The construct state construction across Semitic languages

From the above discussion it should be clear that the CSC is essentially a head-marking construction,³²as in the following Hebrew example (contrast withתִי ַבּ bayiṯ ‘house.free’):

29Creissels prefers the term construct form over construct state due to the confusion arising from the use of the former as a construction label. I shall stick to the traditional term, but note that the notion of construct state can relate both to the morphological marking, and to the syntactic position of a primary (not necessarily marked as such). When in doubt, I will use the term “construct state marking” or “construct state form”.

30An alternative term, proposed by Dixon (2010: 268) is pertensive “based on the Latin verb pertinēre ‘to belong’”. The term has not gained wide usage, as far as I am aware of. Dixon uses this term, moreover, as designating both simple markers and pronominal markers. It may be for this reason that he does not simply adopt the notion of construct state, although he is aware of the partial equivalence between the two (Dixon 2010: 310, fn. 16.2). I am grateful to Adam Pospíšil, who drew my attention to this term.

31For possible origins of the emphatic state, unique to Aramaic among Semitic languages, see König (1901).

32In the Semitic languages which mark case, namely Akkadian and Classical Arabic, it is a double-marked construction; see discussion further down in this section.

(2) Biblical Hebrew: Noun–Noun (Head-marked AC) ךְ ֶל ֶמּ ַה תי ֵבּ

bēṯ house.cst

ham-mɛlɛḵ def-king

‘the house of the king’

In Hebrew the construct state nominals are characterized by “lighter vocali-sation” in comparison with their corresponding free state. Sometimes they are marked by specific suffixes, namely -aṯ for feminine singular nouns (in contrast to free form -ā) and -ē (Modern Hebrew -ey) for masculine plural nouns (in con-trast with -īm). All these changes can be explained by the primary losing word stress, and forming one phonological word with the secondary (Van Hecke 2013:

580). This explanation, however, is diachronic, as in Modern Hebrew these forms prevail even when the primary gets its own stress, as is evident from cases where two construct state primaries are conjoined:³³

(3) Modern Hebrew: Conjoined Nouns–Noun Phrase רפסה תי ֵב י ֵדימלתו י ֵרומ

[mor-ey

teacher-m.pl.cst ve=

and=

talmid-ey]

pupil-m.pl.cst [bet house.cst

ha-sefer]

def-book

‘teachers and pupils of the school’

Biblical Hebrew allows not only nouns as secondaries, but also other elements, such as prepositional phrases and clauses (Cohen 2013: 236).

The CSC of Pre-modern Aramaic is similar in essence to Hebrew. The AC sys-tem of Syriac shall be treated in detail in Chapter 3.

In Akkadian and Classical Arabic, which manifest the old Semitic case system, the secondary is further marked by the genitive case, giving effectively rise to a doubly marked construction. In Akkadian, the construct state is created by re-moving the mimation, i.e. an -m ~ -n suffix, typical of free state nouns. In some texts, the singular construct state forms are further characterized by losing the case endings, though they may surface before pronominal secondaries (Golden-berg 2013a: 144).

33In Biblical Hebrew it is generally accepted that only one nominal can occur as primary, as the counter-examples are extremely rare (Verhej 1989: 210): Gesenius (1909: 433, §128a, note 1) lists 4 such tentative cases, of which only one is really clear (Ezekiel 31:16):ןוֹנ ָב ְל־בוֹ ֽט ְו ר ַח ְב ִמ[miḇḥar wə tō̱ḇ] ləḇānōn ‘the choice and best of Lebanon’ (King James translation). Yet one finds other cases of intervening material between a construct state primary and its secondary; see Freedman (1972). In Syriac too there is a rare occurrence of conjoined primaries; see example (11) on page 52. Similar examples are attested in Standard Arabic (Badawi et al. 2004: 138f.).

(4) Akkadian: Noun–Noun (Double-marked AC) bīt-∅

house-cst

awīl-i-m man-gen-free

‘man’s house’ (Goldenberg 2013a: 232)

Clausal secondaries in Akkadian are marked by a special verbal form, the sub-junctive:

(5) Akkadian: Noun–Clause (Double-marked AC) bīt-∅

house-cst īpuš-u

made.3ms-sbjv

‘house (which) he made’ (Cohen 2010: 80)

The situation in Classical Arabic is similar to Akkadian, in that the nunation (from Arabic

ﻦﻳﻮﻨﺗ

tanwīn), or -n suffix, disappears, while case endings, however, are retained.³⁴ In Arabic, the nunation occurs in complementary distribution with definite article, and therefore is normally seen as an exponent of the in-definite. This analysis, however, is challenged by Lyons (1999: 91–94). He argues that the nunation (in a variant form), can co-occur with the definite article in plu-ral and dual nouns, and thus cannot mark indefiniteness. While he analyses it as

“a semantically empty marker of nominality” (Lyons 1999: 93f.), he notes that it is always dropped in the construct state. Thus, it seems reasonable to conclude that the nunation is a marker of the free state.³⁵The lack of nunation of definite singular nouns may be then tentatively explained as resulting from the principle of economy, as the primary of the CSC cannot in general be determined by a definite article.³⁶Conversely, the absence of nunation coupled with the absence of a definite article is a clear indicator of the construct state, as in the following example:³⁷

34I refer here to the functional similarity between Arabic and Akkadian. Whether the two are historically related is of course a separate question. It is worthwhile noting, in this respect, that also Hebrew free.m.pl suffix -im, as well as Aramaic abs.m.pl suffix -in lose the /m/ or /n/ segments respectively in construct state. Prima facie, it seems reasonable to assume that all these functionally and phonetically elements share a common origin.

35A similar position is maintained by Retsö (1997), who investigates also the origin of this system.

36The exception for this is the CSC headed by adjectives, a construction termed in Arabic Gram-mar impure annexation. See Goldenberg (2002: 204ff.) for further details and analysis.

37The situation, however, is complicated by the fact that a certain class of nouns is never marked by the nunation.

(6) Arabic: Noun–Noun (Double-marked AC)

ِلكﳌا ُﺖﻴﺑ

bayt-u-∅

house-nom-cst

l-malik-i def-king-gen

‘The house of the king.’

In Modern Arabic dialects, both the case endings and the nunation are gone, giving rise to pure juxtaposition of the primary and secondary, the only indicator of the CSC being the lack of definiteness marking on the primary:

(7) Iraqi Arabic: Noun–Noun (Juxtaposition) bēt

house ʿali A.

‘Ali’s house’ (Erwin 2004 [1963]: 370) (8) Maltese: Noun–Noun (Juxtaposition)

omm mother

Pawlu P.

‘Paul’s mother’ (Fabri 1996: 230)

A remnant of the construct state marking is however found in feminine nouns, which in Classical Arabic are written with a tāʾ marbūṭa letter (

ة

) word-finally.

This letter represents a /t/ phoneme, which is however not pronounced at the edge of a phonological word. In the CSC, such a primary forms one phonological word with the secondary, and ends therefore with an -(a)t segment, effectively marking the construct state in opposition to the free state ending -a.

(9) Iraqi Arabic: Noun–Noun (Head-marked AC) sayyār-at

car-fs.cst ʿali A.

‘Ali’s car’ (Erwin 2004 [1963]: 370)

(10) Maltese: Noun–Noun (Head-marked AC) nann-t

grandmother-fs.cst Pawlu P.

‘Paul’s grandmother’’ (Fabri 1996: 232)

(11) Moroccan Arabic: Noun–Noun (Head-marked AC) mədras-t

school-fs.cst Nadya N.

‘Nadia’s school’’ (Benmamoun 2005: 479)

In document Attributive constructions in North-Eastern Neo-Aramaic (Page 56-60)