This is all old news to me. Many years I heard the same from Dr. Bernstein - it is an obvious issue. However, the resolution seemed just as obvious to me. But first, a digression:
In Masechet Sofrim 6:4 (also occurs elsewhere, such as Yalkut Shimoni):
{Update: also, Sifrei Devarim, as well as Yerushamli Ta'anis, 20B 4:2 - courtesy of a post by Greg Gershman about this source}
Three books were found in the [Temple] courtyard - the maon book, the za'atutei book, and the hi book... so they retained the [majority] reading of the two and abandoned the [minority] of one.To cite Gil Student's explanation from Torah Emet:
In other words, when Ezra returned to Jerusalem to rebuild the Temple, he was able to find three reliable scrolls with minor differences. The differences were as follows. In one, the word נערי was written in its Aramaic translation זאטוטי. In another, the word מעונה was written מעון without the final ה. And in the third, the word היא was written as הוא (but vocalized as 'hi') in eleven places.I would add that the scrolls from the Temple Courtyard are not just any scrolls. They were exceptionally accurate scrolls, used to correct and confirm the correct girsa in other scrolls. They served in much the same role as the Allepo Codex and Leningrad Codex in more recent times. So here, they (perhaps Ezra) wanted to establish the correct text, and there were slight differences between codices.
In order to decide between these divergent texts, they seem to have used a principle laid down in Shemot 23:2:
peshat level, this means that one should not go after a multitude to pervert from the correct. However, on a midrashic (hyper-literal, significance-maximalist) level, one can zoom in to just those words, and this phrase, and claim that the word וְלֹא only applies to the immediately following phrase (bear witness in a cause to turn aside) and not to the next phrase. Thus, there are certain instances in which one should go after a multitude (or majority) to pervert.
Thus, the solution adopted was to follow the majority. Each of these scrolls had a unique reading not found in the other two. If so, we should ignore the reading in the single scroll and follow the other two scrolls. (Note that this was also in choosing amongst exceptionally good textual witnesses.)
Thus the final, composite scroll did not entirely match any of the three sources exactly. If one one of them was the original, then the resulting scroll did not contain the original text. However, this did not matter. Since they applied the halachic principle of acharei rabim lehatot, following the majority, even to pervert, the final, composite scroll was an halachically valid scroll. (No comment about the status of a scroll with the true, original text.)
True, the Amoraim were uncertain about certain plene and deficient spellings, but in the end, in the present day, we have a (basically) standard text. And true, one might collect thousands of variants from Torah scrolls, but those are vulgar texts, used by the populace. They were intended to match the original, but at times scribes made mistakes. However, they should have, and frequently did, consult one of the standard Masoretic Codices, either that of Ben Asher of Ben Naftali. These Codices were excellent expert productions from competing Masoretic schools, and when one had a question, he wrote to inquire as to the correct text. Furthermore, the Allepo Codex has halachic validity as the standard text of Tanach. Such halachic validity was granted by the Rambam.
If God knew the disputes of Abaye and Rava all the way back at matan Torah, such that all such discussions were given over to Moshe (at least on a metaphorical level), then God would also know the eventual halachic configuration of his Written Torah. If the claim of those who propound the Torah Codes is that God is relating future events, would not such knowledge of the future include the eventual text of the Torah as it exists in the present day. For it would make no sense to encode messages in a text that would be munged by additions and deletions of matres lectiones, since the messages would then be lost. Rather, if God's Supreme Intellect could give over a plain text while also encoding messages in minimum skip patterns, he could also encode messages such that the eventual configuration of the Torah would contain said patterns. The logical text would be the standard one available when the Torah Code methodology became known.
If so, the objection that knowledge of the original plene and defective spellings was lost is mitigated.
What about variations between Ashkenazic, Sefardic, and Yemenite Torah Scrolls? Shouldn't these make a difference? Well, the Torah Codes folks said the experiment works with each of these. But, one would expect it to work, without having to run the experiment. After all, a Sefardi sefer Torah differs from the Ashkenazi in whether one spells petzua daka with a heh or an aleph. This is a single letter difference, and one that does not disturb and skip patterns. Indeed, the most that could be messed up is a Torah code that utilizes that particular letter, which is statistically insignificant. And the Yemenite Torah differs slightly in only three places. At most, this would mess up Torah codes in these three local areas. While skip patterns range over many letters, the effect of these variants is constrained to very specific locales within a much wider area, such that I would assume there no statistical impact would be felt.
Some Musings:
Had you asked me, I would have thought to take a different approach to deciding the correct reading.
That is, I would agree that זאטוטי is not likely to be the correct text. It is an Aramaic replacement for na'arei, and it means youngster, but without the possible other connotation of servant that na'arei carries (see Jastrow). The Talmud tells us that this was one of the changes the 70 elders made when creating the (actual) Septuagint (=translation of 70). See Megilla 9a.
Note that this textual variant of זאטוטי makes it somewhat difficult to claim that the events described in Megillat Sofrim occurred in Ezra's day. After all, the זאטוטי variant was deliberately introduced by the 70 elders in a translation of the Torah into Greek, commanded by Ptolmey, so that people would not become confused. This happened about 250 BCE, centuries after Ezra, so it seems unlikely that this particular variant would pop up much earlier. Proof from 4 Ezra 14:44 (mentioned by R' Student in that Torat Emet article):
"And it came to pass when the forty days were fulfilled that the Most High said to [Ezra] saying: Publish the twenty four books you have written so that the worthy and unworthy may read them."seems unconvincing - after all, this is an Apocryphal book, of Jewish origin, but with Christian interpretations, and even this verse does not suffice, but one needs to construct a "midrash" on a verse in an Apochryphal book. I am not aware of all the other proofs, but have suspicions that the books of the Azara (Temple courtyard) were read by someone as books of Ezra, and from that sprung the association with Ezra.
In terms of מעונה vs. מעון, I would agree with the single scroll that had מעון. This is because I think that in all of the scrolls, the intended pronunciation (krei) would be מעונה. This is an example of chaser/yeter as mentioned by Rav Yosef in the gemara Kiddushin mentioned above.
Here is a good parallel. In parshat Ki Teitzei, in Devarim 22:13 and on:
heh at the end, such that the ketiv is officially na'ari = young man. But this is not the meaning at all on a peshat level. Rather, this is another instance showing that the final heh is not always necessary to express the kametz at the end of a word. We are familiar with many other instances of this. From the text just cited:
וּסְקָלוּהָ
מִקִּרְבֶּךָ
וּבִעַרְתָּ
Now, it is more likely for the text to be changed to add the final heh than for it to lose the final heh, since the more common spelling at that time would be with the final heh. Thus, מעוןrecommends itself.
In terms of the variant of היא was written as הוא (but vocalized as 'hi') in eleven places, that reading also recommends itself under the same principle of lectio difficilior as above, since it is weirder = the more difficult reading. We would assume one would emend the text to match the gender of the way it was read and the gender of the noun to which היא serves as a pronoun. Both readings are acceptable since, early on, the matres lectiones served to denote different vowels, such that a vav could stand in place of the iy sound. (There are many such examples in Tanach.)
That is the analysis from the perspective of which reading recommends itself. One could also consider the issue from the perspective of textual families.
Imagine there was originally a single, correct sefer Torah from which these three variant scrolls sprung. This scroll would either have נערי or זאטוטי; it would have מעונה or מעון; and היא or הוא. Now, assume that each of the three scrolls were directly copied from this original scroll A (or copied from a flawless copy), but in copying, mistakes were made. The simplest explanation seems to be that scroll Z copied everything right with the exception of substituting זאטוטי for נערי.
Otherwise, we would have to assume that the original scroll read זאטוטי and that two scrolls(H and M) made identical mistakes in isolation, substituting נערי for the correct זאטוטי. The same goes for מעונה/מעון. Better to assume one mistake of substituting מעון for the correct מעונה in scroll M than to assume that two other scrolls (H and Z) made mistakes in the opposite direction. Finally, the same and more goes for היא/ הוא. Better to assume that one mistake was made in scroll H eleven times (while M and Z got it right) than to assume that, in isolation, two scrolls (M and Z) made identical mistakes in eleven separate locations! After all, it was considered a great miracle when 70 elders, each in isolation, intuited to make ten changes in translation. It is surely somewhat improbable for eleven separate mistakes - and the same eleven mistakes, and no others (or only one - the distinguishing marks of M and Z) - to occur in isolation in two separate scrolls. Thus, the idea of nullifying the one and following the two makes sense.
However, this is so only if we assume that the scrolls are not related at all, but each is an isolated spinoff of the one original. Other possibilities exist.
Scroll Z surely is not original, because it is an obviously incorrect variant. But what if scroll H was the original - and had הוא instead of היא? A scribe copied scroll H and made a mistake - he changed the reading to היא in eleven places, thus generating H2. We do not have H2 before us. However, another scribe copied H2 and changed מעונה to מעון, thus generating scroll M. Another scribe copied H2 and did not change מעונה to מעון, but changed נערי to זאטוטי, thus generating scroll Z. Visually, this would look like this:
H
|
H2
| M_ Z
(H2 is the substitution of היא for what we now claim is the original הוא.)
Thus, H could have been the original, and we still are speaking of only three mistakes. A similar process and relationship of texts could be used to justify M as the original, by flipping the order of the introduction of errors:
M
|
M2
| \ _
H Z
(M2 is the substitution of מעונה for what we now claim is the original מעון.)
And of course, we could similarly put forth Z as the original, though we know that to be untrue.
Could one claim that the unique features of M (מעון) + H(הוא) is original? Here difficulties arise:
( ___ M+H )
| \ ___________
H __________M
|
(H2+M2)
|
Z
The problem with a diagram like the above is that the change to lose the H feature must happen in two locations (to create M, and to create Z). The loss of the M feature happens in only one place. But perhaps there is some cause of this loss of feature H in those eleven places.
Slightly less difficult is the following diagram, which requires loss of the M feature by two separate scribes, one to create H, and the other to create Z.
( M+H )
| M ______H
| ____
(H2+M2)
|
Z
All of this is forced by the presence of scroll Z, which lacks the unique features of both scroll H and scroll M.
But perhaps one could claim that Z is not a scroll that should be heeded as a witness to a textual family. That is, why would a scroll have זאטוטי? This is a scroll for popular consumption - it has this Aramaic rendition so that people will not make this incorrect assumption. Perhaps when reading in public there was a tradition among some to read זאטוטי, just as the 70 elders had consensus to change it to זאטוטי in their translation. Similarly, the reading of היא and מעונה reflect the way it is read, and was always read, even though there was an archaic spelling in the written text itself. A scribe who would put in זאטוטי would also put in known changes from other texts which best reflect the way it is publicly read. Then, we could have a tree like:
( M+H__)
| H_____ _M
where H loses the M and M loses the H, and with Z collecting features from each that better match the krei.
Update: Fixed trees. Lets hope they come out well this time.
No comments:
Post a Comment