[89b] Very frequently such interdependent words are also united by Maqqeph (§16a); this, however, is not necessary, but depends on the accentuation in the particular case. On the wider uses of the constr. st. see the Syntax, § 130.
[89c] 2. The vowel changes which are occasioned in many nouns by the construct state are more fully described in §§ 92–5. But besides these, the terminations of the noun in the construct state sometimes assume a special form. Thus:
[89c] (a) In the construct state, plural and dual, the termination is ־ֵי, e.g. סוּסִים horses, סוּסֵי פַרְעֹח the horses of Pharaoh; עֵינַ֫יִם eyes, עֵינֵי הַפֶּ֫לֶךְ the eyes of the king.
[89d] Rem. The ־ֵי of the dual has evidently arisen from ־ַי (cf. יָ֫דַיִם), but the origin of the termination ־ֵי in the constr. st. plur. is disputed. The Syriac constr. st. in ay and the form of the plural noun before suffixes (סוּסַי, סוּסַ֫יִךְ, &c., §91h) would point to a contraction of an original ־ַי, as in the dual. But whether this ay was only transferred from the dual to the plural (so Olshausen, and Nöldeke, Beitr. zur sem. Sprachwiss., Strassb. 1904, p. 48 ff.), or is to be regarded as the abstract, collective termination, as in אִשֶּׁה (see f) and חוֹרַי (so Philippi, ThLZ. 1890, col. 419; Barth, ZDMG. 1904, p. 431 ff.), must be left undecided.
[89e] (b) The original ־ַת is regularly retained as the feminine termination in the construct state sing. of those nouns which in the absolute state end in ־ָה, e.g. מַלְכָּה queen, מַלְכַּת שְׁבָא the queen of Sheba. But the feminine endings ־֫ ־ֶת, ־֫ ־ַת, and also the plural ־וֹת, remain unchanged in the construct state.
[89f] (c) Nouns in ־ֶה (cf. §75e) from verbs ל״ה (§ 93, Paradigm III c) form their constr. st. in ־ֵה, e.g. רֹאֶה seer, constr. רֹאֵה. If this ־ֵה is due to contraction of the original ־ַי, with ה added as a vowel letter, we may compare דַּי, constr. דֵּי sufficiency; חַי, constr. חֵי life; (גַּי) גַּיְא, constr. (גֵּי) גֵּיא valley.
On the terminations וֹ and ־ִי in the constr. st. see § 90.
[90a] 1. As the Assyrian and old Arabic distinguish three cases by special endings, so also in the Hebrew noun there are three endings which, in the main, correspond to those of the Arabic. It is, however, a question whether they are all to be regarded as real remnants of former case-endings, or are in some instances to be explained other-