Gesenius' Hebrew Grammar/27. The Change of the Vowels, especially as regards Quantity
27a The changes in sound through which the Hebrew language passed, before it assumed the form in which we know it from the Masoretic text of the O.T. (see §2k), have especially affected its vowel system. A precise knowledge of these vowel changes, which is indispensable for the understanding of most of the present forms of the language, is derived partly from the phenomena which the language itself presents in the laws of derivation and inflexion, partly from the comparison of the kindred dialects, principally the Arabic. By these two methods, we arrive at the following facts as regards Hebrew:
27b 1. That in an open syllable the language has frequently retained only a half-vowel (Šewâ mobile), where there originally stood a full short vowel, e.g. עֲגָלָה (ground-form ʿăgălăt) a waggon, צְדָקָה (groundform ṣădăqăt) righteousness, קָֽטְלוּ (Arab. qătălŭ), יְקַטְּלוּ (Arab. jŭqattĭlŭ).
27c 2. That vowels originally short have in the tone-syllable, as also in the open syllable preceding it, been generally changed into the corresponding tone-long vowels, ă into ā, ĭ into ē, ŭ into ō (see §9, a–e, k, r). If, however, the tone be shifted or weakened, these tone-long vowels mostly revert to their original shortness, or, occasionally, are still further shortened, or reduced to mere Šewâ mobile, or, finally, are entirely lost through a change in the division of syllables; e.g. מָטָר (Arab. măṭăr) rain, when in close dependence on a following genitive in the construct state), becomes מְטַר; עָקֵב (Arab. ʿăqĭb) heel, dual עֲקֵבַ֫יִם, dual construct (with attenuation of the original ă of the first syllable to ĭ) עִקְּבֵי [on the קּ, see §20h]; יִקְטֹל (Arab. yăqtŭl), plur. יִקְטְלוּ (Arab. yăqtŭlû). For instances of complete loss, as in כַּסְפֵּי, cf. §93m. According to §26, the following details of vowel-change must be observed:
27d 1. The original, or a kindred short vowel reappears—
(a) When a closed syllable loses the tone (§26o). Thus, יָד hand, but יַד־יְהֹוָה the hand of Yahwe; בֵּן son, but בֶּן־הַמֶּ֫לֶךְ the son of the king; כֹּל the whole, but כָּל־הָעָם the whole of the people; so also when a tone-bearing closed syllable loses the tone on taking a suffix, e.g. אֹיֵב enemy, but אֹֽיִבְךָ thy enemy; finally, when the tone recedes, יָקֹם, but וַיָּ֫קָם (wayyāqŏm); יֵלֵךְ, but וַיֵּ֫לֶךְ.
(b) To the same category belong cases like סֵ֫פֶר book, but סִפְרִי my book; קֹ֫דֶשׁ holiness, but קָדְשִׁי my holiness. In spite of the helping vowel, סֵפֶר and קֹדֶשׁ are really closed syllables with a tone-long vowel; when the syllable loses the tone, the original ĭ or ŏ (properly ŭ) reappears.
The same is true of syllables with a virtually sharpened final consonant: the lengthening of original ĭ to ē and ŭ to ō takes place only in a tone-bearing syllable; in a toneless syllable the ĭ or ŏ (or ŭ) remains, e.g. אֵם mother, but אִמִּי my mother; חֹק law, plur. חֻקִּים; but עֹז strength, עָזִּי (and עֻזִּי) my strength.
27e 2. The lengthening of the short vowel to the corresponding long, takes place—
(a) When a closed syllable becomes open by its final consonant being transferred to a suffix beginning with a vowel, or in general to the following syllable, e.g. קָטַל, קְטָ|לוֹ he has killed him; סוּסָ|תִי primarily from סוּסַת. Similarly ă mostly becomes ā even before a suffix beginning with Šewâ mobile; e.g. קְטָֽלְךָ from קָטַל, סוּסָֽתְךָ from סוּסַת.
27g (c) When a weak consonant (א, ו, י) following the short vowel quiesces in this vowel, according to §23a, c, d, §24f, e.g. מָצָא for מָצַא, where the א, losing its consonantal value, loses also the power of closing the syllable, and the open syllable requires a long vowel.
27i 3. When a word increases at the end and the tone is consequently moved forward, or when, in the construct state (see §89), or otherwise in close connexion with the following word, its tone is weakened, in such cases a full vowel (short or tone-long) may, by a change in the division of syllables, be weakened to Šewâ mobile, or even be entirely lost, so that its place is taken by the mere syllable-divider (Šewâ quiescens). Examples of the first case are, שֵׁם name, pl. שֵׁטוֹת, but שְׁמִי my name, שְׁמוֹתָם their names, דָּבָר word, constr. st. דְּבַר; צְדָקָה righteousness, constr. st. צִדְקַת; an example of the second case is, בְּרָכָה blessing, constr. st. בִּרְכַּת. Whether the vowel is retained or becomes Šewâ (דָּם, דָּמִי, but שֵׁם, שְׁמִּי), and which of the two disappears in two consecutive syllables, depends upon the character of the form in question. In general the rule is that only those vowels which stand in an open syllable can become Šewâ.
Thus the change into Šewâ takes place in—
27k (a) The ā and ē of the first syllable, especially in the inflexion of nouns, e.g. דָּבָ֫ר word, plur. דְּבָרִ֫ים; גָּד֫וֹל, great, fem. גְּדוֹלָ֫ה; לֵבָ֫ב heart, לְבָבִ֫י my heart; but also in the verb, תָּשׁ֫וּב she will return, plur. תְּשׁוּבֶ֫ינָה, and so always, when the originally short vowel of the prefixes of the Imperfect comes to stand in an open syllable which is not pretonic. On the other hand, an ā lengthened from ă before the tone is retained in the Perfect consecutive of Qal even in the secondary tone, e.g. וְקָֽטַלְתָּ֫; cf. §49i.
27l (b) The short, or merely tone-long, vowels a, e, o of the ultima, especially in verbal forms, e.g. קָטַל, fem. קָֽטְלָה qāṭelā; יִקְטֹל, יִקְטְלוּ yiqṭelû; but note also יִלְקֹטוּן, תִּדְבָּקִין, &c., according to §47m and o. The helping vowels are either entirely omitted, e.g. מֶ֫לֶךְ king (ground-form malk), מַלְכִּי my king; or, under the influence of a guttural, are weakened to Ḥaṭeph, e.g. נַ֫עַר boy, נַעֲרוֹ his boy. If the tone remains unmoved, the vowel also is retained, notwithstanding the lengthening of the word, e.g. יִקְטֹ֫לוּ pausal-form for יִקְטְלוּ.
27m Where the tone moves forward two places, the former of the two vowels of a dissyllabic word may be shortened, and the second changed into Šewâ. Cf. דָּבָר word; in the plur. דְּבָרִ֫ים; with heavy suffix דִּבְרֵיהֶ֫ם (cf. §28a) their words. On the attenuation of the ă to ĭ, see further, s, t.
27n Rem. 1. An ô arising from aw=au, or by an obscuring of â (see §9b), sometimes becomes û, when the tone is moved forward, e.g. נָקוֹם, נְקוּמ֫וֹתָ (see Paradigm Perf. Niph. of קוּם); מָנוֹס flight, fem. מְנוּסָ֫ה, with suffix, מְנוּסִ֫י. The not uncommon use of וּ in a sharpened syllable, as בְּחֻוּקֵּי Ez 2018 (for בְּחֻקֵּי, cf. also the examples in §9o), is to be regarded as an orthographic licence, although sometimes in such cases û may really have been intended by the Kethîbh.
27o Of the vowels of the U-class, û and tone-long ō stand in a tone-bearing closed final syllable, and ŏ in a toneless syllable, e.g. יָקוּם he will arise, יָקֹם jussive, let him arise, וַיָּ֫קָם and he arose. The only instance of ŭ in an ultima which has lost the tone is וַיָּ֫רֻם Ex 1620 (see §67n). Similarly, of vowels of the I-class, ê, î, and ē stand in a tone-bearing closed final syllable, and ĕ in a toneless syllable, e.g. יָקֵים he will raise, יָקֵם let him raise, וַיָּ֫קֶם and he raised. The only instance of ĭ in an ultima which has lost the tone is וַתָּ֫רִץ Ju 953 (see §67p).
2. In the place of a Pathaḥ we not infrequently find (according to §9f) a Seghôl (ĕ, è) as a modification of ă:
27p (a) In a closed antepenultima, e.g. in the proper names אֶבְיָתָר and אֶבְיָסָף, where LXX Ἀβι- = אַבְי׳, which is certainly the better reading, cf. Ulmer, Die semit. Eigennamen, 1901, p. 12: or in a closed penultima, e.g. יֶהְדֹּף, but also יֶדְכֶם your hand, for yadekhèm. In all these cases the character of the surrounding consonants (see §6q) has no doubt had an influence.
27q (b) Regularly before a guttural with Qameṣ or Ḥaṭeph Qameṣ, where the strengthening has been dropped, provided that a lengthening of the Pathaḥ into Qameṣ be not necessary, e.g. אֶחָיו his brothers, for ’aḥāw; כֶּחָשׁ false, for kaḥāš; פֶּחָה governor, constr. st. פַּחַת; פֶּחָם coal; הֶחָי the living (with the article, הֶ for הַ); יִתְנֶחָם Nu 2319, &c., and so always before הָ and חֳ, as הֶֽחֳדָשִׁים the months, see §35k. Before הָ and עָ Seghôl generally stands only in the second syllable before the tone, e.g. הֶֽהָרִים the mountains; הֶֽעָוֹן the guilt; immediately before the tone Pathaḥ is lengthened into a (pretonic) Qameṣ, e.g. הָהָר, הָעָם; but cf. also הִטֶּהָ֫רוּ Nu 87. Before the weak consonants א and ר (cf. §22c, q), the lengthening of the Pathaḥ into Qameṣ almost always takes place, e.g. הָאָב the father, pl. הָֽאָבוֹת; הָרֹאשׁ the head, pl. הָֽרָאשִׁים. Exceptions, הֶ֫רָה towards the mountain, Gn 1410, in the tone-syllable, for hárrā; יְבֶֽרֶכְיָ֫הוּ (pr. name) for יְבָֽרֶכְיָהוּ. On הֶ as a form of the interrogative הֲ (הַ), see §100n; on מֶה for מָה (מַהּ), §37e, f. Finally, אֲכֶלְךָ֫ Ex 333 also comes partly under this head, in consequence of the loss of the strengthening, for אֲכַלְּךָ, and יְחֶזְקֵאל Ezekiel for יְחַזְּקֵאל = יְחַזֵּקְאֵל God strengthens.
27r (c) As a modification of the orIginal Pathaḥ in the first class of the segholate forms (§93g), when a helping vowel (§28e) is inserted after the second consonant. Thus the ground-form kalb (dog), after receiving a helping Seeghôl, is modified into כֶּ֫לֶב (also in modern Arabic pronounced kelb), yarḥ (month), with a helping Pathaḥ, יֶ֫רַח. The same phenomenon appears also in the formation of verbs, in cases like יֶ֫גֶל (jussive of the Hiphʿîl of גָּלָה), with a helping Seeghôl, for yagl.
27s 3. The attenuation of ă to ĭ is very common in a toneless closed syllable.
(a) In a firmly closed syllable, מִדּוֹ his measure, for מַדּוֹ (in a sharpened syllable); יְלִדְתִּ֫יךָ I hare begotten thee, from יָלַ֫דְתִּי with the suffix ךָ; cf. Lv 1144, Ez 3823, and §44d. Especially is this the case in a large number of segholates from the ground-form qaṭl, when combined with singular suffixes, e.g. צִדְקִי my rIghteousness, for ṣadqî.
27t (b) In a loosely-closed syllable, i.e. one followed by an aspirated Begadkephath, as דִּמְכֶם your blood, for דַּמְכֶם, and so commonly in the st. constr. plur. of segholates from the ground-form qaṭl, e.g. בִּגְדֵי from בֶּגֶד (ground-form bagd) a garment. In most cases of this kind the attenuation is easily intelligible from the nature of the surrounding consonants. It is evident from a comparison of the dialects, that the attenuation was consistently carried out in a very large number of noun and verb-forms in Hebrew, as will be shown in the proper places.
27u 4. Seghôl arises, in addition to the cases mentioned in o and p, also from the weakening of ā of the final syllable in the isolated cases (־ֶה for ־ָה) in 1 S 2815 (? see §48d), ψ 204 (?), Is 595, Pr 2414 (see §48l); for examples of Locative forms in ־ֶה see §90i end.
27v 5. Among the Ḥaṭaeph-sounds ־ֲ is shorter and lighter than ־ֱ, and consequently the vowel group ־ַֽ ־ֲ is shorter than ־ֶֽ ־ֱ; e.g. אֱדוֹם Edom, but אֲדֹמִ֫י (Edomite), shortened at the beginning because the tone is thrown forward; אֱמֶ֫ת (ʾemèth) truth, אֲמִתּ֫וֹ his truth; נֶֽעֱלָ֫ם hidden, pl. נַֽעֲלָמִ֫ים; הֶֽעֱבַ֫רְתִּי but וְהַֽעֲבַרְתִּ֫י; but also conversely נַֽעֲשָׂה fem. נֶעֱשְׂתָה cf. §63f, 3.
27w 6. To the chapter on vowel changes belongs lastly the dissimilation of vowels, i.e. the change of one vowel into another entirely heterogeneous, in order to prevent two similar, or closely related vowels, from following one another in the same word. Hence לוּלֵא for lû lô (unless). Cf. also חִיצוֹן from חוּץ; רִאשׁוֹן from רֹאשׁ; תִּיכוֹן from תּוֹךְ; נִכְחוֹ from נֹ֫כַח; עֵירֹם from stem עוּר; most probably also יִלּוֹד offspring, קִפּוֹד porcupine, for יֻלּ׳, קֻפּ׳, see §68c, note.—On the proper names יֵהוּא and יֵשׁוּעַ, which were formerly explained in the same way, see now Prätorius, ZDMG. 1905, p. 341 f.
- So the LXX write Μελχισεδέκ for מַלְכִּיצֶ֫דֶק.
- Analogous to this attenuation of ă to ĭ is the Lat. tango, attingo; laxus, prolixus; to the transition of ă to ĕ (see above, a), the Lat. carpo, decerpo; spargo, conspergo.
- Cf. Barth, Die Nominalbildung in den semit. Spr., p. xxix; A. Müller, Theol. Stud. u. Krit., 1892, p. 177 f., and Nestle, ibid., p. 573 f.