T-S A36.18 – hole-y megillah it’s the oldest!
Many scholars have remarked on the fact that the scroll of Esther is missing from the Dead Sea Scroll corpus, and whether this was purely a lack of survival or a deliberate choice on the part of the sectarians of the Yaḥad (Community).1 This is particularly notable, since in modern times the scroll of Esther is quite common and is one that would be commissioned for a household to own and to read on Purim, at home and in the synagogue. For scribes like myself, Esther often holds a special place in our hearts since it is the first thing we learn to write, because it does not contain any names of God.
There are Megillot ʾEster present in the Cairo Genizah, which whilst catalogued, are largely ignored, likely since the text is so ubiquitous. This FOTM focuses on one that features in The Illustrated Cairo Genizah: A Visual Tour of the Cairo Genizah Manuscripts at Cambridge University Library, a volume to mark the 50th anniversary of the foundation of the Cambridge Genizah Unit.2 The book consists of the more visually interesting, or historically significant fragments.3
This particular scroll, CUL T-S A36.18, was to appear largely because of its rather interesting damage pattern, such that I named it the ‘hole-y megillah’.4
T-S A36.18 recto.
The catalogue entry is typically brief, noting that it covers Esther 5:14–8:6, is written in unpointed Hebrew and is ‘vellum: 22.5cm. x 38.9cm.; 3 columns; 1 leaf, part of a Megillah scroll; 21 lines; badly mutilated with many large pieces of the text missing’.5 No mention was made of its dating, so I was set the task of producing an estimate of its date of composition to include in the new publication.
Dating kitvey ha-qodeš (holy writings) is a challenge. There are no colophons, as one might have in a text presented in a codex. Thus, to locate and date the work, we are reliant on palaeographical methods to assess the script from known dated comparators, and by reference to certain scribal features (e.g. for a Torah, formats of the songs Širat ha-Yam, Ha-ʾazinu, maleʾ/ḥaser, use or not of wawey ha-ʿamudim, number of lines, deployment of ʾotiyyot mešunnot etc.). C14 dating is also valuable, athough it does give a range of dates, and findings can be questioned. However, as a destructive test, it is rarely used, and some institutions rule it out altogether.6 Thus, there are often arguments over the provenance of scrolls by scholars.
A more positive note, and a key methodological consideration, is that prior to normative scribal halakha becoming more prescriptive as regards the forms of the letters, and certainly before the advent of printing, when specific STa”M scripts (the calligraphic scripts used in Torah scrolls, tefillin and mezuzot) became even more standardised,7 scribes of sacred texts employed the same contemporary book-hand that would be in use for other non-sacred works in their locality.8 Thus, any well-executed, ‘beautiful book-hand’ to borrow Tigchelaar’s phrase,9 can be used as a valid comparator, regardless of the content, sacred or otherwise, giving a wider range of similar scripts, than would otherwise be available.
As is my usual practice, I traced an abecedary with several versions of each letter for comparison.10 The book-hand of T-S A36.18 is a well-executed and consistent square script and has a uniform stroke weight, characteristic of the use of a reed (as opposed to the variant stroke weight that later European manuscripts written with a quill show, with thicks and thins).11 The gagim (roofs) of the letters hang directly from the sirṭuṭ (ruled lines); the script leans ever so slightly to the left, and has a consistent baseline. It is written in 21 lines per ʿamud (column)12 with fully justified text often effected through the dilation of certain letters at the end of the line.13
One of the most prominent features of the script is the downward stroke, a ‘serif’ of sorts, on the left side of the gag (roof) of the letters. This is a deliberately added type of tag (not the later neo-tagin) decoration in the ductus of the letters.
It is a feature that is very evident in early codices and other manuscripts from ʾEreṣ Yisraʿel, Egypt (where our manuscript was likely written) and Babylonia. Calligrapher and sofer Izzy Pludwinski calls this a ‘sloped stroke followed by a horizontal stroke’.14 And Gina Jonas a ‘lozenge’. This device has been adapted into a modern calligraphic script by Fred Pauker and is aptly named “Pauker” script, though his ‘slab-like lozenge stroke’15 is a more stylised, ‘almost diamond-shaped’ element, though with a thorn-like addition.16 Given their shape, I refer to these as shofar serifs.17
Examples of a shofar type tag on letters from T-S A 36.18, and a drawing of my shofar for comparison
A further key marker to the age of the script is the absence of any neo-tagin (either the straight line or the balls on sticks/mini-zayin shapes).18 Esther, like any kitvey ha-qodeš, is supposed to have letters that are adorned with these small crowns. These had been a reinvention of tagin which originally were protrusions from the monoline—that in part had been forgotten, omitted, or became the heads of letters.19
An example line from a Megillat ʾEster I wrote in 2018, that shows the neo-tagin decorations
In the script of T-S A36.18, for example, the regel (leg) of the dalet stroke tends to start from above the gag (roof) crossing over. The lamed generally has a high ascender and is particularly high at the top of the ʿamud (column) where the scribe enjoys the freedom of the marginal space. The descender of the nun sofit is on the right side and not from the centre. The gimel often has a lovely arched form.
Select examples drawn from the full abecedary
Other markers of age include much evidence of nesting, where the foot of letters appear to ‘underline’ the letter that follows (see below for examples). Additionally, that the script ‘breaks’ the halakha for a megillah by having letters that have negiʿot (joins), as opposed to being muqaf gewil (surrounded by parchment), is also an early marker. This is frowned upon by later scribal practice, but scribes of earlier periods seem to have been less concerned about such matters, even though the halakhic prescription appears to hail from Talmudic times.20 We also see that the writing is quite crowded with narrow gaps between words, such that they often approach ke-mila ʾaḥat (like one word), which in modern scribal practice would also make a work pasul (invalid).21
Top line: examples of nesting and negiʿot (joins), characteristic of this Megillah; Bottom line: examples of very tight spacing with words that look like they run together kemila ʾaḥat (as one word)
Even taking a lenient approach, given the overall tight spacing, the two spacing examples would invalidate the scroll, שְׁנֵי֙ סָרִיסֵ֣י הַמֶּ֔לֶךְ (‘two officials of the king’) from Est. 6:2 and לָֽבַשׁ־בּ֖וֹ (‘[which the king] dresses in’) from Est. 6:8. This latter example sparked some lively debate, as below the word there is a dot.22 Is the dot merely a coincidental drip on the parchment or does it have some paratextual significance? Even though halakha states that extra marks on a STa”M article invalidates it for ritual use.23Looking at it closely, it does seem to be a deliberate mark from a qulmus as opposed to a drop of ink.
Detail of T-S A36.18, showing the mark under the last letter of לבש
It is possible that the original scribe was so concerned that the words were too close together that he marked it. Prof. Outhwaite notes that ‘we do see occasional signs in [Genizah] Torah scrolls to mark problematic areas’. However, if that were the case then it would have been more useful to mark under the bet, which starts the second word, since this does not actually remove the safeq (doubt); with the dot under the šin this could be read as לב שבו (heart, that is in it), which would not help the reader. That said, the reader would likely know it is לבש בו anyway. Dr Kim Phillips, layering onto the ‘coincidence’, notes that the shape of the dot resembles the ṭipḥaʾ accent in the Ereš Yisrael trope notation tradition and wonders whether our scribe chose that to mark the separation, since that is the cantillation for this phrase. If so, again, it really ought to be under the bet. However, as Dr Phillips notes: ‘in the Ereš Yisrael system the accent does not necessarily get written adjacent to the accented syllable’ and perhaps, because of the maqqef (hyphen) is being treated as a single accentuation unit. Alternatively, it is simply a random dot! The plot (or should we say the dot) thickens.
Turning to less speculative matters, and looking for immediate comparators for the ductus, this script seems to fall somewhere between Oxford Bodl. Ms. Heb. d.26, which Yardeni in Hebrew Scripts assigns to 8th/9th century because of the early Babylonian vocalisation24 and a more developed script (with a more similar taw), National Library of Russia Firkovich EBR. I B 3, dating from 916. This second comparator has similar shofar serifs in places, though is a much more polished developed calligraphic hand.25 It is also quite crowded and displays considerable nesting. The nun sofit’s descender comes from the right side of the gag (roof), he is joined on the left regel (leg), and on the right side is not as rounded. Pe on the right side, conversely, is a bit more rounded, but nonetheless a good comparator.
The ductus for Bodl. Ms. Heb. d.26 that I have drawn from Yardeni’s suggestions. Multiple forms are provided in Yardeni’s, The Book of Hebrew Script
An example abecedary for RNL Firk. Ebr. I B.3. A near full abecedary with alternatives are also given as Script 3 (Leningrad, PL Firkovitch I B3) in Specimens26
The example of the gimel that Yardeni brings in her book is somewhat different to the rather striking curved/arched gimel in our megillah. However, in a lecture on scripts, Judith Olszowy-Schlanger referenced the Bible from 916 and drew attention to its long gimel, which is much more akin to the curved shape found in our megillah (see above).27 This shows the danger of looking at one example of a letter in a script. For whilst abecedaries with multiple examples are useful, there is still a need to look at allographs of the forms in the original manuscript.
Script 5 in Specimens, National Library of Russia Firk. Ebr. II B 17, which is dated to 929, also has some similarities, but noticeably less prominent shofar serif strokes.
Looking wider for comparators, I found not an Esther scroll but an Aramaic and Hebrew midrash on Esther, Jewish Theological Seminary ENA 5502. This is described on Ktiv as מדרש על מגלת אסתר (קטע) and is datable to the 10th century.28 The script is very similar to our megillah. Even the gimel has the arched regel. The nun is narrower, however, and the nun sofit is also quite different with the descender falling from the centre of the gag rather than the left.
Example details of the script from JTS ENA 5502. Images provided by and © The Library of The Jewish Theological Seminary. Used with permission.
An additional, and particularly useful comparator, with a well-executed script, is ketubba T-S 24.35, for Kuzayr bat Ḥusayn and her groom called Ephraim (his father's name didn't survive), which has very similar features to our script and is explicitly dated to [47]40 AM (= 980 CE).29
Detail from T-S 24.35 recto, ketubba
Just as dated colophons in manuscripts provide a framework for palaeographic dating, dated ketubbot could be a very useful set of comparators for dating scripts. Giṭṭin too can sometimes be useful, but are often less well written, and so may not always be as helpful to compare against generally more calligraphic STa”M scripts.30
Of course, some features of these sorts of scripts carry on through to the 1500s, including crowded spacing, nesting and the use of these shofar serifs, the overlap of the dalet roof and leg (e.g. script 41 in Specimens, Bodl. Or. 23 1248/49), but later megillot will undoubtedly have sported neo-tagin, and would not ‘break’ the halakha regarding joins. The earlier examples I cite above do seem to provide a much better match. Indeed, two of the scripts compare very well to the gimel, which is quite curved and sometimes long in the leg, and it is key. The nun sofit seems longer than most of the other scripts and is a little different (having the descender from the right of the gag) and was perhaps my only reservation against the overall pattern seen, but its presence in the Bible from 916 is reassuring.
One final marker of age is the close-knit loop stitching (see below). Not as tightly packed as you get in Dead Sea Scroll fragments, but much more ‘dense’ than the ‘spaced out’ loop stitching that develops subsequently.31
Example of loop stitching from the left-hand edge of T-S A36.18, recto
Overall, it would seem that this megillah, represented by the single surviving fragment of T-S A36.18, was likely written sometime during the 10th century. This is clear from comparisons with the dated ketubba fragment T-S 24.35, the midrash on Esther, and the gimel, in particular, of the Codex Babylonicus from the National Library of Russia. However, it could even be a little earlier than that based on the Bodleian's Ms. Heb. d. 26.
Looking at the wider corpus of Megillot ʾEster, searches on Ktiv suggests that there are tiny fragments of Megillot ʾEsther dated to the 11th century, Ms. I A 214 (Est. 1:6–10) and Ms. I A 61 (Est. 2:21–3:8) held by the Library of the Alliance Israélite Universelle. The next oldest is Austria D II, dated to the 13th century and described as Ashkenazi. However, I do not think it is, since the script is more Oriental. Additionally, a subsequent sofer has quite obviously added three line-only tagin onto the roofs of the Šeʿaṭnez Gaṣ letters to conform to the later standard. So, it may be earlier than the date given. More certain is the Ashkenazi manuscript from Herzogenburg, Austria Ms. BD 9 1460, dated to the 13th or 14th century, discovered in a binding. Other general searches for Megillat ʾEsther suggest articles or auction sites with megillot that are no earlier than the 14th century. While ʾEsther is represented in early codices, like Leningrad B19a and Sassoon 1053, though missing from the Aleppo Codex. This is, however, not the same as a liturgical scroll such as we have here with T-S A36.18, though people sometimes conflate the two; I have seen articles referring to the ‘oldest Torahs’ that then include references to the early codices – these have the status of a ḥumaš and not a Torah.
Given this not unreasonable dating to the 900s, T-S A36.18 could thus be the earliest ritual STa”M megillah extant, since Esther is not represented in the Dead Sea Scrolls. Of course, proper study of the rest of the undated megillot in the Genizah could always yield something earlier.
As a result, and to mark this, I created a font from my abecedary and a reconstruction of the three megillah ʿamudim on the single surviving yeriʿa to give an indication of what it looked like in its full glory.
My reconstruction of the fragment T-S A36.18.
In conclusion, T-S 36.18 may have originally attracted attention because of its visually interesting damage pattern, but a full examination of this fragment has yielded what may be the oldest Megillat ʾEster fragment found to date.32 This advocates for the continued study of often neglected STa”M manuscripts, regardless of how well-known their contents might be.
Footnotes
1 For what might have been accepted as authoritative scripture at that time see Lim, Timothy H., Authoritative Scriptures and the Dead Sea Scrolls from the Oxford Handbook of the Dead Sea Scrolls, OUP, 2010, pp. 303–322. Also see b. Megillah 7a, and for a brief discussion over the status of Esther as a sacred text, see Michaels, Marc, ממקום אחר “From another place”. Finding God in the Megillah, from Henoch Journal, Vol. 44/1, 2022.
2 My thanks to Professor Ben Outhwaite and Dr Melonie Schmierer-Lee for the opportunity to view and date the scroll, and to Professor Outhwaite for his comments and suggestions. This FOTM is a shortened version of a fuller piece, Cambridge University Library, T-S A36.18—a fragment from the oldest known Megillat ʾEster in Comparative Oriental Manuscript Studies (COMSt) Bulletin, Issue 10, 2024, dedicated to the memory of the late Professor Malachi Beit-Arié.
3 Edited by Dr Nick Posegay and Dr Melonie Schmierer-Lee, published by the Cambridge Genizah Unit, 2024.
4 On p. 43 where it sits aside CUL T-S AS 16.166, which I dated as likely being from the 13th century (perhaps early 14th). This contains a very early example (possibly the earliest) of illustrative borders on a megillah, done by a relative amateur (likely the scribe himself), as opposed to the much more elaborate illuminations that followed in subsequent centuries. Decorating megillot does not really become established till at least the 1500s. However, given this Genizah megillah, perhaps this was happening, to an extent, prior to that. The fleur de lis-like lily imagery in the margins is considered by some to also be a Jewish symbol and the other image is also some kind of flower, though it is not clear what it is. It may be specific flowers native to Israel, the White Lily (see Hosea 14:5) or Narcissus.
5 David, M. C., Hebrew Bible Manuscripts in the Cambridge Genizah Collections, Volume 2, CUP, p. 195.
6 Occasionally, this will be institutions’ or conservators’ concern regarding ‘desecrating’ a religious object (such as a Torah) even when it is clearly pasul (invalid) and would not be used for ritual purposes. In other cases, it is simply lack of interest in dating these. For a good example of C14 testing helping establish the age of a Torah scroll that had been previously seriously mis-catalogued, see Rendsburg, Michaels, Young, Sommerfeldt, Levchenko & Barry, A Complete Old Torah Scroll (Nicholson Ms. 37) Held by Fisher Library, University of Sydney, from Australian Biblical Review 71, 2023, pp. 15–54.
7 See, for example, Yardeni, Ada. The Book of Hebrew Script: History, Palaeography, Script Styles, Calligraphy & Design. Carta, Jerusalem, 2010, pp. 268–271.
8 With the exception of the addition of neo-tagin to later works.
9 Tigchelaar, Eibert, 8th Annual Rabbi Tann Memorial Lecture: Beautiful Bookhands and Careless Characters.
10 See the full article for this detail.
11 The term often employed for this is ‘shading’ from chiaroscuro, meaning ‘light-dark’, and refers to the balance and pattern of light and shade in a painting or drawing’ (https://www.tate.org.uk/art/art-terms/c/chiaroscuro, accessed 14/05/24). This is not an entirely useful term to describe this variant stroke weight caused by the use of the full nib or the edge of the nib.
12 Line lengths for a megillah can vary with 11, 14, 18, 21, 28, 30 and 42 being common through the centuries. For example, eleven lines allows the ten sons of the villainous Haman to be written in one ʿamud (column) without having to be enlarged. 14 lines probably represents יד (i.e. the hand of God operating behind the scenes) and 18 is likely חי (life). A standard developed for Yemenite megillot of 21 lines, and Baer’s Tiqqun ha-Sofer ve-haQore (Rödelheim 1900) is deliberately 30 lines, since he saw this as half the height of a Sefer Torah (i.e. 60 lines), which was the custom of many sofrim in his day. By this reasoning 21 lines would be a good measure nowadays since the standard accepted tiqqun for a Torah is 42 lines.
13 Technically one is only allowed to elongate certain letters such as dalet, he, lamed, mem sofit, reš and taw. However, in the past scribes were less particular with this and would elongate a number of other letters, such as ʾalef, šin etc. The key is to ensure the letter is not stretched so much that it loses its ṣuraʾ (form).
14 Pludwinski, Izzy, The Beauty of the Hebrew Letter: From Sacred Scrolls to Graffiti, Profile Editions, 2023, p. 222. There are also other artists who use the Pauker Script type featured in this volume.
15 Jonas, Gina, Hebrew Calligraphy Styles, 1996 (no page numbers).
16 Pludwinski, Izzy, Mastering Hebrew Calligraphy, The Toby Press, 2012, pp. 52–53.
17 See the full article for details.
18 Which I now refer to as neo-tagin, to distinguish this later re-interpretation from the original meaning, as is discussed further in my PhD dissertation. These are special decorations used in STa”M objects. Some letters take three שעטנ״ז ג״ץ, some have one, בדק חיה and some none, מלאכת סופר (work of the scribe).
19 Dealt with in detail in my PhD dissertation (forthcoming).
20 The rule muqaf qewil where every letter must be surrounded by parchment and not touch one another is derived from b. Menaḥot 29b, אמר רב כל אות שאין גויל מוקף לה מארבע רוחותיה פסולה (Rav said any letter that is not surrounded by [blank] parchment on all four of its sides, is invalid). Additionally, there it dictates that ובין אות לאות כמלא חוט השערה ([the space] between [one] letter and [and the next letter, is equal [lit. like] to a full hairbreadth). For a short but pertinent discussion of this key scribal concept, see קצת כללי מוקף גויל מבעל פמ”ג (Some of the Rules Concerning the Need for ‘Surrounding by Parchment’ from the Author of the Peri Megidim), p. 309 of the Feldheim Mishnah Berurah on Hilkhot Tefillin, Pisgah Foundation, 1992, where it is noted that this rule is derived from מנחות דף כ”ט לפיר”ת דפסקו כוותיה רוב הפוסקים (b. Menaḥot 29b according to the explanation of Rabbeynu Tam, whose view is accepted as the halakhic ruling by most authorities), which suggests it may not have been actively ‘enforced’ by all until at least the 12th century, well after our megillah was written.
21 Menaḥot 30a also explains that ובין תיבה לתיבה כמלא אות קטנה ([the space] between [one] word and [the next] word is equal to [lit. like] a full small letter), i.e. a yod.
22 My thanks to the Genizah debate participants, Prof. Ben Outhwaite, Dr Kim Phillips, Rabbi Maurice Michaels and Mrs Eileen Michaels, that took place on Halloween 2024. This somewhat more speculative element is not in the full article.
23 See Qeset Ha-Sofer 16:6 and also 28:8 where there is a suggestion to be lenient bediʿeved (after the event). Shulḥan ʿArukh ʾOreḥ Ḥayyim 691:9 says similarly, that you should not do it, but after the event and if it is your only available megillah, then it is better than nothing. The GRA in his commentary to this says it is forbidden, but some modern leniencies suggest you can write the trope and vowels in pencil if you require it to read.
24 Yardeni, Ada. The Book of Hebrew Script: History, Palaeography, Script Styles, Calligraphy & Design. Carta, Jerusalem, 2010, pp. 214–215.
25 Ibid., pp. 216–218.
26 Beit-Arié, M., (in collaboration with Engel E. and Yardeni, A.), אסופות כתבים עבריים מימי-הביניים כרך א: כתב מזרחי וכתב תימני (Specimens of Medieval Hebrew Scripts, Part 1; Oriental and Yemenite Scripts), The Israel Academy of Sciences and Humanities, Jerusalem, 1987 (Hebrew).
27 Oxford Hebrew Manuscript Studies Workshop, June–July 2022, Olszowy-Schlanger, Judith, lecture on the Oriental Scripts.
28 For full images see https://www.nli.org.il/en/discover/manuscripts/hebrew-manuscripts/viewerpage?vid=MANUSCRIPTS#d=[[PNX_MANUSCRIPTS990001099370205171-1,FL213724124]] accessed 14/05/24. There are documented joins to this manuscript with those in Cambridge University Library: T-S C2.184 and Mosseri VIII.440.1–2.
29 For the full image and description see https://cudl.lib.cam.ac.uk/view/MS-TS-00024-00035/1, accessed 14/05/24.
30 For example, A search on CUDL for geṭ yields some 104 results in the Cambridge part of the Genizah. 85 of these are complete or partial giṭṭin, dated between 1024 and 1399, most of which are from the 1100s, which I used as a corpus for a Fragment of the Month, Geṭ Groundwork from the Cairo Genizah: practising writing a Jewish divorce document (T-S 10J2.34), April 2024, https://www.lib.cam.ac.uk/collections/departments/taylor-schechter-genizah-research-unit/fragment-month/fotm-2024/fragment-2
31 ‘Looped stitching’ was a largely replaced by ‘blind stitching’ in the mid-1800s.
32 It also has some interesting consonantal variants, as detailed in the full article.
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/