Biblical Hebrew Forum

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

On 8/22/07, Joel Hoffman wrote:
> >> Many people don't know that they are referring to a medieval
> >> language when they use the phrase "Biblical Hebrew."
> >
> >I think it is inappropriate to label the language Medieval as Hebrew
> >was spoken only until about the 3rd century CE.
>
> I think you've missed the point.  By "medieval" I mean the Masoretes,
> who gave us Masoretic Hebrew.

Well, you said "Medieval language," and in any case, the Massoretes are
dated to the 9th/10th century CE which is generally before what people
consider Medieval anyway.

> >historical study of this period.  If a letter from Ezra/Nehemiah is quoted,
> >however, it should ideally retain both the Massoretic vowels and accents
> >because that is the raw evidence that we have from this source.
>
> Here I guess we disagree.  I do not believe that of all scholars
> throughout the last 2,500 or so years, the Tiberian Masoretes are the
> most qualified ---  from a linguistic or other scientific point of
> view --- to interpret Hebrew from the first millennium BCE.

The Massoretes did not interpret the text, they were passing on a
consonantal and vocalization tradition as they had heard and seen the
text.  While some interpretation may have gone on, -- both in the
consonantal and vocalization traditions they had, -- this does not mean
that they invented the entire or even significant parts of the tradition or
grammar.  Furthermore, their consonantal tradition does not magically
become a 1st millennium BCE text if we strip it of the vocalization.  Yes,
much of the differences are hidden if we do this but this is only because
we have stripped away the vowels and certain consonantal differences
that would have mattered.  Even so, there are still some small differences
in the consonantal tradition from the way it would have been spelled and
written in the 1st millennium BCE by a true 1st millennium BCE scribe.
Taking their consonantal tradition and presenting it as 1st millennium
BCE is about as correct as taking their vocalization tradition and
presenting it as 1st millennium BCE.  We might want and wish it to be
so, but it isn't.

> >> Yes, I do know.  For example, some of the cross-word beged kefet
> >> changes cannot be part of a spoken language.
> >
> >But this is not the way most Hebrew linguists view this issue.  The fact
> >of the matter is that bgdkpt is exhibited not only by Hebrew but Aramaic
> >as well.  The Aramaic shift is to be dated "sometime between the end of
>
> Again, you have missed the point.  I refer not to intraword Beged
> Kefet changes, but to *cross-word* changes.

I am not sure why you kept this response.  You must have realized that I
was setting up the discussion for the word boundary spirantization issue.
It makes most sense that word boundary spirantization is related to
general spirantization.

> >of the Second Temple.  But we also have Biblical books in Aramaic, and
> >these show spirantization across word-boundaries (what you call "cross
> >word beged kefet changes") -- for example vavloye in Ezra 4:9.  It would
>
> No, we don't.  We have a Masoretic record of these changes.  That's my
> whole point.  We do not have any evidence from the 1st millennium.

Well, isn't the Massoretic record 1st millennium CE? :)

That is what I mean in saying Biblical books.  The book is part of the Bible
-- hence Biblical.  It doesn't mean that it was written in the 1st millenium
BCE.  Any evidence we reconstruct of the language from the MT Bible -- is
based on 1st millenium CE evidence.  Even DSS evidence (both biblical
books present in the DSS and non-biblical evidence) involves reconstruction
a few centuries back.  This includes reconstruction of most of the language,
as well as the consonantal evidence for this reconstruction.

However, this is how linguistics operate.  We reconstruct an earlier stage of
the language which is either poorly represented or not represented at all,
based on more complete evidence in even earlier more documented stages,
later more documented stages, or cognate languages.

In this case, I brought evidence to make several points:
1) The Aramaic books of the MT Bible include word-boundary spirantization.
2) Spirantization as a phonetic rule was lost by the time of the Massoretes,
although words which were earlier influenced by the spirantization rule
maintained the spirantization.  I brought examples of a case where a vowel
was lost, and a vowel was added, after the spirantization rule ceased.
3) Spirantization is a current phenomenon in Spanish -- including in
"moderately fast, casual speech [where] spirantization does not occur
utterance initially but does take place word-initially ..."  This is a direct
parallel to the Massoretic conjunctive cantillation marks, which identify
a smaller pause between words than disjunctive cantillation marks.
4) The spelling changes that took place between Old Aramaic and Official
Aramaic are evidence that by the time of the latter, the spirantization
rule was already in full operation.

Now, only the last is really 1st millennium BCE evidence, but linguistics
operate by making informed conclusions based on the state of the
language at a later stage.  Here, we can see that various phonetic
processes had taken place without the spirantization rule being
activated for the resulting word -- showing us that the spirantization rule
had ceased to take effect before these phonetic processes began.  We
thus know that there was a spirantization rule that had begun very early
and ended operation before the Massoretes.  We know that a
comparable spirantization rule currently operating in Spanish does
operate across non-phrase-initial word-boundary in casually fast speech.
We know that there are instances of such word-boundary spirantization
in the Biblical Aramaic of the MT Bible.  In light of all of these, it makes
most sense that this word-boundary spirantization is a relic -- like the
rest of the spirantization -- of the time when spirantization was in effect
and that when it was in effect, word-boundary situations were also
affected.  It makes less sense that spirantization as a whole was an
invention of the Massoretes because then we would be at a loss to
explain how they knew to apply it in situations where a vowel was lost
or a vowel added in phonetic processes that began before their time.
It similarly makes less sense that given that spirantization as a whole
was not their invention, and that there are many cases of exceptions
(because of the mentioned loss of vowels or addition of vowels) that
they independently decided ("invented") to extend it to cases of very
short word pauses ("conjunctive cantillation marks"), especially in
light of Spanish that shows us that this does happen when the
spirantization rule is in effect for a spoken language -- and the time
where spirantization was in effect by the above evidence does place
it during the time when Hebrew was a spoken language.

If all this is not enough, I will quote part of a paragraph from Faber's
article, which brings lots of evidence despite its unlikely thesis:
"Lerner (1981, Leshonenu 45) has studied <d>-<z> interchanges in
the Elephantine documents and come to the conclusion that it was
the post-vocalic pronunciation of /d/ that enabled the merger of *dh
with *d.  Of the more than 1000 tokens containing inherited *dh, 57
are written with <d> and the remainder with <z>, the standard
spelling for earlier Aramaic.  Out of the 57 tokens with <d>, 43 are
post-vocalic, an environment in which there are only 76 instances of
<z>.  Phrase-initially, an environment in which the stop variant would
be expected, *dh is spelled <z> over 900 times, but there are eight
examples with <d>."
I draw attention to the words "phrase-initially" which suggests that
word initial but non phrase initial examples are part of the post-
vocalic statistics, and that the above evidence for post-vocalic
pronunciation of d applies equally to word-boundary situations.  For
a more precise conclusion, however, one would have to look up
Lerner's article.  Thus, there may be evidence from 1st millenium
BCE texts that this rule was active at word boundaries as well.

Now, as for your "whole point," I want to reiterate the development
of this argument -- you said that the Massoretes "invented some
of what they recorded."  I said that I'm not sure what you mean.
You suggested, "some of the cross-word beged kefet changes
cannot be part of a spoken language."  I provided evidence of
word-boundary bgd changes in Spanish (a spoken language)
as well as a basic outline of the above argument, now spelled
out in more detail and with a little bit additional information.
Now, you say, your whole point is that we don't have evidence
from the first millenium.  On the one  hand, I see no attempt
to maintain the assertion that "some of the cross-word beged
kefet changes cannot be part of a spoken language."  Does this
mean you now accept that the word-boundary spirantization can
be part of a spoken language, it is just that you now argue that
we have no contemporary evidence from when Hebrew was a
spoken language?  If your argument is that the Massoretes
invented what they wrote because we have no evidence from the
1st millenium BCE, then you are making some methodological
errors:  1) you can't argue that the Massoretes invented something
based on a lack of evidence -- you must argue from evidence.
2) basic linguistic methods mean that some evidence for earlier
stages of a language can be derived from later stages -- I'm not
sure why you ignore these methods or view such conclusions as
illegitimate.

Yitzhak Sapir