Scrinium 18 (2022) 49–84

brill.com/scri

Syriac Spells for a Mill and Their Historical Context

Anna Cherkashina | ORCID: 0000-0003-4412-4788
Beth Mardutho: The Syriac Institute, Piscataway, NJ, USA
Corresponding author
[email protected]

Dmitry Cherkashin | ORCID: 0000-0002-2387-3483

Faculty of Humanities, Institute for Oriental and Classical Studies (IOCS),
HSE University, Moscow, Russia
[email protected]

Ortal-Paz Saar | ORCID: 0000-0002-1000-0763

Department of History and Art History, Utrecht University, Utrecht,
The Netherlands
[email protected]

Abstract

This article considers a text-unit known in five Syriac codices and consisting of up
to three magical recipes. The target of all these recipes is a mill: two of them are
curses (ˀassārā ‘binding spell’) and intend to stop the mill, while the third one is a
counter-spell (šeryānā ‘loosening spell’), which aims to annul the curse. One of the two
binding spells includes a rare example of an Arabic incantation written in Garshuni.
The main purpose of this article is to make these texts available via critical editions. In
addition, light is shed on the broader context of magical practices, by drawing atten-
tion to Syriac recipes for an oven and their Jewish parallels, and by presenting two
Jewish parallels of spells related to a mill: a Judaeo-Arabic text from the Cairo Genizah
and a spell from a Byzantine manuscript. We offer a reconsideration of the interpreta-
tion of the Judaeo-Arabic text, as our reading differs from that of the Editio Priceps.

Keywords

Syriac charms – Book of Protection – binding spell – magic – Syriac-Arabic – Garshuni –

Cairo Geniza – Judaeo-Arabic – Aramaic

© Anna Cherkashina et al., 2022 | doi:10.1163/18177565-bja10062

This is an open access article distributed under the terms of the CC BY 4.0 license.
50 Cherkashina, Cherkashin and Saar

1 Introduction

Syriac charms1 collected in codices widely known as Book of Protection did

not receive much scholarly attention and as noted by A. Pearson, “the field
is still in its early stage”.2 Pearson states that in order for the field to progress,
“our first task is to publish more source material”.3 An edition of these three
recipes is thus valuable, not only to bring more source material to light, but
for three more reasons. First, although they have not yet been truly edited,
they have been interpreted incorrectly by previous scholars, H. Gollancz and
K. Fr. Krämer (see below). Second, two of the recipes belong to the category of
aggressive magic,4 which is even less studied than other types of Syriac magical
texts. Third, one of the binding spells contains an Arabic incantation written in
Syriac script. Only two Syriac recipes of this kind have been edited before the
current publication.5
Though this article focusses on Syriac recipes for the mill, attention is paid
to the broader context of related magic practices. A separate section is devoted
to Syriac recipes for the oven and their Jewish parallels, and another one to
Jewish and Coptic recipes for the mill. Such contextualization assists in better
understanding the place of magic within Syriac culture, and sheds further light
on connections between Syriac Christianity and other cultures.

2 The Manuscripts

Below we provide short catalogue descriptions of the Syriac manuscripts6 in

which the recipes for the mill are found.

1 Our credits go to Maria C. H. Cioată for proof-reading this article and Nikita Kuzin (Freie
Universität Berlin) for his valuable remarks.
2 Abigail Pearson, “Syriac Magic: an Overview of Previous Approach and Prospects for the
Future,” in: Studies in the Syriac Magical Traditions, eds. Marco Moriggi & Siam Bhayro
(Magical and Religious Literature of Late Antiquity, 9), Leiden, 2022, p. 13.
3 Ibid., p. 25.
4 Since the usage of the term ‘magic’ remains a subject of discussions, we affirm that it is being
used in this paper for the sake of convenience and without implying negative connotations.
5 See Cod. A §19 (Hermann Gollancz, The Book of Protection: Being a Collection of Charms, Now
Edited for the First Time from Syriac mss, London, 1912, p. 14) and a dream-request recently
published by Zellmann-Rohrer (Michael Zellmann-Rohrer, “A Syriac-Arabic Dream-Request
and Its Jewish Tradition,” Journal of Near Eastern Studies, 78/1 (2019), pp. 59–74.
6 Images of the relevant folios of the manuscripts at our disposal are provided in the Appendix.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 51

Houghton Syr. 1607 (H)

Harvard University (Cambridge MA), Houghton Library, ms. Syriac 160. Former
owner: I. H. Hall (1837–1896). Title: kṯīḇtā da-nṭuryā ‘Protective Amulet’.8 Paper,
49 leaves, bound, 12 × 8 cm. Up to 18 lines to page. Script: unvocalized East
Syriac. 74 chapters, 17 coloured illustrations. Date and place: 1804, the village of
Shibāni, Tergawar district (Hakkari, Turkey). Scribe: Gewargis bar Zayˁa from
Shamsdin.

BL Or. 66739 (BLa10)

London, British Library Ms. Or. 6673. Paper, 12.5 × 9.0 cm, 48 leaves. Script:
unvocalized East Syriac. Date and place: 1804, the village of Shibāni, Tergawar
district (Hakkari, Turkey). 67 chapters and illustrations. Up to 18 lines to page.
Scribe: Gewargis.11 Described by Gollancz, who collated the ms. with Cod. A,
and edited most of the additional content, which was not in Cod. A. This part
of the manuscript is available to us only via Gollancz’s edition.

7 A recent description with further references can be found in Michael Zellmann-Rohrer,

“More on the ‘Book of Protection’ and the Syriac ‘Charms’: New Texts and Perspectives for
the Study of Magic and Religion,” in: Studies in the Syriac Magical Traditions, eds. Marco
Moriggi & Siam Bhayro (Magical and Religious Literature of Late Antiquity, 9), Leiden,
2022, p. 83; and Anna Cherkashina & Alexey Lyavdansky, “Syriac Love Charms. Part I. The
Recipe-Type,” Scrinium 17 (2021), p. 85.
8 We prefer this translation instead of the common “Book of Protection”. Syriac magic
codices are very rarely entitled with the Syriac term for “book”, i.e., kṯāḇā (for which see,
e.g., StP18 in Michael Zellmann-Rohrer “More on the ‘Book of Protection’,” p. 88). Instead,
most of them are titled with ‫( ܟܬܝܒܐܬ‬so, e.g., H and I), while in NH3 we find its supposedly
orthographic variant ‫( ܟܬܒܐܬ‬NH3). Following Alexey Lyavdansky’s suggestion, we regard
these words as rendering the Neo-Aramaic terms for “amulet, talisman”: kṯəwta/kṯuṯa
(and similar), see e.g., Geoffrey Khan, The Neo-Aramaic Dialect of the Assyrian Christians
of Urmi, Vol. 3, Leiden–Boston, 2016, p. 123; Shabo Talay, Neuaramäische Texte in den
Dialekten der Khabur-Assyrer in Nordostsyrien, Wiesbaden, 2009, p. 444; Hezy Mutzafi,
The Jewish Neo-Aramaic Dialect of Betanure (Province of Dihok), Wiesbaden, 2008, p. 358.
9 Former “BM Or. 6673”. For the description of the manuscript see Hermann Gollancz,
Book of Protection, p. 93; more recently Michael Zellmann-Rohrer, “More on the ‘Book of
Protection’,” pp. 85–86.
10 Hermann Gollancz, The Book of Protection, pp. 101–102, §64.
11 Michael Zellmann-Rohrer suggests, that BLa and H can be written by the same scribe,
which seems improbable. The two manuscripts are visibly written in two different hands.

Scrinium 18 (2022) 49–84

52 Cherkashina, Cherkashin and Saar

BL Or. 528112 (Blb)

London, British Library Ms. Or. 5281. Paper, 38 × 23 cm, 146 leaves. A convo-
lute volume consisting of three manuscripts written in three different hands.
It contains magical, divinatory, and astrological sections. The magical section
has no beginning. The first two manuscripts contain magical texts (ff. 1–2 and
ff. 2–41). This part consists of 62 chapters with 1 illustration in black. Up to
16 lines to page. Script: sporadically vocalized East Syriac. Date: unknown,
paleographically dated to the 18th century. The codex also contains 47 chapters
of the Book of the Bee.

IOM 413 (I)

Russia, St. Petersburg, Institute of Oriental Manuscripts,14 Osnovnoy fond,

ms. Sir. 4. Title: kṯīḇtā d-nuṭṭār bnaynāšā ‘Amulet for protecting people’. Paper,
10.5 × 6.5 cm, 40 leaves, bound. Script: sporadically vocalized East Syriac. Up
to 14 lines on page. 46 sections, 13 colored illustrations. Colophon lacking.
Paleographically dated to the 17th–18th century by Pigulevskaya; later owner-
ship notes of Yoḥannan bar Tērō, and of Saggō bar Merzā.

NH315

New Haven, CT, Beinecke Library, Hartford Seminary collection no. 3. Title:
kṯāḇtā16 d-nuṭṭār bnaynāšā ‘Amulet for protecting people’. Paper, 11 × 7 cm, 41
leaves. Script: vocalized East Syriac. Colophon lacking; assigned to the 19th cen-
tury. 50 chapters, 14 illustrations. Up to 14 lines to page. The ms. belongs to the
same textual tradition as I but is written in a less skilled hand.

12 For the most recent and full description and further references see Michael Zellmann-
Rohrer, “More on the ‘Book of Protection’,” p. 93.
13 Described by Нина Викторовна Пигулевская, “Каталог сирийских рукописей Ленин-
града” [Nina Viktorovna Pigulevskaya, Catalogue of the Syriac Manuscripts in Leningrad],
Палестинский сборник [Palestinian Collection], 6 (69) (1960), pp. 124–128 (ms. XXXIX).
For a Russian edition see the MA thesis of Мария Чамурлиева, Сирийская заклинатель-
ная традиция на примере рукописи из рукописного фонда ИВР РАН Сир. 4 [Mariya
Chamurliieva, Syriac Magical Tradition by the Example of the Manuscript Syr. 4 from the
IOM RAS Manuscript Collection], РГГУ [RSUH], 2009. For the most recent mention see
Michael Zellmann-Rohrer, “More on the ‘Book of Protection’,” p. 88.
14 Contra Zellman-Rohrer, ibid.: “Institute of Oriental Studies”.
15 See further Michael Zellmann-Rohrer, “More on the ‘Book of Protection’,” p. 87.
16 Ms.: ktdtˀ.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 53

3 Text and Translation

Among the five manuscripts at our disposal, we have chosen H as the main
manuscript. Since NH3 follows I, but is less reliable, we do not address it sepa-
rately. The text-critical discrepancies between the four compared manuscripts
are provided in the footnotes. A synopsis of all four versions is presented in the
next section.

MS Syriac 160, Houghton Library, Harvard University, f. 44r, ll. 1–17

1 ̣ a‫ ܐܣܪܐ ܕܪܚܝܐ‬Binding of a mill. Say

‫ܐܡܪ‬
2 ̈ b‫ ܗܠܝܢ‬these words from the psalm
‫ ̣ܡܢ ܡܙܡܘܪܐ‬c‫ܦܬܓܡܐ‬
3 ‫܆‬f‫ ܘܫܕܝܬܢܝ‬e‫ ܕܐܪܝܡܬܢܝ‬d‫ ܕܕܘܝܕ‬of David: “For You have lifted me up
and cast me away”g –
4 i‫ ܒܪܟ ܥܠ ܥܦܪܐ‬h‫ ܙܒܢܝ̈ܢ‬:‫ܓ‬ three times. Blessj upon the dust
5 m‫ ܫܩܘܠ‬l‫ ܕܪܚܝܐ܀‬k‫ܘܫܕܝ ܒܥܝܢܐ‬ and throw [it] into the eye of the mill.
[Another recipe] Take
6 ‫ ܕܚܛܐ ܒܪܟ‬n‫ܫܒܥܐ ܦ̈ܪܕܐ‬ seven grains of wheat. Bless

a Here and in all the following occurrences of this word Gollancz reads ‫ܪܝܚܐ‬.
b Absent from BLa.
c These three words are absent from I. In BLb they are placed after the citation from the Psalm
ending in ‫ܘܫܕܝܬܝܢܝ‬.
d The last three words are absent from BLb.
e BLa: ‫ ܐ̈ܪܝܡܬܢܝ‬with an odd syāmē.
f BLa: ‫‘ ܘܫܪܝܬܢܝ‬and you have loosened me’, which contradicts the text of Ps. 102:11. In I after
these we find a prescription similar to that of the second spell (ll. 5–13):
ܵ ‫ܓܠܗ‬
‫ܕܪܚܝܐ܆‬ ̇ ‫ ܘܫܕܝ ܒܪ‬:‫ ܦ̈ܪܕܐ ܕܚܛܐ‬:‫ ܙ‬:‫ܫܒܥܐ ݂ܙܒ ̈ܢܝܢ ܥܠ‬
݂ ‫݂ܘܬ ܼܢܝ‬
ܸ ܼ
“And recite [it] seven times above seven grains of wheat and throw them in the lower part of
the mill”. On the meaning of reḡlā see commentary to ˁayna d-raḥyā in section 5 below. The
text of the charm in I ends here.
g Ps 102:11 according to the Peshitta numeration.
̈
h BLb: ‫ܓ܀ ܼܲܙܒܖ܀‬, the words are placed after ‫ܦܬܓܡܐ‬ .
i BLa: ‫ ܥܦ̈ܪܐ‬with an odd syāmē.
j On the meaning of this word see section 5.
̇ ‫ ܲܒ‬with the same meaning.
k BLb: ‫ܥܝܢܗ‬ ܼ
l BLa has no punctuation mark, while BLb puts a colon and adds ‫‘ ܐܚܖܢܐ‬another [charm]’ after
it.
ܲ with an unusual mark of abbreviation.
m BLb: ‫ܫܩܘ‬
ܼ prīḏē. The word does not mean ‘grain’ or ‘seed’ in Syriac, but a very close mean-
n BLb: ‫ܦ̈ܪܝ ܹܖܐ‬
ing is attested for its cognate in Jewish Babylonian Aramaic, see prydˀ ‘pomegranate seed’
in Michael Sokoloff, A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic
Periods, Ramat-Gan, 2002, p. 932.

Scrinium 18 (2022) 49–84

54 Cherkashina, Cherkashin and Saar

(cont.)

MS Syriac 160, Houghton Library, Harvard University, f. 44r, ll. 1–17

7 ̈ o‫ܥܠܝܗܘܢ ܥܠܝܗܘܢ‬
p‫ܦܬܓܡܐ‬ upon them <<upon them>> [with the
following] words
8 s‫ ܒܢܕܘܟ‬r‫ ܙܒܢܝ̈ܢ܆ ܣܢܕܘܟ‬q:‫ܙ‬ seven times: “sndwk bndwk
9 w:‫܆ ܗܕܐܠܠ‬v‫܆ ܟܠܡܘܝ‬u‫܆ ܐܓܒܪ‬t‫ܘܐܠܗ‬ and God is the greatest. These words
of mine
10 z:‫܆ ܗܠܡܘܢ‬y‫܆ ܒܗ‬x‫ܘܠܕܝ܆ ܬܟܝܠ‬ and this [word] with which was
entrusted? Solomon
11 ‫܆‬ab‫܆ ܕܘܘܕ܆ ܚܬܡܬ܆ ܒܚܬܝܡܬ‬aa‫ܐܒܢ‬ son of David. I have sealed [the mill?]
with the seal
12 af‫ ܕܦܘܗܐ‬ae‫ ̣ܡܢ ܣܐܝܬ‬ad.‫܆ ܐܪܒܥܐ‬ac‫ܐܠܓܝܢ‬ of the four jinn from the master of its
opening?”.
13 ah‫ ܒܥܝܢܐ ܕܪܚܝܐ ܘܒܛ‬ag‫ܘܐܪܡܝ‬ And throw [the grains] into the eye of
the mill, and it will stop.

o Absent from BLa, which has ‫‘ ܗܠܝܢ‬these’ instead.

p Starting from ‫ܒܪܟ‬, BLb reads instead: ‫ܘܐܡܪ ܗ ܹܠܝܢ ܸܦ ݂ܬ ݂ܓ ̈ܡܐ‬ ̣ .
q BLa: ‫‘ ܗ‬five’.
r BLa has a colon.
s BLa has a colon.
t Absent from BLb.
u BLb: ‫ܐܠܟ ݁ܒܪܐ‬ ݁ for which see section 5.
v BLa: ‫ܟܠܡܝܬ‬.
w BLb: ‫ܒܗܕܐܠܠ‬ ܵ . BLa: ‫ܗܠܠ‬.
x BLa: ‫ܬܟܠ‬.
y BLb: ‫ ܵܒ ܹܗܗ‬.
z BLa: ‫ܗܣܠܡܢ‬. BLb: ‫ ܸܣܠ ܵܡܢ‬.
ܵ
aa BLa: ‫ܐܒ‬. BLb: ‫ ܼܐܒܢ‬. ܵ ܲ
ab BLa reads the same but has no punctuation mark. BLb: ‫ܒܟܬܡܬ‬ ݂ ‫ܘܟܬܡܬ‬.
ac BLa reads the same, but Gollancz provides a note: “or ‫”ܐܐܓܝܢ‬.
ad BLa: ‫ܐܪܒܥ‬.
ae BLa reads the same but has no punctuation mark; Gollancz provides a note: “or ‫”ܣܬܝܐ‬.
BLb: ‫ ܲ ܼܣ ܹܝܕ‬.
af BLa: ‫ܘܦܘܗܐ‬. It adds:
‫ܘܐܢ ܡܨܐ ܐܢܬ ܬܢܝ ܡܬܚܐ ܕܐܪܒܥܝܢ … ܙܒܢܝ̈ܢ‬
‘And if you can, repeat [this for] the period of forty … times.’
BLb adds:
̈ ܲ ܲ ܵ ݁ ܲ ̈ ܵ ‫ܘܚܕ ܸܦ ݂ܬ ݂ܓ ܵܡܐ ݇ܐ‬
‫ܬܢܝ܀‬ ܼ ‫ ݂ܙܒܢܝܢ‬.‫ܪܒܥܝܢ‬ ܼ ‫ܚܪܢܐ ݂ܙܒ ̈ܢܐܬ‬
ܹ ‫ ܹܒܗ ܼܐܝܬܝܢ ܘܐܢ‬.‫ܣܓܝܐ ݂ܬܐ‬
ܼ ‫ܡܨܐ ܼܐܢܬ ܸܡ ݂ܬܚܐ ܼܕܐ‬
‘And [say] another word many times: “Give me this!” And if you can repeat [it] forty times.’
In BLb the text ends here.
ag BLa provides a grammatically correct form ‫ܐܪܡܐ‬.
ah Absent from BLa.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 55

(cont.)

MS Syriac 160, Houghton Library, Harvard University, f. 44r, ll. 1–17

14 ‫ ܫܪܝܢܐ ܕܪܚܝܐ ܫܪܐ ܐܣܝ̈ܪܐ‬Loosening of a mill. [Recite:] “The

Lord releases
15 ‫ ܥܕܡܐ ܕܫܠܡ ܡܙܡܘܪܐ‬ai:‫ ܡܪܝܐ‬the prisoners”aj till the psalm ends,
16 ‫ ܒܪܟ ܥܠ ܚܛܐ ܫܕܝ ܒܥܝܢܐ‬bless upon the wheat and throw it
into the eye of the
17 ‫ ܕܪܚܝܐ‬mill.
ai BLa: ‫ܐܣܝܪ ܐܡܪܝܐ‬. One can suspect a typographic mistake, rather than a scribal one.
aj Ps 146:7.

4 Synopsis of the Four Versions

In the table below, we provide the text of the four manuscripts divided into
eight verses and presented synoptically.
In what follows, we outline the most significant discrepancies between the
four manuscripts. Two of them (H and BLa) preserve three recipes: two bind-
ing spells (verses 1, 2 and 3–7) and one loosening spell (verse 8). In BLb, the text
stops at verse 7, before the end of the second recipe. In I, only one binding spell
has been preserved. This textual variant contains a few remarkable traits. First,
its beginning is similar to the other manuscripts, alluding to Psalm 102, but
what follows does not correspond to the text of the first binding spell. Instead,
it proceeds with instructions similar to the second binding spell (verses 3 and
7), prescribing to recite the psalm above 7 grains of wheat 7 times. Then the
grains are to be thrown into reḡlāh d-raḥyā ‘… of the mill’ which refers either to
the torrent of the watermill, or to some unidentified part of the mill.
Each of the other three manuscripts (H, BLb and BLa) contains scribal mis-
takes, but, as a whole, the texts of H and of BLb are more reliable, while BLa
contains mistakes even in the biblical citations. It appears that BLa has only
one considerable advantage: it provides a better reading for klmwy in H and
BLb – klmty (Arb. kalimātī ‘my words’).
The difference between the ways in which this text-unit has been segmen-
tated in the different manuscripts is significant for its interpretation: in H,
the beginning of each recipe is marked in red, while in BLa only the first and
the third spell have separate headings. At the same time, in H, there are some
other words written with red ink: barreḵ ‘bless’ (ll. 4, 16), wa-ˀrmī? ‘throw’ (l. 13,
instead of wa-ˀrmā), which do not correspond to the segmentation of the text.

Scrinium 18 (2022) 49–84

 ‫‪56‬‬ ‫‪Cherkashina, Cherkashin and Saar‬‬

‫‪Verse‬‬ ‫‪H‬‬ ‫‪BLa‬‬ ‫‪BLb‬‬ ‫‪I‬‬

‫‪number H 160 f. 44r‬‬ ‫‪BL (BM) Or. 6673‬‬ ‫‪BL Or. 5281 f.‬‬ ‫‪IOM 4‬‬
‫‪(Gollancz 1912,‬‬ ‫‪40v–41r‬‬ ‫‪f. 35v‬‬
‫)‪101–102, §64‬‬ ‫)‪(= NH3 f. 35‬‬

‫‪1‬‬ ‫ܐܡܪ ܗܠܝܢ‬

‫ܐܣܪܐ ܕܪܚܝܐ ̣‬ ‫ܕܐܪܝܡܬܢܝ ܐܣܪܐ ܕܪܝܚܐ ܐܡܪ‬ ‫ܚܝܐ ܼ‬ ‫ܐܣܪܐ ܕܪܚܝܐ‪ :‬ܕܐܪܝܡܬܢܝ ܼܲܐ ܵܣ ܵܪܐ ܕ ܹܪ ܵ‬
‫̈‬
‫ܦܬܓܡܐ ̣ܡܢ ܡܙܡܘܪܐ‬ ‫̈‬
‫ܼܲܘܫܕܝܬܢܝ‪ܸ ..‬ܐ ̣ܡܪ ܗ ܹܠܝܢ ܦܬܓܡܐ ̣ܡܢ ܡܙܡܘܪܐ‬
‫̇‬ ‫ܘܫܕܝܬܢܝ‬
‫ܕܕܘܝܕ ܕܐܪܝܡܬܢܝ‬ ‫ܕܕܘܝܕ‪ :‬ܕܐ̈ܪܝܡܬܢܝ‬ ‫̈‬
‫ܸܦ ݂ܬ ݂ܓܡܐ‪.‬‬ ‫̈‬
‫ܫܒܥܐ ݂ܙܒܢܝܢ ܥܠ‪:‬‬‫݂ܘܬ ܼܢܝ ݂‬
‫ܘܫܕܝܬܢܝ‬ ‫ܘܫܪܝܬܢܝ‬ ‫ܙ‪ :‬ܦ̈ܪܕܐ ܕܚܛܐ‪ܼ :‬‬
‫ܘܫܕܝ‬
‫ܵ‬
‫ܕܪܚܝܐ܆‬ ‫ܒܪ ̇‬
‫ܓܠܗ‬ ‫ܸ‬
‫ܲ‬
‫‪2‬‬ ‫ܓ‪ :‬ܙܒܢܝ̈ܢ ܒܪܟ ܥܠ‬ ‫ܓ܀ ܼܲܙܒܖ‪a‬܀ ܒܪܟ ܼܥܠ ܓ‪ :‬ܙܒܢܝ̈ܢ ܒܪܟ ܥܠ‬
‫ܒܥ ̇‬ ‫ܲ‬ ‫ܲ‬
‫ܥܦܪܐ ܘܫܕܝ ܒܥܝܢܐ‬ ‫ܥܦ̈ܪܐ ܘܫܕܝ ܒܥܝܢܐ‬ ‫ܝܢܗ‬ ‫ܼܥܦܪܐ ܘܫܕܝ ܼ‬
‫ܕܪܚܝܐ܀‬ ‫ܕܪܝܚܐ‬ ‫ܚܝܐ܆‬‫ܕ ܹܪ ܵ‬
‫‪3‬‬ ‫ܫܩܘܠ ܫܒܥܐ ܦ̈ܪܕܐ‬ ‫ܫܩܘܠ ܫܒܥܐ ܦ̈ܪܕܐ‬ ‫ܫܒܥܐ‬ ‫ܐܚܪܢܐ ܫܩܘ܆ ݂‬ ‫ܲ‬
‫ܕܚܛܐ ܒܪܟ ܥܠܝܗܘܢ‬ ‫ܘܐܡܪ ܕܚܛܐ ܒܪܟ ܥܠܝܗܘܢ‬ ‫̣‬ ‫ܐ‪.‬܀‬ ‫ܛ‬
‫ܦ̈ܪܝ ܹܖܐ ݂ܕܚ ܹ‬ ‫ܼ‬
‫ܥܠܝܗܘܢ ̈‬
‫ܦܬܓܡܐ ܙ‪:‬‬ ‫̈‬
‫ܗܠܝܢ ܦܬܓܡܐ‪ :‬ܗ‪:‬‬ ‫̈‬
‫ܗ ܹܠܝܢ ܸܦ ݂ܬ ݂ܓܡܐ‪..‬‬
‫ܙܒܢܝ̈ܢ‪:‬‬ ‫ܙܒܢܝܢ‬
‫‪4‬‬ ‫ܣܢܕܘܟ ܒܢܕܘܟ ܘܐܠܗ‪:‬‬ ‫ܣܢܕܘܟ‪ :‬ܒܢܕܘܟ‪ :‬ܘܐܠܗ‪:‬‬ ‫ܣܢܕܘܟ‪ܼ ܲ ..‬ܒܢܕܘܟ‪.‬‬
‫ܐܓܒܪ‪ :‬ܟܠܡܘܝ‪ :‬ܗܕܐܠܠ‪:‬‬ ‫ܐܓܒܪ‪ :‬ܟܠܡܝܬ‪ :‬ܗܠܠ‪:‬‬ ‫ܘܐܠܟ ݁ܒܪܐ ܟܠܡܘܝ‪.‬‬ ‫݁‬
‫ܘܠܕܝ‪ :‬ܬܟܝܠ‪ :‬ܒܗ‪:‬‬ ‫ܠܕܝ ܬܟܠ ܘܠܕܝ‪ :‬ܬܟܠ‪ :‬ܒܗ‪:‬‬ ‫ܲ‬ ‫ܵ‬
‫ܒܗܕܐܠܠ‪ܼ .‬ܘ ܼ‬
‫ܗܠܡܘܢ‪ :‬ܐܒܢ‪ :‬ܕܘܘܕ‪:‬‬ ‫ܗܣܠܡܢ‪ :‬ܐܒ‪ :‬ܕܘܘܕ‪:‬‬ ‫ܵ‬ ‫ܲ‬
‫ܵܒ ܹܗܗ ܸܣܠܡܢ‪ܼ .‬ܐܒܢ‬ ‫ܵ‬ ‫ܵ‬
‫ܘܘܕ‪.‬‬ ‫ܲܕ ܲ‬
‫ܼ‬
‫‪5‬‬ ‫ܚܬܡܬ‪ :‬ܒܚܬܝܡܬ‪:‬‬ ‫ܚܬܡܬ‪ :‬ܒܚܬܝܡܬ‬ ‫ܘܟܬܡܬ ݂ܒܟܬܡܬ‪.‬‬
‫ܐܠܓܝܢ‪ :‬ܐܪܒܥܐ‪̣ .‬ܡܢ‬ ‫ܐܪܒܥܐ ̣ܡܢ ܼܲܣ ܹܝܕ ܐܠܓܝܢ ܐܪܒܥ‪̣ :‬ܡܢ‬ ‫ܵ‬ ‫ܐܠܓܝܢ‪.‬‬
‫ܣܐܝܬ ܕܦܘܗܐ‬ ‫ܣܐܝܬ ܘܦܘܗܐ‪:‬‬ ‫ܕܦܘ ܵܗܐ‪.‬‬
‫ܼ‬
‫ܵ‬ ‫ܵ ݇ ܵ‬
‫‪6‬‬ ‫ܘܐܢ ܡܨܐ ܐܢܬ ܬܢܝ‬ ‫ܘܚܕ ܸܦ ݂ܬ ݂ܓܡܐ ܐܚܪܢܐ‬
‫ܡܬܚܐ ܕܐܪܒܥܝܢ …‬ ‫ܣܓܝ ̈ܐ ݂ܬܐ‪ܹ .‬ܒܗ‬ ‫݂ܙܒ ̈ܢܐܬ ܼ‬
‫ܼܲܐ ݁ܝܬ ܵܝܢ ܘܐܢ ܡܨܐ ܼܲܐܢܬ ܙܒܢܝܢ̈‬
‫ܹ‬ ‫ܲ‬
‫ܪܒܥܝܢ‪݂ .‬ܙܒ ̈ܢܝܢ‬ ‫ܸܡ ݂ܬܚܐ ܼܕܐ ܼ‬
‫ܬܢܝ܀‬ ‫ܼ‬
‫‪7‬‬ ‫ܘܐܪܡܝ ܒܥܝܢܐ ܕܪܚܝܐ‬ ‫ܐܪܡܐ ܒܥܝܢܐ ܕܪܝܚܐ‬
‫ܘܒܛܠܐ‬
‫‪8‬‬ ‫ܫܪܝܢܐ ܕܪܚܝܐ ܫܪܐ‬ ‫ܫܪܝܢܐ ܕܪܝܚܐ ܫܪܐ‬
‫ܐܣܝ̈ܪܐ ܡܪܝܐ‪ :‬ܥܕܡܐ‬ ‫ܐܣܝܪ ܐܡܪܝܐ ܥܕܡܐ‬
‫ܕܫܠܡ ܡܙܡܘܪܐ ܒܪܟ ܕܫܠܡ ܡܙܡܘܪܐ ܒܪܟ‬
‫ܥܠ ܚܛܐ ܫܕܝ ܒܥܝܢܐ ܥܠ ܚܛܐ ܫܕܝ ܒܥܝܢܐ‬
‫ܕܪܚܝܐ‬ ‫ܕܪܝܚܐ‬

‫‪ ‘(three) times’.‬ܙܒܢܝܢ ‪, an abbreviation for‬ܙܒܢ ‪a Sic! Instead of‬‬

‫‪Scrinium 18 (2022) 49–84‬‬

Syriac Spells for a Mill and Their Historical Context 57

The crucial place is l. 5, where we think the second spell begins. In BLa the
first two spells are not divided even by a punctuation mark. In H the second
piece can be distinguished as a separate spell not only due to the red ink, high-
lighting the words šḳol šaḇˁā perdē d-ḥeṭṭē ‘take seven grains of wheat’, but also
because of the mark of the end of a paragraph (‫)܀‬. However, it is BLb which
distinguishes the second spell in the most explicit way, introducing it with Syr.
ˀḥrēnā ‘another [recipe]’.
Where BLb is concerned, a few important features of this variant can be
noticed. Though it lacks the ending of the ritual instruction in the second rec-
ipe and omits the whole third recipe, it expands the text of the second recipe
with a few lines (verse 6). This reading is partly supported by BLa. This verse
expands the ritual instruction by adding “another word” in BLb (see section 5
under * below) and increasing the number of times the spell should be recited
to up to forty (in BLb and BLa). While in BLa this last instruction contradicts
the recommendation to recite the spell five times (compared with seven in H),
in BLb no such contradiction is found, since this is the only place where the
number of utterances is mentioned. Although the assumptions on the recon-
struction of the proto text are usually vulnerable, here it seems highly probable
that this contradiction was present in the original text of the recipe, which
in this aspect was close to the text as preserved in BLa, but it was eliminated
in the course of transmission. H and BLb represent two different decisions:
to keep only the first prescription (H), or only the second one (BLb). Second,
BLb provides partial vocalization of the text. Third, its manner of rendering
Arabic words seems closer to the Garshuni scribal system than the manner of
BLa and H. The clearest case is rendering of Arb. ‫ خ‬with Syr. ‫( ܟ‬with or without
�
rukkāḵā) which is found twice in w-ktmt b-ktmt ˀlgyn (wa-xatamtu bi-xātimat?
al-ǧinni ‘I seal with the seal of jinn’) instead of ḥtmt b-ḥtymt ˀlgyn in H and
BLa.17 Finally, rendering the Arabic incantation, H and BLa use a colon sign as
a word divider, while BLb uses either no word divider at all, or uses it rarely.
The only certain case is sndwk bndwk: BLb places a horizontal colon between
the two words, and a dot after the second word – most probably to mark the
end of the phrase.

17 Other possible cases are ˀlkbrˀ which probably stands for *ˀlh ˀkbr (ˀaḷḷāhu ˀakbar ‘God is
the greatest’). Here Syriac ‫ ܟ‬renders Arabic ‫ك‬, while in H and BLa ‫ ܓ‬is used instead.

Scrinium 18 (2022) 49–84

58 Cherkashina, Cherkashin and Saar

5 Philological Commentary

ll. 4, 6, 16: brk (barreḵ) ‘bless’

In what follows we place the word between quotation marks, because its
meaning is not easy to establish. Two questions arise here:
1) Does the word refer to the words mentioned before/after it, or does it
refer to the separate blessing which is to be pronounced besides the other
words?
2) What does “blessing” mean in the context of the two supposedly aggres-
sive spells?
Concerning the first question, if we look at all the three passages, it appears
that barreḵ refers to the words mentioned before or after it. A schematic repre-
sentation of the three passages seems to be helpful.

ll. 1–4: “say these words from the psalm” – [the words from the psalm] – “bless
upon the dust” – “throw it”

ll. 5–13: “take 7 grains” – “bless upon them [with the following] words” – [the
Arabic incantation] – “throw them”

ll. 14–17: [the words from the psalm] – “bless upon the wheat” – “throw it”

The second passage presumably gives a clue to our question, because it allows
us to equate the following incantation with the “blessing words”. If we expand
this interpretation to the other two passages, we can argue that in the three
texts the object of the verb barreḵ is the words mentioned before or after it.
Turning to the second question, the problem is that when barreḵ occurs in
two passages that we attribute to aggressive magic, we hardly can render it as
“bless”. Similarly, if we ascribe to the verb a euphemistic meaning and render
it as “curse”,18 the meaning does not fit the third passage from the loosening
spell. We suggest, that in these three texts barreḵ designates an action of pro-
nouncing words which have magic power, be it a citation from the Bible or an
Arabic incantation – with a good intention, same as with an evil one. However,
the issue requires further investigation based on the corpus of Syriac charms.

18 Cf. the Rabbinic corpus, where both Heb. brk and Aram. brk have a meaning “to blas-
pheme”, see Marcus Jastrow, A Dictionary of the Targumim, the Talmud Babli and Yerush-
almi, and the Midrashic Literature, Vol. 1, London–New York, 1903, p. 195.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 59

ll. 5, 13, 16–17: ˁynˀ d-rḥyˀ (ˁaynā d-raḥyā) ‘the mill’s eye’, i.e., the opening in a
hand mill where the grain is put

Our main interpretation of this expression is based on metaphorical usages

of ˁaynā ‘eye’ in Aramaic languages, such as JPA ˁyyn ‘draft hole of an oven’.19
Though we are unable to confirm or refute this interpretation based on the
Syriac lexical data, a similar usage is found in JBA: bt ˁyn ‘pupil of the eye’, ‘hole
in the millstone’.20
Another interpretation worth considering is based on the second core
meaning of ˁaynā, i.e., ‘spring’. In this case, raḥyā would mean ‘watermill’ and
ˁaynā d-raḥyā would denote a spring, or a water source, which refers to this
watermill.21 A support for this interpretation may be found in I, where instead
of ˁaynā we see reḡlā. The term’s core meaning is ‘foot’, but it can also denote
a stream, which would fit this interpretation. However, we cannot exclude
that reḡlā refers to some other part of the mill, since the term can also mean
‘foundation’ or ‘leg (e.g., of a table)’. Also, our main interpretation is sup-
ported by the external evidence, the recipe from the Cairo Geniza, which also
requires throwing dust into the opening in the upper mill, which is denoted by
Arb. ḥalq-.

l. 8: sndwk bndwk (sundūk bundūk)

All three manuscripts give the same writing, and, as it seems, here we have
a beginning of an Arabic spell. Initially this phrase probably was an Arabic
paronomastic construction called itbāˁ. According to Pellat, this figure of
speech is used in Classical Arabic when one wants to reinforce the meaning
of a word. Usually, the phrase consists of two words the second of which has
no existence in the language and is formed by alteration of mainly the first
radical, for example ḥasan basan ‘wonderfully attractive’.22 The reconstruction

19 Michael Sokoloff, A Dictionary of Jewish Palestinian Aramaic of the Byzantine Period.

Ramat-Gan, 1990, p. 403, cf. also īnu in Akkadian: meaning 2. ‘hole (of a kiln)’, ‘interstice (of
a net)’, ‘hub (of a wheel)’ (The Assyrian Dictionary of the Oriental Institute of the University
of Chicago, Vol. 7 (I and J), eds. Ignace J. Gelb, Benno Landsberger & A. Leo Oppenheim,
Chicago, 1960, p. 157).
20 Michael Sokoloff, A Dictionary of Jewish Babylonian Aramaic of the Talmudic and Geonic
Periods, Ramat-Gan–Baltimore, 2002, p. 249, see also Gustaf Dalman, Arbeit und Sitte in
Palästina, Vol. 3, Gütersloh, 1933, p. 227.
21 We owe this idea to Sergey Minov (personal communication).
22 See Charles Pellat, “Muzāwajah,” in: Encyclopaedia of Islam, 2nd edition, ed. P. Bearman,
Th. Bianquis, C. E. Bosworth, E. van Donzel, and W. P. Heinrichs, Vol. 7, Leiden, 1960–2007,
p. 823.

Scrinium 18 (2022) 49–84

60 Cherkashina, Cherkashin and Saar

of the original Arabic phrase would supposedly give ṣundūq bundūq, which
indeed is attested in Classical Arabic texts.23 Given this the phrase should
mean ‘an impressive chest’ which is not supported by the context and does not
fit our text. However, Schorch in his study of puns in the prophetic books of the
Hebrew Bible in a wider context argues that paronomasia has more functions
than merely creating emphasis.24 He also states that some kinds of “paronoma-
sia are related in a certain way to the use of paronomasia in imitative magic,
well attested in the neighbouring cultures of ancient Israel”.25 Moreover, this
phrase occurs in a children’s game, in which a player exclaims ṣundūq bundūq
beside other incomprehensible words and phrases in a colloquial Arabic,
loosely connected with each other.26

l. 9: wˀlh ˀgbr (wa-ˀaḷḷāh ˀakbar) ‘and God is the greatest’

The BLb’s reading ˀlkbrˀ could be interpreted as the Arabic word al-kubrā,27 but
this does not seem to fit the context. Two other manuscripts agree with each
other and give what we assume to be a more spontaneous ‫ّٰ أ‬ rather than a sys-
tematic rendering of the Arabic exclamation (‫���خر‬ ‫)� �ل��ل�ه � ك‬.28 If it were transliter-
ated according to the Garshuni scribal system, the second word would have
looked as ˀkbr (‫)ܐܟܒܪ‬.29 Krämer gives the whole Arabic part of the spell under
the title “Ein unverschtändlicher Text zur Beschwörung des Windes” where he
‫”و� �ل�ه(!) خ‬.30 This rendering is obscure to us.
renders this phrase in Arabic as “‫����خر‬

23 See, e.g., a commentary to ibn al-Farīd’s verse in Šarḥ Dīwān al-Fāriḍ liš-Š ayḫ Hasan
al-Būrīnī wa-liš-Š ayḫ ˁAbda al-Ġanī an-Nābulusī, Arnaud, 1853, p. 508.
24 Stephan Schorch, “Between Science and Magic: the Function and Roots of Paronomasia
in the Prophetic Books of the Hebrew Bible,” in: Pun and Pundits. Word Play in the Hebrew
Bible and Ancient Near Eastern Literature, ed. Scott B. Noegel, Maryland, 2000, pp. 206–7.
25 Ibid., p. 218.
26 See, e.g., Muḥammad Rajab al-Sāmarrāˀī, “ˀIṣdārāt Nādī Turāṯ al-ˀImārāt,” al-Ṯaqāfah al-
Š aˁbiyyah 5 (2009), p. 190, https://www.folkculturebh.org/upload/issues/issue5.pdf (2022,
May 4).
27 But for use of y instead of al-ˀalif bi-ṣūrat al-yāˀ in Garshuni see, e.g., Joseph Moukarzel,
“Le Garchouni. Remarques sur son Histoire et son Évolution,” in: Scripts beyond Borders: A
Survey of Allographic Traditions in the Euro-Mediterranean World, ed. Heijer, J., Schmidt, A.
& Pataridze, T., Louvain-la-Neuve, 2014, p. 135.
28 On phonetic rendering of the word aḷḷāh in Garshuni see, e.g., ibid.
29 See, e.g., ibid., p. 133.
30 Karl Friedrich Krämer, Textstudien zu Ostsyrischen Beschwörungsgebeten (unpubl. diss.),
Berlin, 1924, p. 84.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 61

l. 9: klmwy (kalimātī) ‘my words’

BLa has klmty, while H and BLb read klmwy which we assume to be a mistake.
Considering the possibility of incomprehension of the scribe, which will be
mentioned below, this mistake could also be attributed to peculiarities of the
scribal hand, in which shapes of t and w are alike.

l. 9: hdˀll ‘these?’

All three manuscripts give different sequences of letters, none of which makes
any sense in this context (H: hdˀll, BLa: hll, BLb: bhdˀll). The syntax of the
phrase suggests that we have here a demonstrative pronoun with kalimātī ‘my
words’ as its antecedent. It is hard to equate this with Classical Arabic hāḏihī
or hāˀulāˀī. The probable interpretation is to consider these letters as a dis-
torted form of a colloquial demonstrative pronoun, cf. hadōl(e)31 in dialects of
Levant, hāḏ̣ōḷ(ḷah)/hāḏoḷḷayn, hāḏōḷ/hāḏēl, hāḏōl in Bedouin Arabic dialects of
Arabian Peninsula and Near East.32

l. 10: wldy (wa-llāḏī) ‘and which’

Both manuscripts agree and the reading of the phrase is clear.33

l. 10: tkyl bh (ittakal bihi) ‘relied on this?’

Two manuscripts read tkl (BLa, BLb) and one has tkyl (H). It is quite hard
to understand this word especially with the next word (H and BLa: bh, but
BLb: bhh, the last h of which is without doubt taken from the next word). The
context suggests that we have here a verb with the meaning “to rely upon”.
The writing of the assumed verb is not only unclear, but also it cannot be eas-
ily construed with the following alleged preposition b-. Although the verb
ittakala occurs in Arabic with the preposition fī, in other stems of the root
w-k-l the preposition b- is used in reference to things which someone is put in
charge of.34

31 Peter Behnstedt, “Árabe Levantino,” in: Manual de Dialectología Neoárabe, ed. Federico
Corriente & Ángeles Vicente, Zaragoza, 2008, p. 163.
32 Peter Behnstedt, “Árabe Beduino (Península Arábiga y Oriente Próximo),” p. 81.
33 For this kind of spelling see, e.g., Joseph Moukarzel, “Le Garchouni. Remarques sur son
Histoire et son Evolution,” p. 135.
34 For Classical Arabic see, e.g., Albert de Biberstein-Kazimirski, Dictionaire Arabe-Français,
Vol. 4, Barrois, 1860, p. 1599 (ˀittakala ‘Se fier entierement à quelqu’un, s’en remettre de

Scrinium 18 (2022) 49–84

62 Cherkashina, Cherkashin and Saar

ll. 10–11: slmwn ˀbn dwwd (sulaymān ˀibn dawūd) ‘Solomon, son of David’

Every manuscript has different writing (H: hlmwn ˀbn dwwd, BLa: hslmn ˀb
dwwd, BLb: slmn ˀbn dwwd, note the previous commentary about h), but its
reading is obvious.

l. 11: ḥtmt (xatamt) ‘I hereby seal’

Two manuscripts have ḥtmt (H and BLa) and the last has wktmt (BLb). It is
remarkable that in these manuscripts both recorded possibilities for render-
ing Arabic x are present, namely with the letters ḥ and k.35 Here we apparently
have the performative use of the verb ḥtm (‘I hereby seal’).36

ll. 11–12: bḥtymt ˀlgyn (bi-xātimat? ˀal-ǧinn) ‘with the seal of the jinn’

H and BLa agree with each other (bḥtymt ˀlgyn) against BLb (bktmt ˀlgyn).
The context suggests “with the seal of the jinn”, but the Arabic word xātimah
which fits the writing usually means ‘end, conclusion’37 with the only excep-
tion we found being ‘signature, caractère’.38 Both meanings do not allow to get
the assumed meaning of the phrase. The interpretation of ˀlgyn as Arabic �ّ‫�� خ‬
‫� �خل‬
‘the jinn’ is of no doubt.39

‫خ‬
tout à sa volonté, av. ‫ ع��لى‬de la p. et ‫ �ى‬de la ch.,’ wakkala ‘Constituer quelqu’un son agent,
‫ي‬
son fondé de pouvoirs, son avocat, son homme d’affaires, av. acc. de la p. et �‫ خ‬de la ch.’);
Reinhart Pieter Anne Dozy, Supplément aux Dictionnaires Arabes, Vol. 2., Leyden, 1881,
p. 837 (tawakkala bi- ‘se charge de’), and for the colloquial Arabic see, e.g., D. R. Woodhead,
Dictionary of Iraqi Arabic, Washington, D.C., 1967, p. 502 (ttikal ‘to rely, depend, place one’s
trust’, ˀāni ˀattikil ˁalēk ib-hal-qaẓiyya ‘I’m depending on you in this matter,’ note the use of
the preposition b-); Hamdi Qafisheh, NTC’s Gulf Arabic-English Dictionary, Chicago, 1997,
p. 637 (twakkal with �‫ خ‬b- ‘to act as a counsel for (a case)’; Adrien Barthélemy, Dialectes de
Syrie: Alep, Damas, Liban, Jérusalem, Paris, 1969, p. 906 (əttákal).
35 Joseph Moukarzel, “Le Garchouni. Remarques sur son Histoire et son Évolution,” p. 133
and Emanuela Braida, “Neo-Aramaic Garshuni: Observations Based on Manuscripts,”
Hugoye, 17/1 (2014), p. 21.
36 See, e.g., Wolfdietrich Fischer, A Grammar of Classical Arabic, New Haven–London, 2002,
p. 103.
37 See, e.g., J. Milton Cowan (ed.), Arabic-English Dictionary: The Hans Wehr Dictionary of
Modern Written Arabic, Urbana, 1993, p. 265; and Martin Hinds & El-Said Badawi, A Dic-
tionary of Egyptian Arabic Arabic-English, Beirut, 1986, p. 241.
38 Reinhart Pieter Anne Dozy, Supplément aux Dictionnaires Arabes, Vol. 1., p. 352.
39 Michael Zellman-Rohrer, “A Syriac-Arabic Dream-Request and Its Jewish Tradition,” p. 65;
and Michael Zellman-Rohrer, “More on the ‘Book of Protection’,” p. 123.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 63

l. 12: ˀrbˁˀ (ˀarbaˁ ) ‘four’

The reading is obvious (H and BLb: ˀrbˁˀ, but BLa: ˀrbˁ ), but the syntax is
obscure. There could be at least two interpretations. First, it could be in appo-
sition to the previous word ‘the jinn’ meaning ‘the four jinn’.40 This interpreta-
tion suggests that the object (the mill) of the verb xatamt ‘I seal’ is omitted. In
this case the loss of the definite article before the number must be assumed.
Second, it could be construed as a direct object of the verbal form “I seal”, but
then it is unclear what four things are sealed especially considering unclarity
of the following phrase.

l. 12: mn sˀyt (min as-sayyid)

The writing of the two manuscripts (H and BLa: sˀyt) does not lend itself to
understanding.41 The interpretation of the third manuscript’s writing (BLb:
syd) is clear (sayyid ‘master, owner’), nevertheless its syntactic position is dif-
ficult to comprehend. Also, the absence of the article in writing complicates
interpretation of the phrase even more although this kind of recording is pos-
sible in Garshuni.42

l. 12: dfwhˀ (ḏī-fūhā?) ‘of its opening (lit. “mouth”)’

Two manuscripts agree in writing (H and BLb: dfwhˀ, but BLa: wfwhˀ). Although
the grammatical rules of Classical Arabic seem to permit such a use of the
Arabic nota genitivi ḏū still it seems here redundant and clumsy.43 One can rea-
sonably suppose contamination with Syriac particle di-. Combination of the
letters fwhˀ according to the context should refer to the mill and prompts to

40 Ibid.
41 For possible reading of sˀyt see the word sāyah in Edward Lane, An Arabic-English Lexicon,
Vol. 1, London–Edinburgh, 1877, p. 1458 (ḍaraba fulānun ˁalā fulānin sāyatan ‘Such a one
did to such a one a thing that caused displeasure to him, such a one made a way to do
what he desired to such a one’). Other possible interpretations that assume corruption
are reading the word saˀt ‘side of the throat’ (J. G. Hava, Arabic-English Dictionary for
the Use of Students, Beirut, 1899, p. 296) or the word saˀaˀah ‘cambrure au bout de l’arc
retourné en dehors’ (Albert de Bibrestein-Kazimirski, Dictionaire Arabe-Français, Vol. 1,
p. 1039). None of these interpretations easily fits the context.
42 Joseph Moukarzel, “Le Garchouni. Remarques sur son Histoire et son Évolution,” p. 135.
43 For the use of ḏū in Classical Arabic see, e.g., Hermann Reckendorf, Arabische Syntax,
‫ِ ُّ خ ُ ْ خ‬
Haidelberg, 1921, p. 152 (note the attributive use: �ِ ‫‘ � �ل�� خ� � و � �ل�ع���ص‬das Korn mit der
ْ ّ ُ ‫خُ خ‬
Hülse,’ and the use with pronominal suffix: ‫‘ � و ��ر�ه‬ihr Schädiger’).
‫ِِم‬

Scrinium 18 (2022) 49–84

64 Cherkashina, Cherkashin and Saar

read it as ‘a mouth (of the mill)’. This naming of the opening of the upper stone
indeed occurs in Arabic,44 but nevertheless the exact form of a used Arabic
word (or a combination with the pronominal suffix hā)45 as well as the syntax
of the whole phrase are obscure.

* verse 6

In BLa we find an additional instruction to repeat the spell given in the previ-
ous lines “for a period of forty times”. In BLb there is one more sentence. It says:
“And one other word [you repeat] for many times”. The interpretation of the
following two words, bh ˀytyn (vocalised as bēh ˀaytyān) is complicated, since
they can be understood either as a Syriac, or as an Arabic phrase – in each case,
the text seems corrupted and requires emendations.

A. If a Syriac hypothesis is considered, there are a few possibilities. Here

we discuss the two which do not require changing the consonantal text.
A1. bēh ˀaytīn ‘which we adduced [above]’
A2. bēh ˀaytyan ‘with which he came upon us’
Both suggestions involve a few problems. First, the verb ˀty in C-stem is
not normally used with b-, instead the object of bringing is introduced as
a second direct object. Second, any Syriac interpretation of this phrase
would imply a disruption in normal Syriac syntax, since a subordinate
clause, i.e., “the word, which …”, is to be marked with a relative pronoun
d-. Third, none of these two makes sense in our context, because none
of them mentions this another word or phrase which should be uttered.
B. Alternatively, the phrase can be explained as another corrupted
rendering of Arabic. The most attractive interpretation is, probably,
bihi ˀtinī (��‫‘ )خ��ه ي�� خ‬give me this’, i.e., ‘do this for me’. In this case we must
‫ي‬
suppose omitting of alif al-waṣl in writing of the imperative of the verb
ˀatā, as well as a specific syntax of the phrase possibly for the emphasis,
when the preposition with the pronoun (bihi) goes before the verb (ˀti).
Nevertheless, assuming an Arabic citation has an advantage compared

44 See, e.g., al-Murtaḍā al-Ḥusaynī al-Zabīdī,

ْ Tāj al-ˁArūs min Jawāhir al-Qāmūs, Vol. 39,
ِ ِّ ِ‫أِ يِ ِْ ُ خ خ‬ ُ‫ُّ ْ ي‬
�‫)� �ل������ ِو� … � �ا � �ل����ي��ي���ه �ى � � �لر‬. Also note the use in Colloquial
al-Kuwait, 2001, p. 498 (‫حى‬
‫ي ِم‬
Arabic of the words ḥalq and tumm both meaning ‘mouth’ for the same part of a hand mill
(Gustaf Dalman, Arbeit und Sitte in Palästina, Vol. 3, p. 222).
45 For possible variants see, e.g., the whole article under fūhun in Edward Lane, An Arabic-
English Lexicon, Vol. 6, p. 2464.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 65

with a Syriac one, because it allows us to avoid the problem of syntactic

disruption. Also, another Arabic phrase in this charm would fit the con-
text better.

6 Interpretation

In this section we provide a short overview of the three Syriac recipes for the
mill followed by our interpretation. As already mentioned, in every place
where we read raḥyā ‘mill’, Gollancz reads ryḥˀ, which corresponds to Syr. rīḥā
‘smell’. Even though this reading is supported by K. Fr. Krämer, to whose inter-
pretation we will turn further on, we would rather suspect an editorial mistake,
made by Gollancz and replicated by Krämer, than a scribal one, since the two
words look very similar in the Syriac script: ‫‘ ܪܝܚܐ‬smell’ vs. ‫‘ ܪܚܝܐ‬mill’. Evidently
following Gollancz’s reading of raḥyā ‘mill’ as rīḥā ‘smell’, Krämer interprets
the whole text-unit as “Psalmen zur ‘Bindung’ und ‘Lösung’ des Windes”.46
Since the Syriac term does not mean ‘wind’, such translation could be possible
only if we interpret rīḥā not as a Syr. term for ‘smell’, but as an Arabic word
rīḥ- ‘wind’.47 In this case, the Syriac ending -ā, typical for the inherited Aramaic
lexica or well-adapted loanwords, requires an explanation.
There are a few reasons to treat Gollancz’s reading as erroneous. First,
in the case of H, the consonantal writing is clear enough to eliminate any
doubts, while in BLb the word is vocalized as reḥyā. Though not found in
Syriac dictionaries,48 this vocalization is supported by the lexical data of other
Aramaic idioms.49 Second, the sphere of usage of the textual block can be
deduced from the content of the recipes, abundant with agricultural lexica.50

46 Karl Friedrich Krämer, Textstudien zu Ostsyrischen Beschwörungsgebeten, pp. 8–9, §13. See
also p. 84, §40 “Ein unverständlicher Text zur Beschwörung des Windes”, which also refers
to our block of texts, namely, the Arabic charm.
47 J. Milton Cowan (ed.), Arabic-English Dictionary: The Hans Wehr Dictionary of Modern
Written Arabic, Urbana, 1993, p. 432.
48 Robert Payne Smith, Thesaurus Syriacus, Oxonii, 1879–1901, p. 3878, Michael Sokoloff,
A Syriac Lexicon. A Translation from the Latin. Correction, Expansion, and Update of
C. Brockelmann’s Lexicon Syriacum, Winona Lake–Piscataway, 2009, p. 1455.
49 Cf. Jewish Babylonian Aramaic ryḥyˀ ‘mill, millstone’ (Michael Sokoloff, A Dictionary of
Jewish Babylonian Aramaic, p. 1075), Jewish Palestinian Aramaic rḥy, det. ryḥyyh ‘mill-
stone, mill’ (Michael Sokoloff, A Dictionary of Jewish Palestinian Aramaic, p. 520). Both
forms go back to *riḥy-.
50 See ˁaynā d-ra/eḥyā ‘mill’s eye’ in ll. 5, 13, 16–17; (perdē d-)ḥeṭṭē ‘(grains of) wheat’ in ll. 6, 16.
Also see presumably pwhˀ ‘its (i.e. the mill’s) opening’ in l. 12. See also what follows on the
Geniza parallel.

Scrinium 18 (2022) 49–84

66 Cherkashina, Cherkashin and Saar

The lexical issues, together with the text-critical data and comparative data
from Jewish and Coptic magic discussed below, makes the reading offered by
Gollancz and Krämer’s interpretation unconvincing.
If the most extensive textual variants are considered (H and BLa), the text-
unit consists of three fragments, each of them containing a ritual prescription.

1) ˀassārā d-raḥyā ‘Binding of a mill’ (verses 1–2)

The first binding spell prescribes reciting Psalm 102:11 three times above the
dust51 and throwing it, most probably, into the opening in the upper mill, or
“mill’s eye” (see the commentary to l. 5 below in section 5). The cited words
from Ps. 102:11 are probably connected with the following prescription to take
and throw dust into the mill’s eye. However, one cannot leave unmentioned the
fact that the previous verse of the psalm (Ps. 102:10 according to the Peshitta
numeration) reminds us of the aim of the spell: “For I was eating ashes like
bread and was mingling tears with my drink”. Although, in the Syriac Peshitta
we find the term qeṭmā ‘ashes’, while in the spell for a mill (verse 2) in all the
manuscripts ˁaprā ‘dust, clay’ is used, we still find probable that alluding to
Ps. 102:11 both verses 10 and 11 were meant here. It is also worth mentioning
that the Hebrew terms for “ashes” (ˀēp̄ är), used in the Hebrew version of the
verse, and “dry earth, dust” (ˁāp̄ ār) can sound similar and have close semantics.

2) ˀḥrēnā ‘Another [recipe]’ (verses 3–7)

The second fragment, also a binding spell, starts with a prescription, involv-
ing manipulations with seven wheat grains that should be thrown into the
mill after being “blessed” five, seven or up to forty times – depending on the
manuscript.
Verses 4–5 contain an Arabic spell written in Syriac script (Garshuni) with a
supposed infusion of colloquial Arabic traits (on which see above). The scribal
manner of H and BLa to separate each Arabic word with a word divider was
discussed above in the text-critical section. It seems, that both scribes were
uncertain as to where the Arabic text ends and ritual instructions in Syriac
begin (verse 5). In H, the last word with a colon is ˀlgyn ‘the jinn’, while in BLa
the last colon is found two words before, after ḥtmt ‘[hereby] I seal’. Though the
language of these last sentences appears to be corrupted and, in some cases,
may represent a mixture of Syriac and Arabic, according to our interpretation
the Arabic text ends no earlier than after the word pwhˀ ‘(its?) opening’.

51 Here all the manuscripts but I agree, the peculiarities of which were discussed above.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 67

Some Arabic words were correctly interpreted by K. Fr. Krämer in his

unpublished doctoral thesis (see section 5 above). We cannot restore the exact
wording of the original Arabic spell based on the current form of the text, but
the general meaning is obvious. The form of the Arabic part indicates that at
some stage the text became completely incomprehensible for the copyists
of the recipes.52 If they would have had any command of Arabic, the scribes
would not have misspelled the name Solomon or written ‘father’ (ˀb) instead of
‘son’ (ˀbn in “Solomon son of David”) not to mention the facts that only a few
words in this spell are clearly readable in Arabic and the syntax is sometimes
impenetrable.
The Arabic spell alludes to the legend of king Solomon’s magic ring, which
frequently occurs in both Syriac53 and Arabic54 magic. According to our inter-
pretation the opening in the upper mill is claimed to be sealed by the seal of
the four jinn. The four jinn here most probably symbolize the four cardinal
points, i.e., the mill is intended to be bound from all four sides. Each cardinal
point is thus governed by its jinn. This reminds us of a protective Jewish for-
mula used in magic and in prayers: “May Michael be at my right hand, Gabriel
at my left; in front of me, Uriel, behind me, Raphael; and above my head the
Presence of God”. Bearing the exact opposite goal, i.e., protection instead of
harm, the formula also implies positioning of the named supernatural beings
according to the four cardinal points. Notably, a similar protective formula is
found in Syriac and in Muslim magic texts.55

52 Zellmann-Rohrer plausibly assumes “successive cycles of copying by uncomprehending

scribes” (Michael Zellmann-Rohrer, “A Syriac-Arabic dream-request and its Jewish tradi-
tion,” p. 70).
53 For the Syriac charms which mention Solomon and their Jewish parallels see Hermann
Gollancz, The Book of Protection, pp. xii–xix; for a Syriac charm, belonging to Solomonic
tradition, specifically close to the Testament of Solomon, and not attested in Gollancz’s
corpus, see Анна Черкашина, “Соломон и демоны в заклинаниях сирийских хрин -
стиан” [Anna Cherkashina, “Solomon and the Demons in the Charms of Syriac Chris-
tians”], Христианский Восток [Christian Orient], 9 (XV) (2021), pp. 149–162.
54 Hans Alexander Winkler, Siegel und Charaktere in der Muhammedanischen Zauberei,
Berlin–Leipzig, 1930, especially pp. 110–114, 127–132, for Solomon’s ring in Arabic literature
see Georg Salzberger, Die Salomo-Sage in der Semitischen Literatur, Berlin–Nikolassee,
1907, pp. 120–129; see also Allegra Iafrate, The Long Life of Magical Objects: A Study in the
Solomonic Tradition, Pennsylvania, 2019, pp. 50–59.
55 Alexey Lyavdansky, “Syriac Charms in Near Eastern Context: Tracing the Origin of For-
mulas,” in: Oral Charms in Structural and Comparative Light. Proceedings of the Confer-
ence of the International Society for Folk Narrative Research’s (ISFNR) Committee on
Charms, Charmers and Charming, 27–29 October, Moscow, eds. T. A. Mikhailova, J. Roper,
A. L. Toporkov, D. S. Nikolayev, Moscow, 2011, pp. 17–18; Hans Alexander Winkler, Siegel
und Charaktere in der Muhammedanischen Zauberei, p. 17.

Scrinium 18 (2022) 49–84

68 Cherkashina, Cherkashin and Saar

Also, it is worth to mention that seven grains occur in a Muslim story about
the origins of first breadmaking.56 In this story God, via Gabriel, gave Adam
after his exile from Heaven seven grains which he then sowed and harvested.
After that Gabriel gave him two stones with which Adam made a hand mill. He
ground the wheat and kneaded it. Gabriel taught him how to kindle a fire, so
Adam was able to bake the first loaf of bread.

3) šeryānā d-raḥyā ‘loosening of a mill’ (verse 8)

The last spell is a counter-spell to the two previous ones or to any other curse
intended to affect one’s mill. The recipe requires the user to recite Psalm 146
over wheat (without specifying the quantity required) which, just as in the two
cursing rituals, is to be thrown into the mill’s eye.

7 The Broader Context

The first two recipes supposedly aim to stop (cf. Syr. bṭl in l. 13) the mill that
belongs to the spell beneficiary’s adversary. The final goal of the two spells is
presumably to harm the adversary. By causing problems with the production
of flour, the spell beneficiary may seek to affect the sustenance of the spell
target and his family. These considerations allow us to attribute the two reci-
pes to aggressive magic and to compare them with various cursing techniques
attested in magical texts from Egypt, both in Coptic57 and in Jewish58 contexts.
The textual variant of the text-unit found in BLb has been mentioned
recently by M. Zellmann-Rohrer, who also ascribes an aggressive intention to
these recipes, assuming their usage in a context of commercial rivalry59 and
comparing them with recipes aimed to hinder the heating of an oven.

56 See, e.g., The History of al-Tabari. An Annotated Translation. General Introduction and from
the Creation to the Flood, Vol. 1., translated and annotated by Franz Rosenthal, Albany,
1989, p. 298 (English translation); Annales quos Scripsit Abu Djafar Mohammed Ibn Djarir
at-Tabari, Vol.1, ed. M. J. De Goeje, Leiden, 1979–1881, p. 127 (original text).
57 Cf. Text 106 (“Curse to bring seventy different diseases upon victim”) in Marvin W. Meyer
& Richard Smith, Ancient Christian Magic: Coptic Texts of Ritual Power, Princeton–New
Jersey, 1999, pp. 215–216, Text 110 (“Curse to harm a person through the use of wax dolls”)
in ibid., pp. 222–223, and other curses against enemies (Texts 88–112) in ibid.
58 Cf., e.g., Geniza 16 7:1–13 (“causing fire”) in Joseph Naveh & Shaul Shaked, Magic Spells and
Formulae: Aramaic Incantations of Late Antiquity, Jerusalem, 1993, pp. 174–181, Geniza 24
2:8–12 (“for hatred”) ibid., pp. 222–225, Geniza 26 2:1–8 (“meant for every need, for causing
disease, for killing, for causing fire, also for hatred”) ibid., pp. 230–232.
59 Michael Zellmann-Rohrer, “More on the ‘Book of Protection’,” pp. 122–123.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 69

There are three text-units related to a furnace found in three manuscripts,

which notably also contain the binding spells for a mill. The first text-unit is
found in BLb (ff. 22v–23v), the first recipe is titled ˀAssārā d-ˀattūnā w-mānay
d-ˀīṯ bēh ‘Binding of a furnace and the implements which are in it’. The second
text-unit is contained in I (f. 37r), where only the ending of the binding spell
has been preserved, but the full text is found in the closely related manuscript
NH3, where it is called ˀAssārā d-ṯannūrā ‘Binding of the oven’ (f. 36v). The
third text-unit for an oven is found in I (f. 35, partly lost due to the torn page)
and NH3 (f. 35r): ˀAssārā d-ḵūrā d-p̄ arzlā ‘Binding of an iron smelter’ (NH3,
I 35). All three text-units contain both a binding and a loosening spell. Sharing
a similar target, they are completely different on the formulaic and structural
level, allowing us to regard them as three different text-units designed for simi-
lar purposes. These texts deserve to be the subject of a separate research, there-
fore, below we cite only a few excerpts from the recipes.
In BLb (the 1st text-unit) the text reads:

(…) I bind this furnace (ˀattūnā) and the smelter (kūrā). Let it be like
snow and hail. (…) I bind the furnace and everything that is in the smelter.
(…) And all the utensils that are in it. (…) And let them not burn but let
them get broken and become like chaff that the wind carries away (BLb
ff. 22v–23r).

The respective loosening spells for a furnace (šeryānā dīlēh ‘Loosening spell for
it’) in BLb reads:

(…) Let this binding of the furnace and the smelter disappear, let [the
furnace] return to its initial state. Let the utensils in it be heated (or
inflamed), let them not be pierced?, nor broken, the small [ones] together
with the big [ones] (BLb f. 23v).

The binding spell of the second text-unit contains a ritual prescription to recite
the incantation above a piece of wood and throw it into the furnace, absent
from BLb. As for the loosening spell, instead of a mid-length text found in BLb,
in I and NH3 there is only a short prescription to recite Psalm 146 three times.
Notably, the same psalm is used in the ‘Loosening of a mill’ (verses 8–9). The
binding spell says:

They are igniting the furnace in vain, let [the fire in it] go out as [the fire
in] the furnace of the house of Hananiah.60 By that power let the fire not

60 Dan. 3.

Scrinium 18 (2022) 49–84

70 Cherkashina, Cherkashin and Saar

burn in it and not be ignited in the furnace of so-and-so son of so-and-so

(I f. 37r).

A prescription to recite the spell over three pieces of wood is found in the third
text-unit, found in NH3 and I. The spell says:

“He rode upon the cherubim and flew61” – three sentences (lit. “words”).
Cut three pieces of wood, recite [the words] over them and throw them
into the furnace (NH3 35r).

Indeed, at least the first and the third text-units appear to refer to an industrial
furnace for firing pottery or smelting metals, while the second can be designed
for a baking oven – either domestic or industrial. One should note, however,
that also at household level the binding of an oven could have had adverse eco-
nomic implications, since this would result in more wood needing to be used
and more work invested in attempting to kindle the fire. This is evident when
we consider the reverse type of magical practices: those intended to hasten the
cooking of food, presumably to require less wood and work. Examples of this
type are found in rabbinic literature, for instance Tosefta Shabbat 7.7–8, where
several practices are adduced (and prohibited) for hastening the cooking of
food.62 Similarly, the Babylonian Talmud Shabbat 67b mentions “A woman
who urinates in front of her pot so it will cook quickly: that action contains
an element of the ways of the Amorite.”63 To conclude, whether the aggressive
charms in the Syriac manuscripts referred to industrial or to home ovens, pre-
venting them from burning would have harmed their owners.
A parallel to these oven-binding charms is found in a recipe for binding the
furnace of a bathhouse from the Jewish composition Sefer ha-Razim (The Book
of Secrets). The recipe begins with the words “If you wish to extinguish (the fire
which heats) a bathhouse so it will neither flare up nor burn”, and continues
with a practice involving the adjuration of a salamander dipped in oil.64 The

61 Ps. 18:10.
62 “She who shouts at the oven so that the bread does not fall (…), she who silences for the
lentils and she who cools/sucks for the rice” (‫המצוחת לתנור שלא תנפל הפת …; המשתקת‬
‫ והמוצצת לאורז‬/ ‫ לעדשים והמצננת‬hmṣwḥt ltnwr šlˀ tnpl hpt…; hmštqt lˁdšym whmṣnnt/
whmwṣṣt lˀwrz). It is not entirely clear what the last two practices, for the lentils and rice,
entailed.
63 .‫( המשתנת בפני קדירתה בשביל שתתבשל מהרה יש בו משום דרכי האמורי‬hmštnt bpny
qdyrth bšbyl šttbšl mhrh yš bw mšwm drky hˀmwry).
64 Sefer ha-Razim, Third firmament, §186–188 (Bill Rebiger & Peter Schäfer, Sefer ha-Razim
i und ii: Das Buch der Geheimnisse i und ii, 2 vols, Tübingen, 2009. Vol. 1, in collabora-
tion with E. Burkhardt, G. Reeg and H. Wels: Edition, pp. *64–*65): ‫אם בקשת לכבות‬

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 71

oil is then dripped at the four corners of the bathhouse, with the intention
to “make it like cold snow or cold water”.65 This spell is followed by a loosen-
ing formula, meant to ignite the furnace again: “I adjure you, angel of fire and
angel of conflagration, that you will undo what I have bound”.66 Obviously,
preventing the furnace from heating the bathhouse would have inflicted eco-
nomic damage on its owners, just as the binding of an industrial or home-oven
would. However, the target addressed in all the above spells is not the owner
but the fire.
Additional parallels to the oven-binding and releasing recipes can be found
in another Jewish composition, Ḥarba de-Moshe (The Sword of Moses), in a rec-
ipe titled “If you wish to close an oven or a basin or a pot so that (foods) will not
be put (in them)”.67 This binding recipe is followed by a counter one, so that
“they will be (released for) cooking”.

8 Historical Context

Aggressive magic, both in antiquity and in later periods, was usually meant to
directly affect human targets. Some examples would be practices for causing
bodily harm, such as disease, fever, impotence, bareness, and inability to speak,
along with more psychologically oriented practices, such as those meant to
cause memory loss and a change of feelings from love to hatred or vice versa.
A smaller number of aggressive magical practices were directed at animals,
with the intention to obliquely affect a human target. Prominent examples of
this category are curses and binding spells directed at racing horses.68 Lastly,

‫( מרחץ שלא תעלה ותלהט‬ˀm bqšt lkbwt mrḥṣ šlˀ tˁlh wtlhṭ). English translation from
Michael A. Morgan, Sepher Ha-Razim: the Book of Mysteries, Chico (CA), 1983, p. 62.
65 ‫( ועשוה כצנת שלג וכמים קרים‬wˀswh kṣnt šlg wkmym qrym).
66 Sefer ha-Razim, Third firmament, §189 (Rebiger & Schäfer, Sefer ha-Razim i und ii, Vol. 1,
p. *66): ‫( משביע אני עליך מלאך אש ומלאך שלהבת שתתיר את מה שאסרתי‬mšbˁ ˀny ˁlk
mlˀk ˀš wmlˀk šlhbt šttr ˀt mh šˀsrt). English translation from Morgan, Sepher Ha-Razim,
p. 63.
67 ‫( אם בעית למיצר תנור או כיור או קידרא דלא נטמיין‬ˀm bˁyt lmyṣr tnwr ˀw kywr ˀw qydrˀ
dlˀ nṭmyyn). For the English translation, see Yuval Harari, “The Sword of Moses (Harba
de-Moshe): A New Translation and Introduction,” Magic, Ritual & Witchcraft, 7 (2012),
pp. 58–98 (90).
68 See, e.g., PGM 111. 1–164 (Hans Dieter Betz (ed.), The Greek Magical Papyri in Translation,
Including the Demotic Spells, Chicago–London, 1992, pp. 18–22). For recipes for bind-
ing wild animals, see Sefer ha-Razim, Second Firmament, §155 (Rebiger & Schafer, Sefer
ha-Razim i und ii, Vol. 1, p. *52}; Ḥarba de-Moshe (Yuval Harari, “The Sword of Moses,”
p. 91).

Scrinium 18 (2022) 49–84

72 Cherkashina, Cherkashin and Saar

some aggressive practices focused on inanimate targets. Here, too, the ultimate
intention was to affect humans, as shown in the examples below. However, the
notion of binding an inanimate entity is particularly intriguing, as it addresses
the inanimate as a living being.
Binding and aggressive magical practices targeting inanimate entities could
include natural formations such as rivers or seas,69 agricultural products,70 and
also objects and installations, such as a ship71 or the mill which is the focus of
this article. The latter type of practices is found already in Late Antiquity, as
evidenced by the recipes for binding the furnace of a bathhouse mentioned
above. Further parallels to this type of practices can be found in Coptic magic,
for instance in a manuscript currently located in the Egyptian Museum in
Cairo, inventory number JdE 42573, roughly dated between the tenth and the
twelfth century.72 The aim of this magical text is to affect a water-wheel and
prevent it from supplying water. A water-wheel is a type of revolving installa-
tion similar to the mill mentioned in the recipes we discuss. In the Coptic text,
the binding of the water-wheel is to be effected by pouring a mixture of liquids
onto it: acacia juice (?) and snake water, the latter possibly meaning venom.
This practice is similar in form to that attested in the Syriac recipes, but lacks
the verbal element.
Aggressive magic against a water-wheel is also mentioned in another Coptic
text currently located in Heidelberg, P. Heid. Inv. Kopt. 408. This magical pro-
cedure – which also involves pouring a mixture of liquids – is accompanied by
a historiola in which a supernatural entity named Khoubin Harpak describes
himself to King Solomon and claims: “My work is destruction. (…) A thresher,

69 See, e.g. Sefer ha-Razim, Second Firmament, §155 (Rebiger & Schafer, Sefer ha-Razim i und
ii, Vol. 1, p. *52): “(if you wish to quell) a river or a sea which is rising and washing against
buildings” (‫נהר או ים העולה ושטף בבתים‬, nhr ˀw ym hˁwlh wšṭp bbtym) (English transla-
tion from Morgan, Sepher Ha-Razim, p. 53).
70 This category may have been alluded to in the Twelve Tables law on “enchanting” crops
(qui fruges excantassit), that was interpreted in later periods either as stealing crops with
the help of magic, or ruining them by invoking storms. For a recent treatment of this law,
with reference to previous literature, see J. B. Rives, “Magic in the XII Tables Revisited,”
Classical Quarterly, 52 (2002), pp. 270–290. Conversely, for examples of magical practices
meant to positively affect agricultural products, see Ḥarba de-Moshe: “For trees that do
not produce fruit” “For white rot that afflicts fruit”, “For blight that afflicts the field” (Yuval
Harari, “The Sword of Moses,” pp. 87–88, 92).
71 See Ḥarba de-Moshe, a binding recipe “to detain a ship at sea”, followed by a releasing
recipe: ‫( למיכלא אילפא בימא‬lmyklˀ ˀylpˀ bymˀ) (Yuval Harari, “The Sword of Moses,” p. 90).
72 See Korshi Dosoo, Edward O. D. Love & Markéta Preininger (chief editors). “KYP T55:
Destruction of a water-wheel,” Kyprianos Database of Ancient Ritual Texts and Objects,
www.coptic-magic.phil.uni-wuerzburg.de/index.php/text/kyp-t-55. Accessed on 07/04/
2022. We are grateful to Korshi Dosoo for this reference and the following one.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 73

I destroy it, a […] I destroy it, a shovel, I lay waste to it, a water-wheel, I destroy
it, a garden, I destroy it, […] I destroy it, a storehouse, I destroy it (…).”73 As can
be seen from the above list, this Coptic text was meant to affect a variety of
inanimate entities (both objects and installations) and presumably ultimately
to harm their human owners.
Moving now to the Jewish parallels to the mill spells, we start with the obser-
vation that two parallels have been found. The first is part of a bifolio with
magical recipes from the Cairo Geniza, T-S Ar. 49.54 1:12–3:5, that has been pub-
lished by Joseph Naveh and Shaul Shaked.74 Below we revise the reading of
Naveh and Shaked.

Cambridge T-S Ar. 49.54 1:12–3:5. Paper bifolio, ca. 13.5 × 19.5 cm, each leaf
measures ca. 13.5 × 10 cm. Inscribed in black ink by one or two hands. Oriental
square script. The text contains Judaeo-Arabic, Hebrew and Aramaic.
Our reading revises that offered by the first editors of this text, Naveh and
Shaked, on several points. In the first recipe, lines 14–15, they read the last let-
ters of the magic names as a final mem, while we suggest reading samekh. The
scribe wrote these two letters in nearly identical form (cf. 4, line 8: ‫)המקום ואסף‬.
However, a comparison with the releasing recipe found in this manuscript,
where similar magic words end with a clear samekh, as well as with the paral-
lel recipe from NYPL 190, tips the scales in favor of a samekh. Second, in line 15
the previous editors read “‫( ”לפנונים‬lpnwnym), but the first nun is mistaken, and
the word is ‫( לפוניס‬lpwnys). Third, in line 16 the last word begins with a clear
ṭet, the first letter in the noun ‫טחונתה‬, but this part was not reconstructed nor
translated by Naveh and Shaked. The last line of the recipe, line 17, is poorly
preserved. It clearly contains the goal of the recipe, namely that the mill should
not function, but the verb is missing. We reconstruct here ‫( תהלך‬thlk) based on
two indications: a partially preserved lamed in the middle of the sequence,
and the verb found in the releasing recipe, which is ‫תהלך‬. The parallel from
NYPL 190 contains two different verbs: ‫( תסוב ותרגיש‬tswb w-trgyš). Neither
seems to fit the lacuna in T-S Ar. 49.54 due to the presence of the lamed. Lastly,
we reconstruct the preposition ‫( ב‬b) ‘at’ in the phrase “at this moment (lit.
hour)”. The expression ‫( בהדא שעתה‬bhdˀ šˁth) is found in other magic recipes,
e.g., T-S Misc. 27.4.11, 1:14 (‫ עבידו לי צביוני בהדא שעתא‬ˁbydw ly ṣbywny bhdˀ šˁtˀ).75

73 See Korshi Dosoo, Edward O. D. Love & Markéta Preininger (chief editors). “KYP T174: Curse
to destroy a business (?),” Kyprianos Database of Ancient Ritual Texts and Objects, www
.coptic-magic.phil.uni-wuerzburg.de/index.php/text/kyp-t-174. Accessed on 07/04/2022.
74 Joseph Naveh & Shaul Shaked, Magic Spells and Formulae, pp. 227–230, Geniza 25.
75 Joseph Naveh & Shaul Shaked, Magic Spells and Formulae, pp. 216–219, Geniza 22.

Scrinium 18 (2022) 49–84

74 Cherkashina, Cherkashin and Saar

T-S Ar. 49.54 1:12–17

12 bˀb l-ṭḥwnh a‫ באב לטחונה‬A section for a mill.b

13 kwd twrˀb mn tḥt rǧlk ‫רגלך‬
ׄ ‫ כוד תוראב מן תחת‬Take some dust from
ˀl-ysˀr w-ˀlqyh ‫ ואלקיה‬c‫ אליסאר‬under your left foot, and
throw it
14 py ḥlqhˀ w-qwl b-šm ‫ אלפוניס‬d‫ פי חלקהא וקול בשם‬into its opening and say:
ˀlpwnys “In the name of ˀlpwnys
15 lpwnys w-ˀkpwnys ˀtwn ‫לפוניס ואכפוניס אתון אתיה‬ lpwnys and ˀkpnwnys!
ˀtyh qd[yšyh] [‫קד]ישיה‬ You, holy letters
16 w-šmhtˀ qdyšyh ˀswrw ‫ושמהתא קדישיה אסורו הדא‬ and holy names, bind
hdˀ ṭ[ḥwnth] [‫ט]חונתה‬ this mill,
17 dl[ˀ thlk b-h]ˀdˀ šˁtˀ: ‫דל]א תהלך)?( ב)?(ה[אדא‬ so that it does not [move,
: ‫שעתא‬ at] this moment!”

T-S Ar. 49.54 2:1–5

1 […] e[…] […]f

2 qwl ˀnpynws ˀpntpynws ‫ קול אנפינוס אפנתפינוס‬Say: “ˀnpynws ˀpntpynws!
3 ˀtwn šmhth qdyšyh ‫ אתון שמהתה קדישיה‬You, holy names,
4 šrwn hdˀ ṭḥwnth xx g‫ שרון הדא טחונתה‬loosen this mill xx
5 w-thlk: mǧrb ׄ :‫ ותהלך‬and let it move”. Proven.
‫מגרב‬

a The title is emphasized with supralinear clusters of three dots/short strokes.

b The translation provided by Naveh and Shaked is as follows:
12) <<A chapter for a lame woman (?)
13) Take dust from underneath your left foot and throw it
14) into her gullet, and say:>> “In the name of (magic
15) words). You, holy characters
16) and holy names, bind this …
17) that he may not … this hour.”
c Both here and in the following word the “al” is written as ligature.
d From this word until the word ‫ הדא‬in line 16 (including) all the words are emphasized with
supralinear clusters of three dots.
e The missing text is centered and was probably the title of the recipe.
f The translation provided by Naveh and Shaked is as follows:
1) <<Say:>> (magic names)
2) you, holy names,
3) release this female lame (?) person
4) that she may walk. <<Tested.>>
g The word is followed by two smaller signs resembling the qamatz of the Babylonian vocaliza-
tion system. These might have functioned as graphic line fillers.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 75

Alternatively the preposition could have been ‫( מן‬mn), “from (this moment)”,
but the space seems rather small for two letters (although it is not impossible
that they were tightly spaced). The scribe could have used the preposition ‫מ‬,
in its Hebrew form, thus mixing Hebrew with Aramaic. This, too, is not impos-
sible, given other imprecisions in the text, such as the spelling of ‫ הדא‬with an
extra aleph, ‫האדא‬.
Naveh and Shaked, while noting that the term ‫( טחונה‬ṭḥwnh) relates to the
Aramaic noun “miller”, interpret it as an Arabic word “meaning something like
‘lame’ or ‘cripple’” based on the context (p. 230). It seems that they were con-
fused by the verb in the releasing recipe, ‫( ותהלך‬w-thlk), “and it shall move (lit.
walk)”, and did not think of the possibility of a mill “walking”, consequently
interpreting the text as referring to a person, not an installation. The Syriac
parallel clearly shows this to be otherwise.
The Genizah recipe for releasing a mill found on the verso of this leaf (2:1–5)
seems to be written by a different hand, or by the same scribe but in a very
different form. The letters are much larger, thicker and less cursive, and there
are no supralinear dots for emphasis. If the recipe was indeed written by the
same person as the previous one, the reason for the change in writing style is
unclear. We should note that the writing changes again on the following leaf,
in the middle of a recipe (3: 3).
The first line of the releasing recipe probably consisted of one word, its title,
which could have been the Aramaic ‫‘ לשריה‬To release it’, or a Judaeo-Arabic
version thereof. Such a title would stress its connection with the previous rec-
ipe for binding a mill. The practice is entirely oral, as opposed to the previous
one, which involved a manipulation of materials. It consists of a simple invoca-
tion of two magical names, that are asked to “release this mill”. The result (or
perhaps part of the spell to be uttered) is that the mill “shall walk”, that is, begin
to turn and function again.
The two magical names in the releasing recipe closely resemble those in the
binding one. However, while the binding recipe contains three variations of
the sequence ‫( פוניס‬pwnys), the releasing spell contains two variations of the
sequence ‫( פינוס‬pynws). It is probable that both spells were meant to invoke
similarly sounding names, built on the basis of ‫( פינוס‬suggesting that the releas-
ing spell preserved the more accurate forms). This assumption is based on the
parallel found in NYPL 190, where the magical names are clearly based on the
sequence ‫( פינוס‬pynws) and not on ‫( פוניס‬pwnys).
Both recipes on T-S Ar. 49.54 contain a mixture of languages: Judaeo-Arabic,
Aramaic, and a couple of Hebrew words. This situation will be discussed fur-
ther below, after presenting the second parallel.

Scrinium 18 (2022) 49–84

76 Cherkashina, Cherkashin and Saar

NYPL 190, fol. 180: 20–2376

https://digitalcollections.nypl.org/items/8cfbe0d0-a5ee-0133-e6a4-00505686d14e

This recipe is part of a paper manuscript of 258 pages, ca. 27 × 17 cm. Date: 1468.
Hand: Byzantine. Scribe: Moshe ben Yaakov ben Mordechai and Marḥaba.
The manuscript contains two works by Abraham Abulafia, Ḥaye ha-Nefesh
and Matsref ha-Ḥokhmah, followed by a long compendium of magical recipes
in a combination of Judaeo-Arabic, Aramaic and Hebrew (pp. 58–258 of the
manuscript). The magic manuscript, including the recipe below, was edited
by Gideon Bohak,77 who noted the partial parallel to T-S Ar. 49.54 but did not
provide an alternative interpretation.
The NYPL 190 recipe consists of a simple invocation of four magical names,
followed by a series of synonym verbs denoting binding effects on the mill. The
subsequent results, “it will not turn and not move” prove that the interpreta-
tion of ‫ טחונה‬as a mill is the correct one. The invocation is concluded by the
name of the great and holy God and an abbreviated sequence of Amen.

20 lˁqd ˀlṭˀḥwn tqwl hdh ‫ תקול הדה‬b‫לעקד אלטאחון‬a To bind a mill. Say these
ˀlˀsmˀ llpynws ˀlpynws ‫אלאסמא ללפינוס אלפינוס‬ names: “llpynws ˀlpynws
21 pynws wkpynws ˀtwn ‫פינוס וכפינוס אתון שמהתא‬ pynws and kpynws”, You
šmhtˀ qdyšyˀ ‫קדישיא‬ holy names,
22 ˀswrw hdˀ ṭḥwntˀ wkptwh ‫אסורו הדא טחונתא וכפתוה‬ bind this mill and tie it
wˀwqmwh wlˀ tswb wlˀ ‫ואוקמוה ולא תסוב ולא תרגיש‬ and stop it and it will
trgš bˀlhˀ ‫באלהא‬ not turn and it will not
move. By God
23 rbˀ wqydšˀ ˀ’ ˀ’ s’ '‫ רבא וקדישא א'א'ס‬the great and the holy.
A(men) A(men) S(ela).

a The recipe is numbered on the margin, ‫שפח‬, “378”. The title is repeated on the margin.
b Under the title a different hand inscribed two letters (?) that resemble ‫( רן‬rn) but their mean-
ing is unclear.

76 Gideon Bohak, A Fifteenth-Century Manuscript of Jewish Magic, Ms New York Public Library
Heb. 190 ( formerly Sassoon 56): Introduction Annotated Edition and Facsimile (in Hebrew),
2 vols., Los Angeles, 2014.
77 Ibid., Vol. 1, p. 222.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 77

As opposed to the Genizah recipe, NYPL 190 only consists of an incantation to

be uttered, whereas T-S Ar. 49.54 also involves a practical action (taking dust
and feeding it to the mill). The magical words to be uttered, however, resemble
each other and clearly reflect the same incantation:

T-S Ar. 49.54 NYPL 190 T-S Ar. 49.54

binding binding releasing

‫אלפוניס‬ ‫ללפינוס‬
‫לפוניס‬ ‫אלפינוס‬
‫פינוס‬ ‫אנפינוס‬
‫ואכפוניס‬ ‫וכפינוס‬ ‫אפנתפינוס‬

The core of the magic words in NYPL 190 and in the releasing recipe from T-S
Ar. 49.54 is ‫( פינוס‬pynws), and (probably) this also was the original form of the
core of the names of the T-S Ar. 49.54 binding recipe, where they were changed
in the course of its transmission. Given the aim of the recipe, namely binding
a mill, an installation meant to produce flour and consequently bread, could
this core have a related meaning? It is perhaps possible to relate it to the Greek
πεινών, meaning “hungry”. However, to date we have not found any Greek
magic recipes that are directed against a mill or a similar installation.
In terms of chronology, both the Aramaic text preserved in NYPL 190 and
that in the middle of the Judaeo-Arabic version from T-S Ar. 49.54 attest to
the antiquity of these spells. In the medieval period most Jews were no longer
using Aramaic. Hence, in both manuscripts the title and instructions of the
recipes are provided in Judaeo-Arabic, while the actual incantation remains in
the original Aramaic.78

78 For a broader discussion of the division of languages in magic texts, see Gideon Bohak,
“Towards a Catalogue of the Magical, Astrological, Divinatory and Alchemical Fragments
from the Cambridge Genizah Collections,” in: “From a Sacred Source”: Genizah Studies
in Honour of Professor Stefan C. Reif, ed. Ben Outhwaite and Siam Bhayro, Leiden, 2011,
pp. 53–79 (62–64, 68–69).

Scrinium 18 (2022) 49–84

78 Cherkashina, Cherkashin and Saar

9 Conclusions

In this study of the Syriac text-unit for a mill we have shown that the magical
practices attested in these recipes have a much broader cultural and historical
context. Comparing the Syriac texts with Jewish recipes, we can try to identify
the core of the binding and loosening practices shared by both magical tradi-
tions. In the case of the binding spell the core-practice is throwing dust into
the opening in the upper mill, while reciting the incantation. In the case of the
loosening spell, in both traditions Psalm 146:7 is used for this purpose.
Notably, both in the Syriac and in the Jewish text-units for a mill a part of
the recipe is written in Arabic. As we have mentioned before, Arabic was fre-
quently used as a language of both written and spoken communication for
Oriental Jews in medieval times (written often in the form of Judaeo-Arabic).
However, as has also been mentioned above, the Arabic language is a highly
marginal phenomenon for Syriac magical manuscripts. It is worth mentioning
here that the Arabic spell (verses 5–7) in the Syriac text-unit shows not only
the kind of scribal mistakes which occur during the transmission of a poorly
understood text, but also mistakes and discrepancies which clearly demon-
strate that this incantation was written down from hearing. In addition to the
arguments presented above, it should be outlined that unlike the Jewish recipes
for a mill, the Arabic spell found in the Syriac manuscripts contains clear indi-
cations of its circulation in Muslim context, cf. ˀaḷḷāh ˀakbar (l. 9), bi-xātimat?
ˀal-jinn (l. 11), sundūk bundūk (l. 8). Taken together this evidence suggests that
the binding practice for a mill was shared by Arabic-speaking Muslims, Jews
and Syriac Christians. The Arabic spell written down in Garshuni attests to oral
practice. To sum up, we do not assume, though it cannot be refuted either, that
the Arabic magical tradition was a source for this practice in both the Syriac
and the Jewish manuscripts. The available evidence allows us only to deduce
that similar practices occurred in Arabic, Syriac and Jewish magical traditions.

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 79

Appendix

Figure 1 MS Syriac 160, Houghton Library, Harvard University, f. 44r (H)

Scrinium 18 (2022) 49–84

80 Cherkashina, Cherkashin and Saar

Figure 2 BL Or. 5281, f. 40v (Blb)

© British Library Board (BL Or. 5281, f. 40v)

Scrinium 18 (2022) 49–84

Syriac Spells for a Mill and Their Historical Context 81

Figure 3 BL Or. 5281, f. 41r (Blb)

© British Library Board (BL Or. 5281, f. 41r)

Scrinium 18 (2022) 49–84

 82 Cherkashina, Cherkashin and Saar

Figure 4 Cambridge University Library T-S Ar. 49.54 (recto)

Reproduced by kind permission of the Syndics of Cambridge University
Library

Scrinium 18 (2022) 49–84

 Syriac Spells for a Mill and Their Historical Context 83

Figure 5 Cambridge University Library T-S Ar. 49.54 (verso)

Reproduced by kind permission of the Syndics of Cambridge University
Library

Scrinium 18 (2022) 49–84

84 Cherkashina, Cherkashin and Saar

Figure 6 NYPL 190, fol. 180: 20–23. Dorot Jewish Division, The New York Public Library
“Ḥaye ha-nefesh” New York Public Library Digital Collections.
Accessed April 19, 2022. https://digitalcollections.nypl.org/items
/8cfbe0d0-a5ee-0133-e6a4-00505686d14e

Scrinium 18 (2022) 49–84