Script and phonology

This file covers what all five varieties share — the script, the consonants and their history — and then the phonology of Biblical Aramaic in the Tiberian vocalisation. The other folders describe only their differences (gutturals in Galilee and Babylonia, the loss of final consonants in JBA, the Yemenite reading tradition, the Zohar's medieval spelling). Transliteration and glossing conventions: Aramaic of the Jewish texts — overview and Biblical Aramaic.

1. The script

All the texts in this manual are written in the square script (כְּתָב מְרֻבָּע, "Assyrian script" כְּתָב אַשּׁוּרִי in rabbinic sources), the Aramaic chancery script of the Persian empire, which Jews adopted for Hebrew as well (Naveh 1982, Early History of the Alphabet, ch. 4; b. Sanhedrin 21b–22a). It has 22 consonant letters, written right to left, and five final forms:

Neither the letter forms nor the order differ between the varieties: what differs is spelling (matres lectionis, final ־א or ־ה) and vocalisation.

1.1 Matres lectionis

Vowels are written by consonant letters (matres lectionis) only in part:

• final vowels: ־א or ־ה for -ā and -ē/-e; ־י for -ī, -ay; ־ו for -ū, -ō. In Biblical Aramaic both ־א and ־ה write the final -ā of the determined state: מַלְכָּא but also קִרְיְתָא beside forms like רָזָה (Dan 2:47 Ketiv). The choice is partly orthographic history and partly scribal habit (Rosenthal 1961 §§ on orthography; Bauer and Leander 1927 §§ 4–6).
• internal vowels: י for ī and ē, ו for ū and ō, used more often the later the text. JPA and JBA manuscripts write much more fully (plene) than the Bible: JBA אִיתְּמַר, קָאָמַר, לֵימָא; JPA אִית, הוּא, הָדֵין with י and ו everywhere (Phonology and orthography of Jewish Babylonian Aramaic; Phonology and orthography of Jewish Palestinian Aramaic).
• double ו and י for consonantal w and y in medieval spelling: JBA אַיְיתֵי "he brought", Zohar צְוָוחָא.

1.2 Ketiv and Qere

The Masoretic text of Daniel and Ezra often has a Ketiv (what is written) and a Qere (what is read). In the Aramaic chapters the Qere usually replaces an older spelling or form with a later one, for example:

Checked 2026-10-18: Kautzsch 1884 §53.2, p. 89 (suffixes on plural nouns, Ketiv with yod, Qere without); Strack 1901 §15b, p. 29 (the 3 f. pl. perfect only in the Qere: נְפַלָה Dan 7:20, also 5:5; 7:8) and §8l, p. 23 (gentilic and ordinal Ketiv ־יא against Qere ־אָה / ־אֵי). Corrected: the Dan 2:4 and Dan 5:5 rows. Second pass 2026-10-18: all nine rows agree letter for letter with the Ketiv and Qere of the WLC as encoded by the Open Scriptures Hebrew Bible (wlc/Dan.xml, wlc/Ezra.xml); the six rows left open are now also in a grammar: Dan 6:17 in Bauer and Leander 1927 §19b, e, p. 69 (Ketiv אנתה, probably read אַנְתָּה, Qere אַנְתְּ everywhere); Dan 3:18 איתינא / אִיתַנָא in §68z, p. 256; Ezra 4:12 אשכללו / שַׁכְלִלוּ in §50s, p. 175 (they point the Ketiv אַשְׁכְלִלוּ and write the Qere without yod, the WLC שַׁכְלִילוּ); Dan 4:31 follows the gentilic rule of Strack §8l (Ketiv ־יא, Qere ־אָה) and Dan 3:18 לאלהיך the Ketiv-yod rule of Kautzsch §53.2. Disagreement on Dan 3:5: the WLC and Krauss 1899 (II, p. 573, s.v. קתרוס III: Ketiv קיתרוס, Qere קַתְרוֹס) spell the Ketiv with ו; Marti 1896 §88, p. 92 (קיתרס) and Bauer and Leander §57f, p. 208 (קִיתָרֹס or קִיתְרֹס "überall Kt.") print it without ו, following Baer's edition.

(Text: WLC via Sefaria "Tanach with Nikkud", public domain, where the Qere is printed in brackets.) For Mova, every Ketiv/Qere pair is two tokens on one position: the reading form goes into the tree, and the written form into MISC (Aramaic in Mova: what it would need §3).

1.3 Vocalisation traditions

The consonantal text is read through one of several vocalisation systems, each a written record of a reading tradition (Khan 2020, The Tiberian Pronunciation Tradition of Biblical Hebrew, OBP, open access, vol. 1 ch. I; Morag 1988, Babylonian Aramaic: The Yemenite Tradition):

The editions Mova has on disk are all printed in Tiberian signs (or unpointed). The Yemenite Taj edition of Onkelos on Sefaria (CC BY-SA) keeps the Yemenite tradition in Tiberian signs.

2. Consonants of Biblical Aramaic

• Emphatics ṭ ṣ q were probably pharyngealised or glottalised; their exact articulation is not known (Rosenthal 1961; Creason 2004).
• ś (שׂ) is historical spelling. It is pronounced like s; Biblical Aramaic keeps שׂ in many words where later Aramaic writes ס (שַׂגִּיא "much", Ezra 5:11, beside later סַגִּי; עֲשַׂר "ten", later עֲסַר). Later Aramaic writes ס throughout (JBA סַגִּי, עֲסַר).
• Spirantisation (begadkefat). After a vowel, the six stops b g d k p t become fricatives ḇ ḡ ḏ ḵ p̄ ṯ, as in Tiberian Hebrew: כְּתַב kəṯaḇ "he wrote", but מִכְתַּב miḵtaḇ. In the Tiberian text a dagesh lene marks the stop. The rule is the same as in Hebrew and is written only in vocalised texts.

3. Vowels and syllables

Tiberian vocalisation of Biblical Aramaic has the same signs as Biblical Hebrew:

Compared with Hebrew, Aramaic reduces more. A short vowel in an open syllable before the stress becomes shewa: Aramaic כְּתַב kəṯaḇ against Hebrew כָּתַב kāṯaḇ; Aramaic דְּהַב dəhaḇ against Hebrew זָהָב zāhāḇ. This "pretonic reduction" is the single most visible difference between a vocalised Aramaic and a vocalised Hebrew page (Rosenthal 1961; Johns 1972, A Short Grammar of Biblical Aramaic, lesson 2).

• Stress falls mostly on the last syllable (kəṯáḇ, malkā́). Segolate nouns of the type מֶלֶךְ in Hebrew are monosyllabic or end-stressed in Aramaic: מְלֵךְ / מֶלֶךְ in Biblical Aramaic is borrowed spelling; the native type is מַלְכָּא with the stem malk-. Some suffixes are unstressed: 3ms -ēh in מַלְכֵּהּ.
• Gutturals (א ה ח ע and partly ר) do not take dagesh and prefer a-vowels and ḥaṭef vowels: עֲבַד ʿăḇaḏ "he made", חֲזָה ḥăzā "he saw", אֱלָהּ ʾĕlāh "god".
• Diphthongs ay and aw contract in some positions: bayt "house" absolute בַּיִת in Daniel (a Hebrew-like spelling), construct בֵּית bēṯ; determined בַּיְתָא baytā. In JBA and JPA contraction is wider (JBA בֵּיתָא; see those folders).

4. Correspondences with Hebrew

These regular correspondences let a reader of Hebrew recognise Aramaic words, and let a model link the two lexicons (Aramaic in Mova: what it would need §6). Proto-Semitic (PS) consonants after Moscati et al. 1964, An Introduction to the Comparative Grammar of the Semitic Languages; Huehnergard 1995, "What is Aramaic?", ARAM 7.

Checked 2026-10-18: Strack 1901 §2, pp. 12–13 (BA ד for Hebrew ז, ת for שׁ, ט for צ, ע and in Jer 10:11 ק for צ, with most of the examples above: דְּהַב, דְּנָה, תְּלָת, תַּמָּה, תִּנְיָן, תְּקַל, תְּלַג, תּוֹר, יְתִב, טוּר, טְלַל, נְטַר, קַיִט, אֲרַע, עֲמַר, אַרְקָא; the targum forms אֻדְנָא, עָן, רְבַע and ס for שׂ are given there as targumic, not Biblical). Nasalised spellings: Strack §18a, p. 32 (יִנְתְּנוּן, תִּנְתֵּן, לְמִנְתַּן) and the WLC מַנְדַּע (Dan 2:21).

Nasalisation: Imperial and Biblical Aramaic often write an n where a doubled consonant is expected: מַנְדַּע for maddaʿ, תִנְדַּע, יִנְתְּנוּן "they will pay" (Ezra 4:13, from נתן), תְּהַנְזִק "it will damage" (Ezra 4:13, haphel of נזק), אַנְתּוּן "you (pl.)". Some of these are historical spellings, some real dissimilation of a geminate (Rosenthal 1961; Folmer 1995, The Aramaic Language in the Achaemenid Period, ch. on nasalisation).

5. Phonological processes worth knowing

1. Pretonic reduction (above, §3): the main marker of an Aramaic vocalisation.
2. Metathesis in the t-stems. In the reflexive stems hit-/it-, a following sibilant changes place with the t, and after ṣ the t becomes ṭ, as in Hebrew: הִשְׁתְּכַח hištəḵaḥ "was found" (from שׁכח), יִשְׁתַּמְּעוּן yištammʿūn "they will obey" (Dan 7:27, from שׁמע), הִזְדְּמִנְתּוּן "you have agreed" (Dan 2:9 Qere, from זמן, with t → d after z).
3. Assimilation of n to a following consonant (יִפֵּל "he will fall", from נפל), and of the t of t-stems to a following dental in later dialects (JBA אִיתְּמַר → אִתְּמַר "it was said"; אִידְּכַר "he remembered").
4. Loss of ʾ in the middle of a word: רֵאשׁ, רֵישׁ "head"; בְּאִישׁ / בִּישׁ "bad" (Ezra 4:12, Ketiv ובאישתא, Qere וּבִישְׁתָּא).
5. Final -ā written ־א or ־ה (§1.1).
6. The guttural weakening typical of Galilean and Babylonian Aramaic does not happen in Biblical Aramaic: ע ח ה א are kept apart.

6. Reading a page: an example

Daniel 2:20 in Tiberian pointing (WLC, public domain):

Aramaic:   עָנֵה דָנִיֵּאל וְאָמַר לֶהֱוֵא שְׁמֵהּ דִּי־אֱלָהָא מְבָרַךְ מִן־עָלְמָא וְעַד־עָלְמָא
Translit.: ʿānē ḏāniyyēl wə-ʾāmar lehĕwē šəmēh dī-ʾĕlāhā məḇāraḵ min-ʿālmā wə-ʿaḏ-ʿālmā
Gloss:     answer.PEAL.PTCP.ACT.MS Daniel and-say.PEAL.PTCP.ACT.MS be.PEAL.IPF.3MS name-3MS REL-god.DET bless.PAEL.PTCP.PASS.MS from-age.DET and-until-age.DET
English:   Daniel answered and said: "Blessed be the name of God from age to age"
Source:    Dan 2:20 (WLC, public domain)

What the line shows:

• ד of דָנִיֵּאל has no dagesh after the vowel of עָנֵה: spirantised ḏ across the word boundary (the accents join the two words).
• לֶהֱוֵא: the imperfect of הוה "be" with prefix l- instead of y-, a Biblical Aramaic peculiarity of this one verb (to avoid a form that looks like the divine name, according to a common explanation; Rosenthal 1961; the prefix l-/n- of JBA is a separate development, The verb in Jewish Babylonian Aramaic).
• שְׁמֵהּ: 3ms suffix -ēh with a pointed final ה (mappiq).
• דִּי־אֱלָהָא: the relative particle written as a separate word, joined by maqqef.
• עָלְמָא: determined state with final ־א.

Sources for this file

Naveh 1982; Rosenthal 1961; Bauer and Leander 1927; Johns 1972; Folmer 1995; Moscati et al. 1964; Huehnergard 1995; Khan 2020 (OBP, in the research library as hbobook-obp0163-khan-tiberian-pronunciation-tradition-biblical-hebrew); Morag 1988; Creason 2004; WLC via Sefaria. See Sources.

The Elamur team

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/languages/arc/script-phonology/