FOTM ‘Cursive, foiled again’: Mezuzot CUL T-S A44.36 and CUL T-S AS 11.52
Figs. 1 and 2: The front and reverse of T-S A44.36
As one might expect, given the main purpose of a Genizah is to prevent the destruction of God’s Name,1 there is a fair corpus of מזוזות mezuzot (pl.) within the fragments.2 This could be anticipated, since mezuzot on an external door can be damaged through poor weather. Indeed, הלכה halakha (Jewish law) dictates that you check your mezuzot for damage twice every 7 years.3
In terms of mezuzot, most of the Genizah corpus is relatively unremarkable, since the vast majority look little different to those in use today, though scribes back then were less concerned to fully justify the left-hand side, and thus the writing was a little ragged.
A מזוזה mezuzah (sing.) comprises two passages of Torah: שמע the Shema (Deut. 6:4-9) and והיה אם שמוע vehaya ʾim šamoaʿ (‘and if you listen’, Deut. 11:13–21),4 written on ruled parchment in כתב אשורית Ketav ʾAšurit (Jewish square script),5 with תגין tagin (decorative crowns),6 in one column, on one sheet. These are the two sections that make reference to the specific commandment of mezuzah, וּכְתַבְתָּ֛ם עַל־מְזוּזֹ֥ת בֵּיתֶ֖ךָ וּבִשְׁעָרֶֽיךָ (‘and you shall write them on the doorposts of your home and your gates’, Deut. 6:9). Originally, the word mezuzah referred to the doorpost itself not the parchment, and indeed mezuzot parchments would also be stored in hollowed-out parts of the doorpost, not affixed in a container. A strong custom exists to make a mezuzah twenty-two lines long, and even defines what word each line should begin with.7 As in the Torah, the dalet of the word אֶחָֽד ʾeḥad (‘one’) is made larger than the other letters. Indeed, the ruling is that it must be four times the size of a normal dalet (though usually this is not the case since mezuzot script is small). It is large to ensure that the word is not read as אחר ʾaḥer, meaning (Heaven forbid) that God is ‘another’.8 In addition, the custom is also to enlarge the letter ʿayin in שְׁמַ֖ע šemaʿ (‘hear’), so that it is not read as שמא šemaʾ (perhaps),9 and that both large letters together read as עד ʿed (witness) to complement the aural idea of hearing. The gap that is between the two sections is technically a סתומה setumah (closed section),10 and the second section begins on line seven. Additionally, one should write the last two words עַל־הָאָֽרֶץ (‘on the earth’, Deut. 11.21) at the start of the last line and not write any more after it.11
Those in the Genizah follow these customs, with few exceptions. The most noticeable, and commented on, is the presence of extra Tetragrammatons, additional verses,12 names of angels, and Magen David symbols, such as those present on L-G Bible 3.47 and T-S NS 290.58.13 Gideon Bohak has undertaken an in-depth investigation of the marginal תוספות מאגיות (magical additions) in these, as well as T-S AS 142.245,14 which, in his view, serve ‘to enhance their apotropaic powers’.15 This was apparently a common enough practice for Rambam to decry the custom, and suggest such scribes would not merit a place in the world to come.16 Even so, with their extra elements the basic structure of a mezuzah is still followed, and the additions purely marginal.
However, there are two other mezuzot in the Genizah that do not have these extra elements, and yet they are still noticeably different.
Our Fragments of the Month (and yes, I am cheating by having two) are T-S A44.36 and T-S AS 11.52. They too follow the basic structure of a mezuzah, but with one significant difference, they are written in a cursive script, instead of the more formal book-hand square script used for סת״ם STa"M items.17 This is odd, given the sages split the word וּכְתַבְתָּ֛ם (‘and you will write them’, Deut. 6:7) midrashically to read וכתב תם (and you will write perfectly), in both b. Shabbat 103b and b. Menaḥot 34a, and all that implies, in terms of following the rules. Additionally, the minor tractate, Massekhet Mezuzah 1:3 notes that או ששנה אותיותיה (or if one altered [the shape] of any of its letters) then it is פסול pasul (invalid) for use.
Mezuzah 1: T-S A44.36
To be fair, the scribe of T-S A44.36 started well, and on the ruled parchment has managed three competent lines from שְׁמַ֖ע יִשְׂרָאֵ֑ל (‘Hear O Israel’, Deut. 6:4) up to עַל־לְבָבֶֽךָ (‘upon your hearts’, end of Deut. 6:6) in the required book-hand. However, he ran out of room to fit in the particle אֵ֖ת (Deut. 6:5) on the very first line and squeezed it in. He then struggled to fit וְשִׁנַּנְתָּ֣ם (‘and you will teach [lit. repeat] them [to your children]’, Deut. 6:7) into the third line line, and with the aura of competence ebbing away, subsequently lapsed into a more semi-cursive script for the rest of the text. This produced an even narrower column, and his work was certainly not fully justified to the left margin, leaving it with a very ragged appearance. Even in the cursive text he had to squeeze in בְּעִתּ֖וֹ (‘in its season, Deut. 11:14). He did, at least, stick to twenty-two lines, and ending with the two words, עַל־הָאָֽרֶץ, on a separate line. He has also rigidly adhered to Rambam’s line breaks, beginning each line with the words we have as standard to this day.18
That said, there are a handful of occasions with minor deviations from the agreed mezuzah orthography:
- MT: בְּבֵיתֶ֙ךָ֙ (‘in your home’); T-S A44.36: בבתיך – Deut. 6:7, with yod and taw switched.
- MT: מְזוּזֹ֥ת (‘doorposts’); T-S A44.36: מזוזת – Deut. 6:9. Whilst this corresponds to the MT as set out in the Leningrad codex, it is not how the consonantal text of a mezuzah is written, which is מזזות.
- MT: לְטוֹטָפֹ֖ת (‘as reminders’); T-S A44.36: לטטפות – Deut. 11:18. Placing the waw as the penultimate letter.
Additionally, instead of the normal more square shape, this mezuzah is long and rectangular, 10.1cm by 21.8cm. Moreover, he has not provided us with a large ʿayin or dalet, nor do any of his letters have tagin, even though Rambam specifies certain letters (five in the first paragraph and six in the second) that must have them in Mishneh Torah Hilkhot Mezuzah 5:2–3, as a minimum (though, nonetheless, does not invalidate a mezuzah that does not have them). These twin omissions are potentially significant for dating.
That our scribe was originally attempting to write a mezuzah in formal script is witnessed by the סרטוט sirṭuṭ (ruled lines) that he adheres to for the square script lines, with the letters on the line. Indeed, the holes on each side of the mezuzah to act as a guide are particularly prominent. However, when he switches from this square script the ruled lines are abandoned completely, and letters are written without any reference to the lines, which often go directly through the words. Indeed, this presents a further halakhic problem, since no more than three letters (a later expansion on the ruled line requirement) may be written in a Sefer Torah or mezuzah without the presence of lines, and only tefillin have a leniency in this regard.19 However, also significant for an earlier dating is the shape of this mezuzah. Unlike the later more squared shapes, even if he had stuck to the formal book-hand, the sirṭuṭ show that this was always intended to be quite a long narrow mezuzah.
The combination of too narrow a shape for the script size, together with the orthographic errors and script switching, most likely may be evidence of a scribe in training, though perhaps not faring so well. Alternatively, if this is a trained scribe, he is lapsing into poor practice. Perhaps adhering to the formal script and all the rules for writing a mezuzah was just too much for him that day. Or he was in a hurry to complete the task. Either way, it was probably not his best work, and he was clearly not giving the task the full כונה kavanna (spiritual intention) it deserved
More importantly, having two scripts done by the same scribe allows us to look at his formal and informal writing in close proximity, and to perhaps better date the scroll. The script is quite damaged, but drawn abecedaries below show the likely original forms.
Fig. 3: A simple abecedary for the formal square script for T-S A44.36, used by our scribe at the start of his mezuzah. Unfortunately, not all letters are present in this short section
Fig. 3 shows a simple abecedary of the square script; there are variants of the forms, but these are typical. That said, we only have four lines to play with and not all the letters are present. Notably, the he is consistently joined with no space and whilst this persisted in codices, till quite late, this was not permitted in STa"M articles. In the scripts shown in Specimens of Medieval Hebrew Script, vol. 1,20 from the early 12th century onwards the right leg of he is more consistently separated, even in these non-holy works.
The lack of even simple line tagin in a formal STa"M book-hand also suggests an earlier date (see discussion below, regarding our second mezuzah), as does the lack of the large letters ʿayin and dalet. The lamed is very sharply angled with a triangular flag. The yod is generally very long too and could easily be confused with a waw. The base of the mem sofit is quite curved. The pe has a more angled head. Based on dated script comparators, together with these the other features mentioned, mid 11th to early 12th century seems likely. Similar dated comparator scripts, from Specimens include Script 19 (NLR Firkovitch II B59, dated 1021/2), Script 35 (NLI Heb 8O 2238, dated 1106/7) and Script 36 (T-S K6.78, dated 1116).21
Fig. 4: A simple abecedary for the cursive element of T-S A44.36.
Among the letters of T-S A44.36, the lamed is noticeably cursive, as is the ‘U’ shape at the top of the ṣade sofit, both of which show the economy of this style of writing. The left leg of the he is always separate; the yod selected here is more obviously a yod, but still long, though we also see short straight lines. The mem has a very short base whilst mem sofit is quite square and has a point emanating from the lower left corner. The nun is a hook with no roš (even in the cursive the nun tended more towards a flatter backwards ‘L’ from the late 11th century), whilst samekh has quite a prominent thick stroke on the left of the gag (roof). The pe and pe sofit has a more pronounced nose curving inwards than many of the cursive scripts where it is often just a straight line down (again, see below). ʿAyin has a very short base protrusion. The qof is joined, generally straight, but one shows a tendency to semi-ligature. The šin is a very simple monoline with the short middle branch coming out of the left branch, also quite ‘U’ shaped. Overall the writing is still attempting a consistency, though more ragged than the formal script the scribe began writing in. With the exception of the ʾalef (which is more ‘K’ shaped – see below) and the more pronounced approach stroke on the dalet, he and taw, Script 54, T-S 10J9.25 (dated 1039) generally has a quite similar feel. The hooked nun and spike on the mem sofit are also present in Script 55, T-S 13J1.11 (dated 1045) but very little else is similar. However, some of these forms do persist, such as the short based ʿayin and a spiked mem sofit, the more ‘U’ shaped šin, in Script 82 Oxford, BL Poc. 256. Overall, Script 94 Berlin, SB Ms Or. 8O 256 (dated 1341) is also quite similar.
The cursive forms do not give us the clearest steer, but in combination with the lack of neo-tagin, large letters ʿayin and, and the four lines of formal script, mid 11th to early 12th century seems feasible for T-S A44.36.
The reverse of this mezuzah is also of great interest. Normally in the reverse of the space before the second paragraph is the word שדי Shaddai, meaning ‘Almighty’. Rambam seemed sanguine about this custom, since he recognises it and permits it, because it is on the outside.22 The letters have come to stand for שומר דלתות ישראל šomer delatot Yisrael (guardian of the doors of Israel) but this is unlikely to be its real origin.23 On the reverse of a mezuzah directly behind the יְהוָ֥ה אֱלֹהֵ֖ינוּ יְהוָ֥ה (‘The LORD is our God, the LORD’, Deut. 6:4) of the first line is the apparently meaningless text, כוזו במוכסז כוזו kuzo bemokhsez kuzo written upside down. This is a variant of a Caesar Cipher,24 where the letter is replaced by the next letter along in the Hebrew alphabet (כ for י, etc), and appears to have some kabbalistic symbolism, perhaps to do with God in motion, i.e. a continuous presence. Rambam does not mention it (I suspect that it had not yet become a prominent custom), but presumably, because it is on the back, he would not have had any objection to this custom either. This too became standard practice.25
Fig. 5: Details from the front and reverse of a much later Genizah mezuzah T-S A44.23 showing the large ʿayin and dalet, Shaddai and kuzo.
On our partly cursive mezuzah, T-S A44.36, we see neither of these two customs. It may be that our scribe was himself even stricter than Rambam about writing nothing additional. Equally likely, having abandoned the formal script, the mezuzah was considered pasul and disqualified for use, and he did not complete it with these additional customs. However, even though it had many instances of God’s Holy Names on it, it does not appear to have been consigned to the Genizah immediately, and the back of a mezuzah has been used for scrap, and possibly handwriting practice.
All three additions were likely done at different times, and certainly the two Hebrew hands are different people. The Arabic lettering could be from the same hand as the first addition, but also a third individual. So potentially three people (one likely a youngster) on three separate occasions decided to write something on back of a partly cursive mezuzah. An interesting afterlife journey before it was consigned by someone (else?) to the Genizah. There are also a couple of few free-floating letters, an upside down ʾalef and nun. The main additions are transcribed and discussed below.
In a slightly different, later and more competent script to our original scribe, we have the words מזוזה/ בשביעותיכם or מזוזה/ בשבועותיכם.27 Unfortunately, there is damage in the second word, and we may have a yod or a curtailed waw. Having seen it in reality, and examined it with my Dino-Lite electronic manuscript, it is clearly בשבועותיכם. This makes little sense translating literally as ‘mezuzah/ in your weeks’ or ‘mezuzah/ in your oaths’. It is possible, being on two short lines, they are not meant to be read together, and we have mezuzah on one line identifying the item, and then the word from Num. 28:26, concerning ‘the first fruits’. Though why this word would be randomly added is odd indeed.
Fig. 6: Infrared of the first waw (left) which does have an odd kink to it at the base, unlike the second waw (centre) but is ink. On the right there is a close up of the dotted v-shaped symbol above the Arabic
The catalogue also notes ‘three words of Hebrew in Arabic script with Tiberian pointing: بشم / ونصليح نعسه = בשם / ונצליח נעשה’. The three words are written in Arabic letters, though with Hebrew letters and vowels – a sort of inverse of Judaeo-Arabic where Arabic is written in Hebrew letters. The phrase is pointed slightly oddly as בְּשֶׁם וְנַצְלִיחַ נַעֲשֶהֿ, though it is difficult to be sure of the vocalisation under the first word since there are a number of random dots on the parchment there. There seems to be an ihmāl sign (absence of dotting)28 under the ح, ḥāʾ, to distinguish it from the ḵāʾ and jīm. That also suggests we should understand the dotted v-shape above the س as an ihmāl sign too (to distinguish sīn from šīn).29 It is possible that the Arabic dots were omitted deliberately as they would confusing alongside the Hebrew נקוד niqqud (pointing). The latter dotted v-shaped symbol could be read as a mark of quotation, but we would then expect that to be marked over every word.
That said, it is indeed a quotation, but an inaccurate rendering of one, since the rabbinic phrase should be, בשם השם נעשה ונצליח (in the name of God we will do, and we will succeed).30 This is not a Tanakh quote, or a statement from the sages. It is simply a saying, the source of which seems obscure, but which is used by scribes or authors at the beginning of their works. See the beginning of Saʿadya’s Tafsīr, or the opening to a selection of piyyuṭim by Joseph al-Baradānī in T-S NS 235.120 .
Fig. 7: Detail from the beginning of Joseph al-Baradānī’s piyyuṭim in T-S NS 235.120.
Likely this misquote was merely a writing exercise. However, one might speculate that this might be the original scribe saying he will do better next time!
Most definitely writing practice, we then have a much simpler hand (possibly an older student), writing without vowels, the educational pangram phrase אַתָּה גֹחִי צוּר מִבֶּטֶן כָּל זֹעֵם סַף קָדְשָׁךְ נַפֵּץ (lit. you drew me [O] Rock from the belly, of every fury the threshold of Your temple explodes), with the first part of the phrase coming from Psalm 22:10. This reads, כִּֽי־אַתָּ֣ה גֹחִ֣י מִבָּ֑טֶן (‘for You drew me from the womb’).31 The phrase on the back of our mezuzah has been constructed to have only the 27 letters of the full alphabet. This is to practise, in a single phrase, all of the letters,32 in much the same way as the common English pangram ‘the quick brown fox jumped over the lazy dog’. There are only a handful of verses in Tanakh that have each letter appearing, but these also have additional letters repeated, so the concept is somewhat submerged. Constructing a meaningful pangram without repeating letters is difficult.33 The presence of this phrase reinforces how every scrap of paper or parchment was used, and re-used, in the pursuit of education, but also more evidence that this mezuzah was likely never actually used on a Jewish doorpost.
Mezuzah 2: T-S AS 11.52
Our other mezuzah T-S AS 11.52 is written in a slightly squarer (11.3cm x 14.1cm), twenty-two line format, though with a large margin at the base, and in an even more cursive, informal script, and in this case, throughout the text.
The קלף kelaf (parchment) has been ruled with sirṭuṭ. However, the writing does not have a consistent baseline, and is somewhat suspended between the two lines. There is absolutely no attempt at justification, and there are also no large letters ʾayin or dalet. There is nothing at all written on the verso. It sticks rigidly to Rambam’s line breaks and (damaged areas aside) the text matches the expected orthography. Moreover, it does have tagin,34 though these are simple lines (rather than ‘balls on sticks’) that often hover above the letters that they are supposed to adorn, rather than being affixed to them (which we do see in later medieval manuscripts). This is likely because there are no real heads of the letters to attach to.35 They also often join to each other, which halakha also does not permit. Microscopy shows they are an integral part of the original mezuzah, and have not been added subsequently.
One other discovery using the microscope is that unlike many of the Genizah documents I have examined, where the ink is iron-gall and fades (or in some cases completely vanishes) under the infrared setting, this script stayed black, suggesting it is carbon based (or primarily so). As Zina Cohen notes, ‘in the case of religious documents, a tendency within the Palestinian community to use parchment and iron-gall inks is contrasted to the use of leather but with any type of ink by the Babylonian community’.36 Indeed, referencing her table 5.3 suggests that in Babylonia carbon ink (or an admixture) was more likely for religious documents (though not legal ones).37 This might hint at a Babylonian origin for our fully cursive mezuzah.
Fig. 8: A close up of T-S AS 11.52 showing the free-floating tagin are in the same ink as the original consonants, and the strong amount of carbon in the ink showing up under infrared.
Again there are slight variants to the forms used in the mezuzah, but a typical simple abecedary has been provided below. However, since the tagin are long and thick, they somewhat interfere with the letter shapes. Later halakha would warn against this.38 Additionally, the letters themselves are not always of a consistent size.
This time we have a ‘K’ shaped ʾalef form present. The pe has a few allographs, but generally has a short line as the nose. The samekh is merely a circle which is a particularly useful pointer. The gagim are even more pronounced curves caused by the thick short curved stroke down joined to the sharp right leg (see bet, dalet, he, kaf, reš and taw). Judith Olszowy-Schlanger refers to this as a ‘concave aspect of the upper horizontal strokes’.39 The mem base is very short and the mem sofit very rounded. The nun base is flatter as a reversed ‘L’. The qof is separated. The šin, and the very cursive ṣade sofit are similar to our earlier mezuzah.
Fig. 9: A simple abecedary for T-S AS 11.52.
The lack of an enlarged ʿayin and dalet could similarly suggest an early date, but here, may just be because the whole work is in a cursive hand, so I am less inclined to use this as a marker.
In terms of manuscripts, the ‘K’ shaped ʾalef can be seen as early as 960 in Mosseri VIII.479.9 and 962 Oxford, BL MS Heb. b.12, folio 25 in two letters from Nehemia Gaon, and an early deployment of both this concave roof/harsh angled right leg, together with the ‘K’ ʾalef can be seen in BL Or. 5538 (Samuel ben Hofni Gaon, dated 1013).40 However overall, quite similar dated scripts are seen in Specimens of Hebrew Script Vol. 1. Script 67, JNM III.1 (dated 1136) and Script 76, Oxford BL Poc. 9, (dated 1210) and Script 82, Oxford, BL Poc. 256 (dated 1262), all having a prominent concave roof/harsh angled right leg, and the ‘K’ shaped ʾalef. The former also having a closer pe sofit, and the second having a close pe form. None of these three, have the completely round samekh, nor quite such a round mem sofit. Again, we see this circle samekh early, such as Script 52, Oxford BL MS Heb. b.12 (dated 1021), but it seems is more commonly applied in several scripts from Script 60, Oxford, BL Heb. c.28 (1093), through to Script 81, T-S 8K22.12 (dated 1253, which also has a similar lamed and ṣade sofit). It is not present in the Specimens dated examples thereafter. This gives quite a large window.
However, my current research activity suggests that the creation of these lines (later to become ‘balls on sticks’), which I call neo-tagin, occurred perhaps some time between 1060–1175. It appears to have been initially to facilitate the reintroduction of the Sefer Tagin tradition, and only by the mid-13th century to adhere to the ruling in b. Menaḥot 29b, regarding שעטנ[ן]ז ג[צ]ץ, known by the mnemonic Šeʿatnez Gaṣ, even though both of these sources are referring to the original tagin that were protrusions from the monoline.41
Comparator scripts, together with the consistent application of neo-tagin on Šeʿatnez Gaṣ suggest that T-S AS 11.52 is a little later than our first partly cursive mezuzah, and instead is the latter half of the 12th century to the mid-13th century.
Fig. 10: T-S AS 11.52. The reverse is blank.
Fulfilling the miṣva?
One must ask, why were such mezuzot written when the strong convention was to write using the square script. Were these used as a mezuzah? Would a mezuzah written in a partly cursive or fully cursive script have been considered כשר kašer (valid) for use at that time? That they were written on ruled parchment in the standard 22-line format might suggest that the writers thought they were. Perhaps the writers were not professional scribes but were in urgent need of a mezuzah for their own homes, and this was the fastest way of creating one. One clue as to whether they were used is sadly denied us. Unfortunately, because the scrolls would have been flattened for conservation, we have no idea of knowing whether they were found rolled as a mezuzah should be.42
As with Sifrei Torah, many collections often know very little about their STa”M articles partly because Sifrei Torah, mezuzot, tefillin, megillot etc. are seen as somewhat ‘samey’. Yet, each of these manuscripts has a story to tell. These two cursive mezuzot bring their own puzzles. Why were they written as they are, and were they ever used on the doorpost of a home to fulfil the מצוה miṣva (commandment)? Or, like the villains in so many Victorian melodramas, was the attempt at adhering to strict scribal halakha rendered pasul, foiled by their cursive nature?
Footnotes
1 See Fragment of the Month: The Torah of St Filippo? Taylor-Schechter Genizah Research Unit, June 2025, https://www.lib.cam.ac.uk/collections/departments/taylor-schechter-genizah-research-unit/fragment-month/fotm-2025/fragment-4
2 A large proportion of those in the Cambridge collection, are to be found in the T-S A44 series.
3 See b. Yoma 11a, Shulḥan ʿArukh Yoreh Deʿah 291:1. For a very comprehensive, more popular guide to the laws and purpose of the Mezuzah, see Hoffman, Yair, ספר זכרון שושנה Mezuzah: A Comprehensive Guide, Israel Bookshop, Lakewood NJ, 2002.
4 See Mishneh Torah Hilkhot Mezuzah 5:1.
5 See Mishneh Torah Hilkhot Tefillin 1:19.
6 Though the letters to be decorated do vary according to different authorities with both minimal and maximal (Sefer Tagin-like) iterations.
7 See Mishneh Torah Hilkhot Mezuzah 5:5. However, b. Menaḥot 31b mentions that a differently arranged mezuzah (even two words per line) would not be invalid. Indeed, it records the word אֶחָֽד (‘one’) as being the end of the first line, and not the particle אֵ֖ת. Tosafot to b. Menaḥot 32a (הא) speaks of a 25-line standard, and if it differs from that it is not invalid. Schecter, Aryeh & Aurbach in Pitḥei Šeʿarim he-Ḥadaš, Israel, 1990, p. 249 provide a mock up of a 25-line mezuzah, though I have never seen one in reality. All of the many Genizah mezuzot that I have viewed, comprise 22 lines.
8 The corresponding large reš is in אַחֵ֑ר (‘other’ (Ex. 34:14)), referring to not worshipping other, so-called gods. For a likely origin of this tradition, which likely referred to larger original tagin (protrusions from the monoline see p. 143 of Michaels, Marc, Sefer Tagin—Olde Face, Slab-Serifs and Semi-Ligatures: Covenantal Warnings and Reassurance. Re-housing the Temple Within the Torah, (Unpublished PhD dissertation; University of Cambridge, 2024) – https://doi.org/10.17863/CAM.117276.
9 In modern Hebrew, the more guttural throat sound of an ʿayin has all but vanished.
10 Though any break is theoretically allowed as they are not contiguous sections of the Torah, see b. Menaḥot 32a.
11 Though again, the Talmud in b. Menaḥot 31b reports that some say it should be at the start of the line (so that the words הַשָּׁמַ֖יִם (‘the heaven’ (Deut. 11:21)) at the end of the penultimate line, and הָאָֽרֶץ are furthest apart). Some say it should be at the end (i.e. left-ranged) and there is also a variant that suggests it should be centred. Additionally, there are later traditions to include a פ' מלופפת pe meluppefet (spiral pe) in פֶּ֥ן (‘lest’ (Deut. 11.16)) according to the Meiri (and echoing Sefer Tagin) and a ח' משוך ḥet mašukh (spread ḥet) in וְחָרָ֨ה (‘and [God ’s anger] flare’ (Deut. 11:17), and also to make the last letter, the ṣade sofit in הָאָֽרֶץ, a large letter (again according to the Meiri). These do become common. See Pitḥei Šeʿarim he-Ḥadaš, op. cit., p. 272. For more on the spiral pe and spread ḥet see Michaels, Sefer Tagin..., op. cit., pp. 84–90 and 105–107. Also, for the ḥet, “Getting the Hump: the Letter Ḥet in Holy Writings” in Review of Rabbinic Judaism 28:2 (2025), pp. 127–163, https://doi.org/10.1163/15700704-20250020.
12 For example, Pitḥei Šeʿarim he-Ḥadaš, op. cit., p. 277, shows a mock up with the last verse of Psalm 121 added in small letters after עַל־הָאָֽרֶץ.
13 Described in Posegay, Nick & Schmierer-Lee, Melonie, The Illustrated Cairo Genizah: A Visual Tour of the Cairo Genizah Manuscripts at Cambridge University Library, Gorgias Press, 2024, pp. 116–117.
14 See Bohak, Gideon, מזוזות עם תוספות מאגיות מגניזת קהיר (Mezuzot with Magical Additions for the Cairo Genizah). pp 387–403 (Hebrew).
15 Bohak, Gideon, “The Shema and Magic in the Cairo Genizah” in Benovitz, Nancy & Dudi Mevorah, Dudi (Eds) Hear O Israel: The Magic of the Shema, Israel Museum, 2021, pp. 72–75.
16 See Mishneh Torah Hilkhot Mezuzah 5:4.
17 STa”M is an acronym for ספר תורה Sefer Torah (Torah scroll), תפילין Tefillin (phylacteries) andמזוזה Mezuzah (lit. doorpost), the main ritual objects written by a Jewish scribe. For more on these ritual objects and other scribal activities, see https://www.sofer.co.uk/activities (accessed 5/12/24).
18 See Mishneh Torah Hilkhot Mezuzah 5:5.
19 See b. Menaḥot 32b, b. Megillah 18b, y. Megillah 171d, and also Qeset ha-Sofer 3:5 for discussion over the ruled lines. One of the issues that I frequently encounter is the back projection of more recent halakhic rules and stringencies into periods where these rules had not yet been adopted by the scribal communities. This is something that one needs to be wary of. One also has to be aware that just because a certain renowned rabbi made a pronouncement in a particular work, this does not mean it was universally adopted by all scribes, across all geographies from that date. There is an interesting interplay between rabbinic rulings and scribal practice and their mutual influence on each other that has not been adequately explored. One example of this is that of the Ashkenazi humped ḥet, examined in Michaels, “Getting the Hump: The Letter Ḥet in Holy Writings”, op. cit. Emanuel Tov has produced a very useful summary of the earliest references to a number of scribal rules scattered in the Mishnah, and Jerusalem and Babylonian Talmuds. Whilst later halakhic pronouncements may have been to try and influence scribal practice, many will just be recording that practice. Of the DSS, for example, he notes that 'the description of the book-making process in rabbinic sources reflects the procedures followed in the last centuries BCE and the first centuries CE for the production of sacred and nonsacred writings', Tov, E., ‘Scribal Practices Mentioned in Rabbinic Sources Compared with the Judean Desert Scrolls’ in Bakker, A., et al. (eds.), Missing Pieces: Essays in Honour of Eibert J. C. Tigchelaar, Brill, 2025, pp. 560–78.
20 Beit Arié, Malachi, Engel, Edna and Yardeni, Ada, אסופות כתבים עבריים מימי-הביניים כרך א: כתב מזרחי וכתב תימני (Specimens of Medieval Hebrew Scripts, vol. I; Mizraḥi and Yemenite Scripts), The Israel Academy of Sciences and Humanities, Jerusalem, 1987 (Hebrew).
21 Ibid.
22 See Mishneh Torah Hilkhot Mezuzah 5:4.
23 According to Mišnat Ḥassidim. The שדי should be seen through a transparent section in the mezuzah casing, the bayit, but since most mezuzot do not do this, the letter šin is often part of the outer casing design. Sadly people do tend to put a lot more thought into the casing than the actual document inside, when that is in fact the most important part.
24 A ‘method in which each letter in the plaintext is replaced by a letter some fixed number of positions down the alphabet. The method is named after Julius Caesar, who used it in his private correspondence’, https://cryptii.com/pipes/caesar-cipher (accessed 14/11/25).
25 See the Ramah in Shulḥan ʾArukh Yoreh Deʿah 288:15. As Bohak, “The Shema and Magic in the Cairo Genizah”, op. cit., 74 notes, ‘the addition of kwzw bmwksz kwzw aroused few objections in the Middle Ages and continues to the present day’.
26 The catalogue entry suggests, ‘on verso are the words מזוזה בשביעותיכם’, Davis, Malcolm C.; Knopf, Henry; Hebrew Bible Manuscripts in the Cambridge Genizah Collections, Cambridge University Press, 1978, p. 255.
27 Witkam notes that, ‘Ihmāl —meaning “neglect ”—is the technical term in Arabic grammar for omitting the dots customary on certain consonants of the Arabic alphabet. It also refers to the use of whole range of signs (the ʿAlāmāt al- Ihmāl) whose role is to indicate that certain letters of the Arabic alphabet do not carry dots. See Witkam, Jan Just, “The Neglect Neglected. To Point or Not to Point. That is the Question”, Journal of Islamic Manuscripts 6 (2015), pp. 376–408. See also Yusuf Umrethwala’s Fragment of the Month (FOTM) for September 2025, Al-Qāḍī al-Nuʿmān’s Legal Legacy in the Cairo Geniza, where he notes the scribe's use of ‘ihmāl on certain letters, especially hamza, ḥāʾ, and ʿayn ’ – his fig. 12 shows the ihmāl marking of “صـ” below the letter “ṣād”, https://www.lib.cam.ac.uk/collections/departments/taylor-schechter-genizah-research-unit/fragment-month/fotm-2025/fragment-6.
28 My thanks to Dr Estara Arrant for her assistance with the Arabic, when I visited the Genizah Unit on 13/11/25 to view the mezuzot. Also, to Dr Nick Posegay for suggesting ihmāl. My thanks, as always, to Prof. Ben Outhwaite for his assistance with Arabic too, and other helpful comments on this FOTM.
29 See https://kipa.co.il/שאל-את-הרב/מקור-הביטוי-בשם-השם-נעשה-ונצליח and https://www.hidabroot.org/question/226812 (both accessed 1/8/25).
30 Though the Jewish Publication Society translation recognises that the Hebrew is difficult to translate. According to the HALOT, the root גחה means ‘to uproot’ or ‘to pull out’.
31 It also appears on the back of a ‘memorandum for opticians’ on T-S Ar.35.212.
32 Another popular one used in copying exercises is הָקֵץ עָצֵל דַּיָּךְ מִנּוּם גָּרֵשׁ כָּזָב פֶּן תֹּסֶף חֵטְא (lit. in the end, you are lazy, because you are a false exile, because you will add to sin). For more on these pangrams see Michaels, Marc, Manuscripts from the Cairo Genizah as Educative Tools, Manuscript and Text Cultures Journal (2025), pp. 59–83, https://doi.org/10.56004/v3.1mm.
33 Noted as tagim in the catalogue reference, Davis, Malcolm C.; Knopf, Henry; Outhwaite, Ben, Hebrew Bible Manuscripts in the Cambridge Genizah Collections III. Cambridge University Press, 2003, p. 127.
34 There are later examples of these ‘free-floating’ or ‘flying’ neo-tagin present in manuscripts that employ Sefer Tagin, such as in manuscripts such as Parma MS 2012 (1375), Parma MS 2669 (14th century.), Vat. ebr. 437 (14th century.), Vat. ebr. 15 (14th century), Vat. ebr. 1 (likely mid-15th to early 16th century) and BL Or. 1465 (16th century).
35 Mishnah Berurah on Hilkhot Tefillin 36:3.
36 Cohen, Zina, Composition Analysis of Writing Materials in Cairo Genizah Documents: Cambridge Genizah Studies Series, Volume 15, Brill, 2021, p. 167.
37 Ibid., p. 152.
38 Mishnah Berurah on Hilkhot Tefillin, 36:3.
39 Olszowy-Schlanger, Judith, Oriental Scripts, Oxford Hebrew Manuscripts Workshop 2022.
40 Ibid.
41 For more details on the original tagin and neo-tagin see Michaels, Sefer Tagin, op. cit.
42 See Mishneh Torah Hilkhot Mezuzah 5:6.
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