The nominal plural morphemes in East Semitic by and large match the morphemes in the rest of Semitic; the branch, however, has a unique masculine plural morpheme to mark attributive adjectives, which is not attested with adjectives elsewhere in Semitic: -ūt-; for example, OAkk išpikī kinūtim ‘real yields’ (Gir 19: 39); OA awīlū išten u šina lamnūttum ‘1–2 evil men’ (AKT 6B, 508:1–3); OB šīrāšî damqūtim ‘good brewers’ (CUSAS 36 95: 7); Ebl. a-(wa-)mu ’à-mu-tum ‘warm days’ (Catagnoti Reference Catagnoti2012: 120). Assyriologists and Semitists have assumed that this morpheme is secondary, but differ on the identification of its origin. Initially, the most common assumption was that the plural morpheme -ūt- is related to the nominal derivational morpheme of abstract nouns -ūt- (e.g. von Soden Reference Soden1995: 94, §61k). According to this hypothesis, the abstract morpheme first shifted to mark collectives and then individual plurals. There are, however, a number of hefty arguments against this derivation.Footnote ² First, the forms differ in gender-number: the abstract noun in -ūt is feminine and singular, while the adjective in -ūt is masculine and plural. Second, in Assyrian, the adjectival morpheme and the abstract morpheme differ: mp ūt vs. fs -utt.Footnote ³ Third, the forms behave differently when possessive suffixes are attached to them. The abstract morpheme behaves as expected with the possessive suffix attached directly to the morpheme: OB ah̬h̬ūt-ka ‘your brotherhood’ (CUSAS 36 169: 5'), rīqūs-sina ‘their emptiness’ (CUSAS 36 165: 17); OA šebūtta-ka ‘your testimony’ (TC 2, 39: 6). The masculine plural adjective before possessive suffixes, however, typically reverts to a substantive plural morpheme (-ū or -ānū), or the suffix is attached after a case ending: OB uppusūtī-ya ‘those who object to me’ (CUSAS 36 147: 44); damqūtī-ka ‘your good (troops)’ (CAD D 71a); OA ina damqūte-ka ‘among your good (textiles)’ (BIN 4. 65: 16). Fourth, it is unclear why and under what circumstances a substantive suffix would be transferred to mark plural in the adjectival paradigm, but not in the substantival paradigm. Fifth, the abstract suffix is a derivational morpheme, while the adjective plural is an inflectional morpheme. The assumption that the masculine plural adjectival morpheme developed from the abstract suffix should, therefore, be rejected.

Several scholars have suggested independently a different direction, namely, that the morpheme developed in analogy with the feminine plural, a solution which has now become widely accepted (Bravman Reference Bravman1947; Correll Reference Correll1990; Huehnergard Reference Huehnergard, Deutscher and Kouwenberg2006b: 9).Footnote ⁴ According to this scenario, in proto Akkadian the masculine plural adjectival morpheme developed as a result of a four-part analogy with the feminine plural morphemes, on the basis of the morphological relationship between the suffixes of the predicative and attributive adjectives. The following formal analogy was suggested by all three scholars: pred. damqā : att. damqātum :: pred. damqū : att. X = damqūtum. This solution is quite elegant and convincing and it has an added advantage since by appealing to typical adjectival functions, this analogy also explains why masculine substantives remain unaffected.

There are, however, other forms with the masculine plural ending -ūt: some demonstratives and the relative marker. The relative marker is attested in both major Semitic branches with this peculiar masculine plural suffix: Old Akkadian θūt and Ugaritic dt /dūtu/.Footnote ⁵ Huehnergard (Reference Huehnergard, Hurvitz and Fassberg2006a: 112, fn 59) argues that this form must be original primarily because a masculine plural morpheme with a -t is unusual in Semitic and since there is no obvious internal source for it in Ugaritic. The occurrence of this unique suffix in both East and West Semitic makes a reconstruction of the masculine plural relative marker *θūtu to proto-Semitic highly likely. While most scholars assume that the relative marker is a pronoun, Huehnergard and Pat-El (Reference Huehnergard and Pat-El2018) have argued that on the basis of the relative marker's morphosyntax, specifically its ability to serve as the head in a construct chain, and its agreement pattern, it must have been originally an adjective, and not a pronoun. If the relative marker is indeed an adjective, a masculine plural suffix -ūt-, rather than a pronominal plural, is expected in East Semitic, but not in West Semitic.

Hasselbach (Reference Hasselbach2007: 20) suggests that the final -t on the Ugaritic relative marker is an oblique morpheme.Footnote ⁶ The use of oblique -t, however, is restricted to demonstratives and pronouns with an initial PS sibilant.Footnote ⁷ It is attested in East Semitic, Sabaic, Ugaritic, Phoenician, some Arabic dialects, and several Ethiopic languages (Ge'ez, Tigre, Tigrinya, Gafat, etc.). Other demonstratives do not typically carry oblique -t, with the exception of Ethiopic, where it appears that the ending of the personal pronouns may have spread to other pronominal elements.Footnote ⁸ It is, therefore, unlikely that the -t on Ugaritic /dt/ is oblique. Furthermore, the oblique suffix is likely *-ti, in view of the Akkadian forms 3ms šu’āti, fs ši’āti and the Eblaite form šuwāti (su-wa-ti ARET XVI 2 rev. IV 6).Footnote ⁹ However, since final short *-u# in East Semitic was lost, but not final short -i#, and since the form of the relative marker in Old Akkadian is θūt,Footnote ¹⁰ we should conclude that the oblique suffix on šu’āti more likely has a different origin from the suffix of the relative marker. The fact that forms with the adjectival plural suffix -ūt- take case after the suffix, while the oblique marker -ti is positioned at the right-most edge of the pronoun supports this conclusion.Footnote ¹¹ It is, therefore, much more likely that Huehnergard (Reference Huehnergard, Hurvitz and Fassberg2006a) is correct to assume that Ugaritic /dt/ is identical to the East Semitic relative marker θūtV.

The masculine plural with -ūt on demonstratives is unique to East Semitic: OA anniūtum, alliūtum OB annûtum.Footnote ¹² The same suffix is also used with the masculine plural form of the independent possessive pronouns:Footnote ¹³ OA šuā’ūtum ‘his’, kuā’ūtum ‘your’, yā’ūtum ‘my’, niā’ūtum ‘our’ vs. fp šuā’ātum, kuā’ātum, etc. (Kouwenberg Reference Kouwenberg2017: 321–2). These demonstratives function primarily as attributes (Hasselbach Reference Hasselbach2007: 6).Footnote ¹⁴ In West Semitic languages, the plural demonstrative (masculine or feminine) either has a different base from the singularFootnote ¹⁵ or is marked for plural with more typical morphemes, such as *ū or *ī. As noted above, the use of -t with masculine plural demonstratives in West Semitic, Phoenician hmt and Ugaritic hnhmt, is associated with the oblique.Footnote ¹⁶

The analogy between the predicative and attributive forms of the feminine plural and the masculine plural, that was suggested to account for the adjectival masculine plural -ūt, cannot work with the relative marker or the demonstrative. Both of these morphemes, the relative marker and demonstrative, lack a specific predicative form and can function as predicates only in very rare and highly marked cases. The triggering syntactic functions that would have motivated the proposed analogy are therefore absent. If indeed the relative markers are adjectives and their masculine plural suffix cannot be logically explained in any other way, it is more likely that the attributive masculine plural suffix on adjectives is original.Footnote ¹⁷ This means that adjectives were originally morphologically marked for predicative and attributive functions.

One piece of evidence to support this hypothesis is the feminine plural, which originally had two sets of suffixes, attributive and predicative. The relics of these morphemes are attested in both branches (see Table 1). East Semitic has a set of two adjectival morphemes: attributive -āt and predicative -ā that is used to mark the 3fp of the stative, a verbal adjective in predicative position.Footnote ¹⁸ In West Semitic the perfect, a reflex of this predicative inflection, had a third person feminine plural morpheme ā. This original suffix was retained in Aramaic, Ugaritic, and Ethiopic, lost in Hebrew, and replaced by an innovation in Arabic (Huehnergard Reference Huehnergard1987: 271). Regardless of the changes in individual West Semitic languages, the feminine plural suffix of the Suffix Conjugation is always distinct from the feminine plural suffix of attributive adjectives in all West Semitic languages.

Table 1 Feminine plural adjectival suffixes

Given the consistent distinction between feminine plural attributive and predicative, the reconstruction of this system to the proto-language is fairly secure. The current assumption that the masculine plural had only one suffix for both functions is inconsistent with the distinction between them in the feminine plural. It is not very plausible that the feminine plural is more sensitive to the distinction between predicative and attributive than the masculine; it is more likely that originally both plurals distinguished these functions. The attributive masculine plural morpheme was eventually abandoned in West Semitic in favour of the substantive plural morpheme, which was used on other bases in the verbal system, and was perhaps more distinct from the feminine plural morpheme than -ūt is.Footnote ¹⁹

That the distinction between predicative and attributive adjectives is salient in Semitic can be seen also in the development of the definite article in Central Semitic, which may have started from a syntactic marker of attributive adjectives. Pat-El (Reference Pat-El2009) suggested that the original function of the definite article in Central Semitic languages was to distinctly mark attributive adjectives and participles (*han-Adj), since the combination N-Adj was ambiguous in these languages and the adjective in this construction could be interpreted either as attributive or as a predicate in a nominal sentence.

To summarize, I have suggested that the East Semitic masculine plural morpheme of attributive adjectives -ūt is not an innovation of East Semitic but rather should be reconstructed to Proto Semitic. To support this proposal, I have pointed to the masculine plural morpheme -ūt on the relative marker in Akkadian and Ugaritic, as well as on the masculine plural demonstrative in East Semitic. In addition, I suggested that the fact that the feminine plural distinguishes between predicative and attributive states supports reconstructing the same distinction for the masculine plural. Table 2 reflects the proposed reconstruction.

Table 2 Suggested reconstruction of plural adjectival morphemes