Attributive constructions

Adjectives in Coptic are not adjoined to the head-noun directly, but by means of the linking element N-. To be exact (and without determining the relative order within the noun phrase), the two (or more) elements of an attributive construction are mediated by this linking morpheme, as illustrated by the examples below.⁹⁰ The examples are intentionally chosen in such a way as to show the different degrees of determination that can freely co-occur with attributive constructions.

(33) ou-sxime N-samariths [John 4:9]

INDF.SG-woman ATTR-Samaritan

‘a Samaritan woman’

(34) te-sxime N-samariths [John 4:9]

DEF.SG.F-woman ATTR-Samaritan

‘the Samaritan woman’

(35) ou-¥thn N-jhqe [John 19:2]

INDF.SG-garment ATTR-purple

‘a purple garment’

90 For the description of this construction and further examples, see Layton (2000: §§96-103); Reintges (2004: 90 §3.1.3); Steindorff (1951: §§147-150); Stern (1880: §§185-188); Till (1961: 67-68 §114); Vergote (1983: §189).

(36) te-¥thn N-jhqe [John 19:5]

DEF.SG.F-garment ATTR-purple

‘the purple garment’

(37) kas N-ouwt [John 19:36]

bone ATTR-single

‘single bone’⁹¹

(38) ou-litra N-soqn N-nadros [John 12:3]

INDF.SG-pound ATTR-ointmentATTR-nard

‘a pound of ointment of nard’

Dealing with the attributive constructions two problems arise immediately: firstly, the mere existence of an ‘adjectival’ category in Coptic has been questioned, seeing that the modifier part of the construction can be filled not only by prototypical property describing lexemes (like beautiful, great, etc.) but practically by any nominal element as well, as it was demonstrated and long discussed in section 2.3.1. The second apparently confusing fact is that the linking morpheme that signals the attributive relationship seems to be identical with the one in the possessive Pattern A. In fact, many scholars think that the same morpheme figures in both constructions and thus they describe it by the same relatively neutral name, such as mark of relationship (Layton 2000: §203) or nota relationis (Shisha-Halevy 1986: 20).⁹² The semantic difference, however, is obvious between the attribution/qualification on the one hand and the possession on the other.

Thus, there must be one or more rules for disambiguation on the formal side of these linguistic expressions too, so that Pattern A and the attributive pattern may be clearly distinguished if necessary.

It is worth observing how the structural characteristics of this pattern differ from those of the possessive structures. The second element of the attributive construction, the modifier introduced by N-, cannot have any determiner (Shisha-Halevy 1988: 6–8;

Reintges 2004: 90). Moreover, it is this zero-determination that contributes to the attributive reading of this type of adnominal modification (Shisha-Halevy 1986: 131 §4.1).

Another crucial syntactic difference is that this noun + N- + noun sequence, for its part, can have any sort of determiner (definite, indefinite, demonstrative, or even a possessive

91 This determinerless phrase is used as the subject of a negated Future III pattern where the intended meaning is “no bone at all shall be bruised of him”.

92 The formal likeness of the linking morphemes is not accidental, of course. From a diachronic point of view they are really identical having the same source (see chapter 5 on this issue), but synchronically they are better to be treated separately.

article), which shows that the attributive expansion does not affect the degree of definiteness of the phrase as a whole in the way it is conditioned in possessive Pattern A.⁹³ To illustrate the syntactic differences between possession vs. attribution, a minimal pair is provided below that contains two phrases with practically the same lexical content coming from the same text (The Life of Apa Onnophrios, after Till 1961: 282):⁹⁴

(39) p-xwb N-nek-qij

DEF.SG.M-thing POSS- POSSART.PL.2SG.M-hand

‘the work of your hands’

(40) xen-xwb N-qij

INDF.PL-thing ATTR-hand

‘handiworks’

While the first phrase is a possessive, the second is an attributive construction. In (39), the possessor by itself constitutes a full noun phrase with a possessive article of its own (nek-

‘your’), and the possessed noun (actually the head of the construction) is simply definite.

In (40) only a bare noun follows the linking element N-. The N-marked complex (N-qij) literally corresponds to ‘of hand’ with the intended meaning as to assign the quality ‘made by hand’ to the head-noun. Since the whole expression (xwb N-qij ‘handiwork’) neither describes a possessive relationship nor is referential by itself, it can easily be indefinite, as it actually is in (40).

Contrary to regular possessive constructions, the members of the attributive structures seem to be more loosely connected. As far as I observed, based on the collection of data from the Gospel of John in P. Palau Rib 183, the noun and its modifier can be more easily separated, that is to say, not only do clitics intervene in the sequence of the nominal elements, but independently accented items as well.

It is to be noted that Coptic makes an extensive use of relative clauses to express an attributive modification. This might be (but is not necessarily) related to the relatively low number of native adjectives in the word-stock. What is more, a few property-denoting lexemes have irregular verboid syntax (e.g. nanou- ‘to be good’) and cannot be inserted

93 I cannot share Antonio Loprieno’s view (1995: 56) according to which the linking element n- may be a determinative pronoun similar to the one used in Hebrew attributive constructions. In Hebrew, in case of a definite noun there appears another definite article (or determinative pronoun) on the adjective (hā-’îš hag-gādôl the-man the-great i.e. ‘the great man’), while after an indefinite noun both the article and the determinative pronoun are absent (’îš gādôl ‘a great man’). The Hebrew phenomenon can be explained in terms of agreement in definiteness and, accordingly, is far from being similar to the Coptic modifier marker n- whose appearance is independent of the definiteness of the head-noun.

94 The text is from around 1000, so does not fit in the time-interval defined in the Introduction. However, I decided to cite these examples as they perfectly illustrate the matter.

into the N-linked modifier pattern, which is restricted to nominal elements. The constructions may also vary with no relevant difference in meaning (cf. (44) and (45)).

(41) pe-pNa et-ouaab [John 14:22]

DEF.SG.M-spirit REL-pure.Q

‘the holy spirit’

(42) ou-pNa e-f-ouaab [John 1:33]

INDF.SG-spirit CIRC-3SG.M- pure.Q

‘a holy spirit’

(43) p-¥ws et-nanou-f [John 10:14 and 10:14]

DEF.SG.M-shepherd REL-be.good-3SG.M

‘the good shepherd’

(44) ou-mhh¥e e-na¥w-f [John 4:41 and 6:2]

INDF.SG-multitude CIRC-be.much-3SG.M

‘A great multitude’

(45) ou-noq M-mhh¥e [John 6:5]

INDF.SG-big ATTR-multitude ‘A great multitude’

Admittedly, from a semantic point of view, the boundary of the patterns sometimes seems to be quite fuzzy, and there is certain permeability between the possessive and attributive constructions. All the three examples cited below intend to express the notion of something being ‘living’ or ‘related to life’, but they perform three different structural configurations in one and the same Gospel text. (46) is a possessive construction in which Jesus is described as the bread of life; a few lines below, however, the same notion is presented by an adnominal relative clause (47), and finally the attribute can appear in a real attributive construction where the linking element is N- and no determiner appears on the second noun (48).

(46) p-oik M-p-wnx [John 6:48]

DEF.SG.M-bread POSS-DEF.SG.M-life

‘the bread of the life’

(47) p-oik et-onx [John 6:51]

DEF.SG.M-bread REL-live.Q

‘the living bread’

(48) xen-¥aje N-wnx ¥a-enex [John 6:68]

INDF.PL-word REL-life for-ever

‘words for eternal life’