1. The original vowels in Hebrew, as in the other Semitic tongues, are a, i, u. E and o always arise from an obscuring or contraction of these three pure sounds, viz. ĕ by modification from ĭ or ă; short ŏ from ŭ; ê by contraction from ai (properly ay); and ô sometimes by modification (obscuring) from â, sometimes by contraction from au (properly aw).[1]
In Arabic writing there are vowel signs only for a, i, u; the combined sounds ay and aw are therefore retained uncontracted and pronounced as diphthongs (ai and au), e.g. שׁוֹט Arab. sauṭ, and עֵינַ֫יִם Arab. ‛ainain. It was only in later Arabic that they became in pronunciation ê and ô, at least after weaker or softer consonants; cf. בֵּין Arab. bain, bên, יוֹם Arab. yaum, yôm. The same contraction appears also in other languages, e.g. in Greek and Latin (θαῦμα, Ionic θῶμα; plaustrum = polostrum), in the French pronunciation of ai and au, and likewise in the German popular dialects (Oge for Auge, &c.). Similarly, the obscuring of the vowels plays a part in various languages (cf. e.g. the a in modern Persian, Swedish, English, &c.).[2]
2. The partial expression of the vowels by certain consonants (ה, ו, י; א), which sufficed during the lifetime of the language, and for a still longer period afterwards, must in the main have passed through the following stages[3]:—
(a) The need of a written indication of the vowel first made itself felt in cases where, after the rejection of a consonant, or of an entire syllable, a long vowel formed the final sound of the word. The first step in such a case was to retain the original final consonant, at least as a vowel letter, i.e. merely as an indication of a final vowel. In point of fact we find even in the Old Testament, as already in the Mêša˓ inscription, a ה employed in this way (see below) as an indication of a final o. From this it was only a step to the employment of the same consonant to indicate also other vowels when final (thus, e.g. in the inflection of the verbs ל״ה, the vowels ā,[4] ē, è). After the employment of ו as a vowel letter for ô and û, and of י for ê and î, had been established (see below, e) these consonants were also employed—although not consistently—for the same vowels at the end of a word.
The suffix of the 3rd sing. masc. in the noun (as in the verb) was originally pronounced הוּ. But in the places where this הוּ with a preceding a is contracted into ô (after the rejection of the ה), we find the ה still frequently retained as a vowel letter, e.g. עִירֹה, סוּתֹה Gn 4911, cf. §91e; so throughout the Mêša˓ inscription אַרְצֹה, בֵּיתֹה (also בֵּתֹה), בְּנֹה, בֹּה, לֹה, הִלְתַּֽחֲמֹה; on the other hand already in the Siloam inscription רֵעוֹ.[5] ימה Mêša˓, 1.8 = יָמָיו his days is unusual, as also רשה 1. 20 if it is for ראשיו his chiefs. The verbal forms with ה suffixed are to be read וַיַּלְפֵהֻ (1. 6), וָֽאֶסְחָבֵהֻ (1. 12f.) and וַיְגָֽרְשֵׁהֻ (1. 19).
As an example of the original consonant being retained, we might also include the י of the constr. state plur. masc. if its ê is contracted from an original ay. Against this, however, it may be urged that the Phoenician inscriptions do not usually express this ê, nor any other final vowel.[6]
(b) The employment of ו to denote ô, û, and of י to denote ê, î, may have resulted from those cases in which a ו with a preceding a was contracted into au and further to ô, or with a preceding u coalesced into û, and where י with a has been contracted into ai and further to ê, or with a preceding i into î. In this case the previously existing consonants were retained as vowel letters and were further applied at the end of the word to denote the respective long vowels. Finally א also will in the first instance have established itself as a vowel letter only where a consonantal א with a preceding a had coalesced into â or ā.
The orthography of the Siloam inscription corresponds almost exactly with the above assumptions. Here (as in the Mêša˓ inscr.) we find all the long vowels, which have not arisen from original diphthongs, without vowel letters, thus אִשׁ, חֹצְבִם, מִימִן (or מִיָּמִן); אַמֹּת, קֹל, שְׁלשׁ, צֻר. On the other hand מוֹצָא (from mauṣa˒), עוֹר (from ˓aud); מימן also, if it is to be read מִימִן, is an instance of the retention of a י which has coalesced with i into î. Instances of the retention of an originally consonantal א as a vowel letter are מָאתַ֫יִם, מוֹצָא, and קָרָא, as also רֹאשׁ. Otherwise final ā is always represented by ה: אַמָּה, הָיָה, זרה, נקבה. To this יֹם alone would form an exception (cf. יוֹם), instead of יוֹם (Arab. yaum) day, which one would expect. If the reading be correct, this is to be regarded as an argument that a consciousness of the origin of many long vowels was lost at an early period, so that (at least in the middle of the word) the vowel letters were omitted in places where they should stand, according to what has been stated above, and added where there was no case of contraction. This view is in a great measure confirmed by the orthography of the Mêša˓ inscription. There we find, as might be expected, דיבן (= Daibōn, as the Δαιβών of the LXX proves), חוֹרֹנָן (ô from au), and בֵּיתֹה (ê from ai), but also even הֽשִׁעַנִי[7] instead of הֽוֹשִׁעַנִי (from hauš-), ואשב = וָֽאוֹשִׁיב, בֵּת four times, בֵּתֹה once, for בֵּית and בֵּיתֹה (from bait); ללה = לַיְלָה, אן = אַ֫יִן or אֵין.
(c) In the present state of Old Testament vocalization as it appears in the Masoretic text, the striving after a certain uniformity cannot be mistaken, in spite of the inconsistencies which have crept in. Thus the final long vowel is, with very few exceptions (cf. §9d, and the very doubtful cases in §8k), indicated by a vowel letter—and almost always by the same letter in certain nominal and verbal endings. In many cases the use of ו to mark an ô or û, arising from contraction, and of י for ê or î, is by far the more common, while we seldom find an originally consonantal א rejected, and the simple phonetic principle taking the place of the historical orthography. On the other hand the number of exceptions is very great. In many cases (as e.g. in the plural endings ־ִים and וֹת) the vowel letters are habitually employed to express long vowels which do not arise through contraction, and we even find short vowels indicated. The conclusion is, that if there ever was a period of Hebrew writing when the application of fixed laws to all cases was intended, either these laws were not consistently carried out in the further transmission of the text, or errors and confusion afterwards crept into it. Moreover much remained uncertain even in texts which were plentifully provided with vowel letters. For, although in most cases the context was a guide to the correct reading, yet there were also cases where, of the many possible ways of pronouncing a word, more than one appeared admissible.[8]
3. When the language had died out, the ambiguity of such a writing must have been found continually more troublesome; and as there was thus a danger that the correct pronunciation might be finally lost, the vowel signs or vowel points were invented in order to fix it. By means of these points everything hitherto left uncertain was most accurately settled. It is true that there is no historical account of the date of this vocalization of the O.T. text, yet we may at least infer, from a comparison of other historical facts, that it was gradually developed by Jewish grammarians in the sixth and seventh centuries a.d. under the influence of different Schools, traces of which have been preserved to the present time in various differences of tradition.[9] They mainly followed, though with independent regard to
- ↑ In proper names the LXX often use the diphthongs αἰ and αὐ where the Hebrew form has ê or ô. It is, however, very doubtful whether the αἰ and αὐ of the LXX really represent the true pronunciation of Hebrew of that time; see the instructive statistics given by Kittel in Haupt’s SBOT., on 1 Ch 12, 20.
- ↑ In Sanskrit, in the Old Persian cuneiform, and in Ethiopic, short a alone of all the vowels is not represented, but the consonant by itself is pronounced with short a.
- ↑ Cf. especially Stade, Lehrb. der hebr. Gr., p. 34 ff.
- ↑ According to Stade, the employment of ה for ā probably took place first in the case of the locative accusatives which originally ended in ־ָה, as אַ֫רְצָה, קָדִ֫ימָג.
- ↑ The form רעו contradicts the view of Oort, Theol. Tijds., 1902, p. 374, that the above instances from the Mêša˓-inscription are to be read benhu, bahn, lahu, which were afterwards vocalized as beno, bo, lo.
- ↑ Thus there occurs, e.g. in Melit. 1, l. 3 שנבן = שְׁנֵי בְנֵי the two sons; elsewhere כ for כִּי (but כי in the Mêša˓) and Siloam inscrr.), ז for זֶה (the latter in the Siloam inscr.), בנת = בָּנִתִי (so Mêša˓) or בָּנִיתִי, &c. Cf. on the other hand in Mêša˓, אנכ = אנכי (unless it was actually pronounced ˒anôkh by the Moabites!). As final ā is represented by ה and א and final î by י, so final û is almost everywhere expressed by ו in Mêša˓, and always in the Siloam inscription. It is indeed not impossible that Hebrew orthography also once passed through a period in which the final vowels were left always or sometimes undenoted, and that not a few strange forms in the present text of the Bible are to be explained from the fact that subsequently the vowel letters (especially ו and י) were not added in all cases. So Chwolson, ‘Die Quiescentia הוי in der althebr. Orthogr.,’ in Travaux du Congrès… des Orientalistes, Petersb. 1876; cf. numerous instances in Ginsburg, Introd., p. 146 ff.
- ↑ השעני is the more strange since the name of king הוֹשֵׁעָ is represented as A-u si˒ in cuneiform as late as 728 b.c.
- ↑ Thus e.g. קטל can be read qāṭal, qāṭāl, qāṭôl, qeṭōl, qôṭēl, qiṭṭēl, qaṭṭēl, quṭṭal, qèṭel, and several of these forms have also different senses.
- ↑ The most important of these differences are, (a) those between the Orientals, i.e. the scholars of the Babylonian Schools, and the Occidentals, i.e. the scholars of Palestine (Tiberias, &c.); cf. Ginsburg, Introd., p. 197 ff.; (b) amongst the Occidentals, between Ben-Naphtali and Ben-Asher, who flourished in the first half of the tenth century at Tiberias; cf. Ginsburg, Introd., p. 241 ff. Both sets of variants are given by Baer in the appendices to his critical editions. Our printed editions present uniformly the text of Ben-Asher, with the exception of a few isolated readings of Ben-Naphtali, and of numerous later corruptions.