Lexicon[1]

“From now on, no one may claim to know the vocabulary of New Persian who has not worked through these [Judaeo Persian] translations from beginning to end”.

Paul de Lagarde, 1884[2]

An in-depth lexical examination of EJP is essential for three reasons. First, according to Lazard,[3] only two points really differentiate between Middle and New Persian: the massive Arabic component in NP[4] and the ergative past in MP. EJP’s Arabic component is thus fundamental in its classification as a form of NP. Second, the Hebrew component is less thoroughly fused than in other Jewish languages, shedding some light on the hierarchy of component-integration. Third, the Iranian component provides evidence for the antiquity of the texts as well as the dialectal differences between them and within the TE.

1. Stock Languages

EJP consists of two main stock languages[5] – Hebrew and Persian. Persian, in turn, has Arabic as its second major stock language[6], as well as smaller Turkic, Mongol and dialectal-Iranian components.

The term “Hebrew component” refers to loans from religious sources and therefore includes Aramaic loanwords as well[7]. The following survey examines the Hebrew/Aramaic, Arabic and non-NP Iranian components. The latter is divided into Middle Persian and dialectal or otherwise unattested elements. Due to EJP’s classification as a dialect of NP and the scope of this work, the NP component will not be discussed.

Any words or forms of Iranian origin that exist in both NP and MP (or in other Persian dialects) are treated as NP, in accordance with a rule given by Maman for recognizing the Hebrew component in Judaeo-Arabic: “If the linguistic framework is Arabic [or, in the case of EJP, New Persian], this item is also Arabic [New Persian], unless otherwise proven”[8]. Following the same rule, when unable to determine[9] whether a Semitic word − e.g.אדמי  “human” (2:31 etc.) נִקְמַת “revenge” (115:9 etc.), מדבח “altar” (21:20 etc.) − originates from the Hebrew/Aramaic or Arabic component of NP, it is assumed to be Arabic.

1.1. The Hebrew Component

The Hebrew component constitutes one of the most significant features of a Jewish language, according to some − the most significant. Its scope varies from one language to another, as well as between different genres and speakers within each language. Judaeo-Iranian languages in general have very small and sometimes nonexistent Hebrew component[10].

Rabin (1981) defines Jewish languages as a sprachbund characterized by a state of diglossia, with Hebrew as an “upper language”. In this particular kind of diglossia, associated mainly with religious sprachbunden, only scholarly elites possess intimate knowledge of the “upper language”. Through this elite, words from the “upper language” penetrate the “lower language”[11]. When investigating the Hebrew component of the TE, one should bear in mind that the authors most probably belonged to this scholarly elite. They may have employed Hebrew words and expressions unknown to the average EJP speaker[12], and reinstated merged elements in their Hebrew form.

1.1.1 Identifying the Merged Hebrew Element[13]

1.1.1.1 Phonology

Vocalization that diverges from standard Hebrew occurs in גידים “tendons”, vocalized גִידִים in Hebrew and גִידֵים in the TE (209:18)[14]; the same is true of אוֵיר (Hebrew אוִיר) in LTG (5:34 etc.)[15]. Both appear in clearly Whole Hebrew quotations and therefore merely represent the local tradition of pronouncing the Whole Hebrew element[16].

As most of the text remains unvocalized, and because of the impossibility of comparing Whole and Merged Hebrew pronunciations, the phonological test becomes barely applicable in the TE. A rare case is the slip זינות ((67:3, Hebrew זנות, which occurs in the commentary. Today’s Persian-speaking Jews would also pronounce this word as zenūt[17] when reading a holy text or when speaking Hebrew. Still, the very fact that the spelling renders the pronunciation instead of adhering to the original Hebrew Orthography, as the author usually does, may indicate that the word had already become part of the speaker’s “unmediated language”. The same applies for פושאט (152:10 etc.) instead of פשט “literal meaning”, שבטה ישראל (91:5) instead of שבטי ישראל “the tribes of Israel”, and שולום[18], Hebrew  שלום[19].

Examples of phonological changes from 20^th century Judaeo-Iranian dialects include Judaeo-Esfahani mešimaδ  “a naughty boy” from Hebrew משומד mešummad “a Jew who converted” (which also exists in this meaning in the form mešūmaδ[20]), and Judaeo-Yazdi owén “sin”, Hebrew עָוֹן[21].

1.1.1.2 Semantic Changes

Other processes include changes in meaning, e.g., Judaeo-Esfahani mešimaδ quoted above, and, in many Jewish languages including JP – (h)exāl “(synagogue) ark”, from Hebrew  היכל “nave, temple”[22]. In the TE עבודה זרה, Hebrew “idolatry”, refers also to “idol” (15:31 etc. 184:9 etc. STE 5:1 etc.), and גוים, Hebrew “gentiles”, is treated (as in contemporary JP and dialects) as a singular, i.e., it always receives a plural morpheme: גוימאן[23].

1.1.1.3 Morphological Fusion

A third, deeper merger occurs when the lexemes from one source language combine with morphemes or (in other Jewish languages) patterns from another.

Morag[24] speaks of “external morphological fusion” and “internal morphological fusion”. The term “external morphological fusion” denotes “an affixation or a prefixation of a dialectal morpheme, while the base of the CC [=Hebrew] word is kept intact[25]”. As New Persian, like other Jewish languages of primarily non-Semitic origin, uses only affixation to form words, this concept is irrelevant to the present discussion (internal fusion applies mostly to Judaeo-Arabic, e.g. tša‘bob “lamenting” from Hebrew תשעה באב tiš‘a be’av “the 9^th of Av”, the date both temples were destroyed[26]). An important distinction, however, should be made between different kinds of affixes: derivational affixes – word-forming affixes resulting in a different lexeme – and inflectional or conjugational affixes – affixes that merely mark the grammatical categories (in Persian also the syntactic role) of a lexeme[27]. A few examples of conjugational affixes: a plural suffix – נבואתיהא (8:19 etc.)[28], גוימאן  and עָוֹנִיהָא (26:33), רשעאן and רשעימאן[29] discussed below; a person suffix: רובשאן “their majority” (STE 3:11), and צדקותיש, discussed below[30]; the ezāfe, e.g. מלכותי in מלכותי טעיף “a fragile kingdom” (76:36).

There is little evidence of EJP’s Hebrew element merging enough to combine with derivational morphemes. The Arabic component, although more recently introduced, shows a more advanced stage of fusion, and combines also with derivational morphemes[31].

A few examples of derivational morphemes with Hebrew components can be found in the text, but most of them are doubtful:

 נָבִיאִי כונאן “(those) prophesying” (48:37 translating Ez 13:2) Hebrew נביא “prophet” + abstract yod, could also originate in Arabic נבי (the Arabic abstract noun נְבֻוַת also appears in the TE).

נא מִרְמָה אִי כונד (82:37) “he does not deceive”, lit. “he does not practice deception”. Hebrew מרמה “fraud, deception” is followed by what seems to be a Persian abstract suffix –ī. The suffix is graphically separated from the word, which may indicate that the Hebrew word and the Persian suffix do not “stick”, i.e., the speaker still sense a distinction between Hebrew and Persian words in his natural language[32]. The suffix אי could also be interpreted as yāy-e vahdat, but an abstract (generic) noun fits the context better than an indefinite.

גלותיאן (187:32) “people of exile”. This word most probably results from an independent Hebrew innovation, but the yod may alternatively be NP yāye nesbat (MP –īg or Arabic –ī)[33].

Another example with a yod is גנוזי (7:35), constructed with the auxiliary כרדן meaning “to hide”[34]. The root ג.נ.ז, originally a Persian loan (Esther 3:9 and 4:7, from OP ganza), appears here in the Hebrew passive participle form, with a final yod that could be interpreted in the same manner as the yod in גלותיאן, as well as MP abstract suffix īh.

1.1.1.4 Innovations

The merged Hebrew component also includes new Hebraisms derived from existing Hebrew elements – lexemes and morphemes, roots and structures[35]: e.g., גלותיאן and גנוזי (discussed above[36]), and the plural ראשיתות (102:3) from ראשית “beginning”.

The expression עתיד הַבָּא (160:14) lit. “the future to come” in addition to עתיד (42:37 etc.) “future” may represent a contamination from העולם הבא (82:40 etc.), “the world to come”, or a translation of Persian آینده “future” (present participle of “to come”).

1.1.1.5 Calques

Although the language of traditional Jewish Bible translations is a calque language[37], the Hebrew calques included in the commentary may reveal the authors’ real language. Some of these seem to be influenced by the Bible translation, as in the use of the present participle with a copula where one would expect a simple present, e.g.,אגר אניז לוגת אישאן שנאסא הי “even though you are familiar with their language” (6:31), where one would expect שנאסי. The verb בורידן “to cut” is used in the expression “to make a pact” in the commentary (193:20), imitating Hebrew לכרות ברית, literally “to cut a pact” (Ez 34:25 etc.). Instances imitating the Biblical פעול יפעל construction in tafsīr include עטיל כרדן עטיל בי כונד (5:9) “he stops (lit. stopping he stops)”. This construction abounds, perhaps by force of habit, also in translations of Hebrew verses that employ single-word verbs, e.g. צ'ון גופת י'י כודאה ג'ודא כרדן ג'ודא בי כונום אין סרבנד עימימה וג'ודא כרדן ג'ודא בי כונום אין תאג' אין הסת ונא אין הסת אן אפתאדה רא בולונד כרדן בולונד בי כונום. ואן בולונד רא אבגסתן בי אבגנום  translatingכֹּה אָמַר אֲדֹנָי י'י: הָסִיר הַמִּצְנֶפֶת, וְהָרִים הָעֲטָרָה; זֹאת לֹא זֹאת, הַשָּׁפָלָה הַגְבֵּהַ וְהַגָּבֹהַ הַשְׁפִּיל (110:19-21, Ez21:31 ).

The expression פא כואהישן (16:34) translates Hebrew אהה (Ez 4:14), and probably imitates Hebrew בבקשה “please”[38].

The calque אבר הר רוי “in any event, anyway” (lit. on every face, 72:39-73:1) originates in Middle Hebrew[39]. Further evidence of the connection with the Hebrew speaking scholarly world is provided by the word מוכרת (121:23, 206:25) or מכרת (53:35, 206:23). This word occurs in Karaite Judaeo-Arabic and in mediaeval Hebrew grammar treatises with the same meaning as it does here: ‘disjoined’, an absolute (as opposed to construct) or pausal form of a noun[40].

1.1.1.6 Hybrid Expressions

The expression שבועה כורדן (109:21 etc.) “to take (lit. “to eat”[41]) an oath”, substitutes Persian sōgand with Hebrew שבועה “oath” in the Persian expression sōgand xordan. Calques of sōgand xordan also exist in Turkish, Khorosthi Middle Indian and Armenian[42].  Similarly, צום גריפתן (14:39 etc.) imitates rūze gereftan  “to fast”[43].

The expression עין נבואה (5:18 etc., 209:1) “the eye of prophecy” may seem like a Hebrew construct, but the instance עין ינבואת (208:1) reveals the Persian ezāfe construction, and עין ינבות (206:17) even substitutes the Arabic word for the Hebrew one.

An almost completely Hebrew expression is found in a personal letter from the Cairo Geniza, אלהי ישראל חן וחסד אבזאיאדשאן בעיני אלהים ואדם (L5:3-4) “May the God ofIsraelincrease their grace and favor in the eyes of God and man”. The Persian אבזאיאדשאן demonstrates that the speaker perceives the other parts of the expression to be part of his own language.

1.1.1.7 Bible Translation

Though extremely unreliable when it comes to calques, Bible translations provide the most dependable source for single Merged Hebrew words that defy the phonological, morphological and semantic tests. As the author wishes to translate the verses into Persian and avoid Hebrew words[44], the Hebrew words that do sneak into translation clearly constitute part of the authors “unmediated language[45]”.

Note that this observation applies only when the translation differs from the original Hebrew wording. When the translation uses the same word as the MT, it may suggest that the translator is unable to translate the word; either because he does not understand it or because it defies translation into Persian.

Examples:

צפון “north” (29:37, Ez 8:3 etc. 177:14, Ez 32:2 etc.) also פא סוי צפון translates צפונה in Ez 8:3, and ראה צפון translates דרך צפונה (Ez 8:5, etc.). Twice in Part 1, the Persian word סָאיְה “shade” translates the word צפון (103:20, Ez 21:3 and 104:25, Ez 21:9). In both cases the word appears in conjunction with נגב “south, the Negevdesert”, translated נימרוזה “Sīstān” or/lit.  “midday”, the time of day when the shade is in the north[46].

מזרח “east” (32:17). The word for “east” in the Book of Ezekiel is קדם, never מזרח. The word מזרח occurs often in the commentary but only once in translation –  פא מזרחtranslates קדמה “to the east” (alongside סוי משרק) in Ez 8:16. The usual translation is Arabic  شرق (שרק,, 39:19 etc.).

תימן “south” (doubtful), 138:36, Ez 25:13. This may be the name ofYemen, the country. In the commentary (138:37-38) the word serves both for the country and for “south”.

The word מערב in the TE is always translated as אמיזישן “mixture”, never as “west”. Interestingly, in 103:6, גרב (Arabic غرب “west”) translates the only occurrence of דרום (Ez 21:2)[47].

Other Hebrew words occurring in translation are צדקותיש “his righteousness”, discussed below, אדני “the Lord” (170:5 etc.) translating itself, and גלות “exile” (7:17 etc.) − ג'מאעת (מרדומאן) גלות translates Hebrew גולה (Ez 3:11 etc.), and in 182:25,  גלות אימאtranslates גלותנו.

שבועה “oath”, translates יד (95:10 etc., Ez 20:15 etc.). The word יד translates also as Persian פיאמברי (8:16 etc.), and Hebrew/Aramaic נבואת (183:16 etc.) “prophecy”[48].

1.1.2 The Aramaic Component[49]

The relatively small Aramaic component includes words such as פסוקא (56:1 etc., 170:30 etc.[50]), רובא in the expression רובא מרדומאן (178:30 etc.) “most of the people”[51], דוכרא/דוכרה in the expression דוכרא כרדן “to mention” and ספידא (135:3) in the expression ספידא כון “lament”. The Hebrew root ספד is conjugated in the 1^st conjugation (פָּעַל, e.g. Gen. 23:2 etc.), but it only assumes the verbal noun מספד (Genesis 50:10 etc.) as opposed to ספידה, the regular verbal noun for the first conjugation, attested only in Modern Hebrew. The form ספידא occurs in the Targum and Talmud[52].

The word צדקות “righteousness” (probably *saddaqūt[53]) − featured extensively in exegesis (9:33, 181:7 etc.) − even takes on Persian suffixes: צדקותיש (182:16) “his righteousness” and with ezāfe צדקותי או (57:20). The former appears as the translation of Hebrew צִדְקָתוֹ (Ez 33:18). The authors usually strive to avoid Hebraisms in translation, using Persian ראסתי (9:22 etc. 181:24) and ראסתיגריה (181:5 etc.[54]) to indicate ‘righteousness’. The use of צדקות in translation indicates that the speaker regards the word צדקות as part of his natural language.

Some Hebraized Aramaic words[55] include כשרות in the expression פא כשרות “in a kosher manner”, vocalized כְשֵרות (183:36)[56], as opposed to Hebrew כַּשְׁרות kašrūt. This phonology may alternatively indicate an independent derivation of this Hebrew word or a hypercorrection.

According to Sheynin, כשרות and צדקות appear in Palestinian Aramaic as כשרותה and צדקותה[57] (= Biblical Aramaic כשרותא, צדקותא). The Hebraized forms are derived from these words through omission of the definite suffix ה/א.

The words שבועת (78:27) “oath” and נבואת[58] (14:13 etc. 183:16[59] etc., with a plural morpheme see above) exist in ancient Aramaic inscriptions[60]. The Hebrew feminine suffix -ָה  may appear here as -ת as a result of a process in which the definite form of the word is taken from Aramaic (שבועתה/שבועתא, נבואתה/נבואתא) and its definite suffix -א/-ה removed without further adjustments. A more probable alternative is that the Hebrew word is treated like Arabic, as words ending in ta marbūta often enter NP with a final -at[61]. The occurrence of the alternatives נבואה (5:18 etc., 183:14 etc.), שבועה (63:5 etc., 180:35 etc.) strengthens both assumptions: if they are Arabicized Hebrew, then as Arabic words they may appear both with final hē and final tav. If they are Hebrew, then Aramaic נבואת and שבועת may be regarded as separate words from the Hebrew נבואה and  שבועה, because Hebrew words always appear with final hē.

1.1.3 Classical vs. Integrated Corpus

Morag[62] treats Merged Hebrew words as “non-Hebrew” and further divides Whole Hebrew into Classical Corpus and Integrated Corpus. The term “Classical Corpus” refers to “the communal transmission of Hebrew in the reciting of Scripture and other classical texts”[63], while the “Integrated Corpus” consists of quotations from and allusions to the Classical Corpus, borrowed idioms and single words. Given the scarcity of phonological information, the distinction between Classical and Integrated Corpuses is easier to apply to the majority of Hebrew elements in the TE than that between Merged and Whole Hebrew.

In grammatical discussions, even one-word citations clearly belong to the Classical Corpus, but in exegesis these instances are sometimes debatable.

Consider, for example, the word רשע “wicked” (both noun and adjective). In some cases it has clearly merged with the language, as in גוימאן רְשָעִימָאן (18:35) “the wicked gentiles” (or רשעימאן alone, 40:9 etc., 185:30), where the Hebrew plural רשעים receives a Persian plural suffix -אן. In other cases, when this word appears as part of a verse to be discussed, it qualifies as part of the Classical Corpus. The same applies to Hebrew words appearing after גופתן “saying”, צ'ון גופת “as he said”, and מעני “(the) meaning (of)”. However, the status of single words in an exegetical (as opposed to grammatical) discussion is debatable.

E.g., 9:11-19, Ez 3:19:

ואתה כי הזהרת רשע. ותו כי בי פהריזאני טאלים רא ונא אבאז גרדד אז כואהלי או ואז ראה או אן טאלימי או פא גונאה או בי מירד ותו ג'אן תורא רהאניסתי: גופתן ולא שב מרשעו. רוא כי נא כרדן מצות עשֵׂה הסת. ומדרכו הָרְשָעָה. מצות לא תעשה רוא הסת. ובי נמוד כי אגר רשע אז תו נא פדירד תו כוישתן רא רהאניסתי ואו הלאך בי באשד. …

[2 ½ lines without any Hebrew at all]

ואגר אמרוז נביא ניסת הם באיד פדריפתן קול עילימאן אן צ'יז כי מצות הסת והר אן וקתי כי נא פדירים מוקאם אימא מוקאם אן מרדומאן הסת כי אבר אישאן גופת הוא רשע בעונו ימות:

“And you, when you warn the wicked [Hebrew quotation – Classical Corpus] And you, when you warn the wicked and he does not return from his crookedness and from his way, that wickedness; he, in his iniquity he shall die, and you, your soul [you] have saved.

Saying and he does not return from his crookedness [quotation to be interpreted – Classical Corpus] may refer to not performing the positive commands [Integrated Corpus]. And from his way, that wickedness [quotation to be interpreted – Classical Corpus] may be the prohibitory commands [Integrated Corpus]. And he intimates: ‘if the wicked [using the Hebrew word from the verse in a discussion – Classical or Integrated?] does not accept  (the admonition) from you, you have saved yourself and he will perish’.

[…]

And although there is no prophet [Integrated Corpus/Merged Hebrew] today, it is still necessary to accept the words of those learned in that thing which is precepts [Integrated Corpus/Merged Hebrew]. And each time we do not accept (those words), our position is the position of those people about whom he says he [is wicked], in his iniquity he shall die [supporting quotation – Classical Corpus].”

1.2. The Persian Component

This discussion distinguishes between Hebrew and Persian components, treating the Arabic component as part of the Persian. The source of this non-linguistic division is twofold:

1) A Jewish language takes its vocabulary from two sources − Hebrew/Aramaic and the common tongue − but New Persian, the common tongue underlying EJP, already includes both Iranian and Arabic elements. The speaker, whether aware or unaware of the origins of the words he uses, most probably inserts Arabic elements from the Persian variety he speaks, rather than borrowing directly from Arabic.

2) The second reason is more pragmatic – the Hebrew component appeals mainly to linguists of the Jewish-Languages arena, whereas the Persian component should interest anyone wishing to study the history of New Persian vocabulary (which includes elements of both Iranian and Arabic origins).

The study of the Persian component – both Arabic and Iranian – benefits from the tendency of Bible translations to avoid Hebrew words, and many of the Persian/Arabic words in this discussion appear only in translation of Biblical verses. A (cautious) comparison with the Hebrew source facilitates the deciphering of some difficult words.

However, the translation of the Hebrew verses, potentially a great advantage in determining the meaning of unknown words, proves unreliable in certain instances, an extreme example being the translation of Ez 21:2: “O son of man, set your face to the way towards Tēmān (South and also Yemen) and preach to the south and prophesy to the forest of the field Negev (South and also a name of a desert in the south of Israel)”:

103:4-6 יא פסר אדמי בי ניס רויהא תו פא ראה סוי רוּמִייַה ופיאמברי בי כון פא סוי גרב ו  בי כון פא בֵ"ישָא דשת סוי נֵימְרוּזַה. “O son of man, set your faces to the way towards Rumiya and prophesy towards the west and prophesy to the forest of the field towards Nēmruza.”

Rumiah, i.e. Byzantium, lies west of Iran. Nēmruza regularly denotes “south” in translation, but is also one of the names of Sīstān, in southernIran. This could indicate that the geographic location of the translator lay east of Sīstān (Looking west towardJerusalem instead of South like Ezekiel), but it definitely means that caution should be employed when comparing the translation to the Hebrew original.

1.2.1 The Arabic Component[64]

Although introduced more than a millennium later than the Hebrew element, the Arabic component in the TE is much more extensive and already merged to a much higher degree. This is probably due, at least in part, to its being an element of the common language and to its massive, almost violent entry into Persian.

1.2.1.1 The Extent of the Arabic Component

Although fairly substantial, the Arabic component in Early-Judaeo-Persian constitutes a smaller part of the vocabulary than in Muslim Persian[65]. The extent of the Arabic component largely varies between the different EJP texts. The Dandān Uiliq letter has only one Arabic word[66], but over time the Arabic component becomes more and more noticeable. In personal letters of the Cairo Geniza, Persian-speaking Jews living in Cairo wrote whole sentences in Arabic using Arabic letters. This signifies their awareness of their Persian being a Jewish language, as opposed to non-Jewish Arabic[67]. 

The authors of the TE most probably lived in a Persian rather than Arabic-speaking community, and therefore all instances of employing the Arabic component may be safely ascribed to loans in the common language. The extent of the Arabic component in the text constitutes one of the major differences between the separate parts of the TE, and a distinction between Part 2 of the TE and the STE, written in otherwise very similar language. Generally speaking, Part 2 of the TE exhibits greater Arabic influence than Part 1, and less than the STE.

The following examples illustrate these differences[68]:

When referring to connections between words – verbs and their subjects, attributives and their governing substantives etc. – the authors of Part 2 use Arabic راجع : (גופתן) א' אבא ב' ראג'ע ה(י)סת “(saying) A goes back to B”[69] (174:7 etc. Twice א' ראג'ע ה(י)סת אבר ב' 170:30 etc.), א' אבר ב' ראג'ע כרדן “to refer A back to B” (176:27-28 etc.). In the same context, the authors of Part 1 use Persian: א' אבא ב' פיוסתה/פיוסתא הסת “A is connected to B” (6:12-13 etc.), א' אבא ב' פיוסתה/פיוסתא כרדן (4:18 etc.) or  א' אבר ב' נהאדן “to refer A to B” (3:10 etc.).

Enumeration (of reasons, of possible interpretations etc.) in Part 1 employs mostly Persian ordinals – יכי/אול, דויום, סיום etc., but the writers of Part 2 use the Arabic ordinals: אול, תאני, תלת. While contemporary NP contains a higher rate of Arabisms than any EJP text, it still uses the Persian ordinals, besides Arabic أول “first” – just like Part 1. This instance, along with many others, exemplifies the relative closeness between the language of Part 1 and NP, even in cases where it appears more archaic than Part 2[70].

Other instances are found in Part 1 פאדה (4:7 etc.) and Part 2 תאבת (194:21, STE 2:12[71]) or תבתא (214:5) “firm, stable[72]”, Arabic ثابت.

Intriguingly, words occurring in Bible translations (and then referred to in the tafsīr) provide most of the exceptions to this rule:

e.g., רשע “wicked” translated in Part 1 as טאלים^[73] (9:6 etc.) from Arabic ظالِم, and in Part 2 as בתרי (179:23), occasionally בתר (192:5 etc.), probably battar < *bad-tar^[74]; Part 1 אדמי (Arabic آدم) vs. Part 2 מרדום in Hebrew בן אדם “son of human”, Part 1 פסר אדמי  (4:4)^[75]  Part 2 פסר מרדום ^[76] (170:15 etc.) or פרזנד (י)מרדום (177:33 etc.). The expression “Man and beast”, both in translation and in tafsīr – Part 1 אדמי וצ'ארבא (56:31), Part 2 מרדום וצהארבא(י) (173:21-22 etc.)^[77]. The expression עליה אלסלאם “may he rest in peace” (Arabic عليه السلام) appears only in Part 1 (5:11 etc.), alongside Hebrew זכרו לברכה “blessed be his memory” (7:35 etc.), which in Part 2 serves exclusively (170:14 etc.).

1.2.1.2. Fused Arabic Component

Besides unchanged Arabic[78] words − such as מעני (1:1 etc., 170:10 etc.) “meaning”, עיבארת (1:2 etc., 213:1) and עבארת “expression, explanation”, יעני (1:2 etc. 170:15 etc.) “viz.”, סביל (1:3 etc, 176:26 etc.) “way”[79], etc. − various fused elements are also found in the text[80].

1.2.1.2.1 Phonetic Changes

Certain Arabic words undergo a change in vocalization[81]: e.g.מַעְוַנַתִי[82] (36:16) Arabic (and NP from Arabic) مَعُونة, مَعْوُنة as well as  مُعاوَنة and مَعانة; מוֹגְמְרְא (31:34) “censer”, Arabic (and NP from Arabic) مِجمَرَة  [83].

1.2.1.2.2 Semantic Change

Words attested in Arabic undergo a shift in meaning, e.g., פִ"יגֵ'י discussed below; גַ"ש  “dross” (115 35 etc.) in both translation and tafsīr, from Arabic  غَشّ “fraud, deceit”, NP from Arabic (also “alloy”)[84].

1.2.1.2.3 Innovations

The TE abounds with newly formed Arabic words − Arabic roots in Arabic patterns forming words unattested in Arabic. The following list is a small sampling:

תולַא בודן “to follow” (112:13, 193:10 etc.) from Arabic تَلا, تَلَّى “he followed”; the productive pattern קטלא − representing Arabic قَتلة or قَتال ^[85] − yields בדרא כרדן “to scatter (people, bones)” (21:34 etc.) from Arabic √ب ذ ر “to sow, scatter” (NP from Arabic بَذْر)^[86], and גַ'רְבַּא כרדן “to try” (108:21) from Arabic  جَرَّبَ “he tried”^[87]; עטיל כרדן “to stop, abolish, break (a pact) etc.” (5:9 etc.), unattested in Arabic or NP, derives from √ع ط ل, “to stop, cancel, abolish, be destitute, without work”^[88]; מעמום “general” (1:2 etc.), the unattested passive participle of Arabic عَمَّ “be general, inclusive”.

The root √ب وش “to cry” (بَاشَ, “to vociferate, be noisy, cry out, call for aid or in distress, impatience or fear[89]”) gives the gloss בוש (25:24), which probably also means “a cry”, translating (מֵ)הֱמֵהֶם (Ez 7:11), and בָאשָא (in Part 1a: 36:1 etc.) and באשאה in Part 1b, 148:24 etc.) “a cry” − constructed with כרדן “to cry” − translating all occurrences of Hebrew √ז.ע.ק in the book of Ezekiel.

פִ"יגֵ'י (5:20 etc.) translating Hebrew דֶּבֶר “the Plague” may reflect فِجا(ء)*, formed in the pattern of בִינֵי “building” (21:27) and גְ"נֵי “song” (141:28 etc., Arabic  بِناءandغِناء respectively, with imāla[90]). The Arabic word closest in meaning, فُجاة “sudden death”[91], has a different vocalization on the first syllable and ends in ta marbūta. If the only difference were a change in vocalization (even one that results in imāla), it might still be considered the same word, but an unrepresented ta marb­ūta is far less likely than a new formation.

1.2.1.2.4 Morphological Fusion

Arabic words may also combine with Persian morphemes, both derivational and inflectional/conjugational. Note the extent of Arabic words with inflectional morphemes, taken from less than half a page in the glossary: Arabic broken plural أَحْوَال combines with the Persian plural morpheme to form אחואליהא “circumstances” (72:4) and with a possessive suffix – אחולישאן “their conditions” (190:28); the same occurs with إحسان “favor, benefit”[92] – plural אחסאניהא (205:7) and possessive אחסאנש (181:22) or אחסנש (203:5). Another double plural is אחכאמהא “judgments” (204:9). Possessive suffixes appear with אחתאגאגיש “the need for it/arguing for it[93]” (125:25), and אטראפישאן “their sides” (177:35).

The following sentence demonstrates how deeply the Arabic component is merged in the language: תעלותיה. פא תאזי קנואתיהא גוינד. ופא זואן פארסי מַאדוֹהָא גוינד “(the Hebrew word) תעלותיה: In Arabic they say קנואתיהא and in Persian they say madohā” (165:27-28). The allegedly Whole-Arabic word is brought with a Persian plural morpheme.

Derivational – i.e. word-forming – morphemes include the suffix -גין used to create adjectives, combined with Arabic مُصيبة “affliction” to form מוציבתגין “afflicted” (gloss to 8:23), and with مُراد “wish, will” to form מְרָאדְגִּין (20:11), מוראדגין (20:13) “satisfied”[94].  Apart from periphrastic constructions with auxiliaries such as כרדן “to do” and  בודן “to be”, EJP also derives denominative verbs synthetically: e.g.,  ובי עטימהום ובי כאצהום translating Ez 38:23 “I will become great (denominative from Arabic عظيم) and I will become holy (denom. from Arabic خاص)”[95] (219:31) and ג'ואבהיסתן (55:10 etc.) “to be answered” (Passive with جواب).

Less commonly, Persian words take on Arabic word-forming morphemes, e.g. דורית (probably dūriyyat) “stretching-far” (164:24), from Persian dūr “far” + Arabic abstract ־ِيَّة.

1.2.1.2.5 Calques and Hybrid Expressions

Some common expressions present a mixture of Persian and Arabic words, for example,  פא אין/אן חאל (כי)[96] (37:24 etc., 189:30-31 etc.) “at the time (when)”; פא חקיקה (Part 2, 171:7 etc. and STE 1:1; Part 1 once פא חקיקת 129:3) – a direct calque from Arabic في الحقيقة, and its variety פא תחקיק (Part 1, 2:32 etc. with the verbal noun of حَقّقَ “to investigate”), both meaning “actually, as a matter of fact”.

1.2.1.2.6 Metanalysis

The Arabic مُوَاجَهَة “second person reference of verbs[97]” (in the TE also “addressing”) provides an interesting case study. In Part 2 this word appears, predictably, as מואג'הא (185:1 etc.), but in Part 1, however, it undergoes vocalization changes and apparently also metanalysis. Its spelling varies: מאואג'הא (68:12), closest to the origin word with only one apparent vowel shift u>ā; מאואג'יהא (167:28) with a newly introduced yod; the forms מאואגָ'היהא (72:30) and the most common מָאוָאגָאיְהָא (49:15 etc.), obviously analyzed as plural. In several places this plural undergoes further metanalysis as two words: מאוא ג'איהא (6:15 etc), an expression which resembles (Persian) “dwelling places[98]”.

1.2.2 The Iranian Component

EJP constitutes a form of NP, and therefore, the discussion of the Iranian component concentrates on words unattested in NP, or words attested with different meanings. These include words found in MP and dialectal features (including dialects of Early NP). Elsewhere unattested words are considered − for purposes of the discussion − as dialectal, even if they appear in both parts.

1.2.2.1 Middle Persian

The majority of MP traces are discussed in Phonology, Morphology and Syntax, and below in Word-Formation. This section is concerned only with lexemes. MP words unattested in New Iranian occur mostly in Part 2 of the TE.

Part 1 uses the same verb for “to hear” as NP: שנידן (2:10 etc.) שנו- (4:14 etc.). In Part 2, the 2^nd sg. imperative equals the MMP present stem אכשין (179:19 etc.). The usual present stem undergoes contraction: אכשנ- (178:12 etc. STE 2:16 etc.)[99], and the past stem, unlike MP ’xšyd-, is a secondary derivation from the present stem: אכשניד- (178:12 etc.).

The NP verb nīšīdan “to look” (<ni+√īkš) is used only in the sense of watching the stars. In Manichaean Middle Persian (MMP) nyyšydn means simply “to see” and it also occurs with the same meaning in today’s Tadjik[100]. In Part 1 of the TE, נֵיש- (14:30 etc.)[101] is the suppletive present stem of the verb דידן “to see”, and the form *נישידן is wholly absent. The stem נֵיש- exemplifies the similarity of Part 1 to MMP and Tadjik.

The present copula in the TE and in most other EJP texts is identical to that of MP, with the stem ה-[102].

MP survives also in the verb אוזדן (172:3 etc.) אוזנ- (171:38 etc.) “to kill (people), extinguish (fire)”, which appears only in Part 2.

In Part 1, רונישן (27:10 etc.) translating words from Hebrew √ק.ר.ח related to baldness, and the passive verb רונהיסתן (159:9 etc.) translating מרוטה (plucked) and occurring in the commentary in a context that fits this meaning, lead to a reconstructed present stem רונ- “plucking hair, losing hair”. This corresponds to MP rūdan, rūn– “pluck, pull out”[103], represented in NP by rūda kardan “to clean feathers off a bird or hair from an animal (after killing)”. Part 2 lacks this verb, but it also lacks any context that would require its use.

The word אפסיסת (172:37 etc.)[104] occurs only in compound verbs with kardan and būdan meaning “to destroy” and “to be destroyed” respectively.  It corresponds to MP afsīstan[105] “to be destroyed”, and survives in both form and meaning. Note that rather than the expected NP (and EJP) participial form *אפסיסתה/*אפסיסתא, this form represents the MP participle, identical to the NP and EJP past stem[106].

Another instance of a participle without a final -ה/-א is וִיהִירְד (142:17)[107]. It combines with the auxiliary בודן to translate ונבהלו “they will be scared”. In the commentary it also occurs in the spelling בִ"יהִירְד(142:19). It probably corresponds to MMP whwryd and Parthian whyrd “confused,”[108] and the context in which it appears agrees with “confusion”. Phl. has a present stem wihīr- “to change”, as well as wihīrišn and wihīrēh[109] “change”. All related words in NP and dialects cited by Nyberg in his glossary signify “exchange”. In this case, Phl. and MMP differ in meaning and form, while EJP assumes the Phl. form with the MMP meaning.

MP jādag “cause, reason” appears in the TE as ג'אדה (1:4 etc.), ג'אדא (143:11, 180:13 etc.), צאדה (185:29 etc.) and צאדא (170:22 etc). It occurs alone in the sense “because of”, “for the sake of”, and in the expressions ג'אדה אן רא “therefore” and ג'אדה אן רא כי “because” before a sentence[110].

The hapax ג'הישני (64:16) translates ותצלחי, and probably reflects an unattested *jahišnīg  “fortunate”, from MP jahišn “fortune, destiny (in a positive sense)”. This word has a continuation in the ENP personal name jahšiyārī < jahišyārī.

Note that the word for “thing, something” is צ'יז or צ'יזי (1:5 etc. 171:11 etc. STE2:21 etc.) as in NP, rather than תיס from MP and other EJP texts[111].

1.2.2.2 Dialectal Elements

Both parts of the TE contain words unknown in Middle or Standard New Persian. The vocabulary of Part 1, however, contains more previously unattested dialectal elements than Part 2.

1.2.2.2.1 Part 1.

The adverb אבראז “above” (1:8 etc.)[112] and its derived adjective אבראזין “upper” (33:35)[113] pose a serious puzzle.

Orly Rahimian, a student at the Hebrew University of Jerusalem, suggests the reading abar-az/abar-az-īn, considering it a secondary formation of the preposition abar “over”, with agglutinated az “from”. Although most scholars in the conference “Middle Iranian Lexicography: The Vocabulary of the Middle Iranian Languages” (Rome, April 2001) found Rahimian’s explanation the most plausible, a few instances of the preposition אבראזי in LTG (8:25 etc.) renders this possibility less probable. An alternative analysis still takes abar as the base of this adverb, but the suffix –āz as resulting from ancient ank (Skt. –añc) denoting direction, as in abāz (<*apānk, Skt. apāñc[114]) “back”. Thus אבראז may be reconstructed as < upari-āñc/ānk[115]. The adjectival form may be a further agglutination of the deictic pronoun īn “this”, but more likely represents the adjectival suffix īn[116].

A less probable but still possible suggestion involves √barz, the Eastern counterpart of Persian √bard. This is also the root of bālā, which is used in the same sense as אבראז (a glossאבראז  explains באלא in 2:28, implying אבראז was better known to the writer than Persian באלא).

The last and most plausible alternative would be to classify אבראז as a transposed noun or verb. This transposition may come from either the MP noun abrāz “acclivity”[117] or MP abrāz, NP afrāz, the present stem of MP abrāstan, NP afrāštan/afrāxtan “to elevate, raise”.

The word אבראז may be related to בראזא “tall”, either directly, by omission of the initial aleph[118] or indirectly, by derivation from the same √barz.

אבשנודן אבשנאי-, also unique to Part 1, occurs mainly in translation of Hebrew verbs of compassion – √ח.ו.ס and once √ר.ח.מ.. This word seems to MMP abaxšāyīd (pp) “to pity”, abaxsāyišn “pity”, because the ו in the past stem and the nasal remain unaccounted for. It may reflect the root √xšnav (MP šnāyīdan “to praise, propitiate”, Causative Šnāyēnīdan “to please, propitiate”, adjective xušnūd[119]) “to be satisfied”  accompanied by the preverb apa-/api-/upa-. This verb remains in the semantic field of feelings, more specifically a good feeling caused by another person.

Martin Schwartz’s[120] suggested etymology differs; he notes that √xšnav is unattested with preverbs and its meaning is more distant from “pity” than that of abaxšāy-: “I’ve explained OP ā-xšnaw– ‘to hear’ as based on a conflation of  *(ā)çunau– and xšnau-, cf. Yāv. Surunūiδi …xšnūiδi = ‘hear and give satisfaction‘ (‘hear and obey[121]’), both = ‘hearken’. Thus MP ā/ăšnaw-. Perhaps (after *xšād– became xšāy-) *abaxšāy-:abaxšūd interacted with *(x)šnāy-:(x)šnūd and (semantically related via ‘hearken to prayer’ ~ ‘pity’) – somehow also connected with the mysterious removal of –n– in  ’xšy(n)- vis à vis šny– / ’šnw– etc.

The word גומאר  “a whore’s fee, a gift” (67:6 etc.)[122] remains a puzzle; the present stem גומאר- appears once (16:19) in the context of giving money, and גומאשתן (76:11 etc.) “to appoint” appears only in the past stem. These three separate meanings may be etymologically connected through “entrusting”[123]: entrusting money, entrusting gifts and entrusting authority.

The word אבג'אר translates Hebrew מרקחה, or combines with the auxiliary כרדן to translate הרקח (both in 133:13). It also occurs in the commentary as something that “they throw in a cauldron when it heats completely” (133:19) and then “you shake the cauldron so the אבג'אר reaches everywhere” (133:22). This word most probably corresponds to MP/NP abzār (NP also afzār) “spice, spices”, but the spelling suggests a pronunciation avjār or avžār.

In LTG, the homographic (but certainly unrelated) present stem אבג'אר- (49:31), probably avžār-, means “to pay (a debt)”, and thus fits the meaning “redeem”, quoted in the CPD under wizārdan wizār– (NP guzārdan). The EJP equivalent quoted by the CPD, בזאר- “to explain”, appears in Part 2 of the TE. In some cases it bears the meaning “explain” (Hebrew √פ.ר.שׁ), and in others it seems to mean “spread” (Hebrew√פ.ר.שׂ). The latter could (with great semantic flexibility) be tenuously linked to Part 1 אבג'אר.

Some words in Part 1 remain without clear etymologies. Their meanings must be inferred through context and, in the translation, through their Hebrew equivalents: מוֹזא occurs with the auxiliary verb כרדן in the sense “to make, fashion, create”. In 15:38 the context – making a cake – suggests a connection with NP maze (MP mizag) “taste”, but in 44:34 it appears over the line, where כרדן alone could suffice. The context there – the Lord creating ears and eyes – excludes the possibility of taste. It seems unrelated to מוֹזא “suckling” in 75:30 etc.

The word זִיוֵיהְרִיהָא translates Hebrew תועבות “abominations”. Possible etymons would be Arabic broken plural زواهر (with imāla) + Persian plural suffix  –יהא,  but no such broken plural is attested in Arabic. Deh Xodā has زواهر as “lights, bright stars”, which were perhaps worshipped. Another meaning of Arab. √ز هـ ر involves flowers, which could have been used in worship. This rare word could also be related to NP zahr “poison” or to MP zōhr “libation”. The cluster -הר- sometimes parallels NP -xr–[124] or -rx–[125], but none of these substitutes yield plausible etymons.

Martin Schwartz[126] suggests a form with –hr < OIr –aθra– (cf. MP šahr “kingdom, land” < OIr xšaθra). If זי = Classical Persian zi– (e.g. zidūdan), i.e. < OIr uz-, then זיויהר(י) may go back to OIr *uz–wai–θra– from √way/wī “wind, turn, twist”[127], hence “perversion”.

Another possible etymology with the same suffix (-hr < OIr –aθra-) is zīwēhr < OIr *jīviyaθra “mortal offence” attested in Gathic Avestan jōya– < jīvya– “concerning one’s life, capital” (offence)[128].

1.2.2.2.2 Part 2 and STE

The southwestern corpus (Part 2 and STE) is shorter than the northwestern one (Part 1), and the dialect bears a stronger resemblance to MP than its northeastern counterpart, in vocabulary, syntax, morphology etc. However, words unattested in MP occur here too:

In addition to אנון[129], the word אוכון (172:14 etc. STE 5:14 etc.) also seems to mean “now” or possibly “here”. The word אכנון (NP aknūn) occurs only once in Part 2 (211:18) but regularly renders “now” in Part 1 (1:30 etc.)[130].

The word אבזנידן occurs only twice, both in the STE, once אבזנהיד (STE 2:24) in an illegible context, and the other reads: ובָרֵא אתהן יע' אבזנידן וזדן יפא שמשיר היסת (STE 5:24-25) “They shall cut them down (Hebrew, Ez 23:47) – that is, hitting and striking which is by the sword”. The TE uses the words ברידן וזדן (130:24) “cutting and hitting” to explain the same phrase, but the etymology suggests a meaning closer to that of zadan “to hit, strike, kill”, and may reflect Av. Aipi√gan “to hit, strike dead”.

Context and etymology also explain דונינד (STE 5:4), in: וקו' וקטרתי רוא היסת כו אן קטרת היסת כו י' ית' ש' פרמוד בוד כו פא כאנש דונינד אישאן פיש יעב' ז' כירדינד “And his saying and my incense (Hebrew, Ez. 23:41) may be that incense that The Lord, Blessed Be His Name had commanded that they burn in his house, (which) they did before idols”.[131]

The context here calls for “kindling” or “burning”, and this verb clearly originates in Indo-European *d^huh₂ “to smoke” [132]. Like other laryngeal roots, this root forms its present stem with an –n– infix (Skt. dhūnoti, adhvanīt[133]) and has many nominal derivations with a nasal, e.g., GAv. duuąnman, YAv. dunman “cloud” and Skt. dhūma, Latin fūmus “smoke”. The notion of burning incense thus fits the meaning of this root perfectly, and the etymons account for the nasal present. A past stem dūd– may be reconstructed, which survives in NP only in dūd “smoke” (also TE 31:35 etc. 172:1 etc.). Martin Schwartz[134] adds: “A similar meaning of the same verb in Iranian may be preserved in Boir Ahmad Luri dīnešt ‘Esfand’ (whose seeds are burnt apotropaically to produce a pungent smoke)[135]. This form is probably a deverbative in –išt/ešt[136] from a present stem dīn– = EJP dūn–[137]. The fact that OIr  has dūta vs. OInd dhūmá ‘smoke’ is now explained by the former being a PPP of √dū, whose present stem may be reconstructed as dūn(ā)– (cf. Av. frīn(ā)- √frī <PIE preH-)”.

The present stem of “to teach, learn, educate” (MP hammōxtan hammōz– NP āmūxtan āmūz-), corresponds to that of NP אמוז- (5:3 etc.183:12 etc.)[138] in both parts of the TE. The past stem, however, differs in Parts 1 and 2: Part 1 employs the same stem as in NP אמוכת- (9:15 etc.)[139], while Part 2 uses an inchoative past stem, only in the sense “to learn”: אמוכסיד- (185:26 etc. STE 5:15 etc.). This stem is unattested in MP, but attested in Parthian, which contains the inchoative stem ammōxs– in the passive sense “to be taught”.

1.2.2.2.3 Words Occurring in Both Dialects

Some words occur in both dialects of the TE but still seem dialectal for one of two reasons: Part 2a displays evidence of contamination with the northeastern dialect (of Part 1) on many levels[140]. If a word occurs in Part 2a but not in Part 2b or anywhere else in the southwestern dialect, and other factors point to a dialectal element, then it falls into the dialectal category. In other cases, the existence of elements unattested or differently attested in NP[141] suggests that SNP represents a dialect different from the common ENP language, while the TE reflects the common tongue.

The verb כופסידן (12:24 etc. 170:4 etc.)[142] “to lie down, sleep” − known from MP (xwafs- as a present stem of xuftan) − appears in NP in the 11^th century Eastern Persian works of the poet Nāṣ̣̣er Xosro and the historian Al-Bīrūnī[143], i.e., in works contemporaneous with the TE, written in a dialect geographically close to that of Part 1. כופסידן also exists in a later translation of the Book of Ezekiel to EJP[144]. In the TE this verb is found both in Part 1 and in Part 2a[145]. The use of the verb in Part 2a should probably be considered a dialectal contamination from Part 1 rather than a Middle Persian survival, due to the secondary past stem[146].

The word פַ"רְזֵין [4:27 etc., STE2:23(?)] occurs in the translation and commentary of Ez 2:6. A gloss and the contents give away the meaning – “thorn, briar”. This word may be related to MMP prčyn “hedge”, Kandulaī pärčin “thorn hedge” and Baxtiārī parzīn, “bramble”[147]. The EJP word still differs from these dialects in the vocalization of the second syllable and in the affrication of the first consonant (in both occurrences in the TE). The use of the same word in Part 1 of the TE and in the STE suggests a word in the common Early New Persian language, which is replaced by another in SNP.

1.2.2.2.4 Homographs

פרסתידן and פרסתאדן: both parts include verbs spelled פרסתידן and פרסתאדן. In Part 1, פריסתאדן (5:16 etc.) usually means “to send”, NP ferestādan, from fra√stā, and in Part 2 “to stand” (171:31 etc. also four times in Part 1, 35:8 etc., and in 44:17 פריסתאד is crossed out and replaced by dialectal פאד), probably from pari√stā.

Part 1 uses pari√stā- in the form פריסתידן (30:33 etc.), meaning “to worship” (NP parastīdan). In Part 2 (179:16)[148] this spelling renders the same “to send”, MP frēstīdan.

The present stem ראינ- in Part 1 means “to consider” and is probably a denominative derivation[149]. In Part 2, however, the same stem means “to lead” or “to cause (oil) to run”, and may be derived from raftan[150] as a causative. Both derivations are unusual (see Morphology:7.1.1.2.3 and 8.1.1.2.4).

1.2.2.3 Changes of Meaning

Some words in the TE signify different things than those same words in NP or MP. For example, the verb הנג'ידן (Part I only: 15:2 etc.) “to drink” in NP (both هنجيدن and تنجيدن) means “to draw out” (OIr. √θanj). In MP the verb hixtan/hanjīdan means “drawing up”, with “water” and “bucket from a well” as its direct objects[151]. Had this word existed only in translations or only in combination with the verb “to eat”, it could have been assumed to be adapted from an existing word to fill the empty place of the verb “to drink”, as opposed to “to eat”. The occurrence of this word in many distinct contexts in Part 1 indicates that such a verb with the meaning “to drink” indeed existed in 11^th century Northeastern Iran. Interestingly, the present stem תנג' also renders the meaning “to drink” in the later translation of the book of Ezekiel and in Qissa-ye Daniel[152].

The word xomānā (29:27 etc. 170:2 etc.)[153] also exists in NP in slightly different sense and form than in the TE. NP خَمانا/هَمانا and MP homānāg mean “similar, like, resembling, rival”[154], while the TE uses this word as the nominal component of a compound verb כומאנאי א' בודן or כומאנא בודן א' רא/פא א' – “to be like A”, and also as a substantive translating Hebrew דמות “image” (from √ד.מ.י “to be similar”).

1.2.3 Loanwords reborrowed from Arabic

Many words of Iranian origin have passed through Arabic and back to NP. Some have kept their original form and their re-borrowing from Arabic is indicated only by a broken plural (e.g. وزیر – وزرا ); some have undergone phonological changes and survived in both forms (e.g. گوهر- جوهر, گیهان- جهان- کیهان), and some survive in Persian only in the Arabic forms. One fascinating case of the latter in EJP is the plural צטאהריג' “ditches” (165:9) and the double plural צטאהריג'יהא in LTG (3:12). This word, though unattested in previously studied texts, may be safely traced to MP staxr, NP istaxr, but the double plural and the spelling with צט suggest that this word may have passed through the Arabic “phonological filter”[155]. Further support for this possibility is found in the plural pattern: The broken plural pattern CaCāCiC[156] is typical of loan words in Arabic that do not conform to Arabic patterns, e.g. خندق – خنادِق (MP kandag), برنامج – برامِج (NP barnāme, MP –nāmag), etc.. The final j may represent the MP suffix –ag[157]. The singular may be reconstructed as *צטהרג' < MP *staxrag/stahrag. The retention of the cluster (הר) after a long vowel is odd, as it conflicts with Arabic phonology.

2. Technical Terms

As mentioned in the discussion of stock languages, most technical terms in Part 1 derive from Persian, whereas in Part 2 and especially in the STE they derive mainly from Arabic. In the STE, the most common technical terms, namely “routine words” in discussion, appear in shorthand.

The following is a list of routine words and other technical terms in the TE and STE, sorted by stock language.

2.1 Arabic Origin,

2.1.1 Arabic Routine Words:

• יעני (38:23 etc. and STE יע') يعني  “i.e.”.
• (ו)יק' (only STE) (و)ُيقال “(and) it is said” (sometimes followed by איצא أيضًا, “also”.
• (כ)ק' (only STE) كما قال or كقوله “as he says”.
• מעני (12:17 etc. and STE מע')   معنى“meaning”.
• (ו)קו' (only STE) (و)قوله “(and) his saying”.
• כ' (only STE) كما “like”[158].
• ראג'ע (Part 2 and STE) راجع “referring, going back to” (in א' אבא ב' ראג'ע היסת, “A refers to B”).

2.1.2 Arabic Technical Terms

• אמר أمر “imperative”.
• ג'זם جزم “assertion”.
• שרח شرح “explanation”.
• קצת )STE), קיצת (Part 1) قصة  “story, phrase”. Part 1 also “paragraph”.
• מג'אז مجاز  “metaphor”.
• מסלה/מסלא  مسألة “problem”.
• מצדר مصدر  “infinitive”.
• וצף وصف “description”.
• אבר א' אופתאדן “to refer to A” (<a Hebrew expression> refers to <a Persian expression>).
• א' פא ב' כואסתן “to mean A by B” (the subject of the sentence is Ezekiel/the Lord: א' פדיש כואסת = “by this he meant A”).
• רוא [ה(י)סת] “)it is( possible”.
• גונא “(a) kind (of)”. Usually employed without indicating “of what”.
• נמודן “to intimate, to say”.
• א' ב' נהאדן (21:27 etc., 171:22 etc.STE2:9) "to tag A as B”.
• א' אבר ב' נהאדן “to connect A (a word) to B” (e.g. a verb to its subject).

2.2 Persian Routine Words

2.3 Mixed Arabic and Persian

• מתל זדן “to liken”.

2.4 Hebrew/Aramaic

• וג' = וגומר  “etc” (lit. and he ends)[159].
• עבודה זרה (STE ע'ז') “idol(s)” (Heb. idolatry).
• דומ' (only STE) דומיא, “similar (to)”.
• ענין “matter”.
• פסוק (Usually in Part 1), פסוקא (Usually in Part 2, STE) “a verse”.

2.5 Karaite vs. Rabbinite Terminology

Many of the Arabic as well as the Hebrew\Aramaic terms found in the Large TE are known from Karaite grammatical treatises[160]. Some of the Persian terms seem to be calque translations of the Arabic ones, and some of these terms also appear in Rabbinite JP grammars[161].

Words that occur both in Karaite and Rabbinite JP grammars, whether occurring in the TE[162] or absent from it[163], are irrelevant to the present discussion. However, in cases of conflict between Karaite and Rabbinite terminology, the TE invariably follows the Karaite, whether by calque or by direct borrowing.

E.g. Rabbinite חרק,  סגל “hiriq, segol” (the vowels), Karaite נקטה ואחדה “one dot”, תלאת נקט “three dots”, TE יך נקט and ג' נקט respectively[164]; “Agent”: Rabbinite (Hebrew) פועל, Karaite and TE (Arabic) פאעל; “Interrogative”: Karaite תמוה, Rabbinite תמיהה, TE תמוה (70:7 etc.), twice תמיה (55:14 etc.); “Noun”: Karaite (Arabic) אסם, Rabbinite (Hebrew) שם, TE (Persian) נאם.

In the last example, the TE appears closer to Karaite, because it uses the local word rather than the Hebrew one. And indeed, looking at the lists of words present in the other grammars but absent from the TE, the Karaite list includes mostly words not required by TE context[165], while the Rabbinite list includes only Hebrew terms[166].

The word בנין “conjugation” and the names of the conjugations occur only in Rabbinite grammars, while Karaite texts and the TE compare conjugated verbs to other verbs with the same וזן “construction” (lit. “weight”) and use the verb’s imperative form as a reference. Likewise, the term גזרה “paradigm” and the names thereof (such as כפולים, כרותים) are only used in Rabbinical grammars, while the Karaites and the TE again employ comparison and the imperative.

Geoffrey Khan, having kindly read some grammatical portions of the TE, affirms[167] that it appears distinctly Karaite.

3. Word Formation

Word formation distinguishes between the dialects and − in Part 1 − reveals some of the dialectal differences between this dialect and both Middle and New Persian.

As in NP, words may be formed through transposition, affixation, juxtaposition and compounding[168]. Of these, transposition, juxtaposition and compounding are still productive and can theoretically occur with any word in NP. Affixation is still productive, but the individual affixes may be different. Verb prefixes are discussed together in Morphology:7.1.3.2 and thus the current discussion points out only suffixes unattested, unproductive or phonologically different in NP. Complete information on these suffixes can be found in the phonology and morphology chapters.

3.1 Nominal Suffixes

Salemann 1900:271, upon investigating only segments of Part 1 of the TE, states that the only remnants of MP word construction in the text are the suffixes –išn, –omand and the adverbial –ihā. All his examples of the –ihā, however, draw on translations of biblical prepositions that have a plural form in Hebrew[169], with the single exception of פיראמוניהא (38:29) in the commentary, which may be a calque.

3.1.1 Substantivization

The deverbative suffix –išn (=MP, more common in Part 1) /-išt (dialectal, more prevalent in Part 2) / iš (=NP -eš) demonstrates different stages of transition as well as dialectal differences between the different parts of the TE[170].

The two parts also show different stages of transition in the form of the abstract-noun morphemes -יה (=MP), which is more prevalent in Part 2, and -י (=NP), more common in Part 1[171].

3.1.2 Adjectivization[172]

The suffix –(o)mand, attested only in Part 1, probably ceased to be productive at the time of the TE. All TE words ending in –(o)mand occur in MP, some also in NP[173].

Other suffixes include the adjectival –gīn (MP –gēn) found in NP only in a few frozen forms, and the suffix īn (MP ēn), both probably still productive in 11^th century northeastern Iran. The adjectives created by the suffix īn follow the substantive, while in NP they must precede it.

3.2 Verbal Suffixes

3.2.1 Secondary Past Stems[174]

Part 1 usually uses the suffix -יסת to derive secondary past stems from present stems, e.g. כאמיסתנד “they wished” (94:12). This suffix occurs more in MMP and in some northern dialects than in Phl. and NP. Part2 inmost cases follows the Phl./NP model and employs the suffix -יד.

3.2.2 Synthetic Passive [175]

The use of synthetic passives by affixing -הי- or -ה- after the present stem[176] shows the affinity of the TE language to MP, as well as one of the major differences between the dialects of the TE: While Part 1 uses synthetic passives as a rare stylistic option, in Part 2 this is the nearly-exclusive way to express the passive voice.

An identical morpheme creates denominative verbs, e.g. ובי עטימהום ובי כאצהום ובי שנאסהום “I will become great (from Arabic عظيم) and I will become holy and I will become known (passive)” (219:31 translating Ez 38:23).

4. Exegetic Translation

A degree of caution should be practiced when deducing the meaning of words in translations from the original Hebrew. In some cases, especially in metaphors, the translation provides an interpretation rather than a literal translation of the word. Glosses sometimes give a different meaning rather than a different word with the same meaning.

E.g. זונה “whore, fornicator” and the verb זנה “to fornicate, go astray” are usually translated  והאר“an idolater” (22:27 etc.), והאר(י) בודן/כרדן (51:6 etc.), “to practice idolatry”. פיאמברי or more rarely נבואת “prophecy” often translate יד “hand” (8:16, Ez 3:14) or רוח “spirit” (40:23, Ez 11:5), when referring to the hand or spirit of the Lord, i.e. prophecy. When יד appears as a metonym for punishment, it is translated as ט'רבת “stroke, blow” (49:10, Ez 14:9 etc.)^[177], כ"ישם “anger” translates פנים “face” (56:4, Ez 14:8 etc.) in its metaphoric use as “anger”^[178].

אפרוד אמד “came down” explains בי אפתאד “fell” in ובי אפתאד ^{אפרודאמד} אבר מן  אנג'א פיאמברי י'י כודאה “and fell ^{came down} upon me there the prophecy of the Lord God” (29:20-21, Ez 8:1).

חיה, previously translated as “an animal” is translated as פרישתה “angel”, אלהי ישראל “the God of Israel” is translated כבוד כ"ודאה ישראל “the Glory of the Lord of Israel”, and the word כרובים “cherubim” is translated צורתיהא “images” in או אן פרישתה הסת כי דיד בודום אן כי אזיר כבוד כ"וֹדָאה ישראל פא רוד כַּרְבְּלַא ובי שנאכתום כי אן צורתיהא הנד אישאן “It is that angel that I had seen, that which (was) under the Glory of the God of Israel by the river Karbla, and I knew that those images are they” (39:23-25, Ez 10:20).

Thus, the meaning of otherwise unattested words such as לַך (194:18) translating Hebrew מטע “grove” and its denominative verb לכנידום translating נטעתי “I planted” (205:28) can not be deduced from the Hebrew context with complete certainty.

5. Summary

This chapter demonstrates the differences between the Arabic component, which, although more recently introduced, has already merged into the speakers’ language, and the Hebrew component, whose fusion lags behind. New Arabic words are formed by using known roots and patterns, and existing words change their vowels as well as their meanings. Arabic words combine with Persian word-forming morphemes. Hebrew words, on the other hand, change their meanings very rarely (and when they do, the new meaning remains close to the original meaning). Their vowels remain the same, although occasional spelling deviations testify to their absorption into the language. The only examples of Hebrew words combining with word-forming morphemes can be explained as innovations using Hebrew morphemes.

The low rate and extent of the Hebrew component in EJP is exactly what makes it so intriguing in a comparative study of Jewish languages. It allows a glimpse into the hierarchy of incorporating a component into the language, however simple it may sound: the phonology hardly changed, and when it did it remained the same or similar for Whole and Merged Hebrew; semantic shifts were relatively subtle[179]; morphological fusion was limited to inflection; innovations were rare and the vast majority of Hebrew calques apparently emanated from an intimate knowledge of the Hebrew Bible, which was limited to a small scholarly elite.

In the Iranian part, most of the distinct peculiarities and the dissimilarities between Part 1 and Part 2 correspond respectively to differences between New and Middle Persian[180]; Manichaean Middle Persian and Book Pahlavi[181]; Tadjik/northwestern Classical Persian/an unknown dialect and Standard New Persian[182]; Standard Persian and an unknown dialect/some of the known EJP texts written in Khuzistan[183].

The differences within the Iranian component strengthen the assertions made throughout this work with regard to the different dialects of the TE. The next and final chapter summarizes, classifies and discusses all these differences.