Gesenius' Hebrew Grammar/90. Real and Supposed Remains of Early Case-endings
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 otherwise. It can hardly be doubted (but cf. h, Rem.) that the (locative) termination ־ָה is a survival of the old accusative termination a, and that וּ in certain compound proper names is the old sign of the nominative. The explanation of the î as an old genitive sign, which, as being no longer understood in Hebrew, was used for quite different purposes, and the view that וֹ is a form of the nominative termination וּ, are open to grave doubts.
90b In Assyrian the rule is that u marks the nominative, i the genitive, and a the accusative, ‘in spite of the many and various exceptions to this rule which occur’ (Delitzsch, Assyrische Gramm., § 66). Similarly, the Arabic case-endings in the fully declined nouns (Triptotes) are: -u for the nominative, -i for the genitive, and -a for the accusative; in the Diptotes the ending -a represents the genitive also. In modern Arabic these endings have almost entirely disappeared, and if they are now and then used, as among the Beduin, it is done without regularity, and one is interchanged with another (Wallin, in ZDMG. v, p. 9, xii, p. 874; Wetzstein, ibid., xxii, p. 113 f., and especially Spitta, Gramm. des Arab. Vulgärdialekts von Ägypten, Lpz. 1880, p. 147 ff.). Even as early as the Sinaitic inscriptions, their regular use is not maintained (Beer,Studia Asiatica, iii. 1840, p. xviii; Tuch, ZDMG. iii. 139 f.). Ethiopic has preserved only the -a (in proper names -hâ), which is, however, still used for the whole range of the accusative, and also (the distinction of case being lost) as a termination of the constr. st. to connect it with a following genitive.
90c 2. As remarked above, under a, the accusative form is preserved in Hebrew most certainly and clearly in the (usually toneless) ending ־ָה, originally ă, as in the old Arabic accusative. This is appended to the substantive:
(a) Most commonly to express direction towards an object, or motion to a placer, e.g. יָ֫מָּה seaward, westward, קֵ֫דְמָה eastward, צָפ֫וֹנָה northward, אַשּׁ֫וּרָה to Assyria, בָּבֶ֫לָה to Babylon, חֶ֫רָה (from הַר) to the mountain, Gn 1410, אַ֫רְצָה to the earth, בַּ֫יְתָה to the house, תִּרְצָ֫תָה to Tirzah (תִּרְצָה) 1 K 1417, &c., עַזָּ֫תָה to Gaza (עַזָּה) Ju 161; with the article הָהָ֫רָה to the mountain, הַבַּ֫יְתָה into the house, הַחַ֫דְרָה into the chamber, 1 K 115; הָאֹ֫הֱלָה into the tent, Gn 186, &c.; similarly with adverbs, as שָׁ֫מָּה thither, אָ֫נָת whither?; even with the constr. st. before a genitive בֵּ֫יתָה יוֹסֵף into Joseph’s house, Gn 4317.24; אַ֫רְצָה הַנֶּ֫גֶב toward the land of the south, Gn 201; אַ֫רְצָה מִצְרַ֫יִם to the land of Egypt, Ex 420; מִדְבַּ֫רָה דַמֶּ֫שֶׂק to the wilderness of Damascus, 1 K 1915; מִזְרְחָ֫ה שֶׁ֫מֶשׁ toward the sun-rising, Dt 441; and even with the plural כַּשְׂדִּימָה into Chaldea, Ez 1124; הָשָּׁמַ֫יִמָה towards the heavens. Rem. The above examples are mostly rendered definite by the article, or by a following genitive of definition, or are proper names. But cases like יָ֫מָּה, הֶ֫רָה, בַּ֫יְתָה show that the locative form of itself possessed a defining power.
90d (b) In a somewhat weakened sense, indicating the place where something is or happens (cf. §118d), e.g. מַֽחֲנָ֑יְמָה in Maḥanaim, 1 K 414; שָׁ֫מָּה there (usually thither, see c), Jer 182, cf. 2 K 238, and the expression to offer a sacrifice הַמִּזְבֵּ֫חָה, properly towards the altar for on the altar. On the other hand, בָּבֶ֫לָה Jer 2915, and זְבֻ֫לָה Hb 311, are to be regarded as ordinary accusatives of direction, to Babylon, into the habitation; also expressions like פְּאַת צָפ֫וֹנָה the quarter towards the north, Jos 155 (at the beginning of the verse, גְּבוּל קֵ֫דְמָה the border toward the east), cf. 1815.20, Ex 2618, Jer 238.
90e (c) The original force of the ending ־ָה is also disregarded when it is added to a substantive with a preposition prefixed (cf. also עַד־אָ֫נָה how long?), and this not only after לְ, אֶל־ or עַד־ (which are easily explained), e.g. לְמַ֫עְלָה upwards, לְמַ֫טָּה downwards, לִשְׁא֫וֹלָה to Sheol, ψ 918; עַד־אֲפֵ֫קָה unto Aphek, Jos 134, אֶל־הַצָּפ֫וֹנָה toward the north, Ez 814, cf. Ju 2016; but also after ב, and even after מִן, e.g. בַּנֶּ֫גְבָּה in the south, Jos 1521, cf. Ju 142, 1 S 2315.19, 3113, 2 S 2015, Jer 5210; מִבָּבֶ֫לָה from Babylon, Jer 2716; cf. 113, Jos 1036, 1510, Ju 2119, Is 456.
90f Rem. Old locative forms (or original accusatives) are, according to the Masora, still to be found in
(a) לַ֫יְלָה, in pause לָ֫יְלָה, the usual word in prose for night, which is always construed as masculine. The nominative of this supposed old accusative appeared to be preserved in the form לַ֫יִל, only used in poetry, Is 163, constr. st. לֵיל (even used for the absol. st. in pause Is 2111). Most probably, however, לַיְלָה is to be referred, with Nöldeke and others, to a reduplicated form לילי; cf. especially the western Aramaic לֵילְיָא, Syr. lilya, &c.—Another instance is מְא֫וּמָה something, probably from מְאוּם, מוּם spot, point, generally with a negative=nothing. Similarly אַ֫רְצָה Is 823 and (in pause) Jb 3413, סוּפָ֫תָה Ho 87, and the place-name יַ֫הְצָה 1 Ch 663, might be explained as accusatives. Elsewhere, however, the toneless ־ָה can be regarded only as a meaningless appendage, or at the most as expressing poetic emphasis; thus אָ֫רְצָה (in pause) Jb 3712; הַמָּ֫וְתָה death, ψ 11615; נֶגְדָּה־נָּא ψ 11614.18; נַ֫חְלָה stream, ψ 1244; הַחַשְׁמַ֫לָה amber, Ez 82 [in 14 הַחַשְׁמַל, cf. §80k], &c. In Jos 1512 הַיָּ֫מָּה is probably only a scribal error (dittography). In Ju 1418 instead of the quite unsuitable poetic word הַחַ֫רְסָה (towards the sun??) read as in 151 הַחַ֫דְרָה to the bride-chamber. 90g (b) In the termination ־ָ֫ תָה often used in poetry with feminines, viz. אֵימָ֫תָה terror (=אֵימָה), Ex 1516; עֶזְרָ֫תָה help (=עֶזְרָה), ψ 4427, 638, 9417; יְשׁוּעָ֫תָה salvation (=יְשׁוּעָה), ψ 33, 803, Jon 210; עַוְ֫לָתָה unrighteousness (=עַוְלָה), Ez 2815, Ho 1013, ψ 1253; עֹלָ֫תָה ψ 9216 Keth. Jb 516; צָרָ֫תָה ψ 1201; עֵיפָ֫תָה darkness, Jb 1022; הַמְזִמָּ֫תָה Jer 1115 is corrupt, see the LXX and Commentaries. These cases are not to be taken as double feminine endings, since the loss of the tone on the final syllable could then hardly be explained, but they are further instances of an old accusative of direction or intention. In examples like עֶזְרָ֫תָה for help (ψ 4427) this is still quite apparent, but elsewhere it has become meaningless and is used merely for the sake of poetical emphasis.
90h This termination ־ָה usually has reference to place (hence called ־ָה locale); sometimes, however, its use is extended to time, as in מִיָּמִים יָמִ֫ימָה from year to year. Its use in חָלִ֫ילָה properly ad profanum!=absit! is peculiar.
90i As the termination ־ָה is almost always toneless (except in מִזְרְחָה constr. st. Dt 441; גִּתָּה and עִתָּה Jos 1913) it generally, as the above examples show, exercises no influence whatever upon the vowels of the word; in the constr. st. מִדְבַּ֫רָה Jos 1812, 1 K 1915, and in the proper names גַּ֫תָה 1 K 240, דַּ֫נָה 2 S 246 (so Baer; ed. Mant. and Ginsb. דַּ֫נָה), צְפַ֫תָה 2 Ch 149, צָֽרְפַ֫תָה 1 K 179, צָֽרְתַ֫נָה 1 K 412, an ă is retained even in an open tone-syllable (cf., however, הֶ֫רָה Gn 1410, פַּדֶּ֫נָה Gn 282 from פַּדַּן, with modification of the a to è; also בַּרְמֶ֫לָה 1 S 255 from בַּרְמֶל). In segholate forms, as a general rule, the ־ָה local is joined to the already developed form of the absol. st., except that the helping-vowel before ־ָה naturally becomes Šewâ, e.g. בַּ֫יְתָה, הָאֹ֫הֱלָה Gn 186, &c.; הַיַּֽ֫עֲרָה Jos 1715, הַשַֹּֽׁ֫עֲרָה Ju 2016, &c., but also נַ֫חְלָה Nu 345 (constr. st.; likewise to be read in the absolute in Ez 4719, 4828) and שָֽׁעְרָה Is 286 (with Silluq); cf. נֶ֫גְבָּה Ez 4719 and גֹּ֫רְנָה (Baer, incorrectly, גֹּֽרְנָ֫ה) Mi 412 (both in pause).—In the case of feminines ending in ־ָה the ־ָה local is added to the original feminine ending ־ָת (§80b), the ă of which (since it then stands in an open tone-syllable) is lengthened to ā, e.g. תִּרְצָ֫תָה.—Moreover the termination ־ָה is even weakened to ־ֶה in נֹ֫בֶה to Nob, 1 S 212, 229; אָ֫נֶה whither, 1 K 236.42 and דְּדָ֫נֶה to Dedan, Ez 2513.
90k 3. Of the three other terminations וּ may still be regarded as a survival of the old nominative ending. It occurs only in the middle of a few (often undoubtedly very old) proper names, viz. אֲחוּמַי (if compounded of אחו and מי), חֲמוּטַל (for which in Jer 521 Keth. חֲמִיטַל), מְתֽוּשָׁאֵל and מְתוּשֶׁ֫לַח (otherwise in Hebrew only in the plur. מְתִים men; to מְתוּ corresponds most probably בְּתוּ in בְּתוּאֵל), פְּנוּאֵל Gn 3231 (but in ver. 32 פְּנִיאֵל) face of God (otherwise only in the plur. פָּנִים constr. st. פְּנֵי).—גַּשְׁמוּ Neh 66 (elsewhere גֶּ֫שֶׁם), is the name of an Arab, cf. 61. On the other hand the terminations ־ִי and וֹ are most probably to be regarded (with Barth, l. c., p. 597) as having originated on Hebrew soil in order to emphasize the constr. st., on the analogy of the constr. st. of terms expressing relationship.
In view of the analogies in other languages (see b) there is nothing impossible in the view formerly taken here that the litterae compaginis ־ִי and וֹ are obsolete (and hence no longer understood) case-endings, î being the old genitive and ô for the nominative sign u. Barth objects that the î and ô almost invariably have the tone, whereas the accusative ־ָה is toneless, and that they are long, where the Arab. ĭ and ŭ are short. Both these objections, however, lose their force if we consider the special laws of the tone and syllable in Hebrew. The language does not admit a final ĭ or ŭ, and the necessarily lengthened vowel might easily attract the tone to itself. On the other hand a strong argument for Barth’s theory is the fact that these litterae compaginis are almost exclusively used to emphasize the close connexion of one noun with another; hence especially in the constr. st. Consequently it seems in the highest degree probable that all these uses are based upon forms in which the constr. st. is expressly emphasized by a special termination, i.e. the constr. st. of terms of relationship, אֲבִי, אֲחִי, חֲמִי from אָב father, אָח brother, חָם father-in-law (cf. § 96). The instances given under l and m followed this analogy.
Like î, וֹ is also used only to emphasize the constr. st. (see o), and must therefore have a similar origin, but its exact explanation is difficult. According to Barth, this וֹ corresponds to a primitive Semitic â (cf. §9q) and is traceable to ʾabâ, ʾaḥâ, the accusatives of terms of relationship in the constr. st., which have â only before a genitive. Against this explanation it may be objected that there is no trace of the supposed Hebrew accusatives אֲבוֹ, אֲחוֹ, חֲמוֹ, and only of the analogous בְּנוֹ. It is also remarkable that so archaic a form should have been preserved (except in בְּנוֹ) only in two words and those in quite late passages. However we have no better explanation to offer in place of Barth’s.
Finally we cannot deny the possibility, in some cases, of Barth’s explanation of the וּ in compound proper names like בְתוּאֵל, &c. (see above), as due to the analogy of terms of relationship with nominative in וּ. But this in no way militates against the view expressed above, that in some very old names, like פְּנוּאֵל, בְּתוּאֵל, &c., the original common nominative sign has simply been preserved. The instances found are:
90l (a) Of the ending ־ִי: בְּנִי אֲתֹנוֹ his ass’s colt, Gn 4911; עֹֽזְבִי הַצֹּאן that leaveth the flock, Zc 1117 (cf. the preceding רֹעִי הָֽאֱלִיל); שֹֽׁכְנִי סְנֶה the dweller in the bush, Dt 3316 (on שֹֽׁבְנִי cf. below Jer 4916a, Ob 313); appended to the feminine גְּנֻבְֽתִי יוֹם וּגְּנֻבְֽתִי לְ֫יְלְה whether stolen by day or stolen by night, Gn 3139 (in prose, but in very emphatic speech); מְלֵֽאֲתִי מִשְׁפָּט , Is 121; רַבָּ֫תִי עָם full of people, La 11 (on the retraction of the tone before a following tone-syllable, cf. §29e; in the same verse the second רבתי and שָׂרָ֫תִי, see below, follow the example of רַבָּ֫תִי, although no tone-syllable follows; cf. also Ho 1011 below); עַל־דִּבְרָתִי מַלְכִּי־צֶ֫דֶק after the order of Melchizedek, ψ 1104; cf. also ψ 1139, Jer 4916b. To the same category belong the rather numerous cases, in which a preposition is inserted between the construct state and its genitive (cf. §130a), without actually abolishing the dependent relation, e.g. רַבָּ֫תִי בַגּוֹיִם she that was great among the nations, שָׂרָ֫תִי בַמְּדִינוֹת princess among the provinces, La 11; אֹהַ֫בְתִּי לָדוּשׁ that loveth to tread, Ho 1011; cf. also Jer 4916a, Ob 313.—In Ex 156 נֶאְדָּרִי can only be so explained if it is a vocative referring to יהוה, but perhaps we should read נֶאְדָּרָה as predicate to יְמִֽינְךָ.
Further, the Ḥireq compaginis is found with certain particles which are really also nouns in the constr. st., as זֽוּלָתִי (=זוּלָת) except, מִנִּי (poetical for מִן) from, בִּלְתִּי not, אַפְסִי not (thrice in the formula אֲנִי וְאַפְסִי עוֹד I am, and there is none else beside me; but many take the ־ִי as a suffix here), Is 478.10, Zp 215. [The above are all the cases in which this ־ִי is attached to independent words in the O.T.; it occurs, however, besides] in compound proper names (again attached to the constr. st.), as מַלְכִּי־צֶ֫דֶק (king of righteousness), גַּבְרִיאֵל (man of God), חַנִּיאֵל (favour of God), and others (cf. also the Punic name Hannibal, i.e. חַנִּיבַ֫עַל favour of Baʿal).
90m Otherwise than in the constr. st. the Ḥireg compaginis is only found in participial forms, evidently with the object of giving them more dignity, just as in the case of the construct forms in î. We must distinguish, however, between passages in which the participle nevertheless does stand in close connexion, as Gn 4911, Is 2216 (חֹֽצְבִי and חֹֽקְקִי, also in impassioned speech), Mi 714 (probably influenced by Dt 3316), ψ 1015, 1137; and passages in which the î added to the participle with the article merely serves as an ornamental device of poetic style, e.g. in the late Psalms, 1126.96.36.199 (on verse 8 see n), 1148, 1231.
90n In Kethibh the termination î also occurs four times in יושבתי, i.e. יוֹשַׁבְתִּי, Jer 1017, 2223 (before בְּ), Ez 273 (before עַל־), La 421 (before בְּ). The Qere always requires for it יוֹשֶׁ֫בֶת (or ישׁ׳), except in Jer 2223 ישַׁבְתְּ; cf. ibid. מקננתי Keth., מְקֻנַּנְתְּ Qere, and finally Jer 5113 שׁכנתי Keth., שֹׁכַנְתְּ Qere. Perhaps ישַׁבְתִּי and שֹׁכַנְתִּי are formae mixtae, combining the readings ישֶׁ֫בֶת, &c. and יָשַׁבְתְּ (2nd fem. perf.), &c., but מְקֻנַּנְתִּי may be merely assimilated to ישַׁבְתִּי which immediately precedes it.
The following are simply textual errors: 2 K 423 ההלכתי Keth., due to the preceding אתי, and to be read הַֽהֹלֶ֫כֶת as in the Qere; ψ 308 (read הַֽרֲרֵי), 1138 (read לְהֽוֹשִׁיבוֹ), 1161 (read קוֹל תח׳, as in five other places). On בְּרִיתִי, thrice, in Lv 2642, cf. §128d.
90o (b) Of the ending וֹ (always with the tone): in prose only in the Pentateuch, but in elevated style, Gn 124 חַיְתוֹ־אֶרֶץ the beast of the earth (=חַיַּת הָאָרֶץ ver. 25); similarly in ψ 5010, 792, 10411.20, Is 569 (twice), Zp 214; otherwise only in בְּנוֹ צִפֹּר son of Zippor, Nu 2318; בְּנוֹ בְעֹר son of Beor, Nu 243.15; and מַעְיְנוֹ מַ֫יִם a fountain of waters, ψ 1148.
- This rule is almost always observed in the Tell-el-Amarna letters (see §2f); cf. the instances cited by Barth, l. c., p. 595, from Winckler’s edition.
- On this meaning of the accusative see the Syntax, §118d, and cf. the Latin accusative of motion to a place, as in Romam profectus est, domum reverti, rus ire.
- הָאֹֽהֱלָ֫ה in Baer’s text, Gn 186, is an error, according to his preface to Isaiah, p. v.
- Brockelmann, Sem. Sprachwiss., p. 113, also takes it as such, láylā being properly at night, then night simply. Barth, however (Sprachwiss. Abhandlungen, p. 16, note 1), refers it to an original לַיְלֶה, like אָנָה from אָנֶ֫ה.
- The form clings also to a few place-names, as גֻּדְגֹּ֫דָה Dt 107; שָׁלִ֫שָׁה 1 S 94, 2 K 442; קְהֵלָ֫תָה Nu 3322 f.; יָטְבָּ֫תָה verse 33 f.; תִּמְנָ֫תָה Jos 1943, &c.; אֶפְרָ֫תָה Mi 51, &c.]
- Cf. Sarauw, ‘Der hebr. Lokativ,’ ZA. 1907, p. 183 ff. He derives the ־ָה from the adverbs שָׁ֫מָּה, אָ֫נָה and holds that it has nothing whatever to do with the old accusative.
- So Qimḥi, and the Mant. ed. (Baer הַשַֹּׁ֫עְרָה), i.e. locative from שַׂ֫עַר (Is 720). The reading הַשַּֽׂעֲרָה (Opit., Ginsb.) implies a feminine in ־ָה.
- Cf. the list in L. Kaila, Zur Syntax des in verbaler Abhängigkeit stehenden Nomens im alttest. Hebr., Helsingfors, 1906, p. 54.
- The name שְׁמוּאֵל formerly regarded as a compound of שְׁמוּ=שֵׁם name and אֵל, is better explained with Prätorius, ZDMG. 1903, p. 777, as a name of affection, for שְׁמוּעַ אֵל=יִשְׁמָעֵאל [but see Driver on 1 S 120]; similarly, according to Prätorius, פְּתוּאֵל=פְּתוּחַ אֵל and many others.
- Cf. Kaila, l. c., p. 59 ff.