Post: There is a Samaritan variant text of a pasuk in Vayakhel, in which five types of jewelry are donated to the Mishkan instead of the four mentioned in our Masoretic text.
This would not be troublesome in the slightest, since the Samaritans falsified their Torah. But the Septuagint also lists five items. And most troublesome, a midrash rabba makes a derasha over the fact that five items are donated!
Doesn't this mean that Chazal had this text of the chumash? And if we have a different text, doesn't it mean that our Masoretic text is in error?
Not necessarily. While Shemot Rabba is post-Talmudic, it could be encoding an earlier midrash, from the time of Chazal. And in the time of Chazal, there were various variants floating about. While there was a masoretic text, there were also variants floating about. And for example, we see that Chazal take note of the (meaningful) variant readings in the Torah of Rabbi Meir.
And I recall an explanation of the Haggadah from Rabbi Dr. Leiman's class (Intro to Bible II). The big deal in the words of the rasha is that he said lachem, which implies not him. But the chacham says etchem, which should also imply not him! There are plenty of famous teirutzim, but if you look in the Tosefta (IIRC), it cites the pasuk as stating otanu. And this is no mere error. The Dead Sea Scrolls have the same variant reading. Based on the Torah before the midrashist, the question does not even arise.
These variants can be from "vulgar" texts, not in accordance with the standard, which nevertheless were circulating and in some cases were drashened. Or else perhaps they were an alternate standard. Check into the discussion of this matter. But that we have a derasha of a "vulgar" text, even from Chazal, does not necessarily mean that that is now our traditional text, and that it should supplant what we have in the Masorah. (On the other hand, if there is a popular, universal derasha that Chazal make, how could this not have been noticed as having been against the masoretic text? And how can we have a masorah against Chazal?)
So we have established the Samaritan text as an extant text before some midrashist, perhaps a member of Chazal. Does that mean that our Masorah is wrong? How can it be a Masorah? Well, we might also assume that our variant was before Chazal, as well. And if our Masoretic text was the true, original one, then, it does not really matter than some midrashist made a derasha on a vulgar text.
In this particular instance, I believe that the Masoretic text is more convincing, and is the original. Consider the two versions once again:
The Samaritan version strips out the vav of tabaat, and inserts the word agil. Thus, we have tabbaat, agil, vechumaz. The agil also appears by the donated spoils of Midian, in Bemidbar 31:50:
|נ וַנַּקְרֵב אֶת-קָרְבַּן יְהוָה, אִישׁ אֲשֶׁר מָצָא כְלִי-זָהָב אֶצְעָדָה וְצָמִיד, טַבַּעַת, עָגִיל וְכוּמָז--לְכַפֵּר עַל-נַפְשֹׁתֵינוּ, לִפְנֵי יְהוָה.||50 And we have brought the LORD'S offering, what every man hath gotten, of jewels of gold, armlets, and bracelets, signet-rings, ear-rings, and girdles, to make atonement for our souls before the LORD.'|
Not only does agil appear, but it appears in the middle of the words tabaat and ve-chumaz. Now, there are two choices in play.
The first is that the Samaritan text is original. If so, the Masoretic text would have developed in it by first losing the entire word agil, and then adding a vav to the beginning of the word tabaat. Deletion is generally easier than insertion, and so while agil could have been omitted by a careless scribe, the additional vav it seems more difficult, though still possible.
The second choice is that the Masoretic text is original. If so, the Samaritan text would have developed from it. But how? Deletion of a vav connective is indeed not that difficult, but how would an entire word be inserted? There is a ready answer, which would simultaneously account for the deletion of the vav. That is that kumaz is an extremely rare word, and it appears in both places in the context of tabaat. A scribe could have been copying a text, and forgot what he saw and instead relied on memory. That memory would be, accidentally, from the run of text in Bemidbar 31. Thus, he would copy the entire phrase tabaat agil vechumaz, in one fell swoop! The vav of tabaat would disappear since that is how the phrase occurs in Bemidbar; and agil would be inserted in the middle, because that is how it appears in Bemidbar. One operation, which is quite "easy", and which does not have the difficulty of either insertion or deletion.
Furthermore, there is more entropy in the masoretic text than the Samaritan text. In the Samaritan text, we have the same phrase in two places. In the masoretic text, they differ. This is a related reason for considering the masoretic text as original.
This is enough to persuade me that the Masoretic text is indeed the original, in this particular instance, and that the Samaritan text derives from it. If so, even if a midrash darshens the Samaritan text, this does not impugn the integrity of the Masoretic text.