[b-hebrew] vowel length

Yitzhak Sapir yitzhaksapir at gmail.com
Wed Nov 8 22:31:02 EST 2006

On 11/8/06, omar nshea wrote:
> The argument seems to have
> been inferred not only from the external sources like
> Hidayat al-qari etc etc but also from within the system itself: distribution
> of gay'ot etc.

That is a perfectly legitimate technique.  But one should look out not to
assume things ahead of time.  For example, in suggesting your question
about the preposition and quantity vs quality, you seemed to assume that
the segol/tsere/hataf patah distinction is one of length rather than quality.
But a language could differentiate different forms or conjugations of a word
using either quality or quantity or both.  So just the distinction between the
above isn't sufficient to suggest that it is necessarily a quantity distinction
and not a quality distinction.

> What struck me most was the fact that non-Karaite sources revealed
> distinctly further divisions beyond the usual binary-ness short vs long.

What do you mean?  Are you referring to Uri Hurwitz's comments?

Note that Geoffrey Khan notes in his works a variation of lengths in
vowels, much more than "two"/"binary".  However, phonologically only
the hataf-qamats/segol have significance.  That is, in Geoffrey Khan's
view, the word "you (pl.) will count" should be written tis-'pru even though
it was realized [tis-pa'ru:], because the vowel length distinction between
the various vowels is related to the syllable structure and does not add
any additional information.  Similarly, note that he identifies an extra-long
vowel in a word like "guards" [,$o:-om-'ri-im] (or [,$o:-om-'ri:-im]?).  So
there were many degrees of length but these are categorized into
short and long.  I think this syllabic breakup is related to the reading
tradition where long vowels were broken up by the reading tradition when
stressed to allow for multi-note cantillation of the syllable.

> intrigued by your idea on accents as indicators of length retained on the
> plane of cantillation (did
> I read correctly the idea that sometimes an accent in cantillation extends
> the vowel to four peaks?)

The accent mark "rabia(" is named, essentially "four-fold".  Now, one
might argue that the name refers to the shape of the mark which is a
diamond in modern printed editions but it appears that in manuscripts
this was just a dot.  Representing it so also matches in some ways the
modern way this accent mark is sung in Ashkenazi reading.  So that is
why I think that the cantillation mark would imply a four-fold vowel
peak structure.  It is important to understand that the Massoretic
Hebrew was sung, not spoken.  So a realization like [vaj-jav-'de:-e:-e:-el]
might seem strange in speech but makes perfect sense in the context
of a sung text.

> if, owever, "It even appears to me that when
> Hebrew ceased to be a spoken language after the early
> centuries CE, that vowel length was gradually lost as the
> reading tradition developed Hebrew in ways more suitable for
> the cantillation of the liturgical reading",
> then that would undermine the transcriptions of the Hebrew Bible into
> Arabic by the Karaites etc and
> their reliablity as sources for the reconstruction of quantity would
> be put into question.

When I said vowel length was gradually lost, what I meant
is that the use of vowel length as a language feature that
distinguishes meanings of words (like the "my blood" vs.
"silence" example I brought in that description) was lost.
The Massoretes still distinguished short vs. long vowels
but this distinction usually had little consequence for
telling one word apart from the other.  So more accurately,
the "vowel length distinction" was lost in places where
it made a difference in the meaning of a word, and in its
place, quality distinction arose.

One interesting this is that due to Nesiga, and perhaps
other bits of evidence, we know that the cantillation
system was first applied to the text when stressed
syllables were not always long.

Yitzhak Sapir

More information about the b-hebrew mailing list