The Yemeni Origin of the Vienna Scroll

 

I have found good evidence that the Torah scrolls copied by the Kaifeng Jews in the mid-seventeenth century were based on Yemeni models.   At this point, I can only say this about the Vienna Scroll.  As I get images of other scrolls from Kaifeng, I will examine them as well.  

The Jews of Yemen have distinct traditions in their Torah scrolls that differ from other communities.  Here are twelve of these differences:

Verses	Content	Change
	Genesis (בראשית)
4:13	גדול עוני מנשא[20][21]	The word מנשא is written without a "waw" (defective scriptum)
7:11	נבקעו כל מעינת[21][22][23]	The word מעינת is written without a "waw" (defective scriptum)
9:29	ויהיו כל ימי נח[21][22][24]	The word ויהיו is written as a plural with a final "waw"[25]
41:45	פוטיפרע[26]	Every פוטיפרע is written as one word
	Exodus (שמות)
25:31	תעשה המנורה[21][27]	The word תעשה is written without a "yod" (defective scriptum)
28:26	אל עבר האפד[21][28][29]	The word האפד is written without a "waw" (defective scriptum)
	Leviticus (ויקרא)
7:22–23	פרשת כל חלב[28][30]	This section is written as an "Open Section"[31]
	Numbers (במדבר)
1:17	אשר נקבו בשמת[21][28][33]	The word בשמת is written without a "waw" (defective scriptum)
10:10	ובראשי חדשיכם[21][28][34]	The word חדשיכם is written with a "yod" (plene scriptum)
22:5	בלעם בן בער[28]	The word בער is written without a "waw" (defective scriptum)
25:12	בריתי שלום[28][35][36]	The letter "waw" in שלום is written as all other "waws" (without shortening)
	Deuteronomy (דברים)
23:2	פצוע דכא[21][28][37]	The word דכא is written with an "aleph", instead of "he."[38]

Let's take these in turn:

Genesis 4:13

VS

SMU

HUC 951

Here, three Kaifeng texts, VS, SMU and HUC conform to the Yemeni spelling, which does not contain a vav, as in other traditions:

Genesis  7:11

VS

SMU

No extant HUC

Here, two Kaifeng examples do not conform to the Yemenite spelling, which does not contain the final vav.  The VS and SMU conform to the majority spelling tradition:

Genesis 9:29

VS

SMU

No extant HUC

Here, VS and SMU conform to the Yemeni spelling, which has a vav, making the verb plural.  Other traditions do not:

Genesis 41:45, 50

VS

SMU

No extant HUC

In Yemeni scrolls, each time this personal name is written, spell it as one word.  In other traditions, it is two:

Exodus 35:31

VS

HUC 960

The VS and HUC 960 do not contain the yod of other traditions:

Exodus 28:26

VS

No HUC

The VS does not contain the vav of other traditions:

Leviticus 7:22-23

VS

HUC 965

Both VS and HUC 965 have this section as open, in that it starts a new line or there is a break in the text.  Other traditions have this section as closed:

Numbers 1:17

VS

HUC 967

Both the VS and HUC 967 do not contain the vav of other traditions:

Numbers 10:10

VS

No extant HUC

Here, the VS has an additional yod not found in other traditions:

Numbers 22:5

VS

enhanced VS

HUC 971

Here is an example, one of many in the VS, where a vav was placed in this word sometime after the word was written.  In the HUC text, there is an editorial note to add the vav.  This makes this word conform to non-Yemenite traditions:

I consider this a half an example of the VS conforming to Yemenite tradition as this example was adjusted by a Kaifeng scribe.

Numbers 25:12

VS

No extant HUC

There is a textual tradition outside of Yemeni Torch scrolls of writing a broken or incomplete vav in this word:

In Michael Pollak's work on the textual details of the Kaifeng Torah scrolls, this is one of his examples of textual variants found in the Kaifeng texts:

Pollak is usually generous with the abilities of Kaifeng scribes, but he did not believe that they would have noticed the broken vav and copied it, or they would noticed the broken vav they would have presumed that is was a "scribal slip" and corrected it.  It seems neither of these scenarios is correct.  The most likely explanation is the simplest: the Kaifeng scribes using Yemeni texts, or ancestors of these texts, which did indeed lack a broken vav.

Deuteronomy 25:12

VS

HUC 978

VS and HUC have an aleph in this final word, whereas other traditions have a he:

Of these examples, only 1.5 do not conform to Yemeni scribal tradition, making the VS conform in 83% of these examples.  Why not 100%, or close to it, if the was the model the Kaifeng scribes used in the mid-seventeenth century was Yemeni?

In Padoll's 1970 masters thesis on the liturgy of the Kaifeng Jews, he found a Yemini manuscript that almost completely conformed to the Kaifeng liturgy he investigated.  He calls this MS Y:

Padoll was concerned about the "very few expectations" in HUC 944, a book of Purim prayers, and even through they seemed insignificant to him, they cast a "shadow" over his work.

In fact, texts that are mixed or hybrid are not surprising, however.  There other Jewish communities who have Torah scrolls and siddurim that illustrate the arrival of many influences over their long history, such as the Jews of Cochin, India:

The city of Cochin, on the Malabar Coast in Southern India, aside from its Dravidian languages-speaking Hindu and Muslim population and its various Christian minorities, has been a place of Jewish settlement since at least the early medieval period. Though these Jews, who speak the local language, Malayalam, are usually reputed to be of mostly indigenous origin, their synagogue liturgy stems from the Sephardi (Spanish) rite. The pronunciation of Hebrew that is common among them,1 their liturgical music,2 as well as the very form of prayers used in Cochin is cognate to those used by the various Jewish communities of Iberian background in Europe, North Africa and Asia. However, on that common basis, a later local tradition developed of adding circumstantial poetry for special holidays, life-cycle celebrations and various other occasions to perform liturgical services.

From Two Judeo-Spanish ‘Marrano’ hymns in the liturgy of the Jews of Cochin, by Peter Nahon. 

Like the Jews of Cochin, the Kaifeng Jews were relatively isolated from other Jewish communities, and would have replenished their Torah scrolls and books from visitors.  The origins of the Jews of Cochin was not Sephardic, but because of migration patterns, this Medieval Jewish community, founded around the turn of the first millennium like the Jews of Kaifeng, would have adopted and adapted the customs of other communities into their own over the centuries.

The foundational Jewish culture of the Kaifeng Community were of Judeo-Persian speaking groups from Central Asia.  But it would make sense that Arabic speaking Jews also made their way to China, perhaps from the coast, and eventually may have shared their Torah scrolls and siddurim with the Kaifeng Community. 

Before we saw how the two Torah scrolls in the Kaifeng Synagogue from the mid-1400s were from Chinese coastal cities:

We read in the 1489 and 1519 stone inscriptions that after the flood of the middle 1400s, the community was gifted one scroll a replacement scroll from Jews (or Jewish communities) in Ningbo and Yangzhou.  Are all the Torah scrolls of the Kaifeng Jews the descendants of these two scrolls?  If so, where did they come from?

Were the Ningbo and Yangzhou scrolls, which were probably used to create both the Scroll of Moses and the 12 other copies of the Torah after 1642, Yemenite scrolls, or based off Yemenite scrolls?  Given the evidence above, that seems likely.  The Jews of Kaifeng then, over the centuries, combined them with other books, from other places, in their possession. So we find traces of influences from other eastern Jewish communities in their scrolls as well.  See here.

It is also possible that the Kaifeng texts were already hybrids by the time they reached that interior Chinese city.  Either way, the Kaifeng Scrolls, and their liturgy, have a predominately Yemenite influence.  This, despite the fact that the foundation of the community of Kaifeng was Persian.   

Sometime in the history of the Kaifeng Community, texts from Persia were replaced with Yemenite books.  Is this the reason for the Arabic female names found in the Memorial Book?  Perhaps some Arabic speaking Jews arrived in Kaifeng.  It is possible, but after the Muslim conquest of Persia, many loan words made it into Persian.  Are the names there are example of this trend, or due to an influx of Arabic speaking Jews into the Kaifeng Community?

Next we will examine the columnar words employed in the Vienna Scroll (and by extension, the other scrolls, as Pollak examined these variants as well) and how the convention of having words begin with six distinctive letters helps us illustrate the wanderings of the Kaifeng Torah scroll exemplars.