749 Journal of Semitic Studies LXX/2 Autumn 2025 doi: 10.1093/jss/fgaf023 © The Author(s) Published by Oxford University Press on behalf of the University of Manchester. This is an Open Access article distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial reproduction and distri- bution of the work, in any medium, provided the original work is not altered or transformed in any way, and that the work is properly cited. For commercial re-use, please contact journals.permissions@oup.com A PARTIALLY VOCALIZED HYBRID JUDAEO-ARABIC GLOSSARY-TRANSLATION FRAGMENT OF EXODUS 14.25–15.25; 21.7–30 FROM THE CAIRO GENIZAH BENJAMIN KANTOR UNIVERSITY OF CAMBRIDGE Abstract While Saʿadia’s Tafsīr may be the most well-known Judaeo-Arabic translation of the Bible from the Middle Ages, it belongs to a much wider world of the ‘phenomenon’ of medieval Judaeo-Arabic Bible translation. Alongside Saʿadia’s Tafsīr, numerous fragments from the Cairo Genizah attest to Judaeo-Arabic translations of the Bible from unknown authors. The fragment T-S Ar. 28.170 + 27.60, which con- tains portions of a Judaeo-Arabic translation of Exodus 14, 15, and 21, is interestingw for several reasons. In terms of format and codicol- ogy, it transitions between a columnar glossarial text and a more con- tinuous translation (with Hebrew incipits). Linguistically/scribally, it exhibits classical spelling features for consonants and more phonetic ones for vowels. Translationally, it exhibits features characteristic of the Karaite translation tradition, but in its own specific form. Together, these particular characteristics shed light on both the enter- prise of medieval Judaeo-Arabic Bible translation and medieval Arabic dialectology more generally. 1. Introduction1 In the world of Judaeo-Arabic Bible translations, Saʿadia Gaon’s Tafsīr has undoubtedly garnered the most attention. The Tafsīr’s rel- evance is apparent not only in its place in modern scholarship but also in its impact on subsequent Judaeo-Arabic translations com- posed during the Middle Ages. Nevertheless, recent decades have 1 I would like to thank various colleagues, namely Geoffrey Khan, Phillip Stokes, and Nick Posegay for comments on earlier iterations of this paper. I would also like to thank my students Samuel Frith, Siân Morris, and Joshua Meynell for insightful discussion of bits of this text and related texts that have made this article better. Any remaining errors or inaccuracies are my own. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 750 witnessed an increasing interest in some of the other early Judaeo- Arabic translations of the Bible, both before and contemporary with Saʿadia’s Tafsīr. Although these early translations are often only par- tially preserved in fragments from the Cairo Genizah, they have great potential to contribute to our understanding of the phenomenon of medieval Judaeo-Arabic Bible translation and the linguistic analysis of Judaeo-Arabic (and ‘Middle Arabic’) on the whole. While numerous anonymous texts partially preserved in fragments from the Cairo Genizah are more akin to running translation texts, some of the earliest constitute mere glossaries of the biblical text. In between these two ends of the spectrum, however, are found what may be regarded as texts reflecting the intermediate or transitional stages between glossary and translation (Maman, 2003; Vollandt, 2015, 75). There are numerous Genizah fragments, for example, in which so many words are awarded a gloss—stripped of their preposi- tions and particles—that the text actually begins to resemble transla- tion itself (Polliack and Somech, 2000, 16, 43).2 Though many of these early glossarial fragments have already been published, the more that such texts can be subjected to orthographic, linguistic, and exe- getical analysis, the more we will understand about this crucial initial period in the enterprise of Judaeo-Arabic Bible translation.3 A particularly fascinating Genizah fragment belonging to this cat- egory is T-S Ar. 28.170, which contains a sort of hybrid glossary- translation text for Exod. 14.25–15.25. To this same manuscript (and perhaps scribe?) belongs the fragment T-S Ar. 27.60 (Vollandt, 2015, 75), which contains a similar translation text for Exod. 21.7– 30. What is particularly exceptional about T-S Ar. 28.170, however, is that it contains within the same fragment both columnar glossarial lists of words and what appears to approach continuous translation. While the relevance of this feature of T-S Ar. 28.170 has been touched on in the literature (Vollandt, 2015, 75), this particular frag- ment has yet to be published and anlyzed in its entirety. Aside from a brief note about its gloss of מִדְבַּר שׁוּר ‘the Wilderness of Shur’ in a 2 Polliack and Somech (2000, 43) list the following glossaries as approaching the category of translation: T-S Ar. 5.51 (Ps. 122–143; Job 1–15); T-S Ar. 28.116 (Exod. 1.20–4.14; 22.5–23.15); T-S Ar. 28.132 (Dan. 4.30–5.14; Num. 24.22– 29.18); T-S Ar. 28.137 (Deut. 4.37–5.31); T-S Ar. 28.151 (Gen. 31.13–49); T-S Ar. 28.170 (Exod. 14.25–15.25); T-S Ar. 31.52 (Lev. 15.23–16.8). 3 For some recent publications describing the linguistic features of Middle Ara- bic texts based on data from manuscripts and fragments, see Stokes; 2021; 2022; 2023a; 2023b; 2023c; 2023d; 2024; Posegay 2020; van Putten 2020. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 751 recent publication (Butbul and Hopkins, 2024, 9, 23), T-S Ar. 28.170 has not been subjected to extensive analysis. Another special feature of T-S Ar. 28.170 + T-S Ar. 27.60 is the relatively large proportion of Judaeo-Arabic words written with Hebrew vocalization (i.e., niqqud). From a historical-linguistic per- spective, especially as it concerns phonology, morphology and mor- phosyntax, these vocalized words can shed more light on medieval Arabic dialectology. While other scholars have already called atten- tion to the relevance of vocalized Judaeo-Arabic texts for this purpose (see Blau and Hopkins, 1985; Khan, 1992; 2010; 2017; Blau and Hopkins, 2017–2024; Posegay, 2020), the field as a whole can only benefit from more such texts being published and subjected to lin- guistic analysis. After presenting a complete transcription of the fragments T-S Ar. 28.170 + T-S Ar. 27.60, we proceed to subject the text to ortho- graphic, linguistic, exegetical, and textual analysis in the remainder of this article. Although most, if not all, of the features we will look at below have already been documented in Judaeo-Arabic more broadly (see, e.g., Blau, 1980; Blau and Hopkins, 2017–2024), it is still worth describing and outlining their specific distribution in our frag- ment and thus the specific linguistic profile thereof. As touched on above, the scribal conventions exhibited in our fragment—extensive use of matres lectionis and some Hebrew niqqud—afford us a bit more insight into the phonology and morphology underlying the text than is common in some other Judaeo-Arabic texts. We will con- clude by outlining the ramifications of our findings for the wider fields of Judaeo-Arabic Bible translation and Medieval Arabic on the whole. 2. Fragments and Transcription T-S Ar. 28.170 The fragment T-S Ar. 28.170 contains Judaeo-Arabic in an oriental Hebrew script over 22–24 lines on a roughly 19.7 × 21.2cm piece of vellum (Baker and Polliack, 2001, 174). Based on the palaeography, it may be loosely dated to between the eleventh and thirteenth cen- tury CE.4 The fact that, at least with respect to the consonants, it 4 I would like to thank Geoffrey Khan, Nick Posegay, and Estara Arrant for consulting with me on this question. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 752 reflects Classical Judaeo-Arabic spelling/script (CJAS) also points to the time of Saʿadia as the terminus post quem. Beginning on the recto of this single-leaf fragment we find a bicolumnar list of words from Exod. 14.25–29. Rather than one column containing only Hebrew and the opposite column only Arabic, each column of this glossarial list typically contains both the Hebrew word and its Arabic gloss. From verse 14.30, however, the text shifts—both in its context and its codicology—to more of a running continuous translation (with Hebrew incipits). Even within the continuous text, however, there are often ‘double glosses’ for certain words and phrases that remind us that we are not dealing with pure translation per se. This pattern continues on the verso all the way until verse 15.21, at which point it takes the form of a bicolumnar glossarial list once again until the Figure 1: T-S Ar. 28.170, fol. 1r (Reproduced by kind permission of the Syndics of Cambridge University Library) A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 753 end of the fragment (v. 15.22). This codicological pattern may indi- cate that continuous translation—rather than a mere glossarial list— was more necessary for poetry than narrative, as we see in the figures below, and their transcription: וראכיבוה רמא פי אלים׃ עזי׃ עיזי דו אלעיז אלאזלי ו[א]ל מיגִידּ ○ עיזי ודו אלמגִדּ אלאזלי וכאן לי ללמג́ותה ׳גיאת הדא אלטאיק מוטיקי ואנויהוה אלאה י̇̇ אלרב [אר]דּ ולי י̇̇ אבי ואורפיעוה ○ י̇̇ אלמלאחים אלרב אסמוה׃ מרכבות׃ מראכיב כיל דּואב פרעון וגִחפלוה זג זגִ זדּא פי אלים ומבחר וא׳כיאר קואדּוה טוביעו פי ים אַלקֻלזום ○ לוולב יעני ניעל מראכיבוה ׳תִקַל בכבדות בֵ נלחם אלמולאחים א חדּוה לאתנו אלִַ לקראתם תלקאהום יעני יפירו חידאהום הרכב אלרוכאב הבאים אלדאכילין אפן מרכבתיו דואבוה וינהגהו וקאדּוה אנוסה אפיר לפנות לאתגִאה נסים פארין וינער ונפץ̇̇̇ ו׳גַרַק5 ויכסו ו׳גטת חיל גִחפל תהמות ○ אלא׳גמאר ׳גטתּהום ורדּו פי אלקער מתל אלחגר ○ ימינך ○○ ימינךּ יא רב בטל דו ג́וזר פי אלקווה ימינךּ יא רב תקמע אלבא׳גיץ̇̇ ○ וברב ○ ובכּתרת אקתדּארךּ תהדּים אלמונאציב מוקאוימיך6ּ תרסיל דּרדּךּ חרדּךּ יאכולהום כאלקש׃ וברוח׃ וברוח בךּ תערם צאר ערם ערם אלמא ׳גצ̇̇ אלמיאה תערמת אנתצבת מתל אלטוודּ אלמנהטילה גִמדּת אלא׳גמאר פי קלב אלים׃ אמר׃ קאל אלבא׳גיץ̇̇ אכלב אדּריךּ אקסים סלב תתמלאהום [נ]פסי אוגִרידּ ○ תסתביחהום וישע ○ וא׳גאת ונצר אלרב פי דיאליך אליום איא יסראיל מן יד אלמצריין ורא יס יסראיל איא אלמצרייִן7 מותא עלא שפיר אלים׃ וירא׃ ורא יסראיל איא אלמקדּורה אל׳צרבה אלכבירה אלדי עמיל אלרב בלמצרייִן8 ענדּ דיאליך אתקוו אלקוום איא אלרב ואמנו באלרב ובמוסי עבדּוה ○ אז׃ חיניד אנשא וסבח מוסי ובני יסראיל איא הידיה אלתסבחה לאלרב וקאלו קאילין אֻנשי אוסביח לאלרב אן אקתדּאר אקתדּר אלפרץ דו אלגליל ואלעיז ○ 5 Curved stroke above gimel. 6 There are two dots under the first vav but of a different color and could be mere marks on the fragment. 7 Two dots above the יי in this word. 8 Two dots above the יי in this word. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 754 א יגִוז הדא אלשעב אלדי אקתנית אלִַ איא אליר׳דּן׃ תביאמו׃ תדּכילהום ות׳גרסהום פי גִבל נחלתךּ מוהיא מקעדּ למגִליס אמרך לתבאתּ אמרךּ פעלתּ יא רב מקדס יא רב הית ידּאךּ רחמתךּ מקדּורתךּ׃ י̇̇ אלרב ימלוך9ּ אלא אלדהר ̇○̇ י̇̇ ואלאבדּ ○ כי בא ○ פלמא א׳ד ׳כַיְלוה ברוכאבוה דּ׳כל פרץ פרעון וַ וברגִאלתוה פי אלים ורד אלרב הום ידּי ויקאל תצטאצִלהום סיפי תקרצ̇̇ ידּי׃ נשפת׃ [נש]פת ○ נַשַפְתּ ○ ׳גלטת יעני א׳טלמת פאספית בריחך עליהים בִאַמְרךּ פ׳ג פ׳גַטִאהום אלים רסבו כִלאַבַאר כִאלריצאץ פי אלמִיאה אל׳גזירֶה ○ מי כמוכה׃ מן מתלך פי אלמוטיקין אלקאדּירין ליס מִתלך יא רב מוטיק קאדיר ○ מן מתלך ׳גזיר בטל פי אלקודס ליס מתלך יא רב ׳גזיר בטל פי אלקודּס אלמותּקא 9 Curved stroke above mem. Figure 2: T-S Ar. 28.170, fol. 1v (Reproduced by kind permission of the Syndics of Cambridge University Library) A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 755 עליהום איא מיאה אלים ובני יסראיל סלכּו פי אליבוסה פי וסט אלים׃ ותקח׃ ואכדֿת מרים אלנבִיֶה ׳כת הרון איא אלדּף בידּהא או וכרגִן כול אלניסא וראהא באלדופוף וַבִאלַטובול ○ ותען׃ ואגִאבת להום מרים אנשו סביחו אן אקתדּאר אקתדּר אלפרץ וראכיבוה רמא פי אלים ○ דֿוּ אלמַדְח עאמיל אלעגִב׃ נטית׃ ׳ץ׃ נָחִיתָ׃ מדּת ימינךּ עיזך בלעתּהום אַלאַרְ לךּ הדא אלשעב אלדי ל בפצ̇̇ סַיְרתּ יאדו אלפצ̇̇ פכּכּתּ אנהלתוה בעיזךּ אִלַא מאוא חוסן קודּסךּ׃ שמעו׃ סמיעו אלַשוּעֻוב דיליך פארתגִזו אחאז אלטלק גִולאס פליצטין ○ אז׃ חיניד דּהישו ואתבלדּו צנאדּידּ אלרום מוטיקי מאב יחוזהום אלארתעאדּ מאגִו ??ח?? יק גִיפאר?מצ̇̇ עואצ̇̇ מרה אלמורה ויורהו ואוראה והַדַּאה וימתקו וחַלַא מדבר שור בַרִיֶה מרים הם מור וילנו ולאמו ואפפו ודַּלוה עלא עודּ וישלך וטרח כול גִולאס כנעאן׃ תפל׃ תקע עליהום אַלהיבה ואלפזע בִּכּובר ׳טם דִראעך סולטאנך יסכותו כלחגִר בעו אלא יגִוז שע[בך] יא רב ואדּי אלַגִד T-S Ar. 27.60 The fragment T-S Ar. 27.60 contains Judaeo-Arabic in an oriental Hebrew square script over 11–15 lines on a roughly 13.3 × 20.2cm piece of vellum (Baker and Polliack, 2001, 158). Similarly, based on the palaeography, the fragment may be loosely dated to between the eleventh and thirteenth centuries CE.10 Note, however, that the top of the fragment is torn off. While the verso is quite clear, the writing on the recto is faded. There are no columns apparent as in the first fragment—just the continuous translational text. It begins with pieces of Exod. 21.7 on the top right of the verso, stops in the middle of verse 10 at the bottom of the page, skips over verses 10–12 in the torn portion above before continuing on the top left of the recto from verse 13 to verse 17 at the end of the page. On the recto, bits of verse 19 are discernible through the torn fragmentary portion on the top right. The text continues on this page until verse 22 on the bottom of the page. The torn-off portion of the top left of the page presumably would have contained verses 23–25. Bits of 26 are dis- cernible in the fragmentary portion. This continues until the middle of verse 30 at the bottom of the page, as we see in the images and transcription below. 10 I would like to thank Geoffrey Khan, Nick Posegay, and Estara Arrant for consulting with me on this issue. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 756 ואשר: [ואן לם ……….. ] אללה א[ ……………….. ] יע [… ] [ע לך] מוצ̇̇ פקד צ̇̇ וכי: ואד [י]תּקיח [אמר עלא] צאחיבוה ליקתול[וה] בע[?ראמה?] ׳כובת פאן א[?סתגִ?]אר במ[דב]חי בַ פמן ענד מדב[חי ת]א׳כודוה ללמות ריב פתמיתוה ומכה: ואלדי יצ̇̇ אבוה ואומוה מוות יומאתּ: וגנב: ואלדי יסריק אִמר מן יסראיל ויביעוה פיוגִדּ פי ידּוהיהו יביעוה מות יומ[א]ת אוו שוהיד עליה א בא׳נה באעוה פמוות איצ̇̇ יומאת: ומקלל: ואלדי [י]סוב ו׳י ואד[א י]ביע [אמר איא בנתו]ה על[א אמה] [לא תכרוג] כּ׳כורוגִ אלעבידּ: [אדא כאן  ???? פ]י עיניי11 סיידּהא אלדי לוה לנפסוה [..]מאהאזַיְנהא ופַדַּהא לקוום ׳גריב מִנוה מא לוה סי 12 סולטאן ליביעהא מנהום בעדּן ׳גדּיר בהא: ואם: פאן זינהא וסמאהא ואעדּהא לאבנֻוה פליעמל להא כסירת אלבנאת: [ואם]: אִן אתכֿדֿ אוכרֶי זִהיד[ה?] 11 Two dots above יי. 12 There could be a faint final nun at the end of this word, though apparently from a later hand. Figure 3: T-S Ar. 27.60, fol. 1v (Reproduced by kind permission of the Syndics of Cambridge University Library) A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 757 פא[ יצ[ ]אוו? פי[ וכי: וא?דא? [ינטח תוור איא אִמר אוו] איא מרה פ[ימות  ?אלמנטוח? יח]צב אלתוור ח[?צב? ול]א יוכל לח[מוה ו]צ וצאחיב אלתוור ברי: ואם: וַאִז הו כאן תוור נטאח מן אמס ואול מן אמס וקדּ נושידּ צאחיבוה ואייד? ולם יחתריס בִה ויחפטוה תום אנוה נטח פאמאת אִמר אוו [מרה] [פ]ל[.]חצב אלתוור תום יומאת צא [צאחיב]וה: אם: פאן רו׳צא [ … ] ב̇̇ ̇̇(?)13 י̇̇ ?   ] ותדּו̇̇ ? ]בט[ו]ל א[ ? ]א אַ?ג?ח[] ? ]אוא[  ? ] ועילאגִ ריב [?יועאליגִוה?14 וכי: ו]אדא יצ̇̇ אלמ[רא? א]יא [ע]בדּוה אוו אמתוה יב וימו[ת] תחת ידּוה קיאדּ בלקצ̇̇ יוקאדּ במא[…] ויקאל אנתקאם יונתקם  ?פימא? [???]לן? קוום? אדּבּ? יודב ○ אך: [ואמא] אִן אקאם יום אוו יומ[י]ן לא יודאן? דיליך לאנה מאלוה [ה]וו? וכי: ואדא יתנאצוּ אונאס ויצדּימו מרה חובלא ויכ ויכרוגו אוולאדּהא ולא יכון ת?ח?ציר אלמ?נ?יה פ׳גראמה יגרם כאלדי יגעל עליה ? בעל אלמרה ויעטי [די]ליך פ[י] מחץ̇̇ 13 Two dots above bet, one above yod, and one above vav. 14 This reconstruction and the other lengthy reconstructions on this folio are based on (i) a comparison with the vocabulary and patterns attested elsewhere in Figure 4: T-S Ar. 27.60, fol. 1r (Reproduced by kind permission of the Syndics of Cambridge University Library) A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 758 3. Orthography (and Phonology) From an orthographic perspective, these fragments generally reflect the scribal conventions characteristic of Classical Judaeo-Arabic spell- ing/script (CJAS) for the consonants (see Blau and Hopkins, 1984). There is, however, a higher distribution of matres lectionis for the short vowels /i/ and /u/, a feature more commonly associated with Early Phonetic Judaeo-Arabic script/spelling (EPJAS) (see Blau and Hopkins, 2017–2024).15 Moreover, Tiberian Hebrew niqqud is added to just over thirty words across the fragments. Considering the high frequency of matres and the written vocalization, this fragment affords us valuable insight into both the phonology and the morphol- ogy of the language of the text. 3.1 Consonants While the consonants in the fragments under consideration generally reflect the CJAS conventions (see Blau and Hopkins, 1984), as noted above, there are nevertheless some noteworthy points: 3.1.1 Diphthongs *aw and *ay Although the Arabic consonant w is normally written with a single vav in Hebrew script, when it occurs in the diphthong */aw/, it is written with a double vav (see Blau, 1980, §30–§31):16 Example Arabic Hebrew القَوْم ≈ אלקוום َ◌وْ וו أَوْلادها ≈ אוולאדּהא مَوْت ≈ מוות The same convention does not normally apply to diphthongs from *ay: e.g., הית ≈ There is, however, one case in which the .(?) حَيْث the fragment, (ii) a comparison with the translations of Saʿadia and Yeshuʿah, and (iii) the available space in the lacuna. 15 For background on phonetic spelling in Judaeo-Arabic texts, see Blau and Hopkins (2017, 21–58). For more on phonetic spelling in biblical texts, see from the same authors, vol. I, 59–111. 16 This representation appears to be fairly consistent throughout the text: لَوْلب אתקוו ≈ اتّقَوْا אלטוודּ, ≈ الطَوْد לוולב, ≈. There are, however, some counterexamples: ואוראה ≈ وأَوْراه מותא, ≈ יע and ,مَوْتاً מוצ̇̇ ≈ The fact that it is also absent in .مَوْضع יומין ≈ may indicate that contraction of *aw → [oː] had occurred in some يَوْمين lexemes and/or that the orthography is simply inconsistent. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 759 dual construct form is represented with יי. This may be taken as evi- dence of a lack of contraction of the diphthong to [eː] (see Blau, 1980, §30–§31): Example Arabic Hebrew عينَيْ سيّدها ≈ עיניי סיידּהא َ◌يْ יי Note, however, that geminated yāʾ in the word סיידּהא is represented with two yod consonants in Hebrew (see Blau, 1980, §30–§31). 3.1.2 Hebrew Letters Representing More Than One Arabic Sound Consistent with CJAS, Hebrew letters that represent more than one Arabic sound are often distinguished by means of a supralinear dia- critic. This diacritic typically resembles a short diagonal stroke above the letter,17 though on occasion it is written as the more typical sin- gle dot (see images of fragments above). In the case of the emphatics, this supralinear diacritic is relatively consistent in its distribution (see Blau, 1980, §20–§23; §28): Example Arabic Hebrew طُبِعوا ≈ טוביעו ط ט بعُظم ≈ בעו׳טם ظ ׳ט صار ≈ צאר ص צ الأرض ≈ אַלאַרְ ׳ ץ ض ׳צ It may, however, be absent in the form ויחפטוה ≈ At the .ويحفظه same time, the manuscript is a bit smudged here and the diacritic could be faded. On the other hand, there appears to be one case where ص ≈ צ appears to exhibit a supralinear diacritic (erroneously?): יק .عواصيق ≈ עואצ̇̇ The distribution of the supralinear diacritic is less consistent with the dentals and (voiceless) velars. While it generally distinguishes [ð] from [d], [θ] from [t], and [x] from [k], as is common in CJAS, it is often omitted in particularly common words (see discussion in Blau, 1980, §12; §15–§16; §28): 17 This might be compared to the practice of writing dots with a diagonal stroke in some early Kufic Qurʾān manuscripts. From a palaeographic perspective, it is also worth noting that in the case of dalet, the diagonal stroke may resemble a rafe sym- bol on occasion. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 760 Example Arabic Hebrew دهشوا ≈ דּהישו د דּ وأخذت ≈ ואכ׳דת ذ ׳ד هذا ≈ הדא د/ذ ד الداخلين ≈ אלדאכילין فعلت ≈ פעלתּ ت תּ بثقل ≈ בֵ׳תִקַל ث ׳ת مثل ≈ מתל (ت؟)/ث ת يملك ≈ ימלוךּ ك כּ أخت ≈ ׳כת או خ ׳כ كل ≈ כול ك/خ כ وأخذت ≈ ואכ׳דת Note that tāʾ marbūṭa in construct is normally written with Hebrew tav, which is in fact more consistent with phonetic Judaeo-Arabic spelling (see Blau and Hopkins, 2017, 26): Tāʾ Marbūṭa Example Arabic Hebrew اليُبوسة ≈ אליבוסה (ABS) ة ה الغزيرة ≈ אל׳גזירֶה وبكثرة اقتدارك ≈ ובכּתרת אקתדּארךּ (CST) ة ת كسيرة البنات ≈ כסירת אלבנאת There is also one instance of ḏāl being represented by Hebrew zayin (see discussion in Blau, 1980, §17): Example Arabic Hebrew وَإذِ ≈ וַאִז ز ז This may indicate that in the (vernacular?) pronunciation of the scribe [ð] had merged—or was a free variant—with [z] as is common in numerous modern Arabic dialects: e.g., Palestinian [ʔiza] ‘if; whether’; [ʔizan] ‘so; then’ (Elihay, 2012, 216–217, 253). Note, however, that such words are more frequently written as אד and אדא in T-S Ar. 28.170 + T-S Ar. 27.60. Indeed, וַאִז constitutes a single example and is not sufficient for drawing general conclusions. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 761 In the case of the Arabic jīm (from the historical Semitic plosive *g; see Birnstiel 2019, 374), rather than being written with a plain gimel or with a dagesh in the Hebrew gimel, it is indicated by means of Hebrew gimel with a sublinear dot. The parallel to Arabic ghayin (from the historical fricative *ɣ), on the other hand, exhibits the supralinear diagonal stroke (see discussion in Blau, 1980, §13): Example Arabic Hebrew جبل ≈ גִבל ج גִ غضبك ≈ בךּ ׳גצ̇̇ غ ׳ג If the use of diacritics marks that which is less intuitive graphophon- ically, then the present convention indicates that Arabic ج was not pronounced as [g] in the language of the scribe.18 Otherwise, we would expect it to be represented with a dagesh in the Hebrew letter gimel (i.e., ּג) like the other dental and velar plosives (i.e., ּכ תּ for דּ k t d). That Arabic ج is represented by Hebrew gimel with a sublinear dot (i.e., ִג) may indicate that it was realized as something like a palatal plosive [ɟ] or an affricate [d͡ʒ].19 Then again, even though gimel exhibits a markedly distinct pattern of diacritics, we should be careful in ascribing too much weight to the use of diacritics for reconstructing Judaeo-Arabic phonology. After all, the underdot might simply reflect an imitation of Arabic script (cf. ج).20 3.2 Vowels While vowels are not always indicated explicitly, our manuscript often notates them by two means: (i) the use of matres lectionis and (ii) Hebrew vowel pointing. As noted above, the use of matres lec- tionis is more extensive than in some other contemporary Judaeo- Arabic texts, which may align it more with Judaeo-Arabic texts with phonetic spelling. This, alongside the partial use of Tiberian 18 For evidence that jīm was pronounced as something other than [g] in at least some environments by at least some language users behind various Judaeo-Arabic texts, see the discussion in Connolly (2019, 175–178). Note, however, that Con- nolly also argues that the inconsistencies in the use of diacritics with jīm suggest that ‘the diacritic is a dubious source for recreating the phonetic realizations of ǧīm in pre-modern Egyptian dialects’. Moreover, linguistic diversity in the realizations of ǧīm is likely (2019, 180). 19 For likely realizations of Arabic jīm in roughly contemporary sources, see Al- Jallad (2020, 17). 20 For more on this, see Connolly (2019). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 762 vocalization signs, provides insight into the underlying phonology and morphology. 3.2.1 Vowels with Matres Lectionis In the case of matres lectionis, the Hebrew letters ו ,א, and י may indicate both long and short vowels in open and closed syllables. Note, however, that ʾalef never indicates a short vowel (for back- ground on Judaeo-Arabic scribal conventions for representing the vowels, see Blau, 1980, §3–§10; Blau and Hopkins, 2017, 27–28). 3.2.2 Long Vowels In the case of long vowels, matres lectionis are employed consistently:21 Long Vowels with Matres Lectionis Example Vowel Hebrew غزير ≈ ׳גזיר /ī/ in CV الداخلينי ≈ אלדאכילין /ī/ in CVC موسىى ≈ מוסי /ē/?22 وراكبه ≈ וראכיבוה /ā/ in CV א وكان ≈ וכאן /ā/ in CVC ذو ≈ דו /ū/ in CV ו ويجوز ≈ יגִוז /ū/ in CVC The orthographic conventions corresponding to Classical Arabic ʾalif maqṣūra also require further explanation. In T-S Ar. 28.170 + T-S Ar. 27.60, ʾalif maqṣūra is sometimes represented with a final yod and sometimes with a final ʾalef (cf. discussion in Blau, 1980, §10). In the words إلى ‘to’, على ‘on’, III-y verbs, and the adjective حبلى ‘pregnant’, the ʾalif maqṣūra is represented with the Hebrew letter ʾalef. In the personal name موسى ‘Musa’ and the adjective أخرى ‘(an)other’, it is represented with yod: 21 The lone possible exception is وفَداها ≈ ופַדַּהא. 22 For yod representing ʾalif maqṣūra, see below. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 763 ʾalif maqṣūra Example Vowel Hebrew إلى ≈ אִלַא /ā/? علىא ≈ עלא رمى ≈ רמא حبلى ≈ חובלא موسىى ≈ מוסי /ē/? י أخرى ≈ אוכרֶי Now while EPJAS tends to represent ʾalif maqṣūra phonetically as -the imita (presumably) ,(Blau and Hopkins, 1984, 25–26) [aː] = א tion of Classical Arabic orthography inherent in CJAS proffers more cases of ʾalif maqṣūra written with yod, sometimes even alternating in the same manuscript (see Blau, 1980, 10–11, 24–26; see also Blau and Hopkins, 2017, 24). Christian Arabic also exhibits interchanges between final ا- and ى- for ʾalif maqṣūra bi-ṣūrat al-yāʾ (Blau, 1966, 81–83). It has been argued that, at least in some cases, the inter- change of ʾalif and yāʾ for some cases of ʾalif maqṣūra may point to the phenomenon of ʾimāla (see Hopkins, 1978, 9). The question here, then, is how we ought to explain the inconsistent representation of ʾalif maqṣūra bi-ṣūrat al-yāʾ—with final ʾalef sometimes and with final yod other times—in T-S Ar. 28.170 + T-S Ar. 27.60. Is it merely inconsistent orthography all representing long [aː] or could there be a phonetic dimension to the various orthographic conventions? From a historical-linguistic perspective, ʾalif maqṣūra is likely derived from the sequence *aya (or *ayV, where V represents a vowel). While in some forms of Arabic this eventually came to be realized as [aː], in other forms of Arabic it came to be realized as [eː] (Al-Jallad, 2020, 47–48). In T-S Ar. 28.170 + 27.60, then, we may reasonably posit that at least the spellings מוסי and אוכרֶי reflect something like [muːseː]/[muːsɛː] and [(ʔ)uχreː]/[(ʔ)uχrɛː], especially in light of the added Hebrew niqqud in the latter (see discussion in Blau and Hop- kins, 2017, 26–27; 57, 134). In the case of interpreting the preposi- tions, III-y verbal forms, and the adjective חובלא, on the other hand, there are two options. The orthography could reflect either a distinct pronunciation in which ʾalif maqṣūra is realized as [aː] or merely a scribal convention. The vocalization added in the form אִלַא could suggest that the realization of ʾalif maqṣūra was indeed [aː] in these forms written with final ʾalef.23 23 See footnote 33. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 764 As such, the data from our fragment could indicate that ʾalif maqṣūra in prepositions like إلى and على and III-y verbs like رمى was realized without ʾimāla as [aː]. In III-y nominal/adjectival forms like .on the other hand, it was realized with ʾimāla as [eː] ,أخرى and موسى The one exception here is חובלא, which could reflect the interference of a different dialect and/or scribal (or reading?) tradition.24 Then again, because we only have two examples in our fragment of ʾalif maqṣūra bi-ṣūrat al-yāʾ, there may be other non-phonetic explana- tions, not the least of which being simply scribal inconsistency. Though not identical, the distribution of ʾalif maqṣūra in our frag- ment finds some parallels in early Quranic manuscripts and reading traditions, material cited in Sībawayh, and modern dialects.25 With respect to the final [aː] in prepositions like على, for example, it is noteworthy that early Quranic manuscripts of the Kufic C style often write the preposition as علا rather than على (van Putten, 2022, 250– 251). Pronouncing على and إلى as ʿalā and ʾilā also seems to have been a feature of the Warš ʿan Nāfiʿ reading tradition (van Putten, 2017, 54–55). With respect to the distinct realizations of ʾalif maqṣūra in III-y verbal and nominal morphology, Sībawayh identifies certain speakers who realize III-y verbal forms like رمى as ramā but III-y nominal forms like حبلى as ḥublē (van Putten, 2022, 27). A similar distribution may also be preserved in Mesopotamian Qəltu dialects, which have ramā for رمى but ḥəble for حبلى (Levin, 1992, 87; van Putten, 2022, 27). 3.2.3 Short Vowels While matres lectionis are regular for long vowels, they are not oblig- atory for short vowels. Nevertheless, the data from the text would appear to indicate that it is more common than not—aside from particles and prepositions—for the short vowels /i/ (or /e/?) and /u/ (or /o/?) to be indicated by the matres lectionis י and ו, respectively (see Blau and Hopkins, 2017, 27–28): 24 It is also possible that these two lexemes simply have a lexicalized ʾimāla, similar to what we find in some Quranic qirāʾāt. See van Putten (2022, 68–70). 25 For the relevance of phonetic conditioning for the realization of ʾalif maqṣūra in the Damascus Psalm fragment, see Al-Jallad (2020, 14–15, 18, 28, 47–48). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 765 Short Vowels with Matres Lectionis Example Vowel Hebrew مَراكِبه ≈ מראכיבוה /i/? in CV י العِزّ ≈ ואלעיז /i/? in CVC أُسَبِّح ≈ אוסביח /u/? in CV ו غُزْر ≈ ׳גוזר /u/? in CVC Short Vowels without Matres Lectionis Example Vowel Hebrew بِعُظْم ≈ בעו׳טם /i/? in CV ø مِثْلك ≈ מתלך /i/? in CVC اليُبوسة ≈ אליבוסה /u/? in CV ø الدُفّ ≈ אלדּף /u/? in CVC Because we do not ultimately know how this text would have been pronounced, we cannot determine with certainty whether the use of matres lectionis in representing short vowels might actually be condi- tioned on the basis of some phonetic factors (e.g., raising, centraliza- tion, or lack thereof). Another relevant factor might have been the practical need for clarifying what otherwise might be an ambiguous spelling. Nevertheless, the general conventions outlined above, though somewhat speculative, appear to be consistent with a coher- ent picture of the phonology of the text. 3.2.4 Vowels with Niqqud As noted above, some vowels (typically short vowels) are also indi- cated by means of Hebrew niqqud. Because this convention is not particularly common—it occurs in just over 30 words throughout the fragment26—all attested examples are cited in the charts below (for niqqud in Judaeo-Arabic texts, see, e.g., Khan 2010; Blau and Hopkins 2017, 29; Posegay 2020). 26 Note that almost without exception the use of niqqud in Judaeo-Arabic texts with phonetic spelling is only partial. Continuous vocalization is rare (see Blau and Hopkins, 2017, 29). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 766 Long Vowels Long vowels are rarely represented with niqqud. This occurs most frequently with /ā/, but also with /ī/, /ē/, and /ū/.27 Note that there is one case of the long /ē/ vowel in ʾalif maqṣūra being represented with seghol and yod: Long /ī/ with ḥireq (+ Yod ?)28 Example Vowel Hebrew بالمصريين ≈ בלמצרייִן /ī/ ייִ المصريين ≈ אלמצרייִן Long /ē/ with Seghol + Yod Example Vowel Hebrew أخرى ≈ אוכרֶי /ē/? ֶ◌י Long /ā/ with Pataḥ (+ ʾAlef) Example Vowel Hebrew وهَدَاه ≈ והַדַּאה /ā/ ַ◌א وحَلَا ≈ וחַלַא كلآبَار ≈ כִלאַבַאר إلَِى ≈ א אלִַ إلَِى حده ≈ א חדּוה אלִַ وفَدَاها ≈ ופַדַּהא29 Long /ū/ with Shureq/Qibbuṣ Example Vowel Hebrew ذُو ≈ דֿוּ /ū/ וּ الشعُوب ≈ אלַשוּעֻוב /ū/ ֻ◌ו 27 There may be a ḥireq under the ṭet in the form פ׳גַטִאהום, which could thus indicate raising of long /ā/ so that the form was pronounced as something like [fa-ɣattˁeːhum] ‘and [the sea] covered them’. 28 The two yods in the sequence of ִיי- could be interpreted as a double yod to indicate geminated or consonantal yod or as a sequence of consonantal yod followed by a mater yod. It is also possible, if the double yod indicates consonantal or gemi- nated yod, that writing a third yod for the typical convention of yod mater for long /ī/ was deemed orthographically excessive and thus avoided. 29 Note that the ʾalef mater, which would be expected in a III-y verbal form, is omitted here. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 767 Short Vowels Each short vowel is represented with niqqud occasionally. The cardi- nal phonemes */i/, */a/, and */u/ are represented on occasion with ḥireq, pataḥ, and shureq/qibbuṣ, respectively: Short /i/ with Ḥireq Example Vowel Hebrew بِأمرك ≈ בִאַמְרךּ /i/? ◌ִ بِكبر ≈ בִּכּובר بِخبت ≈ בִּ׳כובת30 وبِالطبول ≈ וַבִאלַטובול كالآبار ≈ כִלאַבַאר بِه ≈ בִּה مِثلك ≈ מִתלך بثِقل ≈ בֵ׳תִקַל ذِراعك ≈ דִראעך إمِر ≈ אִמר إلِى حده ≈ א חדּוה אלִַ إلِى ≈ א אלִַ مِنه ≈ מִנוה إنِ ≈ אִן النبِيّة ≈ אלנבִיֶה برِيّة ≈ בַרִיֶה Short /a/ with Pataḥ Example Vowel Hebrew وحَلا ≈ וחַלַא /a/ ◌ַ وهَداه ≈ והַדַּאה وفَداها ≈ ופַדַּהא وَخَيله ≈ ׳כַיְלוה וַ وَبالطبول ≈ וַבִאלַטובול الشعوب ≈ אלַשוּעֻוב 30 The diacritic under the bet is a bit longer than just a dot and could also reflect a stroke for pataḥ, as in the transcription above. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 768 الجد ≈ אלַגִד القلزم ≈ אַלקֻלזום بثقَل ≈ ׳תִקַל בֵ نَشَفْت ≈ נַשַפְתּ وغَرَق ≈ ו׳גַרַק بأَمرك ≈ בִאַמְרךּ المَدح ≈ אלמַדְח رض َ الأ ≈ ׳ץ אַלאַרְ سَيرت ≈ סַיְרתּ فغَطاهم ≈ פ׳גַטִאהום بَريّة ≈ בַרִיֶה ودَلّه ודַּלוה زَيّنها ≈ זַיְנהא Short /u/ with Shureq/Qibbuṣ Example Vowel Hebrew الشُعوب ≈ אלַשוּעֻוב /u/? וּ أُنشئ ≈ אֻנשי /u/? ◌ֻ لابنُه ≈ לאבנֻוה Coda of Syllable Finally, note that the Hebrew shewa sign is used to indicate the coda of a closed syllable, in some cases corresponding to those in which Classical Arabic has sukūn. This convention is only clearly attested in the nominal pattern 31:فَعْل 31 Hebrew shewa is also attested twice in the second syllable of D-stem verbs in the suffix conjugation: ּסַיְרת ‘you led’ (Exod. 15.13); זַיְנהא ‘adorned her’ (Exod. 21.8). While the Classical Arabic forms of these verbs would be َسَيَّرْت and زَيَّنَها, respectively, it is possible that the form underlying this orthography was pro- nounced differently. One might also suggest that this constitutes an unexpected case of shewa for short [a]. Note also that there are two diagonal dots above the yod in the first example, which could also be relevant. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 769 Hebrew shewa for Coda of Syllable Example Arabic Hebrew المدْح ≈ אלמַדְח ◌ْ ◌ְ الأرْض ≈ ׳ץ אַלאַרְ وخيْله ≈ ׳כַיְלוה וַ بأمْرك ≈ בִאַמְרךּ 3.3 Phonetic Allophones with Niqqud In addition to these more common representations of the cardinal phonemes corresponding with their Classical Arabic values, there are also cases of apparently phonetic realizations represented by other vocalization signs. This could occur with */i/ → [e] and */a/ → [ɛ], represented by ṣere and seghol, respectively: Short /i/ → Phonetic [e] with Ṣere Example Vowel Hebrew بِثقل ≈ ׳תִקַל בֵ /i/ → [e]?32 ◌ֵ Short /a/ → Phonetic [e]/[ɛ] in Tāʾ Marbūṭa with Seghol + Heh Example Vowel Hebrew الغزيرَة ≈ אל׳גזירֶה /a/ → [e]/[ɛ]? ֶ◌ה النبيَّة ≈ אלנבִיֶה بريَّة ≈ בַרִיֶה The vowel on the preposition bāʾ being written with ṣere in the first instance could reflect centralization/lowering of the vowel /i/ → [ɪ]/ [ə]/[e] or simply a different phoneme /e/. Note that this is also attested elsewhere: e.g., الكلام بهذا ≈ כָלַאם ’with this speech‘ בֵהָדֶּל (T-S Ar. 54.63, fol. 3v); בֵּאַחְכָּאמַךּ ≈ بأحكامك ‘by your judgements’ (T-S Ar. 54.63, fol. 3r) (Khan, 2010, 214). Note, however, that T-S Ar. 54.63 is from the Ottoman period and thus later than our text. All three examples of seghol for the vowel for tāʾ marbūṭa occur after a preceding front high vowel, whether /i/ or /ī/. This likely indicates raising of the vowel to [ɛ], a feature attested frequently else- where in Judaeo-Arabic and modern Arabic dialects. This likely belongs 32 One might also suggest that such a niqqud could represent a phonetic schwa sound [ə] or other centralized vowel. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 770 to the wider phenomenon of what is referred to as ‘pausal’ (or word- final) ʾimāla, as is documented in medieval vocalized Judaeo-Arabic texts (see Hopkins, 2005; Khan, 2010, 204) and various modern Ara- bic dialects (see Levin, 1971; Blanc, 1973; Shachmon, 2011). 4. Phonology (and Syllable Structure) While those phonological features most closely related to orthogra- phy were touched on above, there are numerous other phonological features that require further comment, most of which relate to pho- netic variation and syllable structure. 4.1 ʾimāla There appears to be a tendency for ʾimāla when */ā/ or */a/ is fol- lowed by a front high vowel. The relevance of this environment is underscored by its distribution in the demonstratives (see discussion in Blau, 1980, §5; compare also examples in Blau and Hopkins, 2017, 246, 264, 424, 428, 568, 574, 576): ± ʾimāla in the Demonstratives Example Possible Realization?33 هذا ≈ הדא [ha(ː)ðaː]? هذه ≈ הידיה [heːði(h)]? ذلك ≈ (×2) דיאליך [ðeːlik]? ذلك ≈ (×3) דיליך [ðeːlik]? This sort of a distribution has parallels elsewhere in Medieval Arabic. Note how the Damascus Psalm Fragment attests to ἁ(δα) [haːðaː] but δέλικ [ðeːlik]/[ðɛːlik]/[ðæːlik] (Al-Jallad, 2020, 26). Though not particularly common, the spelling הידיה for the FSG demonstrative is 33 In this paper, transcriptions identified under the terms ‘realization’ and ‘Pos- sible Realization?’ (in the tables) should be taken loosely since it is impossible to know the pronunciation of language users of the community of the scribe. Never- theless, extremely tentative suggested transcriptions have been provided to make the potential linguistic data underlying the text more ‘tangible’, but these should be regarded as highly speculative. Note also that rather than transcribing forms like -etc. with explicit transcription of phonetic pharynge ,[tˁ] = ط ,[sˁ] = ص ,[dˁ] = ض alization for the Arabic emphatic series, a more traditional underdot ض = [ḍ], ص = [ṣ], ط = [ṭ], etc. is used for the sake of simplicity and to avoid committing to recon- structing a precise phonetic realization. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 771 attested in several other Judaeo-Arabic texts: אלגיריה هذه ≈ הידיה هذه ≈ הידיה אלשריעה ;(Allony, 1969, 47) (JTS Adler 3316) مثل هذه ≈ מת׳ל היד׳יה ;(Butbul and Hopkins, 2024, 38) (T-S Ar. 27.111) الجارية הידיה ;(Tobi, 1993, 19) (T-S Ar. 28.154, fol. 1r, l. 4) الشريعة الكراسية ≈ אלכוראסיה ,Davis and Outhwaite) (T-S As. 25.272) هذه ,Blau) (Bodl. Heb. c. 18.69) هذه الخصل ≈ הידיה אלכצל ;(358 ,2003 2002, 142). Note also the vocalized form الأمور ≈ הֵדִיה אַלאֻמוּר هذه (T-S Ar. 8.3, fol. 14r) (Khan, 2017, 401). In some of these texts, however, the distribution of ʾimāla is different. In T-S Ar. 28.154 + 168, for example, ʾimāla is not attested in the demonstrative ذلك, even though it is attested in other nominal forms and the G-stem participle: e.g., דאליך [ðaːlik]? ≈ ذلك (T-S Ar. 28.168, fol. 1r, l. 14); but cf. אליהך [(ʔ)ileːhak]? ≈ إلاهك (T-S Ar. 28.154, fol. 1r, l. 20), ,Tobi) (T-S Ar. 28.154, fol. 1r, l. 5) الحاملين ≈ ?[al-ħeːmiliːn] אלחֵימילין 1993, 19).34 Apparent ʾimāla also appears to be attested in the following *faʿīl pattern: ʾimāla in the *faʿīl Pattern Example Possible Realization? مجيد ≈ מיגִידּ [mɛɟiːd]/[mɪɟiːd] The use of a yod mater likely reflects the raising of the first vowel to [æ], [ɛ], or [ɪ]. It is significant to note that this word reflects the same environment for raising of /a(ː)/ as evidenced above, namely before a front high vowel /i(ː)/. Note that Greek transcriptions of Arabic attest to a similar phenomenon, as in the name σεειδ for Arabic saʿīd (Kaplony, 2015, 41). Finally, as noted above, ʾimāla also appears to be attested in the vowel before tāʾ marbūṭa as can be seen from a few examples of Hebrew niqqud in our fragment: ʾimāla in Tāʾ Marbūṭa with Seghol + Heh Example Vowel Hebrew الغزيرَة ≈ אל׳גזירֶה /a/ → [e]/[ɛ] ֶ◌ה النبيَّة ≈ אלנבִיֶה بريَّة ≈ בַרִיֶה 34 Tobi has misread דאליך as דאיך, missing the אל ligature, and אלחֵימילין as .missing the second yod mater ,אלחֵימלין A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 772 This phenomenon of ʾimāla with tāʾ marbūṭa is sometimes attested with ṣere or ḥireq in other Judaeo-Arabic texts from the Cairo Geni- zah (Khan, 2017, 400): e.g., الأخيرة ≈ אַלְאֲכִֿירֵה (T-S Ar. 8.3, fol. 12v); T-S NS) نهاية ≈ נהאיֵה ;(T-S Ar. 53.12, fol. 1r) الحكمة ≈ אַלְחִכְּמִה 91.12, 1b). Elsewhere in T-S Ar. 28.170, this same ending is written without the vowel indicated: e.g., אלכבירה אל׳צרבה ≈ אלמקדּורה It is thus not entirely clear if [ɛ] was the regular .المقدورة الضربة الكبيرة realization of tāʾ marbūṭa—and it was coincidentally indicated in these few words—or if these three words were marked as such because it was an atypical realization of the phoneme. 4.2 Prepositions before the Definite Article When an inseparable preposition is attached to a following word with the definite article, most forms of Arabic do not have a separate vowel for the definite article. Rather, the vowel of the preposition runs right into the /l/ (or assimilated consonant) of the article: e.g., كالبيت ,’bi-l-bayt ‘in the house بالبيت .al-bayt ‘the house’; but cf البيت ka-l-bayt ‘like the house’. In T-S Ar. 28.170, however, there is evi- dence from Hebrew niqqud that both the vowel of the preposition and the vowel of the definite article (/a/ in this case) are both main- tained (perhaps as a hypercorrection or as an orthoepic orthography/ reading) in such sequences, at least on some occasions (see discussion in Blau, 1980, §35–§36): Prepositions + Definite Article Example Possible Realization? وبالطبول ≈ וַבִאלַטובול [wa-bi-(ʔ)aṭ-ṭubūl] In other cases, however, note that the orthography and the addition of Hebrew niqqud seem to indicate that the vowels run together (see discussion in Blau, 1980, §35–§36): Prepositions + Definite Article Example Possible Realization? بالقضيب ≈ בלקצ̇יב [bi-l-qaḍiːb] كالحجر ≈ כלחגר̣ [kV-l-ħaɟar] بالمصريين ≈ בלמצרייִן [bi-l-maṣrijjiːn] كالآبار ≈ כִלאַבַאר [ki-l-(ʔ)aːbaːr] The absence of the ʾalef of the definite article in the consonantal orthography is instructive. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 773 A pair of such orthographies also has a parallel in a Hebrew- Judaeo-Arabic glossary of Samuel documented by Posegay: וּבִּאַלְקְנֵאה ‘and with the spear’ (T-S Ar. 5.58); ִבִּלְמִכְלֵה ‘in the bag’ (T-S Ar. 5.58) (Posegay, 2020, 40–41). It is noteworthy that in both our frag- ment and this glossary of Samuel, it is when the preposition bi- is preceded by the conjunction waw that the non-run-together vocaliza- tion occurs. The idea may be that with the accumulation of preposi- tions/particles and conjunctions, a clearer and more careful vocaliza- tion was added to avoid ambiguity. Note, however, that the consonantal realization of the ʾalif of the article is attested in other similar sequences in vocalized Judaeo-Arabic texts like וְאַלאעלֵא ‘and the highest’ (T-S Ar. 8.3, fol. 16v), פִי אַלחִכְּמִה ‘in wisdom’ (T-S Ar. 53.12, fol. 1v), and אַלאֻמוּר .these matters’ (T-S Ar. 8.3, fol‘ הֵדִיה 14r). This may simply reflect a tendency to treat hamza in such envi- ronments (i.e., the definite article) in certain Judaeo-Arabic reading traditions as hamzat al-qaṭʿ rather than hamzat al-waṣl (Khan, 2017, 399, 401). It could also be a distinct reading tradition of Judaeo- Arabic Bible translation(s) and/or an orthoepic reading convention. The vowel of the preposition -ك ‘like’ in this collocation is repre- sented with ḥireq rather than expected pataḥ (see table above). This could indicate one of two things. For expected /ka-l-/, the ḥireq may indicate a centralization or reduction of the /a/ vowel. Alternatively, it could reflect a situation in which the vowel of the definite article— perhaps in this case a higher front or centralized vowel—rather than the vowel of the preposition, was maintained. Note, however, that the definite article is normally realized as [al-]: Vocalization of Definite Article Example Possible Realization? الأرض ≈ ׳ץ אַלאַרְ [al-(ʔ)arḍ] الجد ≈ אלַגִד [al-ɟidd] الشعوب ≈ אלַשוּעֻוב [aʃ-ʃuʕuːb] وبالطبول ≈ וַבִאלַטובול [wa-bi-(ʔ)aṭ-ṭubūl] القلزم ≈ אַלקֻלזום [al-qulzum] In light of this evidence, then, the word כִלאַבַאר could indicate that the typical realization of the preposition -ك ‘like’ was indeed some- thing like [ki-] or [kə-]. Inconsistent realizations of the definite article in different environments—or some kind of analogy—might also be a possibility. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 774 4.3 Emphasis Spread There are a couple words in which some form of emphasis spread is detectable in T-S Ar. 28.170. First, in the noun فلسطين ‘Palestine; Philistia’, the sin appears to assimilate in emphasis to the following ṭāʾ (see discussion in Blau, 1980, §20): Emphasis Assimilation Example Possible Realization? فلسطين ≈ פליצטין [filiṣṭiːn]/[faliṣṭiːn]? While emphasis spread—or assimilation—among adjacent conso- nants is common, emphasis can also spread across an entire word, as in the case of the verb استأصل ‘to root out’: Emphasis Spread Example Possible Realization? تستأصلهم ≈ תצטאצִלהום [taṣṭaʔṣilhum]/[taṣṭaːṣilhum]? Even though the initial sin-tāʾ sequence is separated from the middle radical ṣād by a vowel and non-emphatic consonant, the emphatic nature of the ṣād appears to spread to the beginning of the word as well, so that it is pronounced as something like [taṣṭaʔṣilhum]/ [taṣṭaːṣilhum] rather than [tastaʔṣilhum]. A similar phenomenon may occur with the word פרץ ≈ فرس ‘horse’, in which the emphatic-like nature of rāʾ appears to bring about the change of the following non-emphatic /s/ > [ṣ]. 4.4 Weakening of ḥāʾ? In one instance, the parallel to (probably?) Classical Arabic ُحَيْث ‘where … ’ is written with Hebrew he instead of ḥet (see discussion in Blau, 1980, §26): Emphasis Assimilation Example Possible Realization? حيث ≈ הית [ħajθ]/[ħeːθ] → [hajθ]/[heːθ] If not a scribal error, this could reflect a case of sporadic weakening of the pharyngeal [ħ], which may be attested in other Judaeo-Arabic texts, as in حتى ≈ התא (Blau, 1980, 40). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 775 5. Morphology Though much of the text is opaque with respect to the morphology of the language, there are still numerous noteworthy elements that can be discerned through the orthography: 5.1 Third-Person Pronominal Suffixes The 3MS suffix is normally represented by וה- on both nouns and verbs, though it can be represented by just ה- when preceded by a long vowel (see discussion in Blau, 1980, §50d–g). This likely indi- cates the realization [-u(h)] generally and [-(h)] before a long vowel: e.g., عبده ≈ עבדּוה ([ˈʕabdu]?); وأوراه ≈ ואוראה ([wa-(ʔ)awˈraː(h)]?).35 The fact that the 3MS suffix is once vocalized as לאבנֻוה may suggest that it was regularly pronounced as something like [-uh] or simply [-u], the heh likely being an orthographic element in imitation of Classical Arabic orthography (see Blau, 1980, 59–60; Khan, 2017, 397–398).36 Presumably, this indicates that the 3MS suffix -u(h) attached to the base noun (or verb, particle, preposition, etc.), which did not inflect for case (see below). While the 3MS suffix is normally attested as וה- on particles/prep- ositions as well, it is attested with the preposition ב as ִה- without a vav mater but with a preceding ḥireq (see discussion in Blau, 1980, §50e; §151): 3MS Suffix on Prepositions/Particles Example Possible Realization? له ≈ לוה [lu] منه ≈ מִנוה [minu]/[minnu]? إنه ≈ אנוה [(ʔ)innu] به ≈ בִּה [bi(h)] 35 If the he is not pronounced in the latter, it would nevertheless likely entail a different stress pattern. 36 This same morphology of the 3MS pronominal suffix is also reflected by the vocalization of several other Judaeo-Arabic texts (cf. Khan, 2010, 202–203, 210– 214). A final [-u] or [-o] is also quite common across modern Arabic dialects. A 3MS [-uh] suffix, on the other hand, occurs, for example, in Najdi Arabic dialects, as in the Shamarri dialect (Porchazka, Jr. 1988) and the Qaṣīmī dialect (al-Rojaie, 2013, 57). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 776 One might compare the form به ≈ ביה alongside the form له ≈ לוה in the same text elsewhere (e.g., T-S Ar. 53.8) (Blau and Hopkins, 2017, 386), but this could also reflect analogy with the preposition This distribution may also have some .(see Blau, 1980, §151) في parallels in the realization of the 3MS suffix in various Christian Ara- bic Gospel manuscripts examined by Stokes.37 For example, Leiden. Or. 561 has ُله alongside ِبه (Stokes, 2023a, 409). Though not exactly the same morphophonological shape as is attested here, such a dis- tinction is reminiscent of T-S Ar. 28.170 + T-S Ar. 27.60. The 3MP suffix is normally represented by הום-, which almost certainly reflects the pronunciation [-hum]/[-hom] (see discussion in Blau, 1980, §50j), as in the following examples from our fragment: e.g., يأكلهم ≈ יאכולהום ([ja(ʔ)kulhum]?); حذاءهم ≈ חידאהום ([ħiðaːhum]?). On prepositions, however, there is some variation between the :suffix following yāʾ -הים suffix and the -הום 3MP Suffix on Prepositions Example Possible Realization? عليهم ≈ עליהום [ʕalajhum]/[ʕaleːhum] عليهم ≈ עליהים [ʕalajhim]/[ʕaleːhim] لهم ≈ להום [lahum] منهم ≈ מנהום [minhum] While most of these examples are straightforward, the orthography (in T-S Ar. 27.60 ×1) עליהים alongside (in T-S Ar. 28.170 ×2) עליהום is curious.38 In the latter case the scribe may have been imitating a register resembling Classical Arabic عَلَيْهِم [ʕalajhim]. 5.2 G Stem 5.2.1 Suffix Conjugation When the second vowel of a G-stem verb in the suffix conjugation is /i/ rather than /a/, this may be marked in the orthography with a yod mater (cf. Blau, 1980, §55): 37 Note also cases in modern Arabic dialects in which the preposition *bi- + 3MS suffix is realized as bī but the 3MS suffix elsewhere is realized as -uh/-oh (e.g., Gali- lean Arabic). I would like to thank Phillip Stokes for bringing this to my attention. Of course, the he could simply be a matter of conservative spelling irrespective of actual pronunciation and/or imitation of Arabic orthography. 38 As seen in line 3 of T-S Ar. 28.170, fol. 1v, the penultimate consonant is clearly yod and not a short vav. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 777 Suffix Conjugation of G Stem with /i/ Second Vowel Example Possible Realization? سَمِعوا ≈ סמיעו [samiʕuː] عَمِل ≈ עמיל [ʕamil] دَهِشوا ≈ דּהישו [dahiʃuː] غَدِر ≈ ׳גדּיר [ɣadir] The orthography of the following geminate-root verb in the G stem may also convey something about the nature of the 2MS ending on suffix conjugation verbs: 2MS Suffix Conjugation of G Stem w/ Geminate Root Example Possible Realization? مدت ≈ מדּת [maddət]? Such a root could be realized in the suffix conjugation of the G stem in a variety of ways: e.g., [madadta], [madadət], [maddeːta], [maddeːt], [maddət]. In theory, if an imitation of Arabic orthogra- phy, the spelling מדּת could also represent [matt] or even [madt(a)], the pattern of which is acknowledged in the writings of the early Arabic grammarians and reflected by the rasm of the Quranic conso- nantal text (see van Putten, 2022, §5.13.1). Given the presence of only a single dalet in the orthography here—and the patterning of the 2MS ending elsewhere—it is plausible that the form [maddət] (or [maddeːt]) is intended. The representation of the 3MS suffix as וה when it attaches to 2MS suffix-conjugation verbs may also indicate that the 2MS verbal ending */-ta/ had come to be realized simply as */-t/ in the language underlying the text (see Blau, 1980, §55). At least in the form אנהלתוה = /(ʔ)anhalt-u(h)/, the 3MS suffix -u(h) appears to attach to a base ending simply in -t rather than -ta. 5.2.2 Prefix Conjugation The following forms constitute likely examples of the prefix conjuga- tion of the G stem in which the theme vowel is represented with a mater in non-weak roots (cf. Blau, 1980, §61): A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 778 Prefix Conjugation of G Stem (Non-Weak Roots) Example Possible Realization? yafʿ ul يملك ≈ ימלוך [jamluk] ويخرجوا ≈ ויכרוגו [wa-jaχruɟuː] يسكتوا ≈ יסכותו [jaskutuː] يأكلهم ≈ יאכולהום [ja(ʔ)kulhum] تأخده ≈ ת]א׳כודוה [ta(ʔ)χuðu] يسب ≈ [י]סוב [jasubb] ليقتله ≈ ליקתול[וה] [li-yaqtulu] yaf ʿil يضرب ≈ ריב יצ̇̇ [jaḍrib] يسرق ≈ יסריק [jasriq] أقسم ≈ אקסים [(ʔ)aqsim] تهدم ≈ תהדּים [tahdim] ويصدموا ≈ ויצדּימו [wa-jaṣdimuː] يفروا ≈ יפירו [jafirruː] While most of these forms are quite reasonably posited as G-stem verbs, there are some more difficult examples like אקסים and תהדּים, which could potentially be interpreted as D-stem or C-stem verbs as well, since an initial vav mater (or /u/ niqqud) only occurs in some 1CS forms of these stems. 5.3 D Stem 5.3.1 Suffix Conjugation There are potentially two cases of Hebrew niqqud on D-stem suffix- conjugation verbs, which likely indicates that its base had two /a/ vowels: Suffix Conjugation of D Stem Example Possible Realization? وغرق ≈ ו׳גַרַק [wa-ɣarraq] نشفت ≈ נַשַפְתּ [naʃʃaft] One might also suggest that the niqqud in ו׳גַרַק could indicate a lack of final vowels on verbs. If there were a final vowel here, we might A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 779 expect the orthography ַו׳גַרַק**. Then again, the niqqud does not have to be comprehensive and may be utilized simply for the sake of dis- ambiguation, in which case a final vowel would not be necessary. Similarly, the fact that the 2MS suffix conjugation verb is written with niqqud on all letters except the final tav may indicate that the 2MS suffix was simply /-t/ rather than /-ta/. At the same time, we should not expect the niqqud always to be comprehensive. II-y verbs exhibit a particular pointing pattern in which the mid- dle radical yāʾ is vocalized with Hebrew shewa: Suffix Conjugation of D Stem Example Possible Realization? سَيَّرْتَ ≈ סַיְרתּ [saj(ja?)rt] or [saj(je?)rt] زَيَّنَها ≈ זַיְנהא [zaj(ja?)nhaː] or [zaj(je?)nhaː]39 The spelling here could reflect some sort of reduction or contraction of the second syllable. In a vocalized Judaeo-Arabic biblical glossary of Samuel (T-S Ar. 5.58), the 1CS form of the D-stem is vocalized with three consecutive pataḥ vowels before the final tav: e.g., עַיַּירַת = [ʕajˈjarat] ‘I condemned’ (Posegay, 2020, 38, 40, 49).40 The final vowel in each of these cases may be an epenthetic. As such, it is curi- ous that shewa here in ּסַיְרת and זַיְנהא in our fragment may reflect a stressed syllable. It is plausible that the shewa sign here is representing a short [a] vowel and that ּסַיְרת reflects [sajˈjart] and זַיְנהא reflects [zajˈjanhaː].41 Note that the shewa sign can represent a stressed short [a] in other vocalized Judaeo-Arabic texts: e.g., מְלִך [ˈmalik] (T-S Ar. 53.12, fol. 1v, l. 10) (see Khan, 1992).42 39 While the shewa sign could reflect a true schwa-type vowel [ə], it is worth noting that already in medieval Hebrew traditions, the shewa sign likely simply reflects a short [a] (e.g., in Tiberian), a short [e] (e.g., in Palestinian), or perhaps a zero value and thus a consonant cluster (e.g., in Babylonian). 40 In the 2MS form of the suffix conjugation of the G stem, a similar pattern is attested: e.g., תְּרַכַּת = [taˈrakat] ‘you left’ (Posegay, 2020, 38, 40, 49). 41 If, on the other hand, the shewa represents a more non-standard Tiberian (or Palestinian) [e] vowel, then this phenomenon may find a parallel in modern Egyp- tian Arabic, in which both faʿʿal and faʿʿil are attested in the derived stems. Note, for example, that the verb زيّن is realized as [zajjin] (see Badawi and Hinds, 1986, 390). I would like to thank Phillip Stokes for bringing this to my attention. 42 Cf. also the treatment of ḥaṭef vowels in Karaite transcriptions of Biblical Hebrew into Arabic script with Tiberian vocalization in Khan (2022, 238). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 780 5.3.2 Prefix Conjugation Given the distribution of matres lectionis, the following spellings imply that at least the 1CS prefix conjugation of the D Stem was of the shape ʾufaʿʿil-: Prefix Conjugation of D Stem Example Possible Realization? أُسَبِّح ≈ אוסביח [(ʔ)usabbiħ] وأُرَفِّعه ≈ ואורפיעוה [wa-(ʔ)uraffiʕu] أُجَرِّد ≈ אוגִרידּ [(ʔ)uɟarrid] In light of these three examples, it might be best to interpret orthog- raphies like אכלב ≈ أكلب ,אדּריךּ ≈ أدرك, and אקסים ≈ أقسم as reflect- ing verbal stems other than the D stem. Although the vav mater in the D stem is only attested in the 1CS form, representing initial /u/ vowels in other parts of verbal paradigms is common, such as in the participles of the L and C stems as well as passive forms of the prefix conjugation throughout the verbal system (see below). 5.3.3 Imperative The imperative of the D stem has a yod mater to indicate the final vowel of the base: Imperative of D Stem Example Possible Realization? سَبِّحُوا ≈ סביחו [sabbiħuː] 5.4 tD Stem 5.4.1 Suffix Conjugation The suffix conjugation of the tD stem has an initial ʾalef before the /t/, which almost certainly indicates the lack of a vowel between the /t/ and the first radical (see Blau, 1980, §75): Suffix Conjugation of tD Stem Example Possible Realization? و(ا)تبلدوا ≈ ואתבלדּו [wa-(ʔ)itballaduː] A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 781 This feature, of course, has parallels in modern Arabic dialects: e.g., Classical Arabic [takallama] ‘he spoke’ vs. [tkallam] ‘he spoke’ in Pal- estinian Arabic (Elihay, 2012, 580) or [itkallim] in Egyptian Arabic (Badawi and Hinds, 1986, 762). 5.5 C Stem 5.5.1 Prefix Conjugation There is one case of Hebrew niqqud on a 1CS D-stem prefix conjuga- tion verb, which indicates that it was likely of the shape ʾufʿil-: Prefix Conjugation of C Stem Example Possible Realization? أُنشئ ≈ אֻנשי [(ʔ)unʃiː] It is not totally clear, however, if the prefix vowel was always realized as such. Note the following spellings of likely C-stem prefix conjuga- tion forms that have a yod mater after the second radical but no vav mater after the prefix consonant: Prefix Conjugation of C Stem Example Possible Realization? ترسل ≈ תרסיל [tursil]/[tarsil]/[tersil]? أدرك ≈ אדּריךּ [(ʔ)udrik]/[(ʔ)adrik]/[(ʔ)edrik]? تدخلهم ≈ תדּכילהום [tudχilhum]/[tadχilhum]/[tedχilhum]? If the normal prefix vowel was /a/ or /e/, then it is possible that such forms reflect the generalization of the prefix vowel of the G stem across various parts of the paradigm, including the C stem. Note that a similar phenomenon occurs in various modern Arabic dialects, though not necessarily with the same vowel.43 The generalization of the prefix vowel in the C-stem might have been facilitated by the fact that the theme vowel would have been different between the G and the C stem.44 43 Note, for example, *yaktub → yikteb/yuktob (Form I) and *yukrim → yikrem in Palestinian Arabic (Elihay, 2012, 52, 281, 755, 760). 44 Note, for example, Jordanian Arabic yigʿud/yugʽud ‘sit’ vs. yigʿid ‘seat’. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 782 5.5.2 Participle All apparent cases of the participle of the C stem have an initial vav mater, which likely represents a short /u/ vowel: Participle of C Stem Example Possible Realization? مطيق ≈ מוטיק [muṭiːq] مطيقي ≈ מוטיקי [muṭiːqiː] المطيقين ≈ אלמוטיקין [al-muṭiːqiːn] Again, the fact that this vav mater is consistent for a short /u/ vowel in the participle may indicate that the prefix vowel was distinct in second- and third-person forms in the prefix conjugation (see above). 5.6 L Stem 5.6.1 Participle There are multiple examples of the L-stem participle, which have an initial vav mater and (usually) a final yod mater in the base stem. This likely reflects the pattern mufāʿil-: Participle of L Stem Example Possible Realization? الملاحم ≈ אלמולאחים [al-mulaːħim] مقاوميك ≈ מוקאוימיךּ [muqaːwimiːk(a)] المناصب ≈ אלמונאצב [al-munaːṣib] 5.7 Passives Generally speaking, passive verb forms in the variety of verbal stems and conjugations are indicated with an initial vav mater (cf. Blau, 1980, §66–§67). The regularity of the vav mater to indicate an initial /u/ vowel in passive constructions might suggest that the absence of a vav mater in the non-passive prefix conjugation of D-, L-, and C-stem forms may be explained scribally as (i) a convention to avoid confusion with passive forms or phonetically as (ii) an indication that the prefix vowel was not always /u/ in these other stems. 5.7.1 Suffix Conjugation Passives are attested in the suffix conjugation of the G stem and the D stem: A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 783 Suffix Conjugation Passives Example Possible Realization? نُشِد ≈ נושידּ [nuʃid] شُهِد ≈ שוהיד [ʃuhid] طُبِّعوا ≈ טוביעו [ṭubbiʕuː] 5.7.2 Prefix Conjugation Passives are attested in the prefix conjugation of the G stem and the Gt stem: Prefix Conjugation Passives Example Possible Realization? يُمات ≈ יומאתּ [jumaːt] يُقاد ≈ יוקאדּ [juqaːd] يُدان ≈ יודאן [judaːn] يُنتقم ≈ יונתקם [juntaqam] I-ʾ يُؤكل ≈ יוכל [juwkal]/[juːkal]? يُؤدب ≈ יודב [juwdab]/[juːdab]? I-w فيوجد ≈ פיוגִדּ [fa-juːɟad] There appears to be only one case of a passive verb in the prefix con- jugation without an initial vav mater, namely יקאל ≈ يُقال . This may be due to its frequency and formulaic nature in exegetical language to introduce an alternate Arabic rendering of a Hebrew word.45 There may, however, be at least one other case of a prefix- conjugation passive without a vav mater in the following example: Prefix Conjugation Passive without vav mater Example Possible Realization? يجرم ≈ יגרם [juɟram] 45 This is common in the translation of Yeshuʿah ben Yehudah, as in Exod. 21.19 (BL Or. 2398, fol. 172r). On the phrases קיל and יקאל to introduce alterna- tive translations, see Polliack (1997, 191–193). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 784 Although by itself this form could be read as an active verb, the fuller context points to a passive rendering: עליה יגעל כאלדי יגרם פ׳גראמה he will surely be‘ בעל אלמרה ≈ فجرامة يجرم كالذي يجعل عليه بعل المرة punished just as the husband of the woman would obligate him’ (v. 21.22). Presumably, if not an intransitive/stative vowel pattern, this form would be realized as something like the passive [juɟram]. 5.7.3 Participle Passives are attested in the participle of the Gt stem and the D stem, each with a vav mater: Passive Participles Example Possible Realization? المتّقى ≈ אלמותּקא [al-muttaqaː] مُهَيَّأ ≈ מוהיא [muhajja(ʔ)] 6. Syntax There are numerous (mopho-)syntactic features of our text worthy of mention. 6.1 Case and Verbal Mood There are numerous pieces of evidence that indicate that the lan- guage underlying our text was not inflected for case or verbal mood. Concentric with this phenomenon is the fact that all available evi- dence from the text points to an absence of word-final short vowels, as in other Judaeo-Arabic texts (see Blau, 1980, §3, §127; §216– §221). 6.1.1 Masculine Sound Plurals with /ī/ in Nominative The masculine sound plural ending is /-īn/ rather than /-ūn/ in the nominative—there are no cases of /-ūn/ in the fragment (see Blau, 1980, §127). Phrases with parentheses indicate unattested material supplied to give context to entries in the glossarial list: Masculine Sound Plural as /-īn/ in NOM (ומצרים) נסים - (والمصريين) فارين ≈ (ומצרים) נסים - (ואלמצריין) פארין A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 785 ‘and the Egyptians were fleeing … ’ (v. 14.27)46 حينئذ أنشأ وسبح موسى وبني إسرائيل ≈ חיניד אנשא וסבח מוסי ובני יסראיל ‘then Moses and the sons of Israel composed/recited and praised… ’ (v. 15.1) Again, a comparison with Saʿadia’s rendering of these same verses in his tenth-century Tafsīr exhibits a distinct inflection for case: .and the Egyptians were fleeing’ (Exod. 14.27; Evr‘ ואלמצריון הארבון II C 1, fol. 145r) and חיניד סבח מוסי ובנו אסראייל (Exod. 15.1; Evr. II C 1, fol. 146r). Indeed, in Classical Arabic, the sound masculine plural form can reflect case even in the consonantal orthography, whether via waw (Hebrew vav) reflecting /ū/ for nominative or via yāʾ (Hebrew yod) reflecting /ī/ for oblique (i.e., genitive or accusa- tive). Elsewhere in our fragment, however, note the true oblique form אלדאכילין ‘who was coming’, which matches the form in Saʿadia’s Tafsīr (Exod. 14.28; Evr. II C 1, fol. 145r). This demonstrates that though Saʿadia inflected for case (or at least some manuscripts reflect- ing his Tafsīr) depending on the context, our text generalized /-īn/. 6.1.2 3MS Suffix as /-u(ː)(h)/ on Nouns in GEN and ACC There are numerous expressions in the continuous translation in which a suffixed form (with the 3MS suffix) exhibits a consistent /u(ː)/ vowel irrespective of its syntactic case (see Blau, 1980, §3; §50): 3MS Suffix Vowel as /u(ː)/ in GEN and ACC الفرس وراكبه رمى في اليم ≈ אלפרץ וראכיבוה רמא פי אלים ‘horse and its rider.ACC he has thrown in the sea’ (v. 15.1) وأخيار قواده طُبِّعوا في يم القلزم ≈ וא׳כיאר קואדּוה טוביעו פי ים אַלקֻלזום ‘the finest of his leaders.GEN were drowned in the Sea of Reeds’ (v. 15.4) إذ دخل فرس فرعون وخيله بركابه وبرجالته في اليم ≈ ׳כַיְלוה ברוכאבוה א׳ד דּ׳כל פרץ פרעון וַ וברגִאלתוה פי אלים ‘when the horses of Pharaoh and his horses.NOM with his riders.GEN and with his foot soldiers.GEN entered the Sea’ (v. 15.19) والذي يضرب أبوه وأمه ≈ ריב אבוה ואומוה ואלדי יצ̇̇ ‘whoever strikes his father.ACC or his mother.ACC’ (v. 21.15) 46 Unless otherwise cited, translations in this article are by the author. Direct translations of the Hebrew Bible are also by the author, but perhaps in consultation with common English translations such as NET and ESV. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 786 That a Judaeo-Arabic translation of the Bible would not inflect for case is by no means a given. In Saʿadia’s Tafsīr of Exod. 21.15 (paral- lel to the first phrase above), case is indeed present: ומן צרב אבאה או ’and whoever strikes his father.ACC or his mother.ACC‘ אמה (Exod. 21.15; Evr. II C 1, fol. 165r), but cf. כלאמה חפט̇̇ ואבוה ‘and his father.NOM kept his word in mind’ (Gen. 37.11; Evr. II C 1, fol. 89r). There is just one case in which an apparently secondary super- script correction might indicate the attempted implementation of case: 3MS Suffix Vowel as /i/ in GEN فيوجد في يده ≈ פיוגִדּ פי ידּוהיהו ‘and it is found in his hand’ (v. 21.16) It is perhaps possible that the scribe originally wrote what would reflect [fiː jadu(h)] and wanted to emend it to reflect [fiː jadihu] so as to reflect a register with case inflection. It might be more likely, however, that this superscript emendation reflects a modification to [fiː jadeːhu] ‘in his (two) hands’, though the Hebrew source ֹבְיָד֖ו ‘in his hand’ (Exod. 21.16) is singular. 6.1.3 3MS Suffix as /-u/ or /-uh/ on Verbs in Suffix Conjugation The 3MS suffix is also consistently /-u/ or /-uh/ when attached to verbs whose parallel in reading traditions of Classical Arabic had a final short /a/ vowel, namely the suffix conjugation and the subjunc- tive (see Blau, 1980, §3; §50): 3MS Suffix as /-u(h)/ on Suffix Conjugation ودله على عود ≈ ודַּלוה עלא עודּ ‘he showed him a stick’ (v. 15.25) شُهِد عليه أيضا بأنه باعه ≈ א בא׳נה באעוה שוהיד עליה איצ̇̇ ‘he was witnessed also to have sold him’ (v. 21.16) Such orthography appears to reflect dialectal pronominal suffixes rather than the expected iʿrāb pattern of prominent Classical Arabic reading traditions. At the same time, it is noteworthy that vocalized Judaeo-Arabic texts with (what are traditionally regarded as) dialectal features do not unequivocally reflect only (what are traditionally A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 787 regarded as) dialectal features. This raises the question about the tar- get register of the authors and/or readers of such texts (cf. Khan, 2017, 397–403). It seems that they were (at least subconsciously) aiming at a register that was neither identical with Classical Arabic nor with any particular vernacular. 6.2 Gender and Number Agreement The gender and number agreement patterns attested in our text are fairly consistent (for a general overview of gender and number agree- ment in Judaeo-Arabic, see Blau, 1980, §182–§198). 6.2.1 Animate Plural Nouns In the case of plural animate nouns—both masculine and feminine— their modifying verbs are (essentially) always plural. This, of course, is the case when the verb comes after the subject:47 Animate Plural with PL Verb after Subject (ומצרים) נסים - (والمصريين) فارين ≈ (ומצרים) נסים - (ואלמצריין) פארין ‘and the Egyptians were fleeing … ’ (v. 14.27) הבאים الداخلين ≈ (הרכב…הפרשים…חיל פרעה) הבאים אלדאכילין ‘who were coming… ’ (v. 14.28) وآمنوا بالرب وبموسى عبده ≈ ואמנו באלרב ובמוסי עבדּוה ‘and they believed in the Lord and Moses his servant’ (v. 14.31) وقالوا قائلين ≈ וקאלו קאילין ‘and they said, saying… ’ (v. 15.1) وأخيار قواده طُبِّعوا في يم القلزم ≈ וא׳כיאר קואדּוה טוביעו פי ים אַלקֻלזום ‘and the finest of his leaders were drowned in the Sea of Reeds’ (v. 15.4) وردوا في القعر مثل الحجر ≈ ורדּו פי אלקער מתל אלחגר ‘and they descended into the depth like a stone’ (v. 15.5) رسبوا كالآبار كالرصاص في المياه الغزيرة ≈ רסבו כִלאַבַאר כִאלריצאץ פי אלמִיאה אל׳גזירֶה ‘they sank as (though in) wells, like lead, in the great waters’ (v. 15.10) 47 For our purposes, verbs without an explicit subject are grouped in the ‘verb after subject’ category. Note also that phrases with parentheses indicate unattested material supplied to give context to entries in the glossarial list. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 788 سمعوا الشعوب ذلك فارتجزوا ≈ סמיעו אלַשוּעֻוב דיליך פארתגִזו ‘the peoples heard this and roared’ (v. 15.14) يسكتوا كالحجر ≈ יסכותו כלחגִר ‘they are quiet like stones’ (v. 15.16) وبني إسرائيل سلكوا في اليبوسة ≈ ובני יסראיל סלכּו פי אליבוסה ‘and the sons of Israel walked on dry ground’ (v. 15.19) This is also the overwhelming norm, however, when the verb pre- cedes the subject (see Blau, 1980, §182): Animate Plural with PL Verb before Subject اتقوا القوم إيا الرب وآمنوا بالرب ≈ אתקוו אלקוום איא אלרב ואמנו באלרב ‘the people feared the Lord and believed in the Lord’ (v. 14.31) سمعوا الشعوب ذلك ≈ סמיעו אלַשוּעֻוב דיליך ‘the peoples heard that’ (v. 15.14) حينئذ دهشوا وتبلدوا صناديد الروم ≈ חיניד דּהישו ואתבלדּו צנאדּידּ אלרום ‘then the brave of the Romans48 were appalled and were confounded ’ (v. 15.15) ماجوا كل جلاس كنعان ≈ מאגִו כול גִולאס כנעאן׃ ‘all the inhabitants of Canaan were agitated ’ (v. 15.15) وخرجن كل النساء وراها ≈ וכרגִן כול אלניסא וראהא ‘and all the women went out after her’ (v. 15.20) ولاموا وأففوا (القوم) ≈ ולאמו ואפפו (אלקוום) ‘and the people blamed and grumbled ’ (v. 15.24) واذا يتناصوا أُناس ≈ ואדא יתנאצוּ אונאס ‘and if people come together in conflict…’ (v. 21.22) There is just one case in which a plural animate subject is preceded by a 3MS verb: Animate Plural with SG Verb before Subject حينئذ أنشأ وسبح موسى وبني إسرائيل ≈ חיניד אנשא וסבח מוסי ובני יסראיל ‘then Moses and the sons of Israel composed/recited and praised…’ (v. 15.1) This exceptional form, which mirrors agreement patterns in some other registers of Arabic (e.g., Classical Arabic), may be due to the fact that there is a compound subject beginning with a singular mas- 48 See section 8.2.2 for more on this translation. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 789 culine noun (i.e., ‘Moses and …’). It is also possible that the 3MS verbal form is resembling the Hebrew source text, which reads ז אָ֣ ל יִשְׂרָאֵ֜ י וּבְנֵ֨ יר־מֹשֶׁה֩ ’Then sang.3MS Moses and the sons of Israel‘ יָשִֽׁ (Exod. 15.1). 6.2.2 Inanimate Plural Nouns Adjectives or verbs modifying plural inanimate nouns, on the other hand, are always and exclusively modified by feminine singular verbs and adjectives in our text. As above, the position of the verb in rela- tion to the subject does not matter. When the verb comes after the plural inanimate subject, the verb is always 3FS (cf. Blau, 1980, §185):49 Inanimate Plural with FS Verb after Subject ויכסו وغطت ≈ (וישבו המים) ויכסו ו׳גטת ‘[the waters] covered’ (v. 14.28) الأغمار غطتهم ≈ אלא׳גמאר ׳גטתּהום ‘the depths covered them’ (Exod. 15.5) This correspondence between inanimate plural and 3FS verbs also applies when the verb precedes the subject (cf. Blau, 1980, §185): Inanimate Plural with FS Verb before Subject تعرمت انتصبت مثل الطود المنهطلة ≈ תערמת אנתצבת מתל אלטוודּ אלמנהטילה ‘the liquids piled up, stood up like a heap’ (v. 15.8) جمدت الأغمار ≈ גִמדּת אלא׳גמאר ‘the depths became solid’ (v. 15.8) There is only one case of an adjective modifying a plural inanimate noun in our text. In this case, the attributive adjective is feminine singular (cf. Blau, 1980, §187–§190): Inanimate Plural with FS Adjective في المياه الغزيرة ≈ פי אלמִיאה אל׳גזירֶה ‘in the great waters’ (v. 15.10) 49 For our purposes, verbs without an explicit subject are grouped in the ‘verb after subject’ category. Note also that phrases with parentheses indicate unattested material supplied to give context to entries in the glossarial list. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 790 6.2.3 Single Relative Particle Also somewhat related to agreement patterns, there appears to be a single relative particle אלדי, related to (or, depending on the pronun- ciation, identical with) Classical Arabic الذي, which is used for both masculine and feminine antecedents (see Blau, 1980, §361): Relative الذي ≈ אלדי for MS and FS الضربة الكبيرة الذي عمل الرب بالمصريين ≈ ׳צרבה אלכבירה אלדי עמיל אלרב אל בלמצרייִן ‘the great blow, which the Lord wrought against the Egyptians’ (v. 14.31) والذي يضرب أبوه وأمه ≈ ריב אבוה ואומוה ואלדי יצ̇̇ ‘and whoever strikes his father or his mother’ (v. 21.15) Other vocalized Judaeo-Arabic texts indicate that, at least among some speakers, the form reflected by אלדי was pronounced differently from in Classical Arabic, whether as something like [iˈladi]/[iˈlaði], [ˈildi]/[ˈilði], or even [ˈaðði]. Note, for example, the pointing אֶלְדִֿי in T-S Ar. 18(1).113 (see Blau and Hopkins, 1985, 466–467). 6.3 Subordinate Clauses 6.3.1 baʿd + Verb without mā/ʾan There is one case of a temporal subordinate clause in which the sub- ordinating particle is not followed by mā/ʾan before the verb (see Blau, 1980, §337): Particle baʿd + VERB بعد غدر بها ≈ בעדּן ׳גדּיר בהא ‘after he dealt treacherously with her’ (v. 21.8) This may be contrasted with Saʿadia’s Tafsīr, in which we find phrases like בעד אן בלית ‘after I have been worn out … ’ (Gen. 18.12; Evr. II C 1, fol. 37v) and בעד מא קלב אלקרי ‘after he overturned the cit- ies … ’ (Gen. 19.29; Evr. II C 1, fol. 40r). At the same time, there appears to be a faint (possibly from a second hand?) final nun at the end of the sequence ּבעד in our fragment, possibly reflecting בעדּן as a shortening of בעד אן (see Blau, 1980, §47). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 791 6.3.2 kallaḏī instead of kammā There is one case of the lexeme كالذي to introduce a subordinate clause where we might expect to find the more typical Arabic كما in some other registers (see Hopkins, 1978, §283; Blau, 2002, §134): Particle كالذي فجرامة يجرم كالذي يجعل عليه بعل المرة ≈ פ׳גראמה יגרם כאלדי יגעל עליה בעל אלמרה ‘he will surely be punished just as the husband of the woman would obligate him’ (v. 21.22) While one could take this as a straight calque of Hebrew כַּאֲשֶׁר ‘as; in accordance with…’—made up of Hebrew ְּכ ‘like’ + אֲשֶׁר ‘that; which’—it is worth noting that this particle is attested in medieval Arabic papyri, particularly in epistolary formulae: e.g., سالمون ونحن يسركم كالذي and we are well and good, as it may please‘ صالحون you’; ونحن كالذي تحب ‘and we are as you would like’; كالذي كان أولا ‘as it had been at first’ (Hopkins, 1978, §283; Blau, 2002, §134). As such, there is no reason to assume that كالذي is not an authentic medieval Arabic form; it need not be regarded as a Hebrew calque. 7. Lexicon There are a number of noteworthy lexical phenomena attested in our fragment. In fact, the lexicon is perhaps the area where the influence of Hebrew—whether loanwords or influence of the source text—is most readily apparent. 7.1 Influence of Realization of Hebrew Cognate on Root Choice Although this might be more appropriately cited in the section on phonology, due to its relevance for lexical choice, we may cite here a couple cases in which an unexpected Arabic verb is used to better match the realization of the cognate root in the Hebrew. This is a recognized feature among many medieval Judaeo-Arabic glossaries and dictionaries (see Ratzaby, 1985; Maman, 1992; Polliack, 1997, 170–174, 179–180; Blau and Hopkins, 2017, 86–90): A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 792 Translation Choice to Match Hebrew Cognate Root Hebrew Source Example Possible Realization? הַיָּֽם׃ (v. 14.30) اليم ≈ אלים [al-jamm] ז אָחַ֔ (v. 15.14) أحاز ≈ אחאז [(ʔ)aħaːz] אחֲזֵ֖מוֹ ֹֽ י (v. 15.15) يحوزهم ≈ יחוזהום [jaħuːzhum] In the first instance, the Hebrew word יָם is rendered by اليمّ ≈ אלים, which is curious given that بحر baḥr is the more common Arabic term, even in Saʿadia’s Tafsīr. Nevertheless, the form ّاليم al-yamm is attested in Quranic Arabic (Q7:13; Q20:39, 78, 97; Q28:7, 40; Q51:40), even if بحر baḥr is more common in the Quranic corpus overall (41×). The form يم ‘sea’ is also attested in a number of early Gospel translations, even those based primarily on Greek Vorlagen: e.g., اذهب قاصداٍ الى يم ‘Go straight away to the sea’ (Leiden Or. 561, 15th c. CE) (Stokes, 2023d, 434–435). Nevertheless, even though يم yamm ‘sea’ is attested in various registers of Arabic in the Middle Ages, in light of the fact that the expected (or more common) term would be بحر baḥr, the translation choice of يم yamm could be at least partially motivated by the Hebrew cognate here. If so, its use in Gos- pel translations could perhaps even point to some degree of interface between Arabic Bible translation traditions of Christians and Jews in the Middle Ages. In the second instance, the word אחאז occurs in the phrase אחאז the agony of childbirth possessed the inhabitants‘ אלטלק גִולאס פליצטין of Philistia’, which translates Hebrew שֶׁת׃ פְּלָֽ י ישְֹׁבֵ֖ ז אָחַ֔ יל .Exod) חִ֣ 15.14).50 In the third instance, the word יחוזהום occurs in the phrase אלארתעאדּ יחוזהום מאב ,the powerful (leaders) of Moab‘ מוטיקי trembling possesses them’ (Exod. 15.15). Given the cognate nature and similar semantics of Arabic أخذ and Hebrew אָחַז, we would expect that utilizing the Arabic cognate would be the most likely translation choice. This is, in fact, exactly what we find in Saʿadia’s Tafsīr in both contexts: פלסטין סכאן אלטלק ד̇̇ ;Exod. 15.14) ואכ̇̇ Derenbourg, 1893) and תהם אלרעדה ד̇̇ לא מואב אכ̇̇ ;Exod. 15.15) ואג̇̇ 50 For طلق as the ‘pain of childbirth’, see the entry ٌطَلْق in Lane’s lexicon and the entry طالقة ‘woman in labour’ in Blau’s lexicon (2006, 408). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 793 Derenbourg, 1893).51 It may be the case, then, that the choice of a much rarer52 verb like Arabic حاز/أحاز could be based on an imita- tion of the realization of the cognate root in Hebrew, which has [ħ] and [z] rather than [χ] and [ð] (see also Vollandt, 2013; 2014, 66). This exact parallel between Arabic حاز/أحاز and Hebrew אח״ז appears to be evidenced in other Judaeo-Arabic Bible translations of the Mid- dle Ages (see Blau and Hopkins, 2017, 89). This phenomenon appears to have been fairly widespread among Biblical Hebrew commentators, translators, and lexicographers writ- ing in Judaeo-Arabic. Note, for example, how Yefet ben ʿEli renders Biblical Hebrew ׁיֵש with איס = [(ʔ)ajs], presumably formed from chopping off the negative etymology of Arabic َلَيْس laysa: e.g., ומלעון כר תאל ואיס פי קטיעה ד̇̇ ר for מג̇̇ ל וְיֵשׁ֤ בְּעֶדְרוֹ֙ זָכָ֔ and cursed be‘ וְאָר֣וּר נוֹכֵ֗ the cheat who has a male in his flock … ’ (Mal. 1.14) (see Polliack, 1997, 174; Blau and Hopkins, 2017, 51; Nadler-Akirav, 2021, 153). A similar sort of ‘seeking after cognates’ appears to have been an impulse undergirding al-Fāsī’s lexicographical work. When providing glosses for the Biblical Hebrew synonyms בָּא ‘to come’ (typical) and to come’ (poetic), for example, he tends to gloss the former‘ אָתָה primarily with א ג̇̇ see) أتى ≈ אתי and the latter more so with جاء ≈ Maman, 1992, 121–122), thus demonstrating a propensity to cite cognates of otherwise synonymous forms where possible. Vollandt (2014, 66) refers to this lexical phenomenon as the employing of ‘cognate homophones’ or ‘pseudo-cognate roots’ (see also Polliack, 1997, 170–174, 179–180; Vollandt, 2013). 7.2 Hebrew Loanwords/Loanroots Perhaps the most obvious area of Hebrew influence on the lexicon concerns what appear to be loanwords or ‘loanroots’. The latter term merely refers to the use of a borrowed Hebrew verbal root within Arabic verbal morphology to convey the semantics associated with the Hebrew root. The following examples are likely candidates for loanwords and ‘loanroots’, in which the meaning of the root in Hebrew—rather than the semantics associated with the root in Ara- 51 Derenbourg (1893) is the source of Saʿadia’s Tafsīr here due to the fact that the relevant portion is missing from Evr. II C 1, which elsewhere serves as the default manuscript source for the Tafsīr in this paper. 52 Note that the Arabic verb حاز/أحاز appears rarely in Saʿadia’s Tafsīr. When it does occur, it translates verbs like ׁיָרַש ‘to inherit; to possess’, נָחַל ‘to inherit; to possess’; הִתְנַחֵל ‘to take possession of’. On one occasion, however, it does translate the nifʿal form נֶאֱחַז ‘to have possession of’ (Gen. 47.27). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 794 bic—clearly attests to the appropriate meaning for the context (for the discussion of roots in the Karaite translation tradition, see Polli- ack, 1997, 170–174; see also Blau and Hopkins, 2017, 86): Loanwords Example Hebrew cf. Arabic وأنويه ≈ ואנויהוה and I‘ וְאַנְוֵהוּ praise him’ n-w-y ‘to intend; to purpose’ (see Blau, 2006, 725–726; but cf. the noun תנויה ‘praise’ in Friedman 2016, 503) طُبِّعوا ≈ טוביעו were‘ טֻבְּעוּ drowned’ ṭ-b-ʿ ‘to seal; to imprint’ (see Blau, 2006, 394) أنهلته ≈ אנהלתוה ’you led‘ נֵהַלְתָּ n-h-l ‘watering place’ (see Lane, 1863–1893, 3039) It is not entirely clear why Hebrew loanwords were used rather than a more natural Arabic cognate in such contexts. While it is tempting to argue that such lexical choices communicate something significant about the translation technique of the author, it is equally possible that they convey something about the use of Hebrew loanwords in the parabiblical traditions of the translator’s community. Alterna- tively, they may also simply call attention to lexical items which, though less common, were current in the world of Medieval Arabic. 8. Exegetical Strategies and Interpretive Tradition(s) Although the orthography and register of the texts make for some particularly intriguing linguistic data on medieval (Judaeo-)Arabic— as we have outlined above—perhaps more interesting for the phe- nomenon of Judaeo-Arabic translation are those exegetical strategies and interpretive traditions reflected in T-S Ar. 28.170 + T-S Ar. 27.60. 8.1 Double Glosses or Alternative Translations One of the most characteristic and regular exegetical strategies reflected in our fragment concerns the presence of ‘double glosses’ or, as Polliack (1993; 1997, 181) refers to it in her work on Karaite Bible versions, ‘alternative translations’ (see also Blau and Hopkins, 2017, 75–84). On frequent occasions, the translator does not furnish just a single gloss for a Hebrew word but presents his readers with A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 795 several translation options. While this begins in the bicolumnar glos- sarial lists alongside a single Hebrew tag-word from the biblical text, it is also a feature of the running continuous translation and may even work on the phrase level. Such ‘double glosses’ are present in almost every verse of the translation. For the sake of space, however, we have included just those in the beginning part of the text by way of illustration (up through verse 15.6): Double Glosses in T-S Ar. 28.170 + T-S Ar. 27.60 אפן מרכבתיו || لَوْلب יעני نعال دوابه || مراكبه ≈ אפן מרכבתיו || לוולב יעני ניעל דואבוה || מראכיבוה ‘ʾPN MRKBTYW: || spiral, that is, shoes (of); his riding animals, his riding vessels’ (v. 14.25) נסים فارين || לקראתם تلقاءهم יעני يفروا حذاءهم ≈ נסים פארין || לקראתם תלקאהום יעני יפירו חידאהום ‘NSYM: fleeing || LQRʾTM: in front of them that is, they flee opposite them’ (v. 14.27) וינער ونفّض وغرّق ≈ וינער ונפץ̇̇ ו׳גַרַק ‘WYNʿR: and he shook, and he drowned ’ (v. 14.27) וישע وأغاث ونصر ≈ וישע וא׳גאת ונצר ‘WYʃ ʿ: and [the Lord] rescued and delivered ’ (v. 14.30) إيا المقدورة الضربة الكبيرة ≈ [את היד הגדלה] איא אלמקדּורה אל׳צרבה אלכבירה ‘(about ʾT HYD HGDLH:) that which is decreed, the great blow’ (v. 14.31) אז׃ حينئذ أنشأ وسبّح ≈ אז [ישיר]׃ חיניד אנשא וסבח ‘ʾZ: Then he/they composed/recited and praised ’ (v. 15.1) أُنشئ أُسبِّح ≈ [אשירה] אֻנשי אוסביח ‘I will compose/recite, I will praise’ (v. 15.1) עזי׃ عِزّي ذو العِزّ الأزلي والمجيد عِزّي وذو المجد الأزلي ≈ עזי [וזמרת יה]׃ עיזי דו אלעיז אלאזלי ו[א]ל מיגִידּ עיזי ודו אלמגִדּ אלאזלי ‘ʿZY: my strength, the one possessing eternal renown and the glorious one my strength, and the one possessing eternal glory’ (v. 15.2) وكان لي للمغوثة غياث ≈ [ויהי לי לישועה] וכאן לי ללמ׳גותה ׳גיאת ‘and he was to me as a deliverance, a help’ (v. 15.2) هذا الطائق مطيقي ≈ [זה אלי] הדא אלטאיק מוטיקי ‘this is the powerful one, my enduring one’ (v. 15.2) A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 796 الرب [أر؟]د ولي الملاحم ≈ [יהוה איש מלחמה] אלרב [אר?]דּ ולי אלמלאחים ‘the Lord, [most proficient?? of], the manager of wars’ (v. 15.3) مراكب خيل دواب فرعون ≈ מרכבות: מראכיב כיל דּואב פרעון ‘MRKBWT: the chariots of the horses of the riding animals of Pharaoh’ (v. 15.4) زج زدى في اليم ≈ [ירה בים] זגִ זדּא פי אלים ‘he threw, he cast(?) into the sea’ (v. 15.4) ימינך يمينك يا رب بطل ذو غزر في القوة ≈ ימינך [יהוה נאדרי בכוח] ימינךּ יא רב בטל דו ׳גוזר פי אלקווה ‘YMYNK: your right hand, O Lord, a hero, possessing abundance in strength’ (v. 15.6) While some of the above examples constitute mere synonyms, there are some cases in which the first gloss appears to be more literal (or etymological) and the second gloss more idiomatic (or contextual), as in the case of ו׳גַרַק ונפץ̇̇ וינער ≈ وغرّق ونفّض WYNʿR: and he‘ וינער shook, and he drowned’ (v. 14.27) (for ‘etymological’ alongside ‘con- textual’ glosses, see Polliack, 1997, 182–183). While ‘to shake off’ is the more ‘literal’ meaning of the root n-ʿ-r in Hebrew, the translator seems to want to elucidate it by specifying that in the context it means ‘to drown’. In other cases, rather than listing the glosses back to back, they are connected with the conjunction vav, as in אנשא وسبح ≈ וסבח composed/recited and praised’ (v. 15.1), which‘ أنشأ could border on actual translation and thus intend a sort of phasal interpretation of יָשִׁיר which appears to be present in Yeshuʿah ,אָז ben Yehudah’s translation of the phrase as ينشد ≈ חיניד אכד מוס[י] פי אן ינשד أن في موسى أخذ then Moses took up (i.e., began)‘ حينئذ singing’.53 This phenomenon of ‘double glossing’ or ‘alternative translations’ is a prominent feature in Karaite translation technique, especially that of Yeshuʿah ben Yehudah (see, e.g., Polliack, 1993/94; 1997, 181–199; Tirosh-Becker, 2007). What is particularly noteworthy here, is that these alternative ren- derings occur far more frequently in T-S Ar. 28.170 than they do in Yeshuʿah’s translation of the same exact passage (see BL Or. 2398, fol. 4r–6v, 48v–49r), even though Yeshuʿah especially has a reputa- tion for utilizing alternative translations. Moreover, they are not always introduced with phrases like ויקאל ‘and it is said’ or או ‘or’, as 53 For the phasal interpretation of ʾaz + yiqṭol … vayyiqṭol, see Khan (forthcom- ing). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 797 are typical in Yeshuʿah’s translation, but sometimes are simply juxta- posed without any obvious introductory phrase. Sometimes they are even connected via the conjunction waw as if part of the translation. This may reflect a typologically more primitive stage of ‘double gloss- ing’ or ‘alternative translations’ and may be related to the fact that our fragment is a hybrid between glossary and translation. Blau and Hopkins (2017, 75) identify a similar tendency, namely a high fre- quency of double- and triple-renderings, in their collection of biblical translations and glossaries with phonetic spelling. 8.2 Interpretive and Grammatical Traditions Behind Various Glosses In numerous cases, the choice of a particular rendering may convey something significant about the interpretive tradition of the transla- tor or his community. In other cases, a particular rendering may convey something about the grammatical understanding of the trans- lator. A selection of instructive examples is presented below. 8.2.1 Metaphorical Interpretations of Hand Throughout our fragment, the Hebrew word יָד ‘hand’ is translated in different ways. While it is most frequently rendered with the Ara- bic cognate يد (v. 14.30; 15.9, 17, 20; 21.16, 20), there is one instance in which it is interpreted more metaphorically. In the phrase ה is rendered as both יָד the great hand’ (v. 14.31), the word‘ הַיָּד֣ הַגְּדלָֹ֗ that which is decreed; that which is invested‘ المقدورة ≈ אלמקדּורה with power’ and الضربة ≈ אל׳צרבה ‘the blow’. A similar translation choice is attested in Saʿadia’s Tafsīr, in which ה הַגְּדלָֹ֗ .Exod) הַיָּד֣ 14.31; Evr. II C 1, fol. 145v) is rendered as ימה الآفة ≈ אלאפה אלעט̇̇ the great plague’. For a similar phenomenon, we might also‘ العظيمة compare the rendering of יָמִין ‘right hand’ as سلطان ‘might’ and ّعِز ‘strength’ in T-S Ar. 28.49 (fol. 1v): בילקווה גזירה רב יא סולטאנך سلطانك يا رب غزيرة بالقوة ≈ סולטאנך עיזך יא רב תרהיב תקמע אלבאגיץ̇̇ -your might, O Lord, is abun‘ سلطانك عِزّك يا رب ترهب تقمع الباغض dant in power, your might, your strength, O Lord, destroys, represses the enemy’ (Exod. 15.6). In Yefet’s commentary on this passage we similarly find יָמִין ‘right hand’ rendered as قدرة ‘strength’ in the phrase your power, O God …’ (Evr. Arab. I‘ قدرتك يا الله ≈ קדרתך יא אללה 125, fol. 31v). This is consistent with a wider phenomenon among Judaeo-Arabic translators of understanding words like יָד and יָמִין as metaphorical, especially when applied to God, to avoid anthropo- morphism. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 798 8.2.2 Edomites → Romans as the Oppressors Another interesting interpretive tradition is found in our fragment’s rendering of the proper noun אֱדוֹם ‘Edom’.54 In Exod. 15.15, the translator renders אֱדוֹם ‘Edom’ as الروم ‘the Romans/the Byzantines’ in the phrase אלרום צנאדּידּ ואתבלדּו דּהישו then the brave ones‘ חיניד of the Romans/Byzantines were astonished, confounded’. This may reflect an interpretive tradition that juxtaposed more contemporary political entities with ancient ones. This interpretation notably con- trasts with Saʿadia’s Tafsīr, in which the proper noun אדום is simply rendered as is (Derenbourg, 1893 in lieu of Evr. II C 1 attestation). Other fragments render it more in line with the root meaning of the word: e.g., חיניד דהישו צנאדיד אל אחמר ‘then the brave ones of the red were astonished’ (T-S Ar. 28.49, fol. 1r). 8.2.3 The Desert ‘Speaks’ In the glossarial list covering Exod. 15.22–25 on fol. 1v of T-S Ar. 28.170, the place name מִדְבַּר־שׁ֑וּר ‘the wilderness of Shur’ (Exod. 15.22) is rendered as גִיפאר יק עואצ̇̇ جِفار ≈ בַּרִיֶה عواصيق55 It is .برية not clear if the latter expression should be taken as a single phrase or as a double gloss of some sort. We may find a clue in T-S Ar. 27.114a, where the phrase ין שׁ֑וּר שׁ וּבֵ֣ .between Kadesh and Shur’ (Gen‘ בֵּין־קָדֵ֖ 20.1) is rendered as بين قدس وبين ≈ בין קדס ובין אלגיפאר ויקאל אלעואציק العواصيق ويقال ,between Kadesh and al-jifār, that is to say‘ الجفار al-ʿawāṣīq’ (Butbul and Hopkins, 2024, 22). The rendering of جفار ‘wells’ is also attested in Saʿadia’s Tafsīr for שׁוּר ‘Shur’ in Gen. 20.1 (Evr. II C 1, fol. 41v) and Exod. 15.22 (Derenbourg, 1893) and in Yeshuʿah ben Yehudah’s translation of Exod. 15.22 (BL Or. 2398, fol. 48v). According to Butbul and Hopkins (2024, 9), al-jifār is identified with a particular place in various geographical literature of the Middle Ages.56 In the case of عواصيق, on the other hand, there is no clearly identifiable place of that name. One possible explana- tion—though admittedly speculative—is that عواصيق is not actually a double gloss for שׁוּר ‘Shur’ but rather for מִדְבָּר ‘wilderness’. Although unattested elsewhere in Arabic literature, the similar word 54 Note that this interpretive tradition of identifying Edom with the Romans may be found already in late antiquity in the Dead Sea Scrolls and in Bereshit Rabba (see Finkel, 1963, 362). It is, in fact, quite common in Rabbinic literature. 55 This is assuming that the diacritic above ṣade in יק .is an error for ṣād עואצ̇̇ 56 Above גִיפאר, T-S Ar. 28.170 also attests to a very faint (and very uncertain) superscript word  ??ח?? which is glossed in ,مضحاة which could be related to ,?מצ̇̇ Lane’s lexicon as ‘a land from which the sun is hardly, or never, absent’. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 799 واللغط is glossed as عصاقياء/عصاقية clamour and noise’ in the‘ الجلبة 14th/15th-century lexicon al-muḥīṭ. It may be, then—speculatively— that the association of מִדְבָּר ‘wilderness’ with the root דב״ר ‘word(s); speaking’ in Hebrew could have brought about a rendering that highlighted the meaning of the root. 8.2.4 Grammatical Gender of neʾdari The first part of the Hebrew text of Exod. 15.6 reads י ה נֶאְדָּרִ֖ ינְךָ֣ יְהוָ֔ יְמִֽ חַ -your right hand, O LORD, (is) majestic in strength’. The gram‘ בַּכֹּ֑ matical (and/or exegetical) issue here is that while most translations prefer to take י ינְךָ֣ majestic’ as applying to‘ נֶאְדָּרִ֖ ,’your right hand‘ יְמִֽ thus rendering the phrase as ‘your right hand … is majestic in strength’, the morphology does not necessarily align with this interpretation. After all, while the noun יָמִין ‘right hand’ is feminine, the nifʿal parti- ciple י appears to be the MS participle with the addition of the נֶאְדָּרִ֖ so-called ḥireq compaginis. If the participle were modifying יָמִין ‘right hand’, we would expect it to be נֶאְדֶּרֶת or—with ḥireq compaginis— /If it is a MS participle, then, it would likely be an appositive .נֶאְדַּרְתִּי vocative with the Lord as the referent: i.e., ‘your right hand, O LORD, (O you who are) majestic in strength’. One suggestion, however, has argued that the final ḥireq could reflect an archaic/poetic feminine morpheme /-ī/ and thus modify יָמִין ‘right hand’ (Butts, 2010). With this grammatical background in mind, it is instructive to take note of the fact that most Judaeo-Arabic translations understand י ,as applying to the right hand. In Saʿadia’s Tafsīr, for example נֶאְדָּרִ֖ we read يمينك يا رب جزيلة القوة ≈ ימינך יא רב גזילה אלקוה ‘your right hand, O Lord, (is) abundant in strength’ (Evr. II C 1, fol. 146v). This interpretation also appears to be reflected in the commentary of Avraham Maimonides, which reads כאנה ת̇א̇ מקאם הנא אלי̇א̇ נאדרי נאדרת ≈ נאדרת كأنه تاء مقام هنا neʾdari, the yāʾ in it is‘ נאדרי الياء instead of tāʾ, as if it were neʾderet’ (Wiesenberg, 1959, 273). In Yefet’s commentary on this passage, on the other hand, we find the phrase rendered as ליל באלקוה قدرتك يا الله يا جليل بالقوة ≈ קדרתך יא אללה יא ג̇̇ ‘your power, O God, O Majestic One in strength’ (Evr. Arab. I 125, fol. 31v). Yeshuʿah ben Yehudah does something quite clever, in that he renders י as a vocative but adds ‘dialectal tanwīn’ to mimic נֶאְדָּרִ֖ the ḥireq compaginis: לילא באלקוה قدرتك يا رب ≈ קדרתךּ יא רב יא ג̇̇ your power, O Lord, O Majestic One in strength’ (BL‘ يا جليلا بالقوة Or. 2398, fol. 5v). In our fragment, however, there appears to be something of a mid- dle ground. Even though the above renderings of Saʿadia and the A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 800 Karaites understand the referent of י differently, they all render נֶאְדָּרִ֖ it as an adjective. In T-S Ar. 28.170, however, at least one of the renderings—it is not clear if the translation is a ‘double gloss’ or a single phrase—appears to substantivize י :and take it as a noun נֶאְדָּרִ֖ Exod. 15.6a in T-S Ar. 28.170 ימינך يمينك يا رب بطل ذو غزر في القوة ≈ ימינך [יהוה נאדרי בכוח] ימינךּ יא רב בטל דו ׳גוזר פי אלקווה ‘YMYNK: your right hand, O Lord, (is) a hero, possessing abundance in strength’ (v. 15.6) At least the translation بطل ≈ reflects a substantivizing of the בטל participle. As potentially a predicative noun, then, it would not have to match the gender of its referent and could thus apply to the ‘right hand’ or to ‘the Lord’. The lack of the vocative particle yā, however, may indicate that the ‘right hand’ as referent is most likely. This rendering could be a clever way of dealing with the grammatical/ exegetical difficulty not attested in other Judaeo-Arabic translations. To put it simply, while the sense of the verse would encourage taking right hand’ as the referent, the grammar does not allow this on‘ יָמִין its face, so long as one takes the participle as directly modifying its referent, essentially as an adjective. By substantivizing the participle and treating it as a noun, however, the translator allows it to be understood predicatively (or in apposition) and apply to יָמִין ‘right hand’. 9. Relationship to Other Judaeo-Arabic Translation Traditions Though the fragment itself is likely post-Saʿadia, being dated to between the 11th and 13th centuries CE, this does not necessarily mean that the compositional history of the text it records or the translation tradition to which it belongs does not precede or coincide with the time of Saʿadia. As such, before concluding, it behooves us to conduct a brief textual comparison of the translation attested in T-S Ar. 28.170 + T-S Ar. 27.60 to other prominent Judaeo-Arabic Bible translations of the wider period, namely that of Saʿadia, on the one hand, and the Karaites, Yefet ben ʿEli and Yeshuʿah ben Yehu- dah, on the other. In this brief comparison, we are utilizing MS RNL Evr. II C 1 for Saʿadia, MS RNL Evr. Arab. I 125, BL Or. 2467, and MS IOM (of RAS) B 367 for Yefet, and MS BL Or. 2398 for A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 801 Yeshuʿah (OPenn 2024/2025; Ktiv, 2024/2025). Select comparisons of their translations with the relevant passages of our fragment are displayed in the following tables: Saʿadia (Evr. II C 1) Yefet (Evr. Arab. I 125) Yeshuʿah (BL Or. 2398) T-S Ar. 28.170 15.4 מראכב פרעון וגנודה זדי בהם פי אלבחר וכיאר קואדה גרקו פי בחר אלקלזם מראכב פרעון וגישה רשק פי אלבחר וכיאר קואדה רסכו פי בחר אלקלזם ישה מואכב פרעון וג̇̇ זג̇̇ פי אלבחר וכיאר קואדה ארסכו פי בחר אלקלזם בקדרתה מראכיב כיל דּואב פרעון וגִחפלוה זג זגִ זדּא פי אלים ומבחר וא׳כיאר קואדּוה טוביעו פי ים אַלקֻלזום 15.5 ואלגמור גטתהם נזלו פי אלקער כאלחגארה טיהם מור תג̇̇ אלג̇̇ אנחדרו פי אלקעור מתל אלחגִר וצאר אלגמור יהם ענד מא תגטّ אנחדרו פי אלקעור ר מתל אלחג̇̇ אלא׳גמאר ׳גטתּהום ורדּו פי אלקער מתל אלחגר Saʿadia (Evr. II C 1) Yefet (BL Or. 2467) Yeshuʿah (BL Or. 2398) T-S Ar. 28.170 15.19 N/A אן ענד מא דכ̇ל כיל פרעון מע מואכבה ופרסאנה פי אלבחר ורד רב אלעאלמין עליהם מא אלבחר ובנו אסראיל סארו פי אליביס פי וסט אלבחר N/A פלמא א׳ד דּ׳כל פרץ ׳כַיְלוה פרעון וַ ברוכאבוה וברגִאלתוה פי אלים ורד אלרב עליהום איא מיאה אלים ובני יסראיל סלכּו פי אליבוסה פי וסט אלים Saʿadia (Evr. II C 1) Yefet (B 367) Yeshuʿah (BL Or. 2398) T-S Ar. 28.170 21.15 ומן צרב אבאה או אמה פליקתל קתלא ארב אבאה או וצ̇̇ ה יקתל קתלא אמّ ארב אבוה ואמה וצ̇̇ קתלא יקתל׃ ריב אבוה ואלדי יצ̇̇ ואומוה מוות יומאתּ 21.16 ומן סרק אנסאנא פבאעה או וגד פי ידה פליקתל קתלא וסארק אנסאנא ויביעה או יגד פי ידה יקתל קתלא ומן סרק אנסאנא ד פי פבאעה או וג̇̇ ידה פליקתל קתלא ואלדי יסריק אִמר מן יסראיל ויביעוה פיוגִדּ פי ידּויהו יביעוה מות יומ[א]ת אוו שוהיד א בא׳נה עליה איצ̇̇ באעוה פמוות יומאת On the whole, at least when compared to Saʿadia and the Karaites, the translation attested in T-S Ar. 28.170 + 27.60 appears to reflect an independent translation tradition not derived from either. There are numerous pieces of evidence for this suggestion. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 802 In terms of the lexicon, our fragment has a few unique features that set it apart from these other medieval translators, such as: man’ (though this portion is‘ אִישׁ for biblical רגל instead of אִמר — unattested in Yefet) ’sea‘ יָם for biblical בחר instead of ים — ’to do; to make‘ עָשָׂה for biblical פעל or צנע instead of עמיל — the able ones; the powerful ones’ instead of‘ אלמוטיקין/אלקאדרין — ’gods‘ אֵלִים the worshipped things’ for biblical‘ אלמעבודאת מאב — מואב the able ones of Moab’ instead of‘ מוטיקי אגלי/אגלא ‘the preeminent ones of Moab’ for biblical אֵילֵי מוֹאָב ‘rams/rulers of Moab’ In terms of morpho-syntax, our fragment frequently introduces defi- nite direct objects with איא (cf. biblical אֵת/אֶת־), whereas such a marker is absent in Saʿadia and Yeshuʿah’s renderings of the passage and, though present, used far less frequently in Yefet’s (note its use in Exod. 15.1).57 Note that the use of איא to introduce direct objects is a common feature of biblical translations with phonetic spelling (Blau and Hopkins, 2017, 50–51). From a religious-linguistic per- spective, while our fragment uses אלרב ‘the Lord’ for the tetragram- maton, Saʿadia and Yeshuʿah use both אללה ‘God; Allah’ and אלרב ‘the Lord’, and Yefet uses the Quranic term רב אלעאלמין ‘Lord of the worlds’. There are, however, numerous cases in which a ‘double gloss’ or ‘alternative translation’ is provided, one of which corresponds to the word choice of Saʿadia and/or the Karaites (Yefet and Yeshuʿah) and one of which appears to be unique to our fragment. Note the follow- ing examples: -and tri‘ ונצר .and rescued’ (w/ Saʿadia and Karaites) vs‘ וא׳גאת — umphed’ for biblical וַיּ֨וֹשַׁע ‘and saved’ (Exod. 14.30) possessing‘ דו אלעיז .my strength’ (w/ Saʿadia and Yefet) vs‘ עיזי — strength’ for biblical י my strength’ (Exod. 15.2)‘ עָזִּ֤ /your adversaries; those who rise against you’ (w‘ מ(ו)קאומיך — Saʿadia and Karaites) vs. אלמונאצב ‘the opposing one’ for biblical יךָ your adversaries; those who rise against you’ (Exod. 15.7)‘ קָמֶ֑ צאר .accumulated’ (w/ Saʿadia and Karaites) vs‘ תערם/תערמת — were piled‘ נֶעֶ֣רְמוּ became heaps and heaps’ for biblical‘ ערם ערם up’ (Exod. 15.8) 57 For the association of this feature with Karaites, see Polliack (1997, 285–287) as well as Blau and Hopkins (2000, 6). A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 803 were astonished’ (w/ Saʿadia and Karaites, who use forms‘ דהישו — of the same root) vs. ואתבלדּו ‘and were confounded’ for biblical .were dismayed; were terrified’ (Exod. 15.15)‘ נִבְהֲלוּ֙ This may point to an awareness of other (perhaps more prominent or mainstream) translation traditions. Juxtaposed with the translation features particular to T-S Ar. 28.170 + 27.60, this may suggest that producing innovative translations, all the while interacting with dif- ferent existing translation traditions, was a means of engaging in the wider exegetical and theological conversation of various Jewish com- munities of the period. Despite its independent textual basis from Saʿadia, Yefet and Yeshuʿah, there are several features that may recommend ascribing a Karaite source to T-S Ar. 28.170 + 27.60. One such feature is the high frequency of double glosses or alternative translations, which is characteristic of the Karaites and especially Yeshuʿah ben Yehudah. Similarly, the unusually high frequency of איא, a feature commonly associated with Karaites (see Polliack, 1997, 285–287; Blau and Hopkins, 2000, 6), is particularly noticeable. A high propensity for loanwords and/or cognate roots in translation is also a feature shared by our fragment with Karaite Bible translations (see Polliack, 1997, 170–174). 9. Conclusions The publication of this fragment is unlikely by itself to drastically alter our understanding of the Arabic language varieties of the com- munities to which the translators of various medieval Judaeo-Arabic translations belonged. On the whole, however, the linguistic analysis of medieval Judaeo-Arabic depends on the cumulative power of many such isolated publications to shed light on both similarity and diver- sity in the language(s) and register(s) of the scribes and their com- munities. Moreover, the data gleaned specifically from Tiberian- vocalized words in such texts may help address the wider question of the nature and/or existence of ‘reading traditions’ for medieval Arabic Bible translations. Finally, the translational features and exegetical strategies reflected in T-S Ar. 28.170 + T-S Ar. 27.60—from codi- cology to glossarial lists, ‘double glossing’, and interpretive tradi- tions—can provide insight into an early and formative stage in the history of Judaeo-Arabic Bible translation as well. In terms of its communal setting, though independent from other attested trans- lations of the Middle Ages, our fragment carries several features A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 804 typically associated with Karaite translation traditions, albeit some- times in somewhat idiosyncratic form. MANUSCRIPTS CITED MS Cambridge, CUL T-S Ar. 5.51 MS Cambridge, CUL T-S Ar. 5.58 MS Cambridge, CUL T-S Ar. 8.3 MS Cambridge, CUL T-S Ar. 18(1).113 MS Cambridge, CUL T-S Ar. 27.60 MS Cambridge, CUL T-S Ar. 27.111 MS Cambridge, CUL T-S Ar. 27.114a MS Cambridge, CUL T-S Ar. 28.49 MS Cambridge, CUL T-S Ar. 28.116 MS Cambridge, CUL T-S Ar. 28.132 MS Cambridge, CUL T-S Ar. 28.137 MS Cambridge, CUL T-S Ar. 28.151 MS Cambridge, CUL T-S Ar. 28.154 MS Cambridge, CUL T-S Ar. 28.168 MS Cambridge, CUL T-S Ar. 28.170 MS Cambridge, CUL T-S Ar. 31.52 MS Cambridge, CUL T-S Ar. 53.12 MS Cambridge, CUL T-S Ar. 54.63 MS Cambridge, CUL T-S AS. 25.272 MS Cambridge, CUL T-S NS 91.12 MS London, BL Or. 2398 MS London, BL Or. 2467 MS New York, JTS Adler 3316 MS Oxford, Bod. Heb. c. 18.69 MS St. Petersburg, RNL Evr. II C 1 MS St. Petersburg, RNL Evr. Arab. I 125 MS St. Petersburg, IOM (of RAS) B 367 WORKS CITED Al-Jallad, Ahmad. 2020. The Damascus Psalm Fragment: Middle Arabic and the Legacy of Old Ḥigāzī. Chicago: The Oriental Institute of the University of Chicago. Allony, Nehemiah. 1969. ‘ספר הניקוד לרב סעדיה גאון’ [The Book of Niqqud of Rav Saʿadia Gaon]. Beit Mikra 15: 19–67. Badawi, el-Said, and Martin Hinds. 1986. A Dictionary of Egyptian Arabic. Beirut: Librairie du Liban. Baker, C. F., and Meira Polliack. 2001. Arabic and Judaeo-Arabic Manuscripts in the Cairo Genizah Collections, Arabic Old Series (T-S Ar. 1a–54). Cambridge: Cam- bridge University Library. Birnstiel, Daniel. ‘Classical Arabic’, in The Semitic Languages: Second Edition, John Huehnergard, and Na’ama Pat-El (eds). London and New York: Routledge, 367–402. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 805 Blanc, Haim. 1973. ‘Le perte d’une forme pausale dans le parler arabe du Caire’. Mélanges de l’Université Saint-Joseph 48: 373–390. Blau, Joshua. 1966. A Grammar of Christian Arabic: Based Mainly on South-Pales- tinian Texts from the First Millennium. Louvain: Secrétariat du Corpus SCO. -A Grammar of Mediaeval Judaeo] דקדוק העברית־היהודית של ימי־הביניים .1980 .—— Arabic]. 2nd Edition. Jerusalem: The Magnes Press. ——. 2002. A Handbook of Early Middle Arabic. Jerusalem: The Max Schloessinger Memorial Foundation, The Hebrew University of Jerusalem. הביניים .2006 .—— מימי ערביים־יהודיים לטקסטים Dictionary of Medieval] מילון Judaeo-Arabic Texts]. Jerusalem: Academy of the Hebrew Language. Blau, Joshua and Simon Hopkins. 1984. ‘On Early Judaeo-Arabic Orthography’, Zeitschrift fü r Arabische Linguistik (12): 9–27. ——. 1985. ‘A Vocalized Judaeo-Arabic Letter from the Cairo Genizah’, Jerusalem Studies in Arabic and Islam 6: 417–476. Ancient Bible Translations into] ’תרגומי מקרא קדומים לערבית־היהודית‘ .2000 .—— Judaeo-Arabic]. Peʿamim 83: 4–14. פונטי .2024–2017 .—— בכתיב הקדומה Early Judaeo-Arabic in] הערבית־היהודית Phonetic Spelling]. 2 Vols. Jerusalem: The Ben-Zvi Institute. Butbul, Sagit, and Simon Hopkins. 2024. ‘קדום ערבי־יהודי תרגום של חדשים דפים New Folia of an Early Judaeo-Arabic Translation of the Book of] ’לספר בראשית Genesis], in ‘An Inspired Man’: Studies in Judeo-Arabic Culture Dedicated to the Memory of Joshua Blau, Miriam Frenkel, and Phillip I. Lieberman (eds). Leiden: Brill, 3–51. Butts, Aaron. 2010. ‘A Note on neʾdārî in Ex 15:6’. Vetus Testamentum 60: 167– 171. Connolly, Magdalen. 2019. ‘Revisiting the Question of Ǧīm from the Perspective of Judaeo-Arabic’. Journal of Semitic Studies 54 (1): 155–183. Davis, M. C., and Ben Outhwaite. 2003. Hebrew Bible Manuscripts in the Cam- bridge Genizah Collections: Volume 3: Taylor-Schechter Additional Series 1–31. Cambridge: Cambridge University Press. Derenbourg, Joseph. 1893. Oeuvres complètes de R. Saadia ben Iosef al-Fayyoûmî. Volume premier: Version arabe du Pentateuque. Paris: Leroux. Elihay, J. 2012. The Olive Tree Dictionary: A Transliterated Dictionary of Conversa- tional Eastern Arabic (Palestinian). Jerusalem: Minerva Publishing House. Finkel, Asher. 1963. ‘The Pesher of Dreams and Scriptures’. Revue de Qumrân 4: 357–370. al-Fīrūzābādī, M. 2005. المحيط :Beirut .[The Encompassing Dictionary] القاموس Muʾassasat al-risāla li-l-ṭibāʿa wa-l-našr wa-l-tawzīʿ. Friedman, Mordechai Akiva. 2016. לתעודות הביניים מימי הערבית־היהודית מילון A Dictionary of Medieval Judeo-Arabic: In] הגניזה של ספר הודו ולטקסטים אחרים the India Book Letters from the Geniza and in Other Texts]. Jerusalem: Ben-Zvi Institute. Hopkins, Simon. 1978. ‘Studies in the Grammar of Early Arabic (Based upon Documentary Material Datable to Before 300 A.H. / 912 A.D.)’. Ph.D. diss., The University of London: School of Oriental and African Studies. ——. 2005. ‘On Imāla of Medial and Final ā in Early Judaeo-Arabic’, in Sacrum Arabo-Semiticum: Homenaje al profesor Federico Corriente en su 65 aniversario, Jorge Aguadé Bofill, Leila Abu-Shams Pagés, and Angeles Vicente Sánchez (eds). Zaragoza: Instituto de Estudios Islámicos y del Oriente Próximo, 301– 331. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 806 Kaplony, Andrea. 2015. ‘The Orthography and Pronunciation of Arabic Names and Terms in the Greek Petra, Nessana, Qurra and Senouthios Letters (Sixth to Eighth Centuries CE)’. Mediterranean Language Review 22: 1–81. Khan, Geoffrey. 1992. ‘The Function of the Shewa Sign in Vocalized Judaeo-Arabic Texts from the Genizah’, in Genizah Research after Ninety Years: The Case of Judaeo-Arabic, edited by Joshua Blau, and Stefan Reif, 105–111. ——. 2010. ‘Vocalized Judaeo-Arabic Manuscripts in the Cairo Genizah’, in ‘From a Sacred Source’: Genizah Studies in Honour of Professor Stefan C. Reif, edited by Ben Outhwaite, and Siam Bhayro, 201–218. Place: Publisher. ——. 2017. ‘Orthography and Reading in Medieval Judaeo-Arabic’, in Arabic in Context: Celebrating 400 Years of Arabic at Leiden University, edited by Ahmad Al-Jallad, 395–404. Leiden: Brill. ——. 2022. ‘Hebrew Vocalization Signs in Karaite Transcriptions of the Hebrew Bible into Arabic Script’, in Studies in the Masoretic Tradition of the Hebrew Bible, Daniel Crowther, Aaron Hornkohl, and Geoffrey Khan (eds). Cambridge: University of Cambridge and Open Book Publishers, 203–242. Khan, Geoffrey, et al. Forthcoming. The Cambridge Grammar of Biblical Hebrew. Cambridge: University of Cambridge and Open Book Publishers. Ktiv. 2025. Ktiv: The International Collection of Digitized Hebrew Manuscripts. Avail- able at: https://www.nli.org.il/en/discover/manuscripts/hebrew-manuscripts (Accessed: 2024/2025). Lane, Edward William. 1863–1893. An Arabic-English Lexicon. 8 vols. London: Williams and Norgate,. Levin, Aryeh. 1971. ‘The Imāla in the Arabic Dialects’. Ph.D. diss., The Hebrew University of Jerusalem. ——. 1992. ‘The Authenticity of Sībawayhi’s Description of the ʾImāla’. Jerusalem Studies in Arabic and Islam 15: 74–93. Maimonides, Avraham. 1959. פירוש על בראשית ושמות [Commentary on Genesis and Exodus], Ephraim J. Wiesenberg (ed.). London: S. D. Sason. Maman, Aaron. 1992. ‘The Lexical Element in David Alfasi’s Dictionary’, in Geni- zah Research after Ninety Years: The Case of Judaeo-Arabic, Joshua Blau, and Stefan Reif (eds). Cambridge: Cambridge University Press, 119–125. מבוא‘ .2003 .—— הערות באמריקה: לרבנים המדרש בית מאוסף וגלוסרים ’מילונים [Dictionaries and Glossaries from the Jewish Theological Seminary Collections: Introductory Notes]. Leshonenu 65 (3/4): 303–314. Nadler-Akirav, Meirav. 2021. The Arabic Translation and Commentary of Yefet ben ʿEli the Karaite on the Books of Amos, Haggai, and Malachi. Leiden: Brill. OPenn. 2025. OPenn: British Library. Available at: https://openn.library.upenn. edu/html/0047.html (Accessed: 2024/2025). Polliack, Meira. 1993/94. ‘Alternative Renderings and Additions in Yeshu‘ah ben Yehudah’s Arabic Translation of the Pentateuch’. Jewish Quarterly Review 84 (2–3): 209–226. ——. 1997. The Karaite Tradition of Arabic Bible Translation: A Linguistic and Exegetical Study of Karaite Translations of the Pentateuch from the Tenth and Eleventh Centuries C.E. Études sur le Judaïsme Médiéval 17. Leiden: Brill. Polliack, Meira, and Sason Somech. 2000. ‘עבריים-ערביים מקראיים גלוסרים שני .[Two Hebrew-Arabic Biblical Glossaries from the Cairo Genizah] ’מגניזת קהיר Peʿamim 83: 15–47. Porchazka Jr., Theodore. 1988. Saudi Arabian Dialects. London and New York: Kegan Paul International. A VOCALIZED JUDAEO-ARABIC GLOSSARY-TRANSLATION OF EXODUS 807 Posegay, Nick. 2020. ‘A Judaeo-Arabic Biblical Glossary as a Source for Arabic Historical Dialectology’. Journal of Arabic and Islamic Studies 20: 33–52. Putten, Marijn van. 2020. ‘A Judaeo-Arabic Letter in Early Phonetic Judaeo-Arabic Spelling: T-S 13J8.7’. Jerusalem Studies in Arabic and Islam 48: 49–78. Ratzaby, Yehuda. 1985. אוצר הלשון הערבית בתפסיר ר׳ סעדיה גאון [A Dictionary of Judaeo-Arabic in R. Saadya’s Tafsīr]. Ramat-Gan: Bar-Ilan University Press. al-Rojaie, Yousef. 2013. ‘Regional Dialect Leveling in Najdi Arabic: The Case of the Deaf- frication of [k] in the Qaṣīmī Dialect’. Language Variation and Change 25: 43–63. ——. 2022. Quranic Arabic: From its Hijazi Origins to its Classical Reading Tradi- tions. Leiden: Brill. Shachmon, Ori. 2011. ‘Pausal Final imāla in Central Palestinian Dialects’. Jerusa- lem Studies in Arabic and Islam 38: 145–161. Stokes, Phillip. 2021. ‘In the Middle of What? A Fresh Analysis of the Language Attested in the Judaeo-Arabic Commentary on Pirqê ʾĀvōṯ (The Sayings of the Fathers), Middle Arabic and Implications for the Study of Arabic Linguistic History’. Journal of Semitic Studies 66 (2): 379–411. ——. 2022. ‘Orthography and Phonology in Vocalized Medieval Christian Arabic Gospel Manuscripts’. Collectanea Christiana Orientalia 19: 131–206. ——. 2023a. ‘bi-hī, bi-him…fī-hu? Pronominal Suffix Harmonization Patterns in Some Vocalized Christian Arabic Gospel manuscripts’. Journal of the American Oriental Society 143 (2): 385–418. ——. 2023b. ‘Nominal Case in Christian Arabic Translations of the Gospels (9th– 15th Centuries CE)’. Arabica 70: 239–322. ——. 2023c. ‘Patterns of Nominal Case Marking in Arabic Gospel Manuscripts’. Zeitschrift für Arabische Linguistik 77 (2): 5–31. ——. 2023d. ‘ʾinğīl-in mubīn: A mixed archaic, Quranic, and Middle Arabic trans- lation of the gospels and its implications for the nature of Middle Arabic’. Bul- letin of the School of Oriental and African Studies 86: 405–445. ——. 2024. ‘The Evolution of Morpho-Syntactic Mood in Arabic: A View from Early Christian Arabic Gospel Manuscripts’. Journal of Semitic Studies: 361–413. Tirosh-Becker, Ofra. 2007. ‘תרגומיהם של המלומדים הקראים יעקוב אלקרקסאני, יוסף המשנה מן לקטעים יהודה בן וישועה Judeo-Arabic Translations of the] ’אלבציר Karaite Scholars Yeshuʿah ben Yehudah, Yūsuf al-Baṣīr, and Yaʿqūb al-Qirqisānī to Excerpts from the Mishnah], in Shaʿarei Lashon: Studies in Hebrew, Aramaic, and Jewish Languages Presented to Moshe Bar-Asher, A. Maman, S. E. Fassberg, and Y. Breuer (eds). Jerusalem: Bialik Institute, 435–466. Tobi, Yosef. 1993. ‘שרידי תרגום ערבי לתורה קודם לתפסיר רב סעדיה גאון’ [Remnants of an Arabic Translation of the Bible prior to the Tafsīr of Rav Saʿadia Gaon]. Massorot 7: 87–127. van Putten, Marijn. 2022. Quranic Arabic: From its Hijazi Origins to its Classical Reading Traditions. Leiden: Brill. Vollandt, Ronny. 2013. ‘Hebraisms in Arabic Versions of the Hebrew Bible’, in Encyclopaedia of Hebrew Language and Linguistics, Vol. II, Geoffrey Khan et al. (eds). Leiden: Brill, 182–186. ——. 2014. ‘Whether to Capture Form or Meaning: A Typology of Early Judaeo- Arabic Pentateuch Translations’, in A Universal Art. Hebrew Grammar across Disciplines and Faiths, Nadia Vidro, Irene E. Zwiep, and Judith Loszowy- Schlanger (eds). Leiden: Brill, 58–83. ——. 2015. Arabic Versions of the Pentateuch: A Comparative Study of Jewish, Chris- tian, and Muslim Sources. Leiden: Brill. << /ASCII85EncodePages false /AllowTransparency false /AutoPositionEPSFiles true /AutoRotatePages /None /Binding /Left /CalGrayProfile (Dot Gain 20%) /CalRGBProfile (Adobe RGB \0501998\051) /CalCMYKProfile (Coated FOGRA39 \050ISO 12647-2:2004\051) /sRGBProfile (sRGB IEC61966-2.1) /CannotEmbedFontPolicy /Warning /CompatibilityLevel 1.3 /CompressObjects /Off /CompressPages true /ConvertImagesToIndexed true /PassThroughJPEGImages true /CreateJobTicket true /DefaultRenderingIntent /Default /DetectBlends false /DetectCurves 0.0000 /ColorConversionStrategy /LeaveColorUnchanged /DoThumbnails false /EmbedAllFonts true /EmbedOpenType false /ParseICCProfilesInComments true /EmbedJobOptions true /DSCReportingLevel 0 /EmitDSCWarnings false /EndPage -1 /ImageMemory 1048576 /LockDistillerParams true /MaxSubsetPct 1 /Optimize false /OPM 1 /ParseDSCComments true /ParseDSCCommentsForDocInfo true /PreserveCopyPage true /PreserveDICMYKValues true /PreserveEPSInfo true /PreserveFlatness false /PreserveHalftoneInfo false /PreserveOPIComments false /PreserveOverprintSettings true /StartPage 1 /SubsetFonts true /TransferFunctionInfo /Apply /UCRandBGInfo /Remove /UsePrologue false /ColorSettingsFile () /AlwaysEmbed [ true ] /NeverEmbed [ true ] /AntiAliasColorImages false /CropColorImages false /ColorImageMinResolution 150 /ColorImageMinResolutionPolicy /OK /DownsampleColorImages true /ColorImageDownsampleType /Bicubic /ColorImageResolution 300 /ColorImageDepth -1 /ColorImageMinDownsampleDepth 1 /ColorImageDownsampleThreshold 1.50000 /EncodeColorImages true /ColorImageFilter /DCTEncode /AutoFilterColorImages true /ColorImageAutoFilterStrategy /JPEG /ColorACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /ColorImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000ColorACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000ColorImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasGrayImages false /CropGrayImages false /GrayImageMinResolution 150 /GrayImageMinResolutionPolicy /OK /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 300 /GrayImageDepth -1 /GrayImageMinDownsampleDepth 2 /GrayImageDownsampleThreshold 1.50000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /GrayImageDict << /QFactor 0.15 /HSamples [1 1 1 1] /VSamples [1 1 1 1] >> /JPEG2000GrayACSImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /JPEG2000GrayImageDict << /TileWidth 256 /TileHeight 256 /Quality 30 >> /AntiAliasMonoImages false /CropMonoImages false /MonoImageMinResolution 550 /MonoImageMinResolutionPolicy /OK /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 1200 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.50000 /EncodeMonoImages true /MonoImageFilter /CCITTFaxEncode /MonoImageDict << /K -1 >> /AllowPSXObjects false /CheckCompliance [ /None ] /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError false /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox false /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (U.S. Web Coated \050SWOP\051 v2) /PDFXOutputConditionIdentifier (CGATS TR 001) /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org) /PDFXTrapped /False /CreateJDFFile false /Description << /CHS /CHT /DAN /DEU /ENU (Ghent PDF Workgroup - 2005 Specifications version3 \(x1a: 2001 compliant\)) /ESP /FRA /ITA /JPN /KOR /NLD (Gebruik deze instellingen om Adobe PDF-documenten te maken die zijn geoptimaliseerd voor prepress-afdrukken van hoge kwaliteit. De gemaakte PDF-documenten kunnen worden geopend met Acrobat en Adobe Reader 5.0 en hoger.) /NOR /PTB /SUO /SVE >> /Namespace [ (Adobe) (Common) (1.0) ] /OtherNamespaces [ << /AsReaderSpreads false /CropImagesToFrames true /ErrorControl /WarnAndContinue /FlattenerIgnoreSpreadOverrides false /IncludeGuidesGrids false /IncludeNonPrinting false /IncludeSlug false /Namespace [ (Adobe) (InDesign) (4.0) ] /OmitPlacedBitmaps false /OmitPlacedEPS false /OmitPlacedPDF false /SimulateOverprint /Legacy >> << /AddBleedMarks false /AddColorBars false /AddCropMarks false /AddPageInfo false /AddRegMarks false /BleedOffset [ 0 0 0 0 ] /ConvertColors /NoConversion /DestinationProfileName (U.S. Web Coated \(SWOP\) v2) /DestinationProfileSelector /WorkingCMYK /Downsample16BitImages true /FlattenerPreset << /PresetSelector /MediumResolution >> /FormElements false /GenerateStructure false /IncludeBookmarks false /IncludeHyperlinks false /IncludeInteractive false /IncludeLayers false /IncludeProfiles false /MarksOffset 6 /MarksWeight 0.250000 /MultimediaHandling /UseObjectSettings /Namespace [ (Adobe) (CreativeSuite) (2.0) ] /PDFXOutputIntentProfileSelector /UseName /PageMarksFile /RomanDefault /PreserveEditing true /UntaggedCMYKHandling /LeaveUntagged /UntaggedRGBHandling /UseDocumentProfile /UseDocumentBleed false >> << /AllowImageBreaks true /AllowTableBreaks true /ExpandPage false /HonorBaseURL true /HonorRolloverEffect false /IgnoreHTMLPageBreaks false /IncludeHeaderFooter false /MarginOffset [ 0 0 0 0 ] /MetadataAuthor () /MetadataKeywords () /MetadataSubject () /MetadataTitle () /MetricPageSize [ 0 0 ] /MetricUnit /inch /MobileCompatible 0 /Namespace [ (Adobe) (GoLive) (8.0) ] /OpenZoomToHTMLFontSize false /PageOrientation /Portrait /RemoveBackground false /ShrinkContent true /TreatColorsAs /MainMonitorColors /UseEmbeddedProfiles false /UseHTMLTitleAsMetadata true >> ] >> setdistillerparams << /HWResolution [2400 2400] /PageSize [612.000 792.000] >> setpagedevice