[b-hebrew] Matres Lectionis and critical analysis

Yitzhak Sapir yitzhaksapir at gmail.com
Tue Aug 28 18:39:06 EDT 2007

On 8/28/07, JAMES CHRISTIAN READ wrote:
> Hi all,
> recent discussions have reminded me of points that my old
> Hebrew lecturer used to bring up in lectures. As well as
> mistakes made by the Massoretes in their vowel pointing
> he also spoke about another level of critical analysis of
> the text. He suggested that the oldest texts from which
> the books of the Hebrew bible were copied had no matres
> lectionis and were written without spaces. He further
> suggested that a number of scribal errors could have
> been introduced by introducing a matres lectionis where
> it shouldn't be and by dividing the purely consonantal
> text into space separated words incorrectly.
> He therefore posited that a completely critical
> analysis of the text would include not only questioning
> the Massoretic pointing but further questioning the
> matres lectionis and the word divisions and the further
> variety of possible readings such an analysis would
> provoke.
> Just thought I'd throw it out there for discussion. Any
> thoughts?

Dear James,

I am still working on a response to your post on the
unpointed vowels discussion.  In general, from the
earliest texts we have, there are word divisions -- in
the Gezer calendar, in the Mesha inscription, in the
Deir Alla inscription.  However, some words were
sometimes written together.  For example, in
the Mesha inscription, we have b(lm(n and wxmth(pl
(the second divided by a line break, but line breaks
also divide words in the inscriptions, and a dot
separator does appear at the end of a line when a
word is conclude).  In the Deir Alla inscription, we
have brb(r and )prxy)nph.  (These are not necessarily
the only examples in both inscriptions.  I just chose
the first two that I found).  Lerner's article on the
dh/z interchanges uses this to determine the
statistics of dh at the start of "closely connected"
words.  For the purpose of the study, he defines
a closely connected pair of words as a pair of
words that elsewhere in the Elephantine documents
appear without a word separator.  In the Massoretic
tradition, such a pair of words is designated with
a maqqef.  There is the example of Samuel 1:24,
which reads "bprym $l$h" but in light of the LXX
and the context should probably read "bpr m$l$."
Thus, saying there were no spaces is probably an
exaggeration, and I don't know here if you just
misunderstood him or whether that was his
position.  Most likely, some words were written
without word separators and this may have
contributed to ambiguity.  As for matres
lectionis, the Talmud itself in one discussion
states that the Rabbis (this is before the
Massoretes, by the way) are no longer aware of
the appropriate places for matres lectionis.  The
example from Samuel above is one such
example.  Another two are Deut 4:41 and Exodus
15:1 where the hiphil likely was originally in the
perfect tense - yavdel, yasher, rather than the
imperfect as it appears in the Massoretic text.
There is evidence for this in the Samaritan
Pentateuch.  However, we have to be careful.
In the case of the recently quoted root of BYN
in Proverbs, Y is a root letter, so that we can't
decide that just because the Massoretic
tradition pronounces the 3ms suffix conjugation
as "bin", the Y in this case is a mater lectionis.
In times earlier than the Bible, it could very well
have been a vocalized root letter.  The same goes
for "beth" ("house of") where the LXX transliteration
appears to me to suggest that the diphthong "ai"
had not yet collapsed to an "e" (tsere).  In the
Gezer inscription it appears that there are no
mater lectionis (though some have suggested
final waw).  Scholars generally think that mater
lectionis began to spread from the 9th century
onward and slowly although I am not entirely
convinced and I think that many examples of
"mater lectionis" during this period -- such as
final he on feminine singular nouns -- are really
consonantal spellings.  The system of mater
lectionis in the Pentateuch appears to be very
early and is for the most part as early as the
Persian period (but no earlier).  Thus, what
you state as a sweeping generalization is
probably incorrect but it is correct in some
specific cases.

PS.  Why do you always have a long column
of newlines at the end of each message?

Yitzhak Sapir

More information about the b-hebrew mailing list