Although scholars have generally treated Ugaritic ltn as a cognate of Hebrew liwyātān, the vocalization of this word and its relationship to the Hebrew form remain debated. In this article, we will argue that ltn should be vocalized /lītan-/ and that Ugaritic ltn and Hebrew liwyātān derive from Proto-Northwest Semitic *lawiy-(a)t-an- through a series of attested sound changes. We will also discuss the morphology of *lawiy-(a)t-an- and the syntax of the Northwest Semitic formula *lawiy(a)tanu baṯnu barīḥu … baṯnu ʕaqallatānu “Leviathan, the fleeing serpent … the twisting serpent” found in KTU 1.5 i 1–3 and Isa 27:1.
In memory of Calvert Watkins
Since the Ugaritic word ltn was first discovered, scholars have generally treated it as a cognate of Hebrew liwyātān.1 But the vocalization of the Ugaritic word and its relationship to the Hebrew form remain debated. In this article, we will argue that ltn should be vocalized /lītan-/ and that Ugaritic ltn and Hebrew liwyātān derive from Proto-Northwest Semitic *lawiy-(a)t-an- “the great twisting one” through a series of attested sound changes. We will also discuss the morphology of ltn/liwyātān and the syntax of the inherited Northwest Semitic formula *lawiy(a)tanu baṯnu barīḥu … baṯnu ʕaqallatānu “Leviathan, the fleeing serpent … the twisting serpent” found in KTU 1.5 i 1–3 and Isa 27:1.
1 The Data
Ugaritic ltn occurs only in the following passage:
Hebrew liwyātān appears in a strikingly similar context:
Wilson-Wright has argued that these two passages preserve reflexes of the same inherited Northwest Semitic formula: *lawiy(a)tanu baṯnu barīḥu … baṯnu ʕaqallatānu “Leviathan, the fleeing serpent … the twisting serpent.”3 They employ a series of cognate terms (ltn ~ liwy-āt-ān, brḥ ~ bārîaḥ, ʕqltn ~ ʕăqallātôn) to describe a serpentine monster with only two minor differences: the Hebrew reflex uses nāḥāš in place of the semantically similar Ugaritic term bṯn and repeats the word liwyātān in the second half of the formula. The first difference provides an example of what Calvert Watkins calls “lexical replacement” under conditions of semantic similarity.4 Because Biblical Hebrew does not preserve an indigenous cognate of bṯn—peten is a loanword from Aramaic—nāḥāš provides an appropriate substitute.5 The second difference will be discussed in section 4 below.
In addition to Isa 27:1, Hebrew liwyātān also appears in Ps 74:14; 104:26, and Job 3:8; 40:25 (Eng. 41:1).6 The form liwyātān almost certainly derives from the common Semitic root lwy “to turn (around), wind, circle, twist,” as is usually suggested.7
2 The Vocalization of Ugaritic ltn and Its Relationship to Hebrew liwyātān
The most common vocalization of the Ugaritic form found in the scholarly literature is /lôtā̆n-/, which seems to have been proposed first by W. F. Albright: “The Hebrew Liwyatân stands for Lawyatân, while the Canaanite [i.e., Ugaritic] Lôtân stands for Lawtân. Which form is more original we cannot say.”8 In his monumental Ugaritische Grammatik, Josef Tropper tentatively suggests a different vocalization: “/lâtān-/ < *lawâtān < *lawayatān(?).”9
In an article published in 1982, however, J. A. Emerton argued for vocalizing the Ugaritic word as /lītān-/.10 Rejecting claims by some scholars that the Hebrew form reflects a later folk-etymology,11 he noted that “It seems desirable … to begin with the traditional Hebrew form liwyātān and to work backwards using it as a clue.” He proposed that *liwyatānu developed in Ugaritic first into *liwyitānu, via assimilation of the second vowel to the first; then to an “intermediate stage,”12 *līyitānu, with assimilation of w to the following y and compensatory lengthening; and finally to *lītānu with the collapse of the triphthong iyī. Unfortunately, there is no certain evidence in Ugaritic phonology for either the first stage of Emerton’s proposed development, progressive assimilation,13 or his third stage, the change of -īyi- to -ī.14
Despite the unlikely development suggested by Emerton, we agree that the Ugaritic form was indeed pronounced /lītā̆n-/. But we argue that this pronunciation developed via a different set of phonological processes, each of which can be documented elsewhere in Ugaritic. Like Emerton, we suggest that Hebrew liwyātān and Ugaritic ltn derive from a common ancestral form—consistent with their appearance in an inherited Northwest Semitic formula. But we believe that form to have been *lawiy-(a)t-an, a feminine verbal adjective from the root lwy “to turn (around), wind, circle, twist” with the suffix -an (discussed in section 3 below). The meaning would thus be “the winding one” or “the twisting one,”15 as is generally believed.16 After the loss of final case vowels and the attendant shift of stress, the Hebrew form developed as follows:*lawiyatán became *lǝwǝyatán and then *liwyatán via propretonic vowel reduction and the rule of shwa, and finally the attested liwyātān;17 compare, for example, niblātô “his corpse” < *nabilatúh.18 The Ugaritic form, by contrast, bore the short allomorph of the feminine morpheme, -t; examples of this variation are forms such as Ugaritic mlỉt /maliʔt-/ “full” vs. Hebrew mǝlēʔā < *maliʔat-, ḥmt /ḥāmīt-/ < *ḥāmiyt- “wall” vs. Hebrew ḥômā < *ḥāmiyat-. Ugaritic *lawiytan- developed as follows: *lawiytan- > *lawītan-, with the contraction of iy to ī as in ḥmt, and then *lawītan- > lītan-, with the collapse of the triphthong awī to ī.19
3 The Morphology of Proto-Northwest Semitic *lawiy-(a)t-an
“Begin[ning] with the traditional Hebrew form and work[ing] backwards,”20 in Emerton’s apt formulation, also requires grappling with a perennial bugbear of Hebrew phonology and morphology: the Canaanite shift of *ā > ō. Emerton and others reconstruct the last syllable of ltn/liwyātān as a reflex of the Proto-Semitic substantivizing suffix *-ān, but this reconstruction leads to a problem.21 Normally, the suffix *-ān becomes -ôn in Biblical Hebrew with the operation of the Canaanite shift (e.g., *ʔabyān- > ʔebyôn “poor”). But liwyātān appears to have escaped this sound change.22
Two explanations have been advanced to account for the apparent retention of *-ān in liwyātān and other forms. Some scholars have suggested that the Canaanite shift was conditioned.23 The inherited Proto-Northwest Semitic formula in Isa 27:1 does not bear this hypothesis out, however.24 If the inherited words ʕăqallātôn and liwyātān both bore the suffix *-ān, it is unclear why ʕăqallātôn was subject to the Canaanite shift and liwyātān was not. There are no obvious conditioning factors that could explain the different outcomes. In both words, the suffix -ān appears at the end of the word after the long form of the feminine morpheme. Jacob Barth by contrast proposed two criteria to account for the retention of *-ān in Hebrew: (1) in words like šulḥān “table,” the presence of an u vowel blocked the operation of the Canaanite shift; (2) while, in the remaining words, the retention of *-ān was an Aramaism.25 These two criteria account for most of the data. Of the seventeen words in Biblical Hebrew that bear the suffix -ān, four (dorbān “cattle goad,” nəḥuštān “DN,” qorbān “offering,” šulḥān “table”) contain an u vowel and an additional nine most likely represent Aramaic loanwords into Hebrew (ʔabdān “destruction,” ʔobdān “destruction,” binyān “building,” bîrāniyyôt “fortresses,” migdānōt
“precious goods,” niṣṣānîm “blossoms,” ʕinyān “business,” qinyān “property,” raḥămāniyyôt “compassionate”).26 But the remaining three forms—kibšān “kiln,” naʕămānîm “pleasantness,” šinʔān “?”—do not meet Barth’s criteria.27 None of them contain a u vowel and none of them are likely to be Aramaic loanwords into Hebrew since they lack Aramaic parallels and appear primarily in early biblical texts.28
In light of these issues, we propose an alternative explanation that better accounts for the data: liwyātān, šulḥān, kibšān, etc., do not bear the common Semitic substantivizing suffix *-ān, but instead preserve a reflex of a Semitic suffix *-an. The latter suffix appears in Classical Arabic in addition to Hebrew and is therefore reconstructible to Proto-Central Semitic at least. In Classical Arabic, this suffix appears on adjectives and the occasional substantive, e.g., raʕšan-un “shivering” and jawšan-un “breast.”29 We believe that the same ending appears in Hebrew on the substantivized adjective *lawiy(a)t- and substantives such as šulḥān.
Hebrew liwyātān exhibits another morphological peculiarity. Although it contains a reflex of the feminine morpheme, it consistently takes masculine agreement in Biblical Hebrew, e.g.,
ʔattâ riṣṣaṣtā rā(ʔ)šê liwyātān tittən-ennû maʔăkāl lə-ʕām lə-ṣiyyîm
You crushed the head of Leviathan, You gave him as food for the people of the wilderness ...Ps 74:14
timšōk liwyātān bə-ḥakkâ û-bə-ḥebel tašqîaʕ ləšōn-ô
Can you draw out Leviathan with a fishhook? Can you press down his tongue with a cord?Job 41:130
It thus belongs to the rare class of words that bear the feminine morpheme, but have a masculine referent (e.g., mōdaʕat “kinsman” in Ruth 3:2, referring to Boaz; the name qōhelet passim in Ecclesiastes).31 In Classical Arabic, such words often connote greatness or distinction, e.g., karīm-un “noble,” karīm-at-un “best one, best part.”32 We can therefore translate liwyātān as “the great winding one.”
4 The Syntax of the Inherited Northwest Semitic Formula
The inherited Northwest Semitic formula in which Ugaritic ltn and Hebrew liwyātān appear displays strange syntax. The word for snake seems to switch genders between the two halves of the formula: in the first colon, it takes masculine agreement (bṯn brḥ/nāḥāš bārîaḥ), while in the second it appears to take feminine agreement (bṯn ʕqltn/nāḥāš ʕăqallātôn). To resolve this discrepancy, we suggest that ʕqltn/ʕăqallātôn is not an adjective modifying bṯn/nāḥāš, but rather a substantive standing in apposition to it.33 Support for this interpretation comes from KTU 1.3 iii 38’–46’:34
In this passage, bṯn ʕqltn appears in a list of divine monsters defeated by Anat, many of whom are identified through appositive titles. Such constructions occur frequently in Ugaritic (e.g., btlt ʕnt “the young woman Anat”) and, less commonly, in Biblical Hebrew (e.g., han-nāhār ʔahăwāʔ “the river Ahava”) and are therefore reconstructible to Proto-Northwest Semitic.37 The inclusion of bṯn ʕqltn in this list suggests that bṯn—like mdd ỉlm, ʕgl ỉl, klbt ỉlm, and bt ỉl—stood in apposition to its referent. If this interpretation proves correct, then the Proto-Northwest Semitic formula *lawiy(a)tanu baṯnu barīḥu … baṯnu ʕaqallatānu and its Ugaritic reflex should read “Leviathan, the fleeing serpent … the serpent Aqallatan”—with Aqallatan serving as an alternative name for Leviathan. This conclusion is particularly appealing given the similar morphology and semantics of Proto-Northwest Semitic *lawiy(a)tanu “the great winding one” and *ʕaqallatānu “the great twisting one” and their Ugaritic and Hebrew reflexes.
Such an interpretation also helps explain the repetition of liwyātān in the Hebrew reflex of the formula. Because nominal apposition is rarer in Biblical Hebrew than in Ugaritic and is largely confined to definite nouns, Hebrew speakers may have analyzed ʕăqallātôn as an adjective modifying nāḥāš rather than a proper name. They may then have added the second liwyātān to make it clear that nāḥāš bārîaḥ and nāḥāš ʕăqallātôn referred to the same entity: “Leviathan the fleeing serpent … Leviathan the twisting serpent” (liwyātān nāḥāš bārîaḥ … liwyātān nāḥāš ʕăqallātôn).
In this article, we have advanced three arguments regarding Ugaritic ltn and Hebrew liwyātān: (1) ltn and liwyātān developed from Proto-Northwest Semitic *lawiy-(a)t-an “the great coiling one” through a series of attested sound changes and, in keeping with this origin, the Ugaritic form should be vocalized as /lītan-/; (2) the proto-form of Leviathan and its reflexes did not bear the substantivizing suffix *-ān, but rather the suffix -an; (3) the term ʕqltn/ʕăqallātôn found in connection with Leviathan in KTU 1.5 i 1–3 and Isa 27:1 is not an adjective modifying bṯn/nāḥāš, but rather a name standing in apposition to it. KTU 1.5 i 1–3 can thus be translated as “Leviathan, the fleeing serpent … the serpent Aqallatan.”
We wish to thank Jo Ann Hackett, Naʿama Pat-El, Ellen van Wolde, and the two anonymous reviewers for their perceptive comments on earlier drafts of the present study. Any remaining errors are our own.
Barker, William D. “‘And thus you brightened the heavens …:’ A New Translation of KTU 1.5 i 1–8 and Its Significance for Ugaritic and Biblical Studies.” UF 38 (2006): 41–52.
Barth, Jacob. Die Nominalbildung in den semitischen Sprachen: Zweite, durch ein Wörter- und ein Sachverzeichnis vermehrte Ausgabe. Leipzig: J. C. Hinrich, 1894.
Bauer, Hans, and Pontus Leander. Historische Grammatik der hebräischen Sprache des Alten Testaments. Halle: Max Niemeyer, 1922. Repr., Hildesheim: Georg Olms, 1962.
Brockelmann, Carl. Laut- und Formenlehre. Vol. 1 of Grundriss der vergleichenden Grammatik der semitischen Sprachen. Berlin: von Reuther & Reichard, 1908.
DUL3 = del Olmo Lete, Gregorio, and Joaquín Sanmartín. A Dictionary of the Ugaritic Language in the Alphabetic Tradition. Translated by Wilfred G. E. Watson. 2 vols. 3rd ed. HdO 1.112. Leiden: Brill, 2015.
Fischer, Wolfdietrich. A Grammar of Classical Arabic. Translated by Jonathan Rodgers. 3rd ed. New Haven: Yale University Press, 2002.
GKC = Gesenius’ Hebrew Grammar. Edited by E. Kautzsch. Translated by A. E. Cowley. 2nd. ed. Oxford: Oxford University Press, 1910.
GMD = Meyer, Rudolf, and Herbert Donner. Wilhelm Gesenius’ hebräisches und aramäisches Handwörterbuch über das Alte Testament. 18th ed. Berlin: Springer, 2013.
Gross, Ben-Zion. The Nominal Patterns פעלון and פעלן in Biblical and Mishnaic Hebrew. Jerusalem: Academy of the Hebrew Language, 1993. (Hebrew)
HALOT = Koehler, Ludwig, Walter Baumgartner, et al. The Hebrew and Aramaic Lexicon of the Old Testament. Translated by M. E. J. Richardson. 5 vols. Leiden: Brill, 1995–2001.
Huehnergard, John. “Biblical Hebrew Nominal Patterns.” Pages 25–64 in Epigraphy, Philology, and the Hebrew Bible: Methodological Perspectives on Philological and Comparative Study of the Hebrew Bible in Honor of Jo Ann Hackett. Edited by Jeremy M. Hutton and Aaron D. Rubin. ANEM 12. Atlanta, GA: Society of Biblical Literature, 2015.
- Search Google Scholar
- Export Citation
. “ Huehnergard, John Biblical Hebrew Nominal Patterns.” Pages 25– 64in Epigraphy, Philology, and the Hebrew Bible: Methodological Perspectives on Philological and Comparative Study of the Hebrew Bible in Honor of Jo Ann Hackett. Edited by . and Jeremy M. Hutton Aaron D. Rubin . ANEM12 Atlanta, GA: Society of Biblical Literature, . 2015
Isbell, Charles D. Corpus of the Aramaic Incantation Bowls. SBLDS 17. Missoula, MT: Society of Biblical Literature/Scholars, 1975.
Joüon = Joüon, Paul, and Takamitsu Muraoka. A Grammar of Biblical Hebrew. 2nd ed. SubBi 27. Rome: Pontifical Biblical Institute, 2007.
Noonan, Benjamin J. Non-Semitic Loanwords in the Hebrew Bible: A Lexicon of Language Contact. LSAWS 14. Winona Lake, IN: Eisenbrauns, 2019.
Smith, Mark S., and Wayne T. Pitard. Introduction with Text, Translation and Commentary of KTU/CAT 1.3–1.4. Vol. 2 of The Ugaritic Baal Cycle. VTSup 114. Leiden: Brill, 2009.
Streck, Michael P. “Simply a Seller, Nothing but Gods: The Nominal Suffix -ān in Old Babylonian.” Pages 233–243 in Babel und Bibel 2: Memoriae Igor M. Diakonoff. Edited by Leonid E. Kogan, Natalia Koslova, Sergey Loesov, and Serguei Tishchenko. Winona Lake, IN: Eisenbrauns, 2005.
- Search Google Scholar
- Export Citation
“ Streck, Michael P. Simply a Seller, Nothing but Gods: The Nominal Suffix.” Pages -ānin Old Babylonian 233– 243in Babel und Bibel 2: Memoriae Igor M. Diakonoff. Edited by . , LeonidE. Kogan , Natalia Koslova , and Sergey Loesov Serguei Tishchenko Winona Lake, IN: Eisenbrauns, . 2005
Suchard, Benjamin D. The Development of the Biblical Hebrew Vowels: Including a Concise Historical Morphology. Studies in Semitic Languages and Linguistics 99. Leiden: Brill, 2019.
Wagner, Max. Die lexikalischen und grammatikalischen Aramaismen im alttestamentlichen Hebräisch. Berlin: Alfred Töpelmann, 1966.
Whitney, K. William. Two Strange Beasts: Leviathan and Behemoth in Second Temple and Early Rabbinic Judaism. HSM 63. Winona Lake, IN: Eisenbrauns, 2006.
Wilson-Wright, Aren M. “Love Conquers All: Song of Songs 8:6b–7a as a Reflex of the Northwest Semitic Combat Myth.” JBL 134 (2015): 333–345.
Wilson-Wright, Aren M. Athtart: The Transmission and Transformation of a Goddess in the Late Bronze Age. FAT 2/90. Tübingen: Mohr Siebeck, 2016.
Wright, William. A Grammar of the Arabic Language: Translated from the German of Caspari and Edited with Numerous Additions and Corrections. Vol. 1. 3rd ed., revised by W. Robertson Smith and M. J. de Goeje. Cambridge: Cambridge University Press, 1896.
So already Virolleaud in the editio princeps (“Mort de Baal”) and in an even earlier note (Virolleaud, “Note,” 356–357).
Presumably also in the parallel lines KTU 1.5 i 27–30, where ltn is lost in a lacuna.
See Wilson-Wright, “Love Conquers,” 337. For a discussion of inherited formulae in general, see the magisterial work of Calvert Watkins (How to Kill a Dragon, 12–16); for more on inherited formulae in the Semitic languages, see Wilson-Wright, Athtart, 128–138; Kaplan and Wilson-Wright, “Song of Songs,” 337–344.
Watkins, How to Kill a Dragon, 10, 15.
Hebrew nāḥāš may have originated as a taboo substitute for earlier *baṯn. According to Leonid Kogan (Militarev and Kogan, Animal Names, 210–211; Kogan, Classification, 300–301), Hebrew nāḥāš and Akkadian nēšu “lion” both go back to a Proto-Semitic noun simply meaning “wild animal,” which was subsequently repurposed to designate different dangerous creatures. Militarev and others, however, have suggested different cognates for nāḥāš (see Militarev and Kogan, Animal Names, 210–211).
Also in a Qumran fragment, 4Q380 3: 1; see Schuller, Psalms, 260. For various suggestions as to the identity of liwyātān, see, e.g., HALOT 524; GMD 602; Lipiński, “Liwyātān”; Barker, “New Translation”; Seow, Job, 325–326.
A similar form also appears in Aramaic texts, where it is presumably borrowed from Hebrew: in Aramaic incantation bowls, as lywyt(w)n tnynʔ “L. the tannin” (see Isbell, Corpus, texts 2:4; 6:8; and 7:7, 9; on the writing lywytwn, see n. 22 below); Jewish Aramaic (see CAL), e.g., Lev. Rabbā 278:1(3): bǝzaʕ l-eh l-lwytn “he split the l.,” Targ. Song 8:2: sǝʕodtā d-lwytn “a meal of l.”; Syriac lēwyātan, lēwyātān (see Payne Smith, Thesaurus, 2.1902–1903); Mandaic liuiatan/lʿuiatan (see Drower and Macuch, Dictionary, 236a). See also below, n. 17, on the Ethiopic borrowing of the form.
Albright, “New Light,” 19 n. 18. Note also Virolleaud, “Mort de Baal,” 308: “Ltn, qu’on prononçait sans doute Lôtan.”
Tropper, Grammatik, 272 §51.46e; in his Wörterbuch, 66, however, Tropper vocalizes ltn as /lôtānu/.
Emerton, “Leviathan and Ltn.” Emerton’s vocalization is followed by, inter alios, Barker, “New Translation”; Smith and Pitard, Introduction, 252; Seow, Job, 325; GMD 602.
E.g., Virolleaud, “Mort de Baal,” 308.
Emerton, “Leviathan and Ltn,” 330.
The divine name ỉlỉb, noted by Emerton, may be such an example, but the assimilation presumably has more to do with the intervening /ʔ/. The examples of regressive assimilation noted by Emerton, /ʔibbīr-/ and /ʔullūp-/, occur only in a narrow phonetic environment, for which see Huehnergard, Vocabulary, 269–270, 401.
See Whitney, Beasts, 8–9, n. 53; Huehnergard, Vocabulary, 292; Tropper, Grammatik, 197 offers some possible examples, but none is secure. There is some limited evidence for the second development proposed by Emerton, viz., -iwy- > -īy- (or, perhaps better, > -iyy-), in the form lyt, now tentatively glossed “retinue(?)” in DUL3 503, which, if correct, presumably reflects /liyyat-/ < *liwyat- (cf. Syriac lewyā “retinue”). Others gloss lyt as “wreath,” and thus cognate with Hebrew liwyā; see, e.g., Tropper, Grammatik, 196, 199.
Note that the parallel Ugaritic term tnn “serpent,” vocalized /tunnan-/, is also, formally, a verbal adjective, albeit from the D stem; see Huehnergard, Vocabulary, 72.
See Tropper, Grammatik, 272: “die Gewundene”; GMD 602a: “d. sich Windende.” Bauer and Leander (Grammatik, 500 §61r
The medial radical w is preserved in this root throughout Semitic (e.g., Old Babylonian Akkadian lawi “it goes around”). Normally, intervocalic y is elided in Hebrew, as in dāwā “ill (fem.)” < *dawiyat- (Lev 15:33); there are, however, instances in which the original triphthong is preserved, such as ḥāsāyā “it (fem.) has taken refuge” < *ḥaṣayat (Ps 57:2), yibkāyûn “they weep” < *yibkayūn (earlier *yabkiyūna; Isa 33:7). At least some of the latter are archaic forms, and the preservation of y in liwyātān is also likely due to its occurrence in an ancient poetic formula. Evidence for an intermediate stage of this development may be reflected in the Classical Ethiopic form of the word borrowed from Hebrew, viz., lewiyātān. Both Classical Ethiopic and Amharic also exhibit a variant form lewātā(n); see Leslau, Dictionary, 321–322.
See Huehnergard, “Nominal Patterns,” 43. The Hebrew noun liwyā “wreath” (Prov 1:9; 4:9), on the other hand, presumably reflects a qiṭl-at form.
See Huehnergard, Vocabulary, 289–290; Tropper, Grammatik, 198.
Emerton, “Leviathan and Ltn,” 329.
This suffix was used to derive nouns from other nominal forms. See Streck, “Simply a Seller”; Huehnergard, “Ashkelon,” 92*.
The only possible exception is Aramaic lywytwn, which may ultimately go back to Hebrew liwyātôn. As noted by Lipiński (“Liwyātān,” 19), the spelling -wn presumably indicates -ôn. We argue, however, that this form represents a later “Hebraification” of earlier liwyātān in which a bilingual Hebrew-Aramaic speaker, knowing the word to be Hebrew, replaced the rarer, Aramaic-looking ending -ān with its Hebrew equivalent -ôn. Compare the spelling nʕmwnym in 1QIsaa 17:11 for MT naʕămānîm “pleasantness” and the alternation between dorbān “ox-goad” in 1 Sam 13:21 and dorbōnôt in Eccl 12:11 (on this second form, see also Suchard, Development, 76).
See, e.g., Bauer and Leander, Grammatik, 192; Blau, Phonology and Morphology, 48.
Rabin, “Hebrew Development,” makes a very strong case that *ā > ō in Hebrew was an unconditioned change. See also Suchard, Development, 83–84, who argues that the Canaanite shift was unconditioned, but was blocked in a handful of words by a preceding u or w.
See Barth, Nominalbildung, 318. Barth’s proposal has been followed by Brockelmann, Laut- und Formenlehre, 395–396; Bauer and Leander, Grammatik, 215; Gross, Nominal Patterns, 215–216; and, in a modified form, Suchard, Development, 83–84.
These words occur primarily or exclusively in later Biblical texts and have clear Aramaic parallels (listed here with their respective earliest attestations): ʔabdānā/ʔubdānā “destruction” (Qumran), benyānā “building” (Imperial), bîrānātā “fortress” (Targumic), magdā “precious goods” (Targumic), niṣṣānīn “blossoms” (Christian Palestinian), ʕenyānā “matter of concern” (Qumran), qenyān “property” (Imperial), raḥmān “merciful” (Old Aramaic).
Two other Hebrew words that do not meet Barth’s criteria are širyān (more often širyôn, also siryôn) “scale armor” and bîtān “palace,” but both terms are loanwords into Hebrew (see Noonan, Non-Semitic Loanwords, 213–214; Mankowski, Loanwords, 47–48). We have also excluded the form ḥelbǝnâ “galbanum” in Exod 30:24 from consideration because of its decidedly non-standard morphology. It appears to be an inner-Semitic loanword into Hebrew from earlier *ḥilbVnat-.
The form liwyātān itself poses a similar problem for Barth’s explanation. As mentioned above in n. 7, the Aramaic forms of this word are most likely loanwords from Hebrew.
See, e.g., Barth, Nominalbildung, 344.
Ugaritic ltn never appears in situations that call for gender agreement.
See GKC 463 §145h; Joüon 244 §89b.
See Fischer, Grammar, 46 §73a; see also Wright, Grammar, 1.139 § 233 remark c; WKAS, K, 146.
Another possible explanation is that bṯn and its Proto-Northwest Semitic or Proto-Semitic antecedent *baṯnu were epicene nouns and that *lawiy(a)tanu baṯnu bāriḥu and *baṯnu ʕaqallatānu referred to two different serpents, one male and one female. The evidence for this theory is slim, however. Although Ugaritic bṯn takes masculine agreement once (KTU 1.5 i 1) and feminine agreement twice (KTU 1.5 i 2; 1.178:4 ʕlk l tʕl bṯn “(so that) the snake does not jump upon you”), both Akkadian bašmu and Aramaic patnā consistently take masculine agreement. We cannot be certain therefore that *baṯnu was epicene.
The Ugaritic text follows Pardee, “Dragon.”
See Hutton, “Ugaritic */š/,” for this reading of line 40’.
See Richey, “Monsters,” for this interpretation of ʕtk.
See Tropper, Grammatik, 839 § 91.12; GKC 425 § 131f; Joüon 450 § 131h, k.