Phonology and orthography of Jewish Palestinian Aramaic
Script and inherited consonants: Script and phonology. JPA texts differ from Biblical and targumic Aramaic in three ways: gutturals are weak, spelling is very plene, and endings are written in more than one way (־ה / ־א, ־ן / ־ם). Vowels below follow Dalman 1905, Fassberg 1991 and Sokoloff 2002b, and the Guggenheimer vocalisation of the Yerushalmi on Sefaria (CC BY); they are reconstructions.
1. Gutturals
Galilean speech weakened the gutturals, and the writing shows it (Kutscher 1976, ch. on phonology; Sokoloff 2002b, introduction):
| process | example | meaning |
|---|---|---|
| ʾ lost at the end of a syllable, and often between vowels | רֵישׁ "head" (BA רֵאשׁ) | |
| ʾ (from ʾe-) lost at the start of common names | לִיעֶזֶר for אֱלִיעֶזֶר (y. Yevamot 12:2: דרבי ליעזר) | |
| ʿ and ḥ confused with ʾ and h | manuscript spellings that interchange ע and א, ח and ה in the same word (Kutscher 1976 lists them; Sokoloff 2002b gives the variants under each headword) |
The rabbis themselves joked about it. b. Eruvin 53b tells of a Galilean in the market asking for אמר, and being answered: "Foolish Galilean, do you want ḥamār (an ass) to ride, ḥamr (wine) to drink, ʿamar (wool) to wear, or ʾimmar (a lamb) to slaughter?" — because he pronounced ḥ, ʿ and ʾ alike. The anecdote is Babylonian and comic, but it agrees with the spellings in Galilean texts.
2. Plene spelling
JPA manuscripts and inscriptions write vowels with ו and י more fully than any other Jewish Aramaic:
| vowel | example | meaning |
|---|---|---|
| short i written י | אִית, לֵית, דִּיהוּדָאֵי | there is, there is not, of the Jews |
| short u written ו | קוֹמֵי "before" (Yerushalmi; targum קֳדָם) | |
| ō, ū written ו | כְּדוֹן "now"; תּוּבָן "again, still" | |
| final -ē written י | מוֹדֵיי "agree (pl.)" (y. Bava Metzia 2:5) |
3. Variable endings
| ending | variants | example |
|---|---|---|
| determined -ā | ־א and ־ה | Yerushalmi and inscriptions both; Neofiti mostly ־א |
| m. pl. absolute | ־ין, and Hebrew-like ־ים in some words | |
| 3 m. sg. suffix | ־יה, ־ה, ־יי | לֵיהּ, לֵה |
| final -n / -m | 3 m. pl. ־הון, 2 m. pl. ־כון; ־הום, ־כום only in the Palestinian targums | |
| "the Rabbi" | רִבִּי (Palestinian vocalisation, vs Babylonian רַבִּי) | (Guggenheimer: רִבִּי throughout) |
Checked 2026-10-18 in part: Stevenson 1924 §4, p. 16 (the 3 m. sg. suffix ־יה, the 3 m. pl. ־הוֹן and ־וֹן).
Second pass 2026-10-18: the rows of §§ 1–3 and 5 against Dalman 1905 §§ 10–15, pp. 56–105, read from the OCR prose and the page images.
- Gutturals: Dalman §15.1aβ, p. 96: inside a word ʾ always loses its consonantal value at the end of a syllable (ריש "head", Jeb. 11b) and often after a vowelless consonant before a long vowel; corrected: the first row said only "between vowels". §15.1aδ, p. 97: initial ʾ with its vowel falls in Galilean names (לעזר for אלעזר, ליעזר for אליעזר) and in common words (נשא "man", חורי "other"). §10.1, pp. 57–60, and §15.1c–d, pp. 98–99: ḥ, ʿ, h and ʾ are confused in Galilean spellings; Dalman tells the b. Eruvin 53b story (the Galilean asking for אמר) and the Berakhot baraita about Haifa, Beth-Shean and Tivon. Agree.
- Plene spelling: Dalman §12.1a, pp. 70–72: in the Palestinian Talmud and Midrash ו marks u, o, ū, ō and י marks i, e, also in closed syllables (איתעביד "was made"); final yod is often doubled. Agree, with one correction: תּוּבָן is "again, still" (from תוב + ־ָן, Dalman §15.3dε, p. 102; Anhang II.1, p. 397), not "straw", and its vowel is ū; the row now says ō and ū.
- Endings: Dalman §12.1aα, p. 70: final ā is written with ה or א, and the preference for ה is typical of the Palestinian Talmud: agrees. §15.3dδ, p. 102: Galilean also drops the final n of the absolute plural (־י for ־ין), so the masculine plural absolute has a third spelling. §15.3dζ, p. 102: ־הום and the like are forms of the Jerusalem targums; corrected the final -n / -m row accordingly (the Yerushalmi has להון 75, לכון 76 and no להום, לכום; counted 2026-10-18). The 3 m. sg. suffix variants and רִבִּי were not compared (unchecked).
- §5: Dalman §§ 13.5, 14.7, pp. 87–90: au has almost always become ō, but ai is kept inside the word more often in Galilean than in the targums (עיינא "eye", חיילא, זייתין), while Onkelos mostly contracts it; corrected the diphthong line, which said the opposite. §15.3dα, p. 101, and §66.1, p. 291: assimilation of n is the rule in the imperfect and infinitive peal and in the aphel, but not regular elsewhere; it is kept before ה and ע (איתתא "woman" beside the rarer אנתתא). §15.3dε, p. 102: the Galilean liking for final ן explains the 3 m. pl. perfect in ־ון. Agree. Spirantisation and pretonic reduction were not compared.
The spelling רִבִּי against Babylonian רַבִּי, and names in short forms (לִיעֶזֶר, לָעְזָר for אֱלִיעֶזֶר, אֶלְעָזָר; יוּדָן for יְהוּדָה; יוֹסֵה for יוֹסֵי), are typical of the Yerushalmi.
4. Inscriptions
Synagogue mosaics and inscriptions of the 4th–7th c. (Hammat Gader, En Gedi, Rehov, Ḥuseifa, Beth Alpha) are the only JPA texts that come down to us without copying. They show: determined ־ה and ־א side by side, the formula דְּכִיר לְטַב "remembered for good" (donor formulas), plene spellings, Greek names and words, and occasional loss of gutturals. Standard edition: Naveh 1978, On Stone and Mosaic: The Aramaic and Hebrew Inscriptions from Ancient Synagogues (Hebrew); for the long halakhic inscription of Rehov, Sussmann 1974. We do not quote these texts here; their editions are in copyright, and the stones themselves are public monuments (Jewish Palestinian Aramaic texts, annotated §5).
5. Phonology in brief
- Spirantisation: as in all Aramaic of the period; not marked in most texts.
- Pretonic reduction: as in other Aramaic (Script and phonology §3).
- Diphthongs: au becomes ō, but ai is often kept inside the word, more than in the targums (Galilean עַיְינָא beside עֵינָא; Dalman 1905 §14.7b, p. 90). The construct בֵּית is contracted, as everywhere.
- Assimilation of n: as elsewhere (יִפּוֹק); but the t of the t-stems often stays: אִתְקְטִיל, and the prefix is spelled אִית־ / אִת־ (Fassberg 1991).
- Final unstressed vowels stay in writing more than in JBA (e.g. the perfect 3 m. pl. ־ון keeps its n).
The Elamur team
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/languages/jpa/phonology-orthography/