Notes from a shiur given by Chaim Weiner
An Introduction to Mediaeval Manuscripts
Rabbi Chaim Weiner
Manuscripts accumulate errors in copying—except for Torah, and to a lesser extent, other books of the Tenach, because of the great care taken in copying them. But in the case of other manuscripts, scribes, who were paid by the word, simply dashed them off as fast as they could, and consequently made mistakes, which accumulated from copy to copy.
By analysing these, it can be seen there are two families of manuscripts of the Mishnah, the Babylonian one and the Palestinian one. The Babylonian one is probably more accurate, because the amora Rav, who went from Babylonia to Palestine and was a student of R. Yehuda the Prince, who compiled the Mishna, took a copy back home to Babylonia with him, and it was treated there as a sacred text and inviolable. In Palestine, the Mishna was viewed not as the finished Mishna of R. Yehuda, but as a living document, the Mishna of the Sanhedrin, and continued to be changed for a while.
There is only one full copy of the Talmud Yerushalmi which has survived—and indeed the mediaevals referred to books they had which have now been lost. There was a big furore when someone claimed a century ago to have discovered Gemara for some of the parts of the Mishna that don't have Gemara; but that turned out to be a fake. By contrast, there is another tractate where the last mishna is missing; a scholar, reasoning that copyists, tending to note down just a few words rather than a whole quotation (compare the quotations in the manuscript included here with those in the printed copy), had omitted the entire mishna because it repeated material told elsewhere. He then reconstructed what he thought the original mishna would have looked like—and was vindicated when a manuscript copy was later found of it.
Consider the following manuscript:
This manuscript, after being passed down in one family for several centuries, had vowels added. The vocaliser added them by consulting a separate manuscript, and where the two manuscripts differed, noted the differences in the margins. Thus this manuscript now contains two manuscript traditions.
Note שֶל ("of") here being spelled as a compound prefix שֶלְ־ ("which [belongs] to"); that's how the word שֶל originated. In later manuscripts the word was regularised (see the printed version below [though that seems still to have שלכהנים in one place as a single word]). The fact it is spelled this way in the manuscript indicates that this is an early manuscript.
During the Middle Ages, copyists tried to harmonise Mishnaic Hebrew with Biblical Hebrew, considering departures from Biblical Hebrew mistakes. For example, in printed editions of the Mishna, Tosefta, Talmud and Midrashim the word for man is אדם, as in Biblical Hebrew. However, one manuscript [or, according to the Encylopaedia Judaica, reliable manuscripts (in general)], it is אדן; and the transmutation of ־ם to ־ן in Mishnaic Hebrew only survived in later manuscripts in the form of the plural ending ־ין. [Similarly, says the Judaica, the name לעזר [Lazar] (as found in the NT) is corrected to the Biblical אלעזר [Eleazar].] These changes can be used as shibboleths to identify reliable manuscripts.
Where is the first chunk of Mishna 3 in this manuscript? (Compare the printed version below.) It turns out this is not attested in any manuscript; the first place it appears is in what is known as the first printed Mishna (it's actually the second, but it was this edition that everyone copied thereafter, rather than going back to copy manuscripts):
[Layout rearranged slightly for ease of webification.]
As can be seen, this constitutes alternating sections of mishna and the Rambam's commentary, the latter set in Rashi script to differentiate it. What had happened is that the typesetter accidentally set the last paragraph of the Rambam's commentary on Mishna 2 in the wrong typeface—and ever since then it's been accepted as part of the Mishna!
This difference is now acknowledged in any critical (academic) editions of the Mishna, but not (surprise surprise) in rabbinical editions such as Artscroll.