Variant recensions of the Talmud Bavli

How much difference do they make?

When looking at differences between versions of a text found in manuscripts, there are two ideal types of what you might be looking for:

1. There is a correct, original reading of the text, and all other versions are a result of scribal error, a mistaken attempt at correction, the accidental insertion of a marginal explanatory note etc. Your job, then, is to identify the original reading, using principles like lectio difficilior potior, stemmatic analysis, sound judgement etc.

2. There are multiple versions of the text, and the different manuscripts preserve different variations. Your job is is to isolate and compile these different versions then analyze them side by side, perhaps theorizing about how these different versions came to be, or about why they say different things in the parallel sections.

Very often, with an actual text, you are trying to do a bit of both. You have to isolate two or more recensions, but also clear up a lot of inaccurate readings within them. It can get really confusing if the two different traditions crossed over at some later point, contaminating each other. Sometimes, it’s not clear whether something really counts as a separate recension or not, and this is so with the ongoing debate over Mishnah manuscripts, which are divided up into Palestinian and Babylonian variants, except they aren’t, not really, sort of, no-one’s sure.

With the Talmud Bavli, most of the time, what you do when looking at manuscripts is (1). This is partly because the differences are not usually big enough and consistently distributed across different manuscripts to be evidence of different versions, and also because many of the differences can conclusively be shown to be mistakes. However, if you ever learn Geonic sources, and compare their talmudic quotations with the received text tradition, you will often find significantly different versions, that, in many cases at least, cannot possibly be mere mistakes. They certainly cannot be merely scribal mistakes, because the Talmud they used was not written down. What they appear to be, rather, is differing versions of an oral text. (This is, of course, made more difficult to identify by the fact that the geonic texts - many of them found only in secondary or tertiary quotations - themselves frequently contain a high proportion of actual scribal mistakes). The degree of variation strongly indicates that the Talmud Bavli was not fixed until near the end of the Geonic period, but that it was possible to add or remove questions and answers to/from sections, and change illustrative details. However, while this is suggestive of substantial variation, it only offers us snippets of it. We don’t have enough quotations to reconstruct an actual variant recension of a single daf, let alone a tractate.

Which is why it is super cool that precisely such a different recension of Moed Katan 7a-11a was found across 3 Genizah manuscripts by Adiel Schremer. Here you can find a full edition of the Geniza version, and here you can find Schremer’s presentation of illustrative differences.

These differences are big. They include the order material is presented, the illustrations of halachic principles, amoraic dicta and baraitot that do not appear in the standard recension, entirely different phrasing of questions, entirely different questions, and expanded explanations. It is clear that this undermines in a fundamental way modern yeshivish learning approaches that are based on close analysis of details that were actually fluid and varied with the geonic academies. When making diyukkim on such details, you certainly are not making a diyyuk about a text compiled by Rav Ashi, nor even about a fixed text owned by the Geonim. Rather, you are making diyukkim on the version that was eventually written down and sent out to the provinces. That’s not nothing, but it’s not the same thing as is generally thought. In addition, this is obviously problematic for Rambamist theories about the acceptance of the Talmud Bavli as lawcode binding on all the Jewish people. How could they accept something that was still in construction and being liberally played around with long after they allegedly accepted it?

Or is it? I went through Schremer’s examples to find the best example that makes a clear halachic difference and I didn’t find one. Perhaps what this really shows is just that the stam is simply not that relevant pound for pound to actual halachic practice.

I don’t know Moed Katan at all well, so I decided to go through all the dapim, use Ein Mishpat and then compare them with the genizah version to see if it makes any difference at all. I’ll hopefully get that done by next week, but, in the meantime, I thought it would be fun if a reader does the same. What are your subjective impressions when learning these side by side? Is this a super big deal, or just a curiosity? Humour me.

(I also recommend reading this genuinely excellent thesis by Ari Bergmann, which is where I read about Schremer’s findings).