Scrolls, Stratigraphy and the Song of the Sea: Re-examining Ashkar-Gilson
Like the majestic strata of the Grand Canyon, Masoretic Bible manuscripts are multi-layered entities. Numerous textual and paratextual layers are laid down over time. Eventually, in the heat and pressure of the worshipping community, these layers coalesce to form the final monolithic artifact. Some of these strata are demonstrably ancient. For example, thanks to the scrolls from the Judean Desert, we can trace the exact consonantal text well back into at least the third century BC/BCE.1 Similarly, recent scholarship continues to reveal that many details of the pronunciation tradition recorded in the vocalisation also have very deep roots—again, well back into the Second Temple Period.2 By contrast, the use of the codex, rather than the scroll, for the writing of the Hebrew Bible constitutes a far later historical stratum.
What of the myriad minor textual and paratextual features that also find their way into the MT? How old, for example, is the tradition of the ‘dotted letters’? In the Tiberian MT there are fifteen locations throughout the Hebrew Bible (ten in the Torah, four in the Prophets, one in the Writings) in which some or all of the letters of a particular word or phrase are written with dots over the top of them (and also underneath them in one instance).3 These dots are considered an important part of the text: any reasonably high quality codex (and even more informal productions) will carefully mark these dots in all the right places. Here, for example, is the instance in Genesis 33:4, as it appears in four different fragments from the Taylor-Schechter Collection. In this case, all the letters of the word וישקהו “And he kissed him” are to be written with a dot over the top.
Fig.1: T-S A1.28: A bifolium from an Eastern Masoretic Codex (c. 10th–12th centuries)
Fig. 2: T-S A1.63: A bifolium from a proto-Sephardic Codex (c. 12th century)
Fig. 3: T-S NS 249.13: A fragment from an early codex from the Eretz-Yisrael tradition (c. 10th century). In this fragment the ‘dots’ are instead represented by oblique lines.
Fig. 4: T-S A21.56: A fragment from an informal, user-produced manuscript. Only the consonantal text is reproduced (no vowels or accents), yet the dots are added as if ‘attached’ to the consonantal layer.
As surprising as it may seem, the tradition of using dotted letters appears to be very ancient. These dots are mentioned in various early rabbinic texts, including the third-century Tannaitic Midrash Sifrei Bamidbar:
בדרך רחוקה נקוד על הה"א… כיוצא בו וישקהו שלא נשקו בכל לבו. רשב"י אומר הלכה בידוע שעשו שונא ליעקב אלא נהפך רחמיו באותה שעה ונשקו בכל לבו.
“On a long journey” (Numbers 9:10)—there is a single overdot on the heh…[here follow other examples of dotted letters, until:] Likewise there are overdots on וישקהו “and he kissed him” (Genesis 33:4)—indicating that he was being disingenuous when he kissed him. Rabbi Shimon bar Yochai claimed the opposite interpretation, saying: ‘It is well known that Esau hated Jacob. But in this particular instance his compassion was stirred and he did kiss him wholeheartedly’ (Sifrei BaMidbar, §69).
This indicates that the use, and position, of these dots was already a feature of the rabbinic tradition by the third century AD/CE. It is likely that the tradition dates back even further—into the Second Temple period. The Dead Sea Scrolls contain many instances of these over-dots, often used as a signal for textual emendation.4 Typically, they are placed over uncertain letters or those that should be omitted. Here, for example, is an instance from column 29 of the Great Isaiah Scroll:
Fig. 5: detail from column 29 of the Great Isaiah Scroll, 1QIsaa (http://dss.collections.imj.org.il/isaiah)
The scribe placed cancellation dots over the phrase מלך יהודה in Isaiah 36:4, and in doing so brought the text into line with the Masoretic Text as we know it.
Presumably, therefore, these fifteen instances of dotted letters in the MT have their roots in this Second Temple scribal practice; the loci became fossilised into the textual stratum, and were meticulously transmitted alongside the consonantal text itself thereafter.5
Of course, not every stratum of the MT will date back to the Second Temple Period—or, at least, we do not have the data to make such a claim. Take, for example, the layout of the Song of the Sea (Exodus 15). In early mediaeval Torah scrolls and codices the scribes carefully adhere to the distinctive layout of the song. Here is an example from a beautiful, early Torah scroll from the Taylor-Schechter collection:
Fig. 6: T-S NS 2.8 recto
To the best of my knowledge the earliest external reference made to the particular layout of the Song of the Sea is in the Jerusalem Talmud (compiled fourth to fifth centuries):
רבי זעורה רבי ירמיה בשם רב. [שירת הים] ושירת דבורה נכתבים אריח על גבי לבינה ולבינה על גבי אריח. עשרת בני המן ומלכי כנען נכתבין אריח על גבי אריח ולבינה על גבי לבינה.
Rabbi Zeira, Rabbi Jeremiah, in the name of Rav: ‘The Song of the Sea and Deborah’s Song are written space-over-brick, and brick-over-space, whereas the list of the ten sons of Haman, and the list of the Canaanite kings is written space-over-space, and brick-over-brick.
We cannot be absolutely certain that the enactment of these stipulations as found in the mediaeval scrolls and codices is in accord with the intentions of the rabbis behind this statement in the Talmud. Nonetheless, the Talmudic statement at least shows that by the fifth century, at that latest, a scribal tradition was in place regarding the correct layout of the Song of the Sea.6
In the case of the dotted letters, we were able to trace the practice found in mediaeval scrolls and codices back to an earlier external rabbinic tradition (Sifrei Bamidbar—ca 3rd century CE). From there, we were able to trace the practice back to a parallel practice found among the Dead Sea Scrolls (c. 2nd century BCE). In the case of the layout of the Song of the Sea, however, the evidence from the Dead Sea Scrolls is less definitive. Of the three relevant fragments (i.e., those containing portions of Exodus 15), only one (4Q365, frag. 6b) reveals anything like the stichometric layout apparently referred to in the Talmud. Thus, while Tov concludes that “[t]he stichographic layout of the writing was probably embedded in the earliest biblical scrolls” (Tov 2004, 162), the evidence from the Dead Sea Scrolls is less clear cut than in the case of the dotted letters.
In large measure, the difficulty of dating the various strata of the Masoretic Text arises from the fact that vanishingly few Hebrew Bible manuscripts survive from the ‘dark’ period between the latest Judean Desert Scrolls (c. the end of the first century AD/CE) and the mediaeval Masoretic Text (relatively few specimens of which remain from the 10th century, but become very numerous thereafter). Due to this gap in evidence, we must rely on external—indirect—sources of information for dating many aspects of the Masoretic Text.
As we have seen, in the cases of the dotted letters and the layout of the Song of the Sea, we have explicit mention of these textual phenomena in the masoretic and rabbinic literature. These give us some sort of external datum on which to begin to try to assess the age of the textual phenomenon under consideration. However, as we are about to see, such external data are not always available.
In Maimonides’ discussion of the halakhot surrounding the writing of Torah scrolls (Mishneh Torah, Sefer Ahavah, Hilkhot Tefillin, Mezuzah ve-Sefer Torah, pereq 7) he notes:
יש דברים אחרים שלא נאמרו בגמרא ונהגו בהם הסופרים וקבלה הוא בידם איש מפי איש… ושיהיה בראש השיטין למעלה משירת הים ״הבאים. ביבשה. השם. מת. במצרים״. חמש שיטין…
"There are other matters [pertaining to the writing of Torah Scrolls] which are not mentioned in the Gemara, yet which, nonetheless, the scribes customarily implement. These matters constitute an oral tradition within their guild…[For example:] The lines above the Song of the Sea should begin with the words ‘Those entering’, ‘on the dry ground’, ‘The LORD’, ‘he died’, ‘in Egypt’: five lines precisely.”
Though it sounds oxymoronic, Maimonides is claiming that there are ‘orally transmitted textual phenomena’—rules for the correct writing of Torah Scrolls that, though implemented in written artefacts, are transmitted only orally.
Maimonides proceeds to mention several such orally transmitted textual phenomena. One of these pertains not to the layout of the Song of the Sea itself (which, as we have seen, is explicitly discussed in rabbinic sources) but to the layout of the lines that immediately precede the Song. He claims that there is an unwritten tradition of beginning the column of text containing the Song with precisely five lines of preceding prose. Those five lines should begin with the words ״הבאים... ביבשה… יהוה… מת… במצרים…״. Sure enough, examination of the mediaeval scrolls and codices reveals that this tradition is indeed well established.7 Here, for example, is T-S NS 2.8 again. Though much of the top of the column is missing, enough data remain to reconstruct the lines with confidence. Sure enough, there are precisely five lines before the song, and they do indeed begin with the requisite words:
Fig. 7: detail from T-S NS 2.8
The previous column ends with the word פרעה (Exodus 14:28), thus guaranteeing that the first word on the first line of the subsequent column is הבאים, as required.
Fig. 8: reconstruction of the text that is only minimally preserved in T-S NS 2.8
How old is this tradition? If, as Maimonides says, this aspect of the text’s layout was an orally transmitted scribal tradition, we obviously cannot hope to find early rabbinic texts discussing the matter. We are reliant on the (limited) manuscript evidence itself. The Dead Sea Scrolls let us down entirely in this instance: the relevant stretch of text has not been preserved. Thus, in this instance the best we can do is to find the oldest mediaeval evidence currently at our disposal. It is with this particular question in mind that we now turn to reexamine the so-called Ashkar-Gilson Torah scroll.
Already in the mid-twentieth century Birnbaum argued (on palaeographic grounds) that the fragment of a Torah scroll from Jews’ College London that he had been asked to examine, dated all the way back to the eighth century and thereby constitutes a rare survival from the ‘Dark Era’ of Hebrew Bible manuscripts (Birnbaum 1959). In the late 1980s at Duke University, James H. Charlesworth wisely sent off small samples of a Torah scroll under his care for radiocarbon dating (Charlesworth 2015). The results showed that the scroll dated back to the seventh-eighth centuries. Eventually, in the first two decades of the 21st century, Engel and Mishor realised, and demonstrated, that the Jews’ College fragment and the fragment from Duke university derive from the same scroll, now generally referred to as the Ashkar-Gilson scroll (Engel and Mishor 2015). Next, in April 2019 Weintraub announced his discovery of thirteen further fragments from the same scroll, eleven of which are preserved in the Taylor-Schechter collection in Cambridge.8 In total, approximately 10% of the entire seventh-eighth century scroll has been preserved, making it one of the most significant Hebrew Bible fragments to survive from the ‘dark’ era.
For our present purpose, the Ashkar-Gilson scroll is highly relevant: the Song of the Sea, together with its preceding and following lines, has been perfectly preserved. Sure enough, in this seventh-eighth century scroll, the column containing the song begins with exactly five lines, beginning with the sequence: ״הבאים... ביבשה… יהוה… מת… במצרים…״:
Fig. 9: detail of an infrared image of the Ashkar-Gilson scroll, MS Durham, Duke University, Ashkar-Gilson #2 (Engel and Mishor 2015: 27)
Does this mean that this particular tradition dates all the way back at least to the seventh-eighth centuries? Well, in a series of articles between 2014 and 2023, Paul Sanders has taken the evidence provided by the Askar-Gilson scroll in a surprising direction (Sanders 2014; 2015; 2023). Rather than arguing that the scroll demonstrates that this particular oral-tradition dates at least back to the seventh-eighth centuries, he argues the opposite: the tradition is at most as old as the Ashkar-Gilson scroll. He claims that the tradition of beginning the column containing the Song of the Sea with the five lines: ״הבאים... ביבשה… יהוה… מת… במצרים…״ actually originates with the Ashkar-Gilson scroll. In other words, the tradition is not found in the scroll, but is founded on the scroll. If correct, this particular stratum of orally-transmitted scribal tradition dates back no earlier than the seventh-eighth century.
Sanders’ bold claim is based on an argument with two main elements. First, he shows that there is a very close relationship between the Ashkar-Gilson scroll and the Aleppo Codex, particularly in terms of their orthography (spellings).9 Second, he focuses specifically on the five lines preceding the Song of the Sea, and the text in the preceding column. He claims that all the text in that preceding column, and in the five lines themselves, is laid out very naturally. There is no evidence that the scribe had to manipulate the layout of the text, by compression or dilation of letters or words, in order to make the text ‘fit’ an already-authoritative layout. Rather, the text gives the impression that the scribe was simply writing, ‘business as usual’, and the resulting layout, with five initial lines each beginning ״הבאים... ביבשה… יהוה… מת… במצרים…״ was entirely unplanned. This contrasts noticeably with many other masoretic manuscripts, which have to resort to substantial manipulation of the layout in order to ‘land’ הבאים at the top of the relevant column, and then to keep the five first lines beginning with their requisite words. Compare, for example, the simple, continuous layout in the scroll, against the layout found in the Leningrad Codex:
Fig. 10: part of an infrared image of the Ashkar-Gilson scroll, MS Durham, Duke University, Ashkar-Gilson #2 (Engel and Mishor 2015: 27)
Fig. 11: f. 39v of Codex Leningrad (L)
The image above is from the page of text in L immediately preceding the page containing the Song of the Sea. Note how the scribe, Samuel, uses an unusually large number of line fillers—particularly in the first column. He is manipulating the layout (in a rather obvious and ungainly fashion, one might think) to ensure that the last word on this page is פרעה (Exodus 14:28), so that the next page can begin, as the tradition requires, with באים.
Samuel’s difficulties are not over, however. Having successfully wrangled the text into position by the end of this page, he then needs to control the first five lines preceding the Song of the Sea so that they begin with the traditional sequence: ״הבאים... ביבשה… יהוה… מת… במצרים…״. As the image below shows, lines two, three and five had to be stretched and padded out in order to maintain the official sequence of first words:
Fig. 12: the five lines before the Song of the Sea, from f. 40r of Codex Leningrad (L)
Sanders then reasons as follows: from (1), the Ashkar-Gilson scroll was known and valued by the Tiberian Masoretes, including the scribe of the Aleppo codex; from (2), the incidental layout surrounding the Song of the Sea in the Ashkar-Gilson scroll, with its beautiful symmetry of five lines above and below the song itself, was known, and came to be imitated by, subsequent masoretes, thereby birthing a new orally-transmitted scribal tradition (Sanders 2023, 16–19).
In some ways, Sanders’ suggestion is attractive; it would be remarkable to be able to pin down this scribal tradition, which became so widespread, to one particular manuscript. Sadly, the argument does not stand up to scrutiny. Much of the weight of the claim is based on the fact that the layout in the scroll looks so natural, so unforced, in contrast to the patently forced layout found in manuscripts such as L. This contrast, though, is based on (i) insufficient and (ii) inappropriate comparisons. Each of these issues will be dealt with in turn.10
As for the number of comparanda, Sanders’ initial study compares the Ashkar-Gilson scroll against eight other manuscripts: two scrolls and six codices. In his subsequent study he adds another codex to the list of comparanda. Sanders’ choice of comparanda was, in part, dictated by his intention to compare various aspects of the Ashkar-Gilson scroll against other early Eastern representatives of the Masoretic Text, not just the layout of the lines preceding the Song of the Sea. However, when one focusses narrowly on the issue of the layout of the lines before the Song, and the columns of text preceding the Song, one is freed to use a far wider set of comparanda. Perez, for example, in his study of 242 Ashkenazi codices, found 38 codices in which the word הבאים (Exodus 14:28) begins the page containing the Song, 22 of which achieved this aim without obvious forcing of the text in the preceding columns (Perez 2012). In the same study, he also examined the lines preceding the Song in 10 early Eastern masoretic codices. Of the seven codices in which the tradition under discussion is manifested, no fewer than five achieve this result without clear signs of ‘forcing’ the text to comply.11 Thus, it seems that—perhaps contrary to our expectations—it was not beyond the skill of at least certain scribes to unobtrusively control the textual layout of the material before the Song of the Sea such that הבאים fell at the top of a new page seemingly by chance.
Regarding the types of comparanda: Sanders mainly compares the Ashkar-Gilson scroll against codices, rather than other scrolls. Again, this was because Sanders’ intent was to compare a variety of the scroll’s features against early Eastern textual witnesses of the Masoretic Text. However, specifically with respect to the issue of the textual management of the lines preceding the Song, comparing a scroll to codices is doubly problematic. With regard to ‘landing’ הבאים at the top of a new column: scrolls typically have (a) more lines per column; (b) wider columns; (c) more words per line, than codices. These three features all work to the scribe’s advantage when seeking to manipulate the layout of the text: tweaks in spacing can be distributed inconspicuously across a much larger body of text vis-a-vis codices. Secondly, with regard to the layout of the five lines preceding the Song itself, the producer of a scroll has far greater freedom to choose the appropriate column width for those lines than the producer of a codex, with its tightly constrained written area per page.
In light of these factors, one would expect to find scrolls post-dating Maimonides, where the five-line layout is intentionally adhered to, yet without any obvious ‘forcing’ of the layout, either in the preceding columns, or in the layout of the five pre-song lines. Sure enough, this is what Perez found in his study of eight Ashkenazi Torah scrolls. No fewer than half of those scrolls achieved the desired layout of the lines preceding either the Song of the Sea, or Moses’ Song (or both) without obvious adjustments to the text (Perez 2012).
Similarly, when perusing the scroll material in the National Library of Israel, I found several examples where the required layout shows no signs of being forced. Here are three of the best performances I happened across:
Fig. 13: Jerusalem, National Library of Israel, Ms. Heb. 6989 Columns 073–075 (The National Library of Israel. "Ktiv" Project)
Fig. 14: Jerusalem, National Library of Israel, Ms. Heb. 8456, image 25 (The National Library of Israel. "Ktiv" Project)
Fig. 15: Jerusalem, National Library of Israel, Ms. Heb. 8°5405 (The National Library of Israel. "Ktiv" Project)
In light of the examples above (and more could be adduced),12 and the results of Perez’ study, there is now no reason to interpret the layout of the text in Ashkar-Gilson as unintentional, rather than a skillfully executed intentional arrangement.
Sanders was quite correct in drawing scholarly attention to the significance of the Ashkar-Gilson scroll for the question of the antiquity of the tradition concerning the five lines preceding the scroll. However, in light of the above, it seems preferable to draw the opposite conclusion to his. Rather than seeing the Ashkar-Gilson scroll as the terminus a quo for the growth of this particular scribal tradition, it makes more sense to see it as the terminus ad quem. In other words, already in the seventh-eighth centuries there was an orally-transmitted scribal tradition to begin the column containing the Song of the Sea with five lines, beginning with the words ״הבאים... ביבשה… יהוה… מת… במצרים…״. This tradition had been passed on, in Maimonides’ words, איש מפי איש, for at least half a millennium, by the time Maimonides wrote about it.
Appendix I: The Orthography of Ashkar-Gilson
Sanders checked the orthography of much of the extant Ashkar-Gilson text. However, some fragments are so discoloured that they are hard to decipher using ordinary photographs, and Sanders did not check the orthography in these fragments against the presumed orthography of A (Sanders 2023, 3–4).
I had the privilege of examining the Cambridge fragments in person, and so was able to read even the darker ones. As Sanders has already shown, the manuscript is a magnificent example of scribal care. Nonetheless, it turns out that the consonantal text is not quite identical to the text prescribed by the Tiberian masora. At Deuteronomy 33:9, in the middle of Moses’ blessing on the Levites, we read:
האמר לאביו ולאמו לא ראיתיו
In the Aleppo Codex (extant at this point), as well as in all the usual codices serving as comparators (Leningrad, Damascus Pentateuch, Sassoon 1053) the word ראיתיו is spelled (as here) plene: with a yod between the alef and the tav. In the Ashkar-Gilson scroll, however, it is spelled defectively (without the yod). Even in such a tiny detail, however, the story may not be as simple as it initially appears.
The relevant part of the scroll is currently preserved as fragment T-S AS 37.10 in the Taylor-Schechter collection (the image below has been lightened to render it more easily visible):
Fig. 16: T-S AS 37.10
Fig. 17: detail of T-S AS 37.10 in visible light and infrared
Looking at the word in question, there is a suspiciously large and pronounced mark directly over the defectively spelled ראתיו. The mark partially disappears under infrared light, and hence appears to have been made with iron-gall ink rather than carbon ink. There are, naturally, many imperfections in the parchment’s surface. Nonetheless, a mark of such unusual size and prominence directly over the exact location of the non-masoretic spelling would be rather a coincidence. It seems more likely that the mark is a clumsily performed correction, or that it was intended as some sort of editorial mark (perhaps, as we find elsewhere, a prompt to investigate the spelling in more detail and emend the text if required).
Appendix 2: The Number of Words per Column
In the process of preparing the diagrams found in Appendix 3, it was necessary to consider the average number of words per column in the extant parts of the Ashkar-Gilson scroll.13 Data were gathered wherever the column incipit and explicit were not in doubt: 28 columns in total:
| Text Range | Words per column |
|---|---|
| Gen 10–13 | 243; 247; 251 |
| Gen 45–Exod 3 | 241; 255; 512 (two columns); 228; 256; 256; 266; 272; 488 (two columns); 255; 263 |
| Exod 9–14:28 | 237; 261; 266; 246; 241; 249; 208; 263; 244; 223 |
| Deut 33–34 | 220; 213 |
From this data, the mean number of words per column is 246.6. The standard deviation is 16.4. In the table above I have underlined all the numbers more than one standard deviation lower than the mean (i.e., any number lower than 230.2). The two low numbers in Deuteronomy 33–34 are simply explained: the scribe intended to close the book of Deuteronomy at the bottom of a column (per b.Menahot 30a). Consequently, he reduced the width of the final columns of text. The average column width over all the extant fragments is 7.9 cm. The two final columns, however, are only 7.0 and 7.2 cm respectively. The columns containing 228 words and 208 words are each the final column of a sheet, and are, once again, narrower than average (7.7 and 7.3 cm respectively). This just leaves the column containing 223 words, which just happens to be the column immediately preceding the column containing the Song of the Sea. This column is not the final column of the sheet, nor are there other external factors that might constrain the column width. The most likely explanation for the unusually low number of words in this column is that the scribe was carefully attending to the layout of the text, seeking to ensure that he began the Song’s column with the word הבאים, in diligent accord with an already extant tradition.
Appendix 3: The Ashkar-Gilson Jigsaw
The fifteen fragments of the Ashkar-Gilson scroll currently known comprise three isolate fragments (Genesis 10-13; Numbers 10; Deuteronomy 2–3), and twelve fragments that cluster into three groups. The first group, consisting of five fragments, contains Genesis 44–Exodus 3. The second group, also consisting of five fragments, contains Exodus 9–18. The third group, consisting of just two fragments, contains Deuteronomy 32:50–34:12. The connections between the fragments in each of these three groups are as follows:
Fig. 18: Ashkar-Gilson, reconstruction of Genesis 44–Exodus 3
Fig. 19: Ashkar-Gilson, reconstruction of Exodus 9–18
Fig. 20 Ashkar-Gilson, reconstruction of Deuteronomy 32–34
Bibliography
Birnbaum, S. A. 1959. “A Sheet of an Eighth Century Synagogue Scroll.” VT 9:122–29.
Charlesworth, James H. 2015. “Ashkar Manuscript 2: Introducing a Phenomenal Witness to the Bible.” Israel Museum Studies in Archaeology 7:66–69.
Diamond, James S. 2019. Scribal Secrets: Extraordinary Texts in the Torah and Their Implications. Edited by Robert Goldenberg and Gary A. Rendsburg. Eugene, Oregon: Pickwick Publications.
Engel, Edna, and Mordechay Mishor. 2015. “An Ancient Scroll of the Book of Exodus: The Reunion of Two Separate Fragments.” Israel Museum Studies in Archaeology 7:24–61.
Hornkohl, Aaron. 2023. The Historical Depth of the Tiberian Reading Tradition of Biblical Hebrew. Cambridge Semitic Languages and Cultures 17. Cambridge: Open Book Publishers.
Ofer, Yosef. 2019. The Masora on Scripture and Its Methods. Jerusalem: Magnes Press.
Perez, Joseph. 2012. “תחבולות סופרים בכתיבת הקטעים שלפני שירת הים ושירת האזינו.” Quntres 3:35–59.
Phillips, Kim. 2017. “Writing Lines: T-S D1.108 and the Song of Moses.” Genizah Research Unit, Fragment of the Month, March 2017.
Sanders, Paul. 2014. “The Ashkar-Gilson Manuscript: Remnant of a Proto-Masoretic Model Scroll of the Torah.” JHS 14:1–25.
———. 2015. “Missing Link in Hebrew Bible Formation.” BAR.
———. 2023. “Rediscovered Fragments Shed New Light on a Proto-Masoretic Torah Scroll.” JHS 23.
Tov, Emanuel. 2004. Scribal Practices and Approaches Reflected in the Texts Found in the Judean Desert. Studies on the Texts of the Desert of Judah 54. Leiden: Brill.
Weintraub, Mordechai. 2019. “More fragments of early Torah scroll come to light.” Genizah Fragments newsletter 77:1–2.
Yeivin, Israel. 2003. The Biblical Masorah. Studies in Language III. Jerusalem: The Academy of the Hebrew Language.
Footnotes
1 See, for example, Tov 2004.
2 Hornkohl 2023. Not to mention, of course, that the actual content of the text derives—at least in part—from far earlier: into the first half of the first millennium BC/BCE, and even the second half of the second millennium BC/BCE.
3 See Yeivin 2003, 45–46, and the literature cited there.
4 Tov 2004, ch. 5.
5 For a fuller, highly readable, discussion of these scribal dots and the secrets they hold—or have been given for safekeeping—see Diamond 2019.
6 For more on the layout of the songs in the Hebrew Bible, see Ofer 2019, 75–84.
7 Another such orally transmitted stipulation—the layout of the lines preceding Moses’ Song—has been dealt with in earlier Fragments of the Month. See Phillips 2017 https://doi.org/10.17863/CAM.28081.
8 The announcement came in the April 2019 edition of Genizah Fragments (volume 77), which can be found here.
9 Sanders also compares the Aleppo Codex (and other early Tiberian codices) with the Ashkar-Gilson scroll at the level of petuḥot and setumot, though the proximity between the two is less pronounced in this regard (Sanders 2023, 15).
10 Moreover, we are now in a position to show that, in fact, there is evidence that the scribe manipulated the text layout in the lead-up to the column containing the Song of the Sea. See Appendix 2 for details.
11 To this total I am happy to add EVR II B 80+, a 10th–11th century Eastern Masoretic codex on which I have been working recently, which manages to position הבאים at the top of a new page with no apparent wrangling.
12 The following manuscripts manage the layout very well: Jerusalem, National Library of Israel, Ms. Heb. 5935; Ms. Heb. 6100*4; Ms. Heb. 7637*4. Other manuscripts also show impressive line management, but perhaps not quite as successfully as the above: Jerusalem, National Library of Israel, Ms. Heb. 4°8457; Ms. Heb. 1407; Ms. Heb. 1459; Jerusalem, Hechal Shlomoh, Scr.41.
13 Mishor had already provided the relevant data for the London and Ashkar fragments (Engel and Mishor 2015, 31). I would like to extend my most grateful thanks to Mordechai Weintraub, whose attention to detail in reading this appendix saved me from a most embarrassing blunder.
If you enjoyed this Fragment of the Month, you can find others here.
Contact us: genizah@lib.cam.ac.uk
The manuscripts in this article are part of the Cairo Genizah Collection in Cambridge University Library. To see more items from this collection visit: https://cudl.lib.cam.ac.uk/collections/genizah/1