Gesenius' Hebrew Grammar/8. The Vowel Signs in particular
P. Haupt, ‘The names of the Hebrew vowels,’ JAOS. xxii, and in the Johns Hopkins Semitic Papers, Newhaven, 1901, p. 7 ff.; C. Levias in the Hebr. Union Coll. Annual, Cincinnati, 1904, p. 138 ff.
The next two sections (§§ 8 and 9) have been severely criticized (Philippi, ThLZ. 1897, no. 2) for assigning a definite quantity to each of the several vowels, whereas in reality ־ֶ, ־ֵ, ־ׄ are merely signs for ä, e, o: ‘whether these are long or short is not shown by the signs themselves but must be inferred from the rules for the pause which marks the breaks in continuous narrative, or from other circumstances.’ But in the twenty-fourth and subsequent German editions of this Grammar, in the last note on §8a [English ed. p. 38, note 4], it was stated: ‘it must be mentioned that the Masoretes are not concerned with any distinction between long and short vowels, or in general with any question of quantity. Their efforts are directed to fixing the received pronunciation as faithfully as possible, by means of writing. For a long time only שִבְעָה מְלָכִים seven kings were reckoned (vox memor. in Elias Levita וַיּאֹמֶר אֵלִיָּהוּ), Šureq and Qibbuṣ being counted as one vowel. The division of the vowels in respect of quantity is a later attempt at a scientific conception of the phonetic system, which was not invented but only represented by the Masoretes (Qimchi; Mikhlol, ed. Rittenb. 136 a, distinguishes the five long as mothers from their five daughters).’
I have therefore long shared the opinion that ‘the vowel-system represented by the ordinary punctuation (of Tiberias) was primarily intended to mark only differences of quality’ (Sievers, Metrische Studien, i. 17). There is, however, of course a further question how far these ‘later’ grammarians were mistaken in assigning a particular quantity to the vowels represented by particular signs. In Philippi’s opinion they were mistaken (excluding of course î, ê, ô when written plene) in a very great number of cases, since not only does ־ָ stand, according to circumstances, for ā̊ or ă̊, and ־ֶ for ā̈ or ă̈, but also ־ֵ for ē or ĕ, and ־ׄ for ō or ŏ, e.g. כָּבֵד and קָטֹן, out of pause kå̄bĕ́d, qå̄:ŏ́n (form קָטַל), but in pause kå̄bḗd, qå̄ṭṓn.
I readily admit, with regard to Qameṣ and Segol, that the account formerly given in §8f. was open to misconstruction. With regard to Ṣere and Ḥolem, however, I can only follow Philippi so long as his view does not conflict with the (to me inviolable) law of a long vowel in an open syllable before the tone and (except Pathaḥ) in a final syllable with the tone. To me כָּבֵד = kā̊bĕ́d, &c., is as impossible as e.g. עֵנָב = ʿĕnab or בֹּרַךְ = bŏrakh, in spite of the analogy cited by Sievers (p. 18, note 1) that ‘in old German e.g. original ĭ and ŭ often pass into ĕ and ŏ dialectically, while remaining in a closed syllable.
|A||1.||־ָ||Qāmĕṣ denotes either ā, â, more strictly å̄ (the obscure Swedish å) and å̂, as יָד yå̄d (hand), רָאשִׁים rā’šîm (heads), or å̆ (in future transcribed as ŏ), called Qāmeṣ ḥāṭûph, i.e. hurried Qameṣ. The latter occurs almost exclusively as a modification of ŭ; cf. c and §9u.|
|2.||־ַ||Páthăḥ, ă, בַּת băth (daughter).|
Also 3. ־ֶ Segôl, an open e, è (ǟ or ä̆), as a modification of ă, either in an untoned closed syllable, as in the first syllable of יֶדְכֶם yädkhèm (your hand) from yădkhèm—or in a tone-syllable as in פֶּ֫סַח pĕsaḥ; cf. πάσχα, and on the really monosyllabic character of such formations, see §28e. But Segôl in an open tone-syllable with a following י, as in גְּלֶ֫ינָה gelènā (cf. §75f), יָדֶ֫יךָ yādèkhā (cf. §91i), is due to contraction from ay.
|I||1. ־ִי Ḥîrĕq with yod, almost always î, as צַדִּיק ṣaddîq (righteous).
2. ־ִ either î (see below, i), as צַדִּקִים ṣaddîqîm, only orthographically different from (צדיקם)צדיקים,—or ĭ, as צִדְקוֹ ṣĭdqô (his righteousness).
|E||3. ־ֵי Ṣerî or Ṣerê with yod=ê, e.g. בֵּיתוֹ bêthô (his house).
־ֵ either ê, but rarely (see below, i), or ē as שֵׁם šēm (name).
Ṣere can only be ĕ, in my opinion, in few cases, such as those mentioned in §29f.
4. ־ֶ Segôl, ă̈, a modification of ĭ, e.g. חֶפְצִי ḥäfṣî (ground-form ḥĭfṣ); שֶׁן־ šän (ground-form šĭn).
|U||1. וּ Šûrĕq, usually û, מוּת mûth (to die), rarely ŭ.
2. ־ֻ Qibbûṣ, either ŭ, e.g. סֻלָּם sŭllām (ladder): or û, e.g. קֻ֫מוּ qūmū (rise up), instead of the usual ק֫וּמוּ.
|O||3. וֹ and ־ֹ Ḥōlĕm, ô and ō, קוֹל qôl (voice), רֹב rōbh (multitude). Often also a defective ־ֹ for ô; rarely וֹ for ō.
On the question whether ־ֹ under some circumstances represents ŏ, see §93r.
4. ־ָ On Qāmĕṣ ḥāṭūph=ŏ, generally modified from ŭ, as חָק־ ḥŏq (statute), see above, a.
8d The names of the vowels are mostly taken from the form and action of the mouth in producing the various sounds, as פַּתַ֫ח opening; צֵרֵ֫י a wide parting (of the mouth), also שֶׁ֫בֶר=ĭ) breaking, parting (cf. the Arab, kasr); חִ֫ירֶק (also חִרֶק) narrow opening; ח֫וֹלֶם closing, according to others fullness, i.e. of the mouth (also מְלֹא פּוּם fullness of the mouth). קָ֫מֶץ also denotes a slighter, as שׁוּרֶק and קִבּוּץ (also קבוץ פּוּם) a firmer, compression or contraction of the mouth. Segôl (סְגוֹל bunch of grapes) takes its name from its form. So שָׁלֹשׁ נְקֻדּוֹת (three points) is another name for Qibbúṣ.
8e Moreover the names were mostly so formed (but only later), that the sound of each vowel is heard in the first syllable (קָמֶץ for קֹמֶץ, פַּתַח for פֶּתַח, צֵרִי for צְרִי); in order to carry this out consistently some even write Sägôl, Qomeṣ-ḥatûf, Qûbbûṣ.
8f 2. As the above examples show, the vowel sign stands regularly under the consonant, after which it is to be pronounced, רָ rā, רַ ră, רֵ rē, רֻ rŭ, &c. The Pathaḥ called (§22f) alone forms an exception to this rule, being pronounced before the consonant, רוּחַ rŭaḥ (wind, spirit). The Ḥolem (without wāw) stands on the left above the consonant; רֹ rō (but לֹ lō). If א, as a vowel letter, follows a consonant which is to be pronounced with ō, the point is placed over its right arm, thus בֹא, רֹאשׁ; but e.g. בֹאָם, since א here begins a syllable.
8g No dot is used for the Ḥolem when ō (of course without wāw) is pronounced after šîn or before šîn. Hence שׂנֵא śônē (hating), נְשׂא neśō (to bear), משֶׁה môšè (not מֹשֶׁה); but שֹׁמֵר šômēr (a watchman). When ō precedes the śîn, the dot is placed over its right arm, e.g. יִרְפּשֹׁ yirpōś (he treads with the feet), הַנּֽשְֹׁאִים hannôśeʾîm (those who carry).
In the sign וֹ, the ו may also be a consonant. The וֹ is then either to be read ōw (necessarily so when a consonant otherwise without a vowel precedes, e.g. לֹוֶה lôwè, lending) or wō, when a vowel already precedes the ו, e.g. עָוֹן ʿāwôn (iniquity) for עָווֹן. In more exact printing, a distinction is at least made between וֹ wo and וֹ (i.e. either ô or, when another vowel follows the wāw, ôw).
8h 3. The vowels of the first class are, with the exception of ־ֶי in the middle and ־ָה, ־ָא, ־ֶה at the end of the word (§9a–d, f), represented only by vowel signs, but the long vowels of the I- and U-class largely by vowel letters. The vowel sound to which the letter points is determined more precisely by the vowel sign standing before, above, or within it. Thus—
י may be combined with Ṣērê, Qameṣ, Segôl (־ִי, ־ֵי, ־ֶי).
ו with Šûrĕq and Ḥōlĕm (וּ and וֹ).
In Arabic the long a also is regularly expressed by a vowel letter, viz. ʾAlĕph (־ָא), so that in that language three vowel letters correspond to the three vowel classes. In Hebrew א is rarely used as a vowel letter; see §9b and §23g.
8i 4. The omission of the vowel letters when writing î, û, ê, ô is called scriptio defectiva in contrast to scriptio plena. קוֹל, קוּם are written plene, קֹלֹת, קֻם defective.
Cf. Bardowitz, Studien zur Gesch. der Orthogr. im Althebr., 1894; Lidzbarski, Ephem., i. 182, 275; Marmorstein, ‘Midrasch der vollen u. defekt. Schreibung,’ in ZAW. 1907, p. 33 ff.
8k So far as the choice of the full or defective mode of writing is concerned, there are certainly some cases in which only the one or the other is admissible, Thus the full form is necessary at the end of the word, for û, ô, ō, î, ê, ē, as well as for è in חֹזֶה &c. (§9f), also generally with â, ā (cf. however §9d), e.g. קָֽטְלוּ, קָטַ֫לְתִּי, יָדִי, מַלְכֵי. (But the Masora requires in Jer 266, 448; Ezr 621; 2 Ch 3213 גּוֹיֵ instead of גּוֹיֵי; Zp 29 גּוֹיִ [perhaps an error due to the following י] for גּוֹיִי; Is 4031 וְקוֹיֵ [followed by י] for וְקוֹיֵי; Jer 3811 בְּלוֹיֵ for בְּלוֹיֵי.) On the other hand the defective writing is common when the letter, which would have to be employed as a vowel letter, immediately precedes as a strong consonant, e.g. גּוֹיִם (nations) for גּוֹיִים, מִצְוֹת (commandments) for מִצְווֹת.
8l That much is here arbitrary (see §7g), follows from the fact that sometimes the same word is written very differently, e.g. הֲקִימוֹתִי Ez 1660: הֲקִמֹתִי and also הֲקִמוֹתִי Jer 234; cf. §25b. Only it may be observed,
(a) That the scriptio plena in two successive syllables was generally avoided; cf. e.g. נָבִיא but נְבִאִים; צַדִּיק, but צַדִּקִים; קוֹל, קֹלוֹת; יְהוֹשֻׁעַ; מְצָאֻ֫הוּ.
(b) That in the later Books of the O.T. (and regularly in post-biblical Hebrew) the full form, in the earlier the defective, is more usual.
8m 5. In the cognate dialects, when a vowel precedes a vowel-letter which is not kindred (heterogeneous), e.g. ־ָו, ־ֵו, ־ִיו, ־ַי, ־ָי, a diphthong (au, ai) is formed if the heterogeneous vowel be a. This is also to be regarded as the Old Hebrew pronunciation, since it agrees with the vocalic character of ו and י (§5b, note 2). Thus such words as וָו, חַי, גּוֹי, עָשׂוּי, גֵּו, בַּ֫יִת are not to be pronounced according to the usual Jewish custom as vāv, ḥay, gôy, ʿāsûy, gēv, bayith (or even as vaf, &c.; cf. modern Greek av af, ev ef for αὐ, εὐ), but with the Italian Jews more like wāu, ḥai, &c. The sound of ־ָיו is the same as ־ָו, i.e. almost like āu, so that ־ָו is often written defectively for ־ָיו.
- In early MSS. the sign for Qameṣ is a stroke with a point underneath, i.e. according to Nestle’s discovery (ZDMG. 1892, p. 411 f.), Pathaḥ with Ḥolem, the latter suggesting the obscure pronunciation of Qameṣ as å. Cf. also Ginsburg, Introd., p. 609.
- Instead of the no doubt more accurate transcription å̄, å̂ we have retained ā, â in this grammar, as being typographically simpler and not liable to any misunderstanding. For Qameṣ ḥaṭuph, in the previous German edition expressed by å̆, we have, after careful consideration, returned to ŏ The use of the same sign ־ָ for å̄ (å̂) and å̆, shows that the Massoretes did not intend to draw a sharp distinction between them. We must not, however, regard the Jewish grammarians as making a merely idle distinction between Qāmeṣ rāḥāb, or broad Qameṣ, and Qāmeṣ hatûph, or light Qameṣ. It is quite impossible that in the living language an ā lengthened from ă, as in dābār, should have been indistinguishable from e.g. the last vowel in וַיָּשָׁב or the first in קָֽדָשִׁים.—The notation ā, ê, ô expresses here the vowels essentially long, either naturally or by contraction; the notation ā, ē, ô those lengthened only by the tone, and therefore changeable; ă, ĕ, ŏ the short vowels. As regards the others, the distinction into i and ĭ, ŭ and ŭ is sufficient; see § 9.—The mark ֫ stands in the following pages over the tone-syllable, whenever this is not the last, as is usual, but the penultimate syllable of the word, e.g. יֵ֫שֶׁב.
- These Segôls, modified from ă, are very frequent in the language. The Babylonian punctuation (see §8g, note 3) has only one sign for it and tone-bearing Pathah.; see also Gaster, ‘Die Unterschiedslosigkeit zwischen Pathach u. Segol,’ in ZAW. 1894, p. 60 ff.
- On the erroneous use of the term melo pum, only in Germany, for šûreq (hence also pronounced melu pum to indicate û), see E. Nestle, ZDMG. 1904, p. 597 ff.; Bacher, ibid., p. 799 ff., Melopum; Simonsen, ibid., p. 807 ff.
- The usual spelling קָמֶץ and פַּתַח takes the words certainly rightly as Hebrew substantives; according to De Lagarde (Gött. gel. Anz. 1886, p. 873, and so previously Luzzatto), קָמֵץ and פָתַח are rather Aram. participles, like Dageš, &c., and consequently to be transliterated Qâmeṣ and Pâthaḥ.
- Since 1846 we have become acquainted with a system of vocalization different in many respects from the common method. The vowel signs, all except וּ, are there placed above the consonants, and differ almost throughout in form, and some even as regards the sound which they denote: tone-bearing ă and In an unsharpened syllable toneless ă and è, and also Ḥaṭeph Pathaḥ; toneless ĕ and Ḥaṭeph Seghôl; and Ḥaṭeph Qameṣ. Lastly in toneless syllables before Dageš, The accents differ less and stand in some cases under the line of the consonants. Besides this complicated system of the Codex Babylonicus (see below) and other MSS., there is a simpler one, used in Targums. It is still uncertain whether the latter is the foundation of the former (as Merx, Chrest. Targ. xi, and Bacher, ZDMG. 1895, p. 15 ff.), or is a later development of it among the Jews of South Arabia (as Praetorius, ZDMG. 1899, p. 181 ff.). For the older literature on this Babylonian punctuation (נִקּוּד בַּבְלִי), as it is called, see A. Harkavy and H. L. Strack, Katalog der hebr. Bibelhandschr. der Kaiserl. öffentl. Bibliothek zu St. Petersb., St. Petersb. and Lpz., 1875, parts i and ii, p. 223 ff. A more thorough study of the system was made possible by H. Strack’s facsimile edition of the Prophetarum posteriorum codex Babylonicus Petropolttanus (St. Petersb., 1876, la. fol.) of the year 916, which Firkowitsch discovered in 1839, in the synagogue at Tschufutkale in the Crimea. The MS. has been shown by Ginsburg (Recueil des travaux rédigés en mémoire... de Chwolson, Berlin, 1899, p. 149, and Introd., pp. 216 ff., 475 f.) to contain a recension of the Biblical text partly Babylonian and partly Palestinian; cf. also Barnstein, The Targum of Onkelos to Genesis, London, 1896, p. 6 f. Strack edited a fragment of it in Hosea et Joel prophetae ad fidem cod. Babylon. Petrop., St. Petersb. 1875. Cf. also the publication by A. Merx, quoted above, §7h, and his Chrestomathia Targumica, Berlin, 1888; G. Margoliouth, in the PSBA. xv. 4, and M. Gaster, ibid.; P. Kahle, Der masoret. Text des A.T. nach d. Überlief. der babyl. Juden, Lpz. 1902, with the valuable review by Rahlfs in GGA. 1903, no. 5; Nestle, ZDMG. 1905, p. 719 (Babylonian ע. According to the opinion formerly prevailing, this Babylonian punctuation exhibits the system which was developed in the Eastern schools, corresponding to and contemporaneous with the Western or Tiberian system, although a higher degree of originality, or approximation to the original of both systems of punctuation, was generally conceded to the latter. Recently, however, Wickes, Accents of the Twenty-one Books, Oxford, 1887, p. 142 ff, has endeavoured to show, from the accents, that the ‘Babylonian’ punctuation may certainly be an Oriental, but is by no means the Oriental system. It is rather to be regarded, according to him, as a later and not altogether successful attempt to modify, and thus to simplify, the system common to all the Schools in the East and West. Strack, Wiss. Jahresb. der ZDMG. 1879, p. 124, established the probability that the vowels of the superlinear punctuation arose under Arab influence from the vowel letters יוא (so previously Pinsker and Graetz), while the Tiberian system shows Syrian influence. A third, widely different system (Palestinian), probably the basis of the other two, is described by A. Neubauer, JQR. vii. 1895, p. 361 ff., and Friedländer, ibid., p. 564 ff., and PSBA. 1896, p. 86 ff.; C. Levias, Journ. of Sem. Lang. and Lit., xv. p. 157 ff.; and esp. P. Kahle, Beitr. zu der Gesch. der hebr. Punktation,’ in ZAW. 1901, p. 273 ff. and in Der masoret. Text des A.T. (see above), chiefly dealing with the Berlin MS. Or. qu. 680, which contains a number of variants on the biblical text, and frequently agrees with the transcriptions of the LXX and Jerome.
- After the example of the Jewish grammarians the expression, ‘the vowel letter rests (quiesces) in the vowel-sign,’ has become customary. On the other hand, the vowel letters are also called by the grammarians, matres lectionres or supports (fulcra).
- Cf. T. C. Foote, The diphthong ai in Hebrew (Johns Hopkins Univ. Circulars, June, 1903, p. 70 ff.).
- In MSS. ו and י, in such combinations as גֵּוּ, חַיּ, are even marked with Mappîq (§14a).