Collectanea Christiana Orientalia 21 (2024): 97-120
Nick Posegay
University of Cambridge
An Early Arabic Translation of Exodus 15 from
a Palestinian Melkite Psalter in the Cairo Genizah
Introduction
1
The Cairo Genizah is famous for preserving vast quantities of Jewish manuscripts from
medieval Fustat (‘Old Cairo’) and the surrounding Islamicate world.
2
For decades, it has also
been known to contain a small, yet significant, subset of Christian manuscripts. The best
studied of these Christian manuscripts are palimpsests of the Gospels and other biblical texts
in Greek and Christian Palestinian Aramaic, some of which were published more than 120
years ago.
3
More recently, Christa Müller-Kessler has surveyed Genizah collections and
compiled a corpus of CPA palimpsests that now numbers several dozen fragments.
4
Syriac and
Coptic manuscripts are also present in the Cairo Genizah in small amounts.
5
The presence of
1
This research was funded by a Leverhulme Early Career Fellowship project, “Interfaith Exchange in the
Intellectual History of Middle Eastern Languages.”
2
A genizah (pl. genizot) is a hidden space used by Jewish communities to dispose of manuscripts, typically sacred
texts, that are too old or damaged for further use. The term ‘Cairo Genizah’ refers to a group of genizot centred
around the Ben Ezra Synagogue in Fustat (‘Old Cairo’) that saw use between the 11th and 19th centuries.
Most manuscripts from these genizot are now in American and European library collections. For the history of
the Cairo Genizah and details about its manuscript corpora, see Stefan C. Reif, A Jewish Archive from Old Cairo:
The History of Cambridge University’s Genizah Collection (London; New York: Curzon, 2000); Adina Hoffman and
Peter Cole, Sacred Trash: The Lost and Found World of the Cairo Geniza (New York: Nextbook, Schocken, 2011);
Nick Posegay, ‘Searching for the Last Genizah Fragment in Late Ottoman Cairo: A Material Survey of
Egyptian Jewish Literary Culture’, International Journal of Middle East Studies 54, no. 3 (2022): 423–41,
https://doi.org/10.1017/S0020743822000356; and Rebecca J.W. Jefferson, The Cairo Genizah and the Age of
Discovery in Egypt: The History and Provenance of a Jewish Archive (London: I.B. Tauris, 2022).
3
Agnes Smith Lewis and Margaret Dunlop Gibson, Palestinian Syriac Texts from Palimpsest Fragments in the Taylor-
Schechter Collection (London: C. J. Clay & Sons; Cambridge University Press, 1900).
4
Christa Müller-Kessler, ‘Recent Identifications among the Palimpsests from the Cairo Geniza: A
Comprehensive List of Christian Palestinian Aramaic Texts’, Collectanea Christiana Orientalia 20 (2023), pp. 118-
24.
5
For Syriac material, see Sebastian Brock, ‘East Syrian Liturgical Fragments from the Cairo Genizah’, in Oriens
Christianus, vol. 68 (Wiesbaden: Otto Harrassowitz, 1984), pp. 58-79; Sebastian Brock, ‘Some Further East
Syrian Liturgical Fragments from the Cairo Genizah’, in Oriens Christianus, 74 (Wiesbaden: Otto Harrassowitz,
1990), pp. 44-61; George A. Kiraz, ‘A Young Syriac Pupil in the Cairo Genizah: Or.1081 2.75.30’, Fragment of
the Month (August), Cambridge University Library: Genizah Research Unit, 2018,
Nick Posegay
98
fragments in these languages raises questions about the relationships between the Jews of
Cairo and the Christian communities that lived alongside them.
6
By far, however, the most
common Christian manuscripts in the Genizah are those written in Arabic, the common lingua
franca of Cairo’s Christians and Jews for most of the second millennium CE. Such manuscripts
include works of science, medicine, and philosophy by Christian authors, as well as theological
treatises and Bible translations.
7
In the last few years, scholars have given even greater attention to the corpus of Christian
Arabic Bible translations that survive in Genizah collections. As part of his landmark study on
Pentateuch translation, Ronny Vollandt identified more than 30 Genizah fragments of various
Old and New Testament books that Christians translated in the Middle Ages.
8
In 2022, Juan
Pedro Monferrer-Sala published a small Genizah fragment of John 19, dating it to
approximately the tenth century and arguing that it is based on a Syriac source text.
9
Then in
https://www.lib.cam.ac.uk/collections/departments/taylor-schechter-genizah-research-unit/fragment-
month/fotm-2018/fragment-6; Alan Elbaum, ‘A New Judaeo-Syriac Fragment from the Genizah: ENA
3846.2’, Fragment of the Month (February), Cambridge University Library: Genizah Research Unit, 2022,
https://www.lib.cam.ac.uk/collections/departments/taylor-schechter-genizah-research-unit/fragment-
month/fotm-2022/fragment-0. The Coptic portions of the Genizah have not been well studied, but see:
Michael Sokoloff and Joseph Yahalom, ‘Christian Palimpsests from the Cairo Genizah’, Revue d’histoire Des
Textes, no. 8 (1978), p. 110, p. 126; Marvin Meyer and Richard Smith, eds., Ancient Christian Magic: Coptic Texts
of Ritual Power (Princeton: Princeton University Press, 1999), pp. 197-199; Renate Smithuis, ‘A Short
Introduction to the Genizah Collection in the John Rylands Library’, in From Cairo to Manchester: Studies in the
Rylands Genizah Fragments, ed. Renate Smithuis and P.S. Alexander, Journal of Semitic Studies Supplement 31
(Oxford: Oxford University Press, 2013), p. 23.
6
On Christian communities in Fustat, see Audrey Dridi, ‘Christians of Fustat in the First Three Centuries of
Islam: The Making of a New Society’, in A Cosmopolitan City: Muslims, Christians, and Jews in Old Cairo, ed. Tasha
Vorderstrasse and Tanya Treptow, Oriental Institute Museum Publications 38 (Chicago: The Oriental
Institute of The University of Chicago, 2015), pp. 33-40; Audrey Dridi, ‘Christian and Jewish Communities in
Fusṭāṭ: Non-Muslim Topography and Legal Controversies in the Pre-Fatimid Period’, in The Late Antique
World of Early Islam: Muslims among Christians and Jews in the East Mediterranean, ed. Robert G. Hoyland (Berlin:
Gerlach Press, 2021), pp. 107-132.
7
Krisztina Szilágyi, ‘Christian Books in Jewish Libraries: Fragments of Christian Arabic Writings from the Cairo
Genizah’, Ginzei Qedem 2 (2006), pp. 107-162.
8
Ronny Vollandt, Arabic Versions of the Pentateuch: A Comparative Study of Jewish, Christian, and Muslim Sources (Brill,
2015), pp. 328-239, https://doi.org/10.1163/9789004289932; Ronny Vollandt, ‘Biblical Translations into
Christian Arabic Preserved in the Cairo Genizah Collections’, Biblia Arabica (blog), 2019, https://biblia-
arabica.com/biblical-translations-into-christian-arabic-preserved-in-the-cairo-genizah-collections/.
9
Writing for the Princeton Geniza Project, Samuel Bassaly and Peter Tarras estimated a date for this fragment
in the ninth century, based on its early script style (Princeton Geniza Project, T-S Misc.27.4.24b,
https://geniza.princeton.edu/documents/35301/, accessed 3 September 2023). Monferrer-Sala rightly points
out that the manuscript is made of paper, so a tenth-century date is more likely; Juan Pedro Monferrer-Sala, ‘A
Fragment of the Gospel of John Preserved in the Taylor-Schechter Genizah Collection’, Collectanea Christiana
Orientalia 19 (2022), p. 209.
An Early Arabic Translation of Exodus
99
2023, Peter Tarras published another fragment, this one from Revelation 4-5, which he argues
is based on a Coptic source and dates no earlier than the eleventh century.
10
This article presents yet another Christian translation from the Genizah, this time of
Exodus 15, preserved in two fragments of an Arabic psalter (MSS CUL T-S NS 305.198 and
T-S NS 305.210). The style of its Arabic script suggests that it was copied by a well-trained
scribe in the late 9th or early 10th century. Such a date makes it the oldest Christian Arabic
Bible translation yet found in the Genizah. Linguistic analysis further indicates that its
translator had access to the Peshitta and the Septuagint of Exodus 15 during their work. Most
likely, this translator was a ninth-century Melkite Christian who spoke both Syriac and Arabic.
Description
The two Genizah fragments of interest here are Cambridge University Library, Taylor-
Schechter New Series (T-S NS) 305.198 and T-S NS 305.210. Together they make up the
innermost bifolium of a parchment quire, measuring 15.3cm x 20.1cm (see figs. 1-2). Citations
from the reconstructed manuscript in this article will take the format 1r.1 (folio 1 recto, line 1).
Each leaf is about 10 cm wide, and both are torn, with several pieces missing from the middle
of f1 and the bottom of f2. Each page has 13-14 lines of Arabic text in a monochrome brown-
black ink (most likely iron gall, given the fading on 2v.13-14). A heading appears in red ink on
1v.6-7, and small red circles separate short textual units (approximately half-verses) throughout
the manuscript. Most of the text is also badly rubbed, in some places to the point of illegibility,
with 1r being practically indecipherable.
10
Peter Tarras, ‘A Fragment of the Book of Revelation in the Taylor-Schechter Genizah Collection (T-S AS
177.202)’, Collectanea Christiana Orientalia 20 (2023), p. 275.
Nick Posegay
100
Figure 1. T-S NS 305.198 and T-S NS 305.210 reconstructed, ff. 2v-1r.
An Early Arabic Translation of Exodus
101
Figure 2. T-S NS 305.198 and T-S NS 305.210 reconstructed, ff. 1v-2r.
11
Due to the extensive damage to the text, previous attempts to identify these fragments have
been unsuccessful. Avihai Shivtiel and Friedrich Niessen described the contents as a
“theological text with allusions to qur’anic [sic] verses”.
12
Consequently, Aleida Paudice
reproduced that description in her studies of Qurʾanic material in Genizah collections, and
11
Thank you to the Syndics of the Cambridge University Library for providing images of these fragments.
12
Avihai Shivtiel and Friedrich Niessen, Arabic and Judaeo-Arabic Manuscripts in the Cambridge Genizah Collections:
Taylor-Schechter New Series, Cambridge Genizah Series 14 (Cambridge: Cambridge University Press, 2006), no.
6795.
Nick Posegay
102
Ronny Vollandt did not examine this manuscript in his work on Christian Arabic Bible
translations.
13
A more accurate description is that these fragments are part of a Melkite Christian Arabic
psalter, specifically the beginning of the section containing the nine ‘canticles’ or ‘biblical odes’
that appear at the end of such psalters in the Orthodox Church.
14
The canticle that occupies
most of this article begins with the red heading on 1v.6-7, which reads: “The first song by
Miriam, the sister of Aaron, and Moses.” This title indicates the “Song of the Sea,” which
Moses and the people of Israel sing first in Exodus 15:1 before Miriam joins them in Exodus
15:20.
15
The rest of 1v.7-14, 2r, and 2v are the the text of Exodus 15:1-16, which breaks off in
the middle of verse 16 (2v.14). This text presumably continued on the next leaf of this codex
through the end of the song in verse 19, which was followed by the “second song” of Moses
from Deuteronomy 32:1-43. The discussion below focuses only on the Exodus translation.
The damage to the manuscripts has prevented me from deciphering and positively
identifying the other texts which precede the Song of the Sea (i.e. 1r.1-14 and 1v.1-5). Lines
1r.1-14 may be the Arabic text of Psalms 149 and 150 (or 150 and 151), with the green circle in
the margin marking the division between the two. The five lines before the rubricated heading
(1v.1-5) would then be a colophon marking the end of the book of Psalms (including one of
the few legible phrases, in 1v.4, ا ‘the resurrection’). This layout would correspond to
the arrangement of other Arabic psalters, with the Song of the Sea also following a colophon
at the end Psalm 150/151 in Bryn Mawr College MS BV 47, ff. 71v-72r (916-17 CE), Sinai
Arabic 32, ff. 116v-117r (ca. eleventh century), and Sinai Arabic 52, ff. 221v-222r (ca. twelfth
century). Such psalters also tend to have rubricated headings that label each canticle with the
13
Aleida Paudice, ‘On Three Extant Sources of the Qur’an Transcribed in Hebrew’, European Journal of Jewish
Studies 2, no. 2 (1 December 2008), p. 241, https://doi.org/10.1163/187247109X454422; Aleida Paudice,
‘Hebrew Translations and Transcriptions of the Qurʾan’, in A History of Jewish-Muslim Relations: From the Origins
to the Present Day, ed. Abdelwahab Meddeb and Benjamin Stora (Princeton: Princeton University Press, 2013),
p. 646; Vollandt, Arabic Versions of the Pentateuch, pp. 328-329; Vollandt, ‘Biblical Translations into Christian
Arabic’.
14
Traditionally, in the Orthodox Church, the First Song of Moses (Exodus 15:1-19), the Second Song of Moses
(Deuteronomy 32:1-43), the Prayer of Hannah (1 Samuel 2:1-10), the Prayer of Habakkuk (Habakkuk 3:1-19),
the Prayer of Isaiah (Isaiah 26:9-20), the Prayer of Jonah (Jonah 2:2-9), the Prayer of the Three Holy Children
(Daniel 3:26-56), the Song of the Three Holy Children (Daniel 3:57-88), and the Magnificat and Benedictus (Luke
1:46-55 and Luke 1:68-79). Some psalters include the Song of Simeon (Luke 2:29-32) as an additional canticle
(e.g. MS Bryn Mawr College Library BV 47, f. 73v). The First Song of Moses is also attested as the First Song
of Miriam (as in these Genizah fragments and Bryn Mawr College Library BV 47, f. 72r).
15
“Then Miriam, the prophetess, the sister of Aaron, took a tambourine in her hand, and all the women went
out after her with tambourines and dancing. And Miriam sang to them…” (Exodus 15:20-21, ESV). See
edition below for the Arabic text of the heading.
An Early Arabic Translation of Exodus
103
term (calqued from Syriac ܬ) and often use dots or red circles to separate
verses.
16
The Genizah fragments accordingly exhibit all of these features.
There are no vocalisation signs in the manuscript, no hamzas, no diacritics like shadda or
sukūn, no ihmāl signs, and only sporadic diacritic dots. Despite what has been suggested for
other Christian Arabic manuscripts in the Genizah, the absence or inconsistent application of
consonantal diacritics in medieval Arabic manuscripts is so common that it should be expected
regardless of whether a manuscript is copied in Qurʾanic, Classical, or ‘Middle’ Arabic.
17
Consequently, the absence of a diacritic feature in a manuscript cannot be taken as evidence
for the absence of a phonological feature in speech. Therefore, the absence of many dots (as
well as signs like shadda and hamza) in this psalter does not tell us anything about the scribe’s
Arabic dialect. The one diacritic system that is somewhat useful for dating is this scribe’s
consistent use of a single supralinear dot for fāʾ and a pair of supralinear dots for qāf. This
system only became widespread among Arabic scribes during the ninth century. The other
diacritics that are present do not demonstrate any of the older Arabic practices – such as the
qāf with a single dot – that can be used to date manuscripts to the period before the 9th
century.
18
As such, we must rely on palaeographic analysis to estimate a date of production
after that. Such analysis in comparison with other dated Christian Arabic manuscripts suggests
that these fragments were copied in the late ninth or early tenth century.
16
See also Sinai Arabic 21, 143r-144v (ca. eleventh century), Sinai Arabic 31 (ca. twelfth century), and Sinai
Arabic 53, ff. 141v-142r (ca. twelfth century).
17
Diacritics can be useful evidence for detecting the interference of dialectal features in the writing of Arabic
scribes, but primarily when scribes include dots that are not ‘correct’ in Classical Arabic. For example, if instead
of ‘three’, a scribe wrote using two dots on each tāʾ, that would be evidence to suggest that their dialect
lost the interdental fricative /th/ typically represented by ث. However, if the scribe left out the dots entirely
and wrote ﻪﯩﻠﯨ, we could not use that as evidence that their dialect had merged the interdental fricative /th/
and plosive /t/. It only shows that the scribe, in that instance, did not think that the reading was sufficiently
ambiguous to warrant the inclusion of distinguishing dots. Similarly, if a scribe wrote the word د ‘he
entered’ incorrectly as ذ, the added diacritic dot on the dāl may be evidence of a ‘hypercorrection’. That is,
the scribe’s Arabic dialect may have lost the distinction between the alveolar stop /d/ and fricative /dh/, but
they were aware that those sounds were differentiated in Classical Arabic. Not knowing exactly when that
difference occurs, they overcorrected by adding a dot where it does not belong. Contrast the discussion of
diacritics in Monferrer-Sala, ‘A Fragment of the Gospel’, 209.
18
Beatrice Gruendler, ‘Arabic Script’, in Encyclopedia of the Qurʾān, ed. J.D. McAuliffe (Leiden: Brill, 2001), 140;
Andreas Kaplony, ‘What Are Those Few Dots for? Thoughts on the Orthography of the Qurra Papyri (709-
710), the Khurasan Parchments (755-777) and the Inscription of the Jerusalem Dome of the Rock (692)’,
2008, 93–94; Miriam L. Hjälm, ‘A Paleographical Study of Early Christian Arabic Manuscripts’, Collectanea
Christiana Orientalia 17 (2020): 68, https://doi.org/10.21071/cco.v17i0.1148.
Nick Posegay
104
Palaeography and Dating
The script of T-S NS 305.198/T-S NS 305.210 is a relatively unadorned semi-cursive style with
features of several different scripts that Miriam Hjälm has identified in her classification of
Christian Arabic palaeography. Some of these features are characteristic of her “semi-angular”
groups as well as the “cursive” subtype of New Style. For example, the script in this Genizah
psalter exhibits significant horizontal extension, particularly in the letters kāf and ṣad/ḍād,
typical of the scripts that Hjälm calls “semi-angular” Group A. Yet the psalter script is also less
angular than Group A, somewhat resembling Group B and more cursive styles.
19
It further
exhibits certain letter forms typical of a cursive New Style – including an s-shaped independent
alif and ṭāʾ with a rightward-leaning shaft that extends far past its belly – but it lacks the vertical
extension characteristic of the New Styles.
20
As Hjälm and others have shown, all of these
related script types are attested in Christian manuscripts already in the late 9th and early 10th
centuries.
21
The best comparison for the Genizah psalter is what Hjälm designates “plain” scripts,
represented by just a few manuscripts which lack the typical features of angular scripts, yet do
not appear to be based on New Style developments: Sinai Arabic 2 (dated 939/40), Sinai
Arabic 151 (dated 867), and Sinai Arabic 597 (dated before 1002).
22
Of these three, Sinai
Arabic 2 is the only one whose date has not been called into doubt. Sinai Arabic 151 is often
regarded as the earliest dated Christian Arabic Bible, but some scholars question whether the
colophon dating it to 867 might be a copy of an earlier manuscript. Noting the similarities
between its script and Sinai Arabic 2, Alexander Treiger has proposed that Sinai Arabic 151
should be redated to the early tenth century.
23
Additionally, he has shown that the 1002 CE
date commonly cited for Sinai Arabic 597 does not belong with the original ‘plain’ script hand
in that manuscript. He thus also redates that manuscript to the early 10th century based on its
palaeographic similarities to Sinai Arabic 2.
24
19
Hjälm, ‘A Paleographical Study’, pp. 52-53, 56 and pp. 64-69.
20
Hjälm, ‘A Paleographical Study’, pp. 64-65.
21
See also Mark N. Swanson, ‘Some Considerations for the Dating of Fī Taṯlīth Allāh Al-Wāḥid (Sinai Ar. 154)
and al- Ǧāmiʿ Wuǧūḥ al-Īmān (London, British Library Or. 4950)’, Parole de l’Orient, no. 18 (1993), pp. 115-
141; Hikmat Kashouh, The Arabic Versions of the Gospels: The Manuscripts and Their Families, Arbeiten Zur
Neutestamentlichen Textforschung 42 (Berlin: De Gruyter, 2012).
22
Hjälm, ‘A Paleographical Study’, pp. 73-74.
23
Joshua Blau, ‘Über Einige Christlich-Arabische Manuskripte Aus Dem 9. Und. 10. Jahrhundert’, Le Muséon:
Revue d’Études Orientales 75, no. 1–2 (1962), pp. 101-108; Vollandt, Arabic Versions of the Pentateuch, 27, n. 24;
Alexander Treiger, ‘From Theodore Abū Qurra to Abed Azrié: The Arabic Bible in Context’, in Senses of
Scripture, Treasures of Tradition: The Bible in Arabic among Jews, Christians and Muslims, Biblia Arabica 5 (Leiden:
Brill, 2017), p. 40; Hjälm, ‘A Paleographical Study’, p. 72, n. 72.
24
Treiger, ‘From Theodore Abū Qurra’, p. 42, n. 128.
An Early Arabic Translation of Exodus
105
More recent analysis by Vevian Zaki suggests that Sinai Arabic 151 is likely not a copy of an
earlier manuscript,
25
and therefore the extant colophon is original to the codex in which it
appears. Hjälm thus accepts its 867 date. She also stresses the differences between Sinai Arabic
151 and the other two manuscripts in this group, suggesting that even if they belong to the
first half of the tenth century, it is plausible that Sinai Arabic 151 is earlier. One detail she
highlights is that Sinai Arabic 151 lacks the top stroke of final kāf, whereas the stroke appears
more often in Sinai Arabic 2 and 597.
26
This is a feature that Sinai Arabic 151 shares with the
Genizah psalter:
Tables 1-8. T-S NS 305.198/T-S NS 305.210 compared to other ‘plain’ hands
27
In the ‘plain’ style, including for T-S NS 305.198/T-S NS 305.210, initial and medial kāf is
typically a pair of parallel horizontal strokes with a short oblique top stroke. There is often
considerable horizontal extension in the base. Final kāf resembles final dāl, with no top stroke
or only a secondary stroke detached from the body of the letter.
Initial
kāf
Final
kāf
T
-
S NS 305.19
8/T
-
S NS
305.210
Sinai Arabic 151
Sinai Arabic 2
25
Vevian F. Zaki, ‘A Dynamic History: MS Sinai, Arabic 151 in the Hands of Scribes, Readers, and Restorers’,
Journal of Islamic Manuscripts, 2020, pp. 219-220, https://doi.org/10.1163/1878464X-01102004.
26
Hjälm, ‘A Paleographical Study’, p. 73, note 74.
27
Samples from Sinai Arabic 151 ff. 6v-8r, Sinai Arabic 2 ff. 106v-107r, and Sinai Arabic 597 f. 12r-v.
Nick Posegay
106
Sinai Arabic 597
The shape of alif varies considerably in the plain scripts. Some demonstrate the s-shape typical
of NS scripts, but others have only a slight righthand return at the baseline or no return at all.
There is also significant variation in the height of alif, but in general, the plain style lacks the
vertical extension seen in the ascending strokes of NS.
Independent
alif
T
-
S NS
305.198/T
-
S
NS 305.210
Sinai Arabic 151
Sinai Arabic 2
Sinai Arabic 597
An Early Arabic Translation of Exodus
107
Plain scripts display the ‘gamma’ shaped lām-alif, even though this ligature is often considered a
later feature.
28
It can lean right or left, though the degree of obliqueness varies between
manuscripts.
Lām
-
alif
ligature
T
-
S NS 305.198/T
-
S
NS 305.210
Sinai Arabic 151
Sinai Arabic 2
Sinai Arabic 597
Like kāf in initial and medial positions, ṣād/ḍād consists of two parallel lines that make up a
narrow belly with considerable horizontal extension. In some cases, especially in Sinai Arabic
151 (and to a lesser extent, the Genizah psalter), the belly may be pinched short instead. Tails
also tend to be short and the typical lefthand denticle is minimal or absent.
ṣ
ād/
ḍ
ād
T
-
S NS 305.198/T
-
S
NS 305.210
28
Hjälm, ‘A Paleographical Study’, 73.
Nick Posegay
108
Sinai Arabic 151
Sinai Arabic 2
Sinai Arabic 597
Ṭāʾ/ẓāʾ has a similar horizontal extension to ṣad and kāf, and the belly of the letter may also be
pinched short. The shaft curves slightly and leans heavily to the right. This obliqueness is
present in all four ‘plain’ manuscripts, but it is most extreme in the Genizah psalter, where the
shaft can extend far past the belly of the letter.
Ṭāʾ
T
-
S NS 305.198/T
-
S
NS 305.210
Si
nai Arabic 151
An Early Arabic Translation of Exodus
109
Sinai Arabic 2
Sinai Arabic 597
Final bāʾ and its homographs display a typical ‘half-bowl’ shape that usually lacks a finishing
band, although this feature is less common in Sinai Arabic 597.
Final/independent
bāʾ/tāʾ/thā’
T
-
S NS 305.198/T
-
S
NS 305.210
Sinai Arabic 151
Sinai Arabic 2
Sinai Arabic 597
The head of initial fāʾ is frequently lifted off the baseline, with a small counter that is often
closed or nearly closed. The fī ligature is consistent throughout the script type, with a sharp
downward stroke that that connects the head to a far-right extended return.
fāʾ
T
-
S NS 305.198/T
-
S
NS 305.210
Sinai Arabic 151
Nick Posegay
110
Sinai Arabic 2
Sinai Arabic 597
Finally, dāl lacks the top serif typical of Early Abbasid script styles. It often appears as a simple
semicircle.
dāl
T
-
S NS 305.198/T
-
S NS 305.210
Sinai Arabic 151
Sinai Arabic 2
Sinai Arabic 597
Based on the comparisons here, it seems that Sinai Arabic 2 is the latest of the four
manuscripts in this group. It has notably less horizontal extension (particularly with initial kāf),
a more modern final kāf shape, a less oblique ṭāʾ, and is generally more curvilinear than the
other hands. This assessment concurs with Treiger’s conclusion that Sinai Arabic 597 predates
Sinai Arabic 2 and can be placed earlier in tenth century, before 939/940. It is also relevant
that both Sinai Arabic 2 and 597 are made of paper, whereas Sinai Arabic 151 and the Genizah
psalter are parchment. While not proof of their chronology, paper only gained widespread
adoption in Iraq, Syria, and Egypt in the ninth and tenth centuries.
29
Paper Arabic manuscripts
thus tend to be later than parchment manuscripts in this period, so the Genizah psalter is likely
older than both Sinai Arabic 2 and Sinai Arabic 597. Hjälm’s survey suggests a general shift
29
Adam Gacek, Arabic Manuscripts: A Vademecum for Readers (Leiden; Boston: Brill, 2009), p. 186; François
Déroche et al., Islamic Codicology: An Introduction to the Study of Manuscripts in Arabic Script, ed. Muhammad Isa
Waley, trans. Deke Dusinberre and David Radzinowicz, 2nd edition (London: Al-Furqān Islamic Heritage
Foundation, 2015), p. 51.
An Early Arabic Translation of Exodus
111
among Arabic monastic scribes from parchment to paper after about 920 CE.
30
We therefore
estimate that the Genizah psalter was produced between 867 and 920 CE.
Linguistic Evidence of Source Text
As we will see, the translation of Exodus 15 that appears in this psalter is most likely based on
a Syriac source mediated by the Septuagint. Ronny Vollandt has identified four of the most
attested Syriac-based Arabic Pentateuch text types from roughly the period of the Genizah
psalter. He designates them Arab
Syr
1, Arab
Syr
2, Arab
Syr
3, and Arab
Syr_Hex
1a.
31
The first three –
Arab
Syr
1, Arab
Syr
2, and Arab
Syr
3 – are all based on the Syriac of the Old Testament Peshitta,
itself originally translated from a Hebrew vorlage in the second century CE. None of them
match the translation that appears in the Genizah psalter.
32
In contrast to these Peshitta types,
Arab
Syr_Hex
1a is an Arabic translation based on the Syriac of the “Syro-Hexapla.” Paul of Tella,
a Syriac Orthodox bishop, produced the Syro-Hexapla in the early seventh century by
translating the Greek Septuagint version from Origen’s Hexapla into Syriac. Then, sometime
before 956 CE, the Melkite al-Harith ibn Sinan ibn Sunbat al-Harrani translated the Syro-
Hexapla into Arabic, producing Arab
Syr_Hex
1a.
33
His version of Exodus 15 also does not match
the Genizah psalter.
34
The present author has also compared the Genizah translation to five
additional Arabic Psalter manuscripts containing the nine canticles which date between the
tenth and twelfth centuries. While all five have considerable lexical and syntactic similarities to
the Genizah psalter fragments in their versions of the Song of the Sea, they are nevertheless
separate translations. The rubricated heading of the Genizah psalter, with its reference to
“Miriam, the sister of Aaron,” is most similar to that of Bryn Mawr College Library BV 47
(f.72r), a Melkite psalter dated 916-17 CE.
35
Further research is needed to understand the
relationships between these related canticle translation traditions and the Genizah psalter.
Even though the Genizah psalter does not correspond to any of these text types, two
circumstantial details support the hypothesis that it is based on a Syriac source. First, the
earliest dated example of the “plain” script style (see “Palaeography” above), Sinai Arabic 151,
is itself a biblical translation for several books from the New Testament. Its colophon specifies
30
Hjälm, ‘A Paleographical Study’, pp. 76-77.
31
Vollandt, Arabic Versions of the Pentateuch, pp. 244-263.
32
Based on comparison with manuscripts containing Exodus 15 from Arab
Syr
1 (Sinai Arabic 2, f.105v) and
Arab
Syr
2 (Sinai Arabic 4, f.85v). See Vollandt, Arabic Versions of the Pentateuch, pp. 244-245.
33
Vollandt, Arabic Versions of the Pentateuch, pp. 60-61.
34
Based on comparison with Sinai Arabic 10, f.74r. See Vollandt, Arabic Versions of the Pentateuch, 253.
35
The five other psalters compared here are Bryn Mawr College Library BV 47, ff. 71v-72r (916-17 CE); Sinai
Arabic 21, ff. 143r-144v (ca. eleventh century); Sinai Arabic 32, ff. 117r-118r (ca. eleventh century); Sinai
Arabic 52, ff. 222r-223v (ca. twelfth century); and Sinai Arabic 53, ff. 141v-142r (ca. twelfth century). On the
earliest extant Arabic psalters, see Treiger, ‘From Theodore Abū Qurra’, pp. 20-21.
Nick Posegay
112
that it was copied from a Syriac vorlage.
36
Second, Sinai Arabic 2, the example of the “plain”
script type dated to 939/40 CE, is one of the oldest witnesses to the Arab
Syr
1 translation type.
37
Textual correspondences between various versions of Exodus 15:1-16 and the Genizah
psalter further suggest that its translator had access to both the Peshitta and the Septuagint.
Due to the damage in the manuscript, it is often difficult to reconstruct the syntax of entire
sentences, so this analysis relies on the comparison of individual words in the Peshitta, Syro-
Hexapla, Septuagint, and Masoretic Hebrew Bible. This discussion abbreviates these sources
with the sigla P (Peshitta), S-H (Syro-Hexapla), LXX (Septuagint), MT (Masoretic text), and
GP (Genizah psalter). An edition of Exodus 15:1-16 from the psalter fragments and parallel
translations from these potential source versions appear in Table 9. Linguistic observations
follow below.
36
Vollandt, Arabic Versions of the Pentateuch, 59, n. 52; Joseph Nasrallah, ‘Deux Versions Melchites Partielles de La
Bible Du IXe et Du Xe Siècles’, Oriens Christianus 64 (1980), pp. 202-215.
37
Vollandt, Arabic Versions of the Pentateuch, p. 245.
An Early Arabic Translation of Exodus
113
Table 9
38
. Edition of Exodus 15:1-16 in the Genizah Psalter Translation and Four Potential
Source Versions.
The links between the Genizah psalter and the Peshitta begin already in Exodus 15:1a, “Then
Moses and the children of Israel sang this song to the Lord” (MT: ל ֵ֜א ָר ְשִׂי י ֵ֨נ ְבוּ ֩ה ֶשֹׁמ־רי ִֽשָׁי ז ֣ ָא
הָ֔וֹהיֽ ַל ֙תאֹ זּ ַה ה ֤ ָרי ִשּׁ ַה־ת ֶא). The Hebrew designates the “song” of the sea with the word ה ָרי ִשּׁ ַה,
literally ‘song’. The translators of the Peshitta and Syro-Hexapla both render this noun with the
Syriac ܬ, meaning ‘hymn’ or ‘praise’, which is also the typical deignation for the nine
canticles. The Genizah psalter applies the cognate Arabic root sbḥ for the Syriac šbḥ, glossing
ܬ as ]اﺘﻟ[ (1v.8). However, while the P and GP also use this ‘praise’ root to
translate the Hebrew verb רי ִֽשָׁי (P: ‘they praised’; GP: ]ﺒﺳ[ﺢ ‘he/they praised’, 1v.7), the
S-H and LXX do not (S-H: ܘܙ ‘they sang’, esp. of Psalms; LXX: ᾖσεν). The same glosses
appear again in verse 15:1b, “I will sing to the Lord, for he has triumphed gloriously” (MT:
ה ָרי ֤ ִשׁ ָא ֙ה ָוֹהיֽ ַל ה ֹ֣א ָג־י ִֽכּ ה ָ֔א ָגּ ). The S-H translates ה ָרי ֤ ִשׁ ָא as ‘we will sing’ (LXX: ασωμεν) while
the P and GP have ‘we will praise’ (P: ــ; GP: , 1v.9).
This correlation between the Syriac šbḥ and Arabic sbḥ roots is already strong evidence that
the GP translation was based primarily on the P, but there are other orthographic, syntactic,
and lexical indicators. In Exodus 15:8a, the MT reads “At the blast of your nostrils the waters
gathered together” (םִי ַ֔מ וּמ ְר ֶע֣ ֶנ ֙י ֶ֙פּ ַא ַוּ ֤ר ְבוּ). Instead of ‘your nostrils’ here, the S-H follows the
LXX’s θυμοῦ σου ‘your wrath’, giving ܕ . By contrast, the P mimics the orthography
and plurality of the MT with
ܐ ‘your faces’, and the GP simply gives the Arabic cognate of
the Hebrew, ا ‘nose’ (2r.9). Then the MT of Exodus 15:8b has “the depths congealed in the
heart of the sea” (םֽ ָי־ב ֶל ְבּ ת ֹ֖מֹה ְת וּ ֥א ְפ ֽ ָק), and the P again mimics the Hebrew with ــ ‘they
congealed, coagulated’. The GP translates this verb as ﺟ]ﻤ[اود ‘they froze, coagulated’ (2r.10),
matching the P in both sense and plurality. Meanwhile, the S-H has the singular ــ ‘it
condensed’ (LXX: ἐπάγη), with a sense more typically applied to fog or vapour. In verse
15:15a, the MT reads, “Then the chiefs of Edom will be terrified, trembling will seize the
leaders of Moab (ד ַע ֑ ָר וֹמ֖ ֵז ֲחאֹֽ י ב ָ֔אוֹמ י֣ ֵלי ֵא םוֹ ֔ד ֱא י֣ ֵפוּלּ ַא ֙וּל ֲה ְב ִנ ז ֤ ָא). The MT, P, and GP all refer to ‘the
leaders of Moab’ here, (MT: ב ָ֔אוֹמ י֣ ֵלי ֵא; P: ܒܐܕ
; GP: ب ارا, 2v.12), but the S-H
has ‘the leaders of the Moabites’ (S-H:
ܖ
ܐܕ ; LXX: ἄρχοντες Μωαβιτῶν).
Additionally, both the P and GP give cognates of the MT’s verb וֹמ֖ ֵז ֲחאֹֽ י ‘it will seize them’ (P:
ܢــܐ ܐ; GP: ا, 2v.12), while the S-H does not (ܢــܐ ; LXX: ἔλαβεν).
While it is likely that the Genizah psalter’s translator based their work on the Peshitta,
several lexical details suggest that they also had access to the Septuagint. In Exodus 15:4a, the
38
For table 9, please see appendix at the end of this article.
Nick Posegay
114
MT states that Pharaoh’s army drowned “in the sea of reed” (ףוּ ֽס־םַי ְב). The P adapts this phrase
directly into Syriac, giving ܦܕ ‘in the sea of reed’. The S-H, however, follows the LXX
ἐν ἐρυθρᾷ θαλάσσῃ ‘in the red sea’, glossing it as ‘in the red sea’. The GP matches
the S-H and LXX here: ا ا ][ ‘[in] the red sea’ (2r.1-2). Then, while the GP does
match the syntax of the MT and P regarding ‘the leaders of Moab’ (see above), it calls those
leaders ارا (2v.12), using an Arabicised broken plural form of the LXX’s ἄρχοντες ‘archons’.
This loan is unrelated to the equivalent Syriac glosses (P:
; S-H:
ܖــ) and can only have
come from the translator knowing a Greek version of this verse. The combination of Greek
and Syriac sources is a hallmark of Melkite Bible translation,
39
and these lexical connections
support the conclusion that the translator of the Genizah psalter was a multilingual Melkite
Christian.
Conclusion: Melkite Provenance and the Cairo Genizah
The combination of paleographic, codicological, and linguistic evidence indicates that the
manuscript made up of T-S NS 305.198 and T-S NS 305.210 is a Melkite psalter produced in
the late ninth or early tenth century. The translator’s primary source text was the Syriac
Peshitta, but, like the Genizah Gospel fragment published by Monferrer-Sala,
40
their
translation was mediated by Greek sources, specifically the Septuagint. At the very least, the
translator was aware of alternate glosses from the Septuagint and incorporated them into their
Arabic translation of the Peshitta. The psalter’s script style is most similar to Sinai Arabic 151
(dated 867 CE), another Arabic Bible manuscript that belongs to a Melkite liturgical tradition
and contains a translation based on a Syriac source.
41
It is thus most likely that the Genizah
psalter comes from from a multilingual Melkite monastery in Palestine that was active during
the ninth century.
42
This origin would be consistent with other Christian material in the
Genizah, particularly the Greek and Christian Palestinian Aramaic palimpsests, that are
39
Vollandt, Arabic Versions of the Pentateuch, p. 53.
40
Monferrer-Sala, ‘A Fragment of the Gospel’, p. 211.
41
Zaki, ‘A Dynamic History’, p. 221 and p. 246. Previous scholars have argued that the translator of Sinai Arabic
151, Bishr ibn al-Sirrī, was also a Melkite, but recent research indicates he was an East Syriac Christian.
Compare Nasrallah, ‘Deux Versions Melchites Partielles de La Bible Du IXe et Du Xe Siècles’, pp. 203-206;
Samir Khalil Samir, ‘Michel Évêque Melkite de Damas Au 9e Siècle. A Propos de Bišr Ibn al-Sirrī’, Orientalia
Christiana Periodica 53 (1987), pp. 439-441; and Sidney Griffith, The Bible in Arabic: The Scriptures of the ‘People of
the Book’ in the Language of Islam (Princeton; Oxford: Princeton University Press, 2013), p. 134; with Zaki, ‘A
Dynamic History’; Habib Ibrahim, ‘Revisiting the Works of Ibn Al-Sirrī (9th Century)’, Scrinium 19, no. 1
(2023), pp. 28-48; and Alexander Treiger, ‘An East-Syriac Scholar in Ninth-Century Damascus’, Scrinium 19,
no. 1 (2023), pp. 388-413, https://doi.org/10.1163/18177565-bja10082.
42
See Kate Leeming, ‘The Adoption of Arabic as a Liturgical Language by the Palestinian Melkites’, ARAM
Periodical 15 (2003), pp. 239-246; Vollandt, Arabic Versions of the Pentateuch, pp. 54-55; Monferrer-Sala, ‘A
Fragment of the Gospel’, p. 209.
An Early Arabic Translation of Exodus
115
suspected to come from Palestinian Christian communities between the seventh and ninth
centuries CE.
43
However, since the psalter does not show any signs of erasure or reuse, it is
also plausible that it came from a Melkite community in the immediate area of Fustat before
ending up in the hands of Cairo’s Jews.
44
The question of why Egyptian Jews would have obtained Christian material and deposited it
into a Cairene genizah is one that remains unresolved. While the surviving psalter fragments
show no signs of recycling, it is possible that Jewish bookmakers repurposed the rest of the
quire as a palimpsest or to reinforce other bindings. It is also possible that an Arabic-speaking
Jew simply wanted a professional copy of the Song of the Sea in Arabic and was not picky
about who produced it. The same can be said for many Arabic scientific and medical works
produced by Christians and Muslims that now reside in Cairo Genizah collections.
45
On the
other hand, one of the numerous collectors who acquired manuscripts for the Cambridge
Genizah Collections could have purchased the psalter fragments from dealers in Egypt or
Palestine, with only tenuous connections to the Jews of Fustat.
46
Regardless of their exact
provenance though, these fragments represent new data for the study of Christian Arabic
paleography, material history, and Bible translation in the ninth and tenth centuries.
43
See Lewis and Gibson, Palestinian Syriac Texts; Sokoloff and Yahalom, ‘Christian Palimpsests’, pp. 110-111;
Müller-Kessler, ‘Recent Identifications’.
44
On Melkites in and around Fustat, see Dridi, ‘Christians of Fustat in the First Three Centuries of Islam: The
Making of a New Society’, p. 38 and p. 40.
45
Szilágyi, ‘Christian Books in Jewish Libraries’; Mark R Cohen, ‘Geniza for Islamicists, Islamic Geniza, and the
“New Cairo Geniza”’, Harvard Middle Eastern and Islamic Review, no. 7 (2006), pp. 129-245; Nick Posegay, ‘The
Long Road to Samarqand: Reverse-Engineering the Travels of a 12th-Century Andalusi Muslim (T-S
Ar.53.39)’, Fragment of the Month (October), Cambridge University Library: Genizah Research Unit, 2023,
https://www.lib.cam.ac.uk/collections/departments/taylor-schechter-genizah-research-unit/fragment-
month/fotm-2023/fragment-8. See also, Magdalen M. Connolly and Nick Posegay, ‘A Survey of Personal-Use
Qurʾ an Manuscripts Based on Fragments from the Cairo Genizah’, Journal of Qur
ʾ
anic Studies 23, no. 2 (2021),
pp. 2-4, https://doi.org/10.3366/jqs.2021.0465.
46
Rebecca J.W. Jefferson, ‘The Trade in Cairo Genizah Fragments in and out of Palestine in the Late 19th and
Early 20th Centuries’, Journal of Ancient Judaism 14, no. 2 (2023), pp. 169-171,
https://doi.org/10.30965/21967954-bja10046.
Edition of Exodus 15:1-16 in the Genizah Psalter Translation and Four Potential SourceVersions
1
Septuagint
Syro-Hexapla
Masoretic Text
Peshitta
Psalter f1 verso
Verse
Line
6
,
1a
7
8
9
᛫
end 1
10
11
end 2
12
.
.
3-4a
13
4a
14
1
The text of the Genizah psalter is my edition of T-S NS 305.198 and T-S NS 305.210. The Peshitta text comes fromSamuel Lee’s Vetus Testamentum Syriace(London,
1823), the Masoretic Text is from the Westminster edition of the Leningrad Codex (http://www.tanach.us/Tanach.xml, accessed 1 March 2024), the Syro-Hexapla
text isfrom Paul de Lagarde’s Bibliothecae Syriacae (Göttingen, 1892; p. 67),and the Septuagint text is from Alfred Rahlfs’ Septuaginta: id est Vetus Testamentum Graece
iuxta LXX Interpretes (Stuttgart, 1935).
Nick Posegay
116
Septuagint
Syro-Hexapla
Masoretic Text
Peshitta
Psalter f2 recto
Verse
Line
.
4b
1
end 4-
5a
2
end 5
3
6a
4
2
end 6
5
ﻟ
7a
6
7b
7
end
7b
8
8a
9
8b
10
9a
11
9a-9b
12
9b
13
2
The final in [] here is likely a scribal error for , and should be read ‘you have blunted, sullied’. This verb corresponds with the Peshitta ‘you have
broken’ (Exodus 15:6).
An Early Arabic Translation of Exodus
117
end
9b-
10a
14
Septuagint
Syro-Hexapla
Masoretic Text
Peshitta
Psalter f2 verso
Verse
Line
.
end
10a-
10b
1
ﻢ
end
10b-
11a
2
11a
3
4
end
11b
5
.
12a
6
12b-
13a
7
13a-
13b
8
13b-
14a
9
14a-
14b
10
14b-
15a
11
Nick Posegay
118
:
:
15a
12
15a-
15b
13
15b-
16a
14
An Early Arabic Translation of Exodus
119
Nick Posegay
120
Abstract:
T
his article presents an Arabic
translation of Exodus 15 from the Cairo
Genizah, preserved in two fragments of a
Christian Psalter (MSS CUL T-S NS 305.198
and T-S NS 305.210). The style of the
Psalter’s Arabic script suggests that it was
copied by a well-trained scribe in the late 9th
or early 10th century. Such a date makes it
the oldest Christian Arabic Bible translation
yet found in the Genizah. Linguistic analysis
further indicates that its translator had access
to both the Peshitta and the Septuagint of
Exodus 15 during their work. Most likely, this
translator was a ninth-century Palestinian
Melkite who spoke Syriac and Arabic.
Resumen:
En este artículo se presen
ta una
traducción árabe de Éxodo 15 de la
Genizah de El Cairo, conservada en dos
fragmentos de un salterio cristiano (MSS
CUL T-S NS 305.198 and T-S NS 305.210).
El estilo de la escritura del salterio árabe
sugiere que fue copiado por un escriba bien
entrenado en el siglo IX o principios del X.
Esta fecha se convierte en la traducción
árabe cristiana de la Biblia más antigua que
se ha encontrado hasta ahora en la
Genizah. El análisis lingüístico indica que el
traductor tuvo acceso a Éxodo 15 tanto de
la Peshitta como de la Septuaginta durante
su trabajo. Lo más probable es que este
traductor fuera un melquita palestino del
siglo IX que hablaba siríaco y árabe.
Keywords:
Exodus 15; Genizah; Christian
Arabic Bible translation; Melkites; MSS CUL
T-S NS 305.198; T-S NS 305.210.
Palabras clave:
Éxodo 15; Genizah;
Traducción árabe cristiana de la Biblia;
Melquitas; MSS CUL T-S NS 305.198; T-S
NS 305.210.
