Tuesday, March 16, 2010

Did an entire word fall out of the Torah?! pt ii

Summary: No, it didn't. Though I'll attempt to prove this in the third segment. Though the idea that this would be the conclusion could bias one's analysis -- after all, one does not want to be a heretic, according to the Rambam's definition!

This segment is dedicated to analyzing the language of the midrash, and is an attempt to further demonstrate that the midrashist indeed was working off a text equivalent to that found in the Samaritan Torah.

Post: I will begin with quick summary of what has been discussed so far, and then transition to further proofs. In a previous post (read it first) on parashat Vayakhel, I discussed a Samaritan variant of our pasuk.

Our Masoretic text is on the right, while the Samaritan text is on the left. Where no change occurs, there is a dash. Where a change occurs, there is text. Where text from the opposite side is omitted, it is marked with a star.

The Samaritan text changes הביאו to הביא, but that is nothing. It is a regular "smoothing" feature of the Samaritan Torah. In this instance, they are trying to "correct" for number, to match nediv lev and the "each" aspect of kol.

But another feature of the Samaritan text is that they add agil, such that there are five types of jewelry donated to the Mishkan, instead of four. I'll argue in the next segment that this arose as a result of transference from a pasuk regarding the spoils of Midian, such that we need not impugn the integrity of the Masoretic text. But regardless, how popular was this variant reading? Is it just the Samaritans?

It appears that the Septuagint was also working off of a Hebrew original which had agil, and thus five instead of four items. Thus,
22 καὶ ἤνεγκαν οἱ ἄνδρες παρὰ τῶν γυναικῶν, πᾶς ᾧ ἔδοξε τῇ διανοίᾳ, ἤνεγκαν σφραγῖδας καὶ ἐνώτια καὶ δακτυλίους καὶ ἐμπλόκια καὶ περιδέξια, πᾶν σκεῦος χρυσοῦν, καὶ πάντες ὅσοι ἤνεγκαν ἀφαιρέματα χρυσίου Κυρίῳ.
22 And the men, even every one to whom it seemed good in his heart, brought from the women, [even] brought seals and ear-rings, and finger-rings, and necklaces, and bracelets, every article of gold.

Thus, there are five items in the pasuk.

This is not surprising, nor is it the only place the Septuagint differs from the Masoretic text. What is surprising is Midrash Rabba appears to be based on the same text:

ומה כתיב? כל נדיב לב הביאו חח ונזם טבעת וכומז. בנזמים חטאו ובנזמים נתרצה להם, ורוח הקדש צווחת על ידי הושע: (שם ב)והיה במקום אשר יאמר להם לא עמי אתם יאמר להם בני אל חי. אמר משה לפני הקב"ה: כתבת: (שמות כא) כי יגנוב איש שור או שה וטבחו או מכרו חמשה בקר ישלם תחת השור, הרי הביאו חמישה: חח ונזם טבעת עגיל וכומז

Just as one who steals an ox and slaughters it must bring five in recompense, so did the Israelites donate to the Mishkan five items: chach venezem, tabaat agil vechumaz. But in the text in our chumashim, we only have four items.

The meforshim on the spot offer two answers. Either it is a derasha of the continuation of the pasuk:

כב  וַיָּבֹאוּ הָאֲנָשִׁים, עַל-הַנָּשִׁים; כֹּל נְדִיב לֵב, הֵבִיאוּ חָח וָנֶזֶם וְטַבַּעַת וְכוּמָז כָּל-כְּלִי זָהָב, וְכָל-אִישׁ, אֲשֶׁר הֵנִיף תְּנוּפַת זָהָב לַיהוָה.22 And they came, both men and women, as many as were willing-hearted, and brought nose-rings, and ear-rings, and signet-rings, and girdles, all jewels of gold; even every man that brought an offering of gold unto the LORD.

based on the ribui of כָּל-כְּלִי זָהָב, or else based on nezem having two meanings. The former seems far-fetched because this is not only an implicit derasha, unmentioned by the midrash (while there is an explicit derasha which is in play), but it is based on a part of the pasuk which was not even cited once in the midrash. The latter is far-fetched because how is a reader of the midrash supposed to know that nezem is supposed to expand like this, and what in the pasuk indicates we should expand it and make such a derasha? There is a jump in the logic. Both the former and latter are farfetched because of the placement of agil in the pasuk. If based on nezem, it should have been placed after nezem. If based on kol kli zahav, it should have been placed after vechumaz.

We could live with farfetchedness, if forced to. But given that there is an alternative girsa of the pasuk which matches this exactly, including placement of the word agil, the obvious conclusion is that the midrash is somehow based on the text as found in the Samaritan Torah.

Now, the slight difficulty in this is that the Midrash Rabba explicitly cites the pasuk without the word agil. That is, it leads off with:

ומה כתיב? כל נדיב לב הביאו חח ונזם טבעת וכומז.

Note the lack of the word agil. It is like our pasuk. And if so, no matter how difficult, perhaps we should revert to the standard teirutzim.

To this, I would respond as follows: Firstly, the midrash does not cite our pasuk, precisely. And secondly, it cites the pasuk twice. To provide the contrast:

Our Masoretic text: הֵבִיאוּ חָח וָנֶזֶם וְטַבַּעַת וְכוּמָז
The Samaritan text: הֵבִיא חָח וָנֶזֶם טַבַּעַת עגיל וְכוּמָז
The beginning of the Midrash: הֵבִיאוּ חָח וָנֶזֶם טַבַּעַת וְכוּמָז
The end of the Midrash: חח ונזם טבעת עגיל וכומז

The difference between our Masoretic text and the pasuk cited in the beginning of the Midrash is the vav pattern. The masoretic text has  וְטַבַּעַת, with a vav, a pattern we might expect of a short list of items, with a vav at the start of each item other than the first. The pasuk as cited in beginning of the midrash has טַבַּעַת, without the leading connective vav! That matches the lack of vav in the Samaritan text. And at the end of the midrash, we have טבעת, again without the leading vav. And the order of this item, including the placement of the word agil, and the vav pattern, exactly matches the Samaritan text. I would therefore assert that the list at the end of the midrash is indeed a citation of the pasuk, rather than a mere listing.

My strong suspicion, and my explanation for this phenomenon, is that the even the midrash's citation at the beginning had the word agil. But while the pasuk citation at the beginning is clearly a citation, the one at the end looks like a mere list. Some overeager scribe saw the pasuk, and saw that the pasuk was "incorrect". And in that context at the beginning the derasha does not depend on the word agil. And so, the sofer "corrected" the pasuk, as cited only at the beginning of the midrash, by simply removing the word agil; and not noticing the lack of vav in the word טבעת. The end result was a pasuk without the word agil but with a vav pattern matching that of the Samaritan text, rather than the masoretic text. And he didn't emend the citation of the pasuk at the end, since it did not look like it was a pasuk. Only we, who have the Samaritan text before us, recognize that it was a pasuk citation.

This correction by soferim is not uncommon. And it explains the textual features quite nicely. And overall, it seems to me much more plausible that the alternative.

It is still strange for a Midrash to darshen a non-masoretic text. But it happens elsewhere. For example, the derasha on totafot. And there is one on karnot in this week's parsha, Vayikra, that I might get around to. And in the seder, the derasha on the rasha's words is based on a non-masoretic version of the chacham's words. And so on, and so forth. Bli neder, I will discuss this in the next post, and why this need not be a contradiction to the ikkarei emunah of the Rambam.


thanbo said...

Well, books like Tov (Textual Criticism of the Hebrew Bible) posit parallel Palestinian and Babylonian texts. The Masoretic text is based on the Babylonian stream, while the Samaritan and LXX derive from a Palestinian text. Wiki on Genesis Rabbah attributes it to the tanna Hoshayah, who was in EY, so he may have been working with a Palestinian text as well. Tov notes that the Greek version of the Tabernacle narrative (Shmot 35-40) differs pretty widely with the Masoretic version.

As for which is "right", there is the idea that "editors smooth out problems". So, with an implicit drasha of the klal (don't most klalim in the Gemara include one extra case?), and the passage about the war with Midian that you cite, maybe the editor added a fifth item.

joshwaxman said...

"Genesis Rabbah"
the slight problem with this is that each of the seforim of midrash rabba was written in a very different time (and place). Shmos Rabba is pretty late (as in post-Talmudic), and Bamidbar Rabba is much later.

but it is indeed likely that they were working off of different texts.

"As for which is "right","

i'm going to suggest something pretty similar in segment 3.

kol tuv,

joshwaxman said...

but not precisely the same. i don;t think this was the work of an editor, but rather of an errant scribe.



Blog Widget by LinkWithin