib.
text the neutral §e§ sign rendered it possible to adhere (at first) to
their traditional habits of reading while accepting the Tib. system of
punctuation on the basis of the numerous other advantages provided by it."

I therefore suggest that the Masoretes made a non-phonemic choice to
represent an "a"-sound by shewa in some contexts and with patah in other
contexts. But because these two vowels *function* very differently inside
their system of pointing, several graphic differences was the consequence.
The first Hebrew grammarians who followed the Masoretes were influenced by
Arabic, Mishnaic-Hebrew and Aramaic and other languages where tense played
a role. When they saw the graphic differences which resulted from the
choice of the Masoretes, and in addition realized that the WAYYIQTOL group
represented past reference, and the WEYIQTOL group represented "future"
reference, they interpreted the system as a four-component one. Thus the
modern view was based on a graphic difference which was not phonemic but
which later was interpreted as a semantic difference!

The points above is just a working hypothesis which need much more work. So
I will appreciate your comments, and I am particularly interested to know
where the weak points are from a synchronic and an Masoretic historical
point of view. I accept that you have made a strong case for your opinions
in your diachronic studies, and your conclusions are definitely one very
good alternative. But could you please in connection with my point avoid a
diachronic view?

>I have what I think is a fair question for you Rolf: since no clear
>systematic distinction is made in Hexplaric transcriptions between
>prepositions l- b- k- with definite article and without definite article
>(Bronno, p. 204 and p. 387) -- i.e. in Masoretic terms the difference
>between _b at bhayith_ "in a house" vs. _babbayith_ "in the house" etc. --
>and I'm pretty willing to bet that no such distinction was made in
>Qumran orthography either, therefore do you think that the Masoretes or
>post-Masoretic grammarians artificially "invented" this distinction?
>And if not, why not?  (Phonologically, the difference between
>_b at bhayith_ and _babbayit_ is quite parallel to that between _w at yiqtol_
>and _wayyiqtol_.)

The answer is much the same as my suggestions above. I think that the
article as to pronunciation had a rather strong strength/marking which
caused gemination. So the Masoretes heard this and indicated it by their
use of patah.

>When all these factors are combined with the very specific historical
>phonological reconstructions which strongly tie Tiberian WAYYIQTOL to
>earlier consonant-final _*yaqtul_, and Tiberian YIQTOL to earlier
>vowel-final _*yaqtulu_ (where the final vowel in _*yaqtulu_ was lost
>around the end of the second millennium B.C.) -- as discussed in my
>dissertation at URL at end -- then your theory of a very late artificial
>differentiation between WAYYIQTOL and W at YIQTOL becomes even more
>P.S. When discussing cross-Semitic evidence, I wish you wouldn't pull in
>languages like Ugaritic and Geez (which are less directly helpful in
>illuminating the Hebrew situation, and which I am consequently less
>familiar with) to use as your first line of evidence, or try to make use
>of forms such as Akkadian IPTARAS which are not directly historically
>comparable to any morphological form in Hebrew according to the
>historical-comparative method accepted since the 19th century.  (If you
>only want to compare the synchronic semantic systems of languages,
>without paying any attention to linguistic "genealogical"
>considerations, then there's no need to confine yourself to the Semitic
>group of languages at all; however, if you want to try to do historical
>reconstruction between related languages, then you must pay attention to
>the established historical-comparative rules.)
 I agree that each language should be studied in its own right. However,
the start of the thread where I used Ge'ez/Accadian as examples was a claim
that particular Hebrew forms resembles Accadian ones. In this context,
Ge'ez is important because it has several parallels with Accadian. But the
problem is that if we stress a morphological similarity in the verbs of
these two languages, we get a difference in semantics; and if the stress a
semantic similarity, we get a difference in morphology. I used this to show
that a comparative historical study can be tricky business.



Rolf Furuli
University of Oslo

More information about the b-hebrew mailing list