Gesenius' Hebrew Grammar/68. Verbs פ״א
The Weakest Verbs (Verba Quiescentia).
a So far as א retains its full consonantal value as a guttural, these verbs share all the peculiarities of verbs primae gutturalis, mentioned in § 63. They are, however, to be treated as weak verbs, when the א loses its value as a consonant, and coalesces with the preceding vowel (originally short) to form one long syllable. This takes place only in the following very common verbs and forms, as if through phonetic decay:—
b 1. In the imperfect Qal, five verbs (viz. אָבַד to perish, אָבָה to be willing, אָכַל to eat, אָמַר to say, אָפָה to bake) regularly make the א quiesce in a long ô, e.g. יֹאכַל. In a few others the ordinary (strong) form is also in use, as יֹאחֵז (18 times) and יֶֽאֱחֹז (3 times) he takes hold; יׄסֵף (see h), also יֶֽאֱסֹף, he collects. This ô has primarily arisen from an obscuring of ô (§9q), and the â from ־ַאְ, the weak consonant א coalescing with ă to â; cf. §23a.
c In the second syllable ō (for original ŭ) never appears, but either ē or ă; and in pause almost always ē, even before the tone-bearing heavy afformative וּן, e.g. יֽאֹכֵלוּן Dt 181, without the pause יֹֽאכְלוּן Dt 428. In the 3rd sing. masc. and 1st sing. of אָמַר, however, ă is always retained in pause, יֹאמַ֫ר and אֹמַ֫ר; but in the 2nd masc. תֹּאמֵ֑ר 1 K 520, in the 3rd fem. תֹּאמַֽר Pr 121; in the plural יֹאמֵ֑רוּ Jer 52, ψ 1456.11, תֹּאמֵ֫רוּ Jer 2338, with Segolta; cf. also תּאֹכַֽל 1 S 17, &c. But with conjunctive accents in the body of the sentence, ă (as being a lighter vowel) is used, e.g. תֹּאבַ֖ד לָעַֽד ψ 919, but in pause תּאֹבֵֽד ψ 16; cf. a similar interchange of ē and ă in §65c. The 3rd fem. plur. impf. always has the form תֹּאכַ֫לְנָה Zc 119.
d When the tone moves back, the final syllable of the imperfects of אָבַד and אָכַל, with a conjunctive accent, also always takes Pathaḥ, e.g. יֹ֣אבַד יוֹם Jb 33, וַיּ֫אֹכַל and he did eat; in אָמַר the loss of the tone from the final syllable only occurs in the form with wāw consecutive (but never in the 1st sing. וָֽאֹמַר; cf. וָֽאֹכַל), and then the final syllable, if without the pause, always takes Seghôl, וַ֫יֹּאמֶר and he said (except וַתֹּ֫אמַר לוֹ Pr 713).
e In pause, however, the imperfect consecutive (except the 1st pers. of אָכַל, see below) always has the form וַיֹּאכַ֫ל (but plur. always יֹאכֵ֫לוּ, וַיֹּאכֵ֫לוּ), וַיֹּאמַ֫ר; except וַיֹּ֫אמַר in the poetic portion of the book of Job, as 3:2, 4:1, &c., but not in 32:6, in the middle of the verse. The weak imperfect of אָחַז is always יֹאחֵז and וַיֹּאחֶז, but in the 1st sing., according to §49e, וָֽאֹחֵ֫ז Ju 206; cf. וָֽאֹכֵ֫ל Gn. 312.13 in pause.—אָבָה and אָפָה are, at the same time, verbs ל״ה, hence imperfect יֹּאבֶה (§75c).
f Before light suffixes the vowel of the second syllable becomes vocal Šewâ, as יֹֽאכְלֵם, תֹּֽאכְלֶ֫נּוּ but תֹּֽאכַלְכֶם.—In a few cases, instead of the ô in the first syllable an ê is found, which is due to contraction from the group ־ֶֽ ־ֱ (or ־ֶ ־ְ) in place of ־ַ ־ְ; e.g. תֵּאתֶה it shall come, Mi 48, from תֶּֽאֱתֶה (from אָתָה); אֵהָ֑ב (for אֵהַב) I love, Pr 817, also (four times) אֹהַב Mal 12, &c., with suffixes אֹֽהֲבֵ֫הוּ Ho 111, 145, &c. (but only in 1st sing., otherwise יֶֽאֱהַב, &c., from אָהֵב, אָהַב); וָֽאֵחַ֫ר and I stayed, Gn 325. The infinitive construct of אָמַר with לְ is always לֵאמֹר dicendo, for לֶֽאֱמֹר.—According to Barth (ZDMG. 1889, p. 179) וַיָּ֫אצֶל Nu 1125 is to be regarded as an imperfect Qal, without the obscuring of ־ָא to ô, not as imperfect Hiphʿîl, since אצל elsewhere occurs only in the perfect Qal and Niphʿal; on the original i in the second syllable, see above, §67p. For תְּאָכְ֫לֵהוּ Jb 2026 we should simply emend תֹּֽאכְל׳; the view that it is imperfect Pôʿēl (which nowhere else occurs) can, as regards the change of ô to ŏ, be supported only by the very doubtful analogies of ψ 624 (see §52q) and ψ 1015 Qerê (see §55b), while the view that it is Piʿēl (תְּאָכְ׳=תְּאָֽכְ׳=תְּאַכְּ׳) rests on no analogy whatever. It would be more admissible to suppose that תְּאָכְ׳ stands for תְּאֻכְּ׳, Puʿal (cf. אֲכֶלְךָ for אֲכַלְּךָ, §27q); but no reason has been discovered for this departure from the natural punctuation תֹּאכְ׳.
g 2. In the 1st pers. sing. imperfect, where two א’s would ordinarily come together, the second (which is radical) is regularly dropped (§23f), as אֹמַר (for אֹאמַר), &c., and even plene וָֽאוֹמַר Neh 27, &c., אֽוֹמְרָה ψ 4210. In the other cases, also, where the א is ordinarily regarded as quiescing in ô or ê, it is only retained orthographically, and on etymological grounds. Hence the possibility of its being dropped in the following cases:—
h Always in the contracted forms of אָסַף, as תֹּסֵף for תֹּאסֵף ψ 10429; וַיֹּ֫סֶף 2 S 61 (but for יֵאָֽסֵףJb 2719 read יֹאסִף=יוֹסִף with the LXX); cf. also in the 1st pers. Mi 46 and אֹסִֽפְךָ 1 S 156, which is apparently (from the Metheg with the i), intended for an imperfect Hiphʿîl: instead of it, however, read, with the Mantua edition, אֹֽסִפְךָ (with ĭ, according to §60f). But תֹּֽאסִפוּן Ex 57 (for תּֽוֹס׳), וַיֹּ֫אסֶף 1 S 1829 (for וַיּ֫וֹסֶף), and יאסף Jb 2719 (see above) are due to a mistake, since all three forms must be derived from the stem יָסַף. Furthermore, יֽׄמְר֫וּךָ ψ 13920 (where certainly יַמְר׳ is to be read); תֹּבֵא Pr 110 (cf. §75hh); וַתֹּפֵ֫הוּ 1 S 2824; יוֹכְלוּ Ez 425; תֹּֽמְרוּ 2 S 1914; וַתֹּ֫חֶז 2 S 209; תֵּֽוְלִי thou gaddest about (from אָזַל), Jer 236; וַיֵּ֫תֵא Dt 3321 (for יֶֽאֱתֶה), according to other readings (on the analogy of the cases mentioned in §75p) וַיֵתֵ֫א, וַיֵּ֫תֶא or וַיֶּ֫תֶא.
Paradigm I shows the weak forms of the imperfect Qal, and merely indicates the other conjugations, which are regular.
i Rem. 1. In the derived conjugations only isolated weak forms occur: Perfect Niphal נֹֽאחֲזוּ Nu 3230, Jos 229; Hiph. וַיָּ֫אצֶל Nu 1125 (but the statement in verse 17 is וְאָֽצַלְתִּי, therefore Qal); equally doubtful is the punctuation of וַיָּ֫רֶב (for וַיַּֽאֲרֵב?) and he laid wait, 1 S 155, and אָזִין I listen, Jb 3211 (on the analogy of verbs ע״וּ); cf. also אוֹכִיל (ô from â) I give to eat, Hos 114; אֹבִ֫ידָה (ô from â) I will destroy, Jer 468; וַיּוֹחֶר 2 S 205 Qerê (for וַיָּאח׳); the Kethîbh appears to require the Piʿēl וַיְיַחֵר, from יחר as a secondary form of אחר; but וַיֵּיחַר=וַיֵּאחַר for וַיֶּֽאֱחַר as imperfect Qal is not impossible. On וָאֽוֹצְרָה Neh 1313, cf. §53n.—Infinitive לְהָכִיל Ez 2133 (=לְהַֽאֲכ׳ unless it is rather infin. Hiph. from כּוּל); Participle מֵזִין giveth ear, Pr 174 (clearly by false analogy of verbs ע״וּ, for מַֽאֲזִין); Imperative הֵתָ֫יוּ bring (from אָתָה) Jer 129. (On the same form used for the perfect in Is 2114, cf. §76d.)
k 2. In the Piʿēl the א is sometimes elided (like ה in יְהַקְטִיל, יַקְטִיל), thus מַלֵּף (as in Aramaic and Samaritan) teaching, for מְאַלֵּף Jb 3511; יַהֵל (if not a mere scribal error) for יְאַהֵל Is 1320; וַתַּזְרֵ֫נִי thou hast girded me, 2 S 2240, for וַתְּאַזְּרֵ֫נִי as ψ 1840; וָֽאַבֶּדְּךָ Ez 2816; cf. §23c.
- So in the modern vulgar Arabic of South Palestine, yaʾkul (he eats) becomes yôkul.
- On this ē (originally ĭ) as a dissimilation from ō (originally ŭ), cf. §27w, and F. Philippi, in the Zeitschrift für Völkerpsychologie und Sprachwissenschaft, xiv. 178. The latter rightly observes that the existence of an original u in the imperfect of אָכַל is indicated by the form of the imperative אֱכֹל, the Arabic yaʾkul and the Aramaic יֵאכֻל, as well as by the fact that יֶֽאֱחֹז and יֶֽאֱסֹף are found along with יֹאחֵז and יֹאסֵף.
- The regularity of this orthography indicates that the contraction of אַאְ to â in this 1st pers. occurred at a time when in the 3rd and 2nd persons the א was still audible as a consonant (which accordingly was almost always retained in writing). Nöldeke (ZDMG. xxxii. 593) infers this from the fact that also in Arabic the 3rd and 2nd persons are still written yăʾkŭlŭ, tăʾkŭlŭ, but the 1st pers. ʾâkūlŭ, not ʾăʾkŭlŭ.