Pronunciations of Hebrew
Pronunciations of Hebrew
PRONUNCIATIONS OF HEBREW
This article is arranged according to the following outline:the transmission of hebrew as a liturgical language
classification of the traditional pronunciations of hebrew
The Yemenite Pronunciation
The Sephardi Pronunciation
The Ashkenazi Pronunciation
Classification of the Pronunciations of Hebrew
the realizations of the consonants and the vowels and the stress patterns in the various pronunciations
qameṢ and pathaḤ
Ṣere and segol
Mobile Šewa and Quiescent Šewa in the Various Communities
The Realizations of the Mobile Šewa
Before its revival at the end of the 19th century, Hebrew existed, for a period of about 1700 years, mainly as a literary and liturgical language. This period in the history of Hebrew probably began around the third century c.e. There is evidence that Hebrew was spoken, at least in some parts of Palestine, in the second century c.e. This is clear from the story in the Talmud about the maid of Rabbi Judah ha-Nasi who knew the meanings of some Hebrew words with which the scholars of that time were not acquainted (rh 26b; Meg. 18a; tj, Meg. 2:2, 73a). Evidence that Hebrew was spoken in the first half of the second century c.e. is also borne out by the Hebrew letters of Bar-Kokhba, some of the grammatical forms of which show that Hebrew was still a living language at that time (c. 135 c.e.). It should be mentioned, however, that at that period, and for centuries before, other languages were spoken concomitantly by the Jewish communities of Palestine, mainly Aramaic and Greek. The use of Hebrew as a spoken language became more and more limited, and finally it was superseded by Aramaic and Greek. Although the exact time when Hebrew ceased to be spoken is not known, there is no unequivocal evidence for the use of Hebrew as the ordinary spoken language of any Jewish community in a period later than the second century. It may be assumed, therefore, that the period in which the use of Hebrew was limited to literature and liturgy only began about the third century c.e.
As a liturgical language Hebrew has been transmitted during this long period, and in fact up to the present day, in a number of forms which are known as the "traditional pronunciations" of Hebrew. This term denotes those pronunciations which have been used by the various Jewish communities in reading the Bible and the post-biblical literature and in prayers. Another term used for "traditional pronunciation" is "reading tradition," or "liturgical reading tradition." A few words explaining these terms are in order here. A "reading tradition" may be defined as a corpus of linguistic information, transmitted orally, upon which the correct reading of a text is based; a "liturgical reading tradition" is a reading tradition that is used in the transmission of those parts of the literature which have particular religious importance. The traditional pronunciations of Hebrew have been transmitted in the various communities over a long period. They still exist in Israel and in various Jewish communities of the Diaspora. In Israel, however, the traditional pronunciations are disappearing at a fast rate, as a result of the mutual contact among the various communities, and of the influence of the current pronunciation of Hebrew.
The traditional pronunciations of Hebrew extant with most communities are of two major categories: (a) the pronunciations used in the reading of the Bible; (b) the pronunciations used in the reading of the post-biblical literature, primarily the Mishnah. In the pronunciations pertaining to the first category, the reading is based upon the vocalized text of the Bible, whereas in the second it is based, in many communities, upon an unvocalized text. This results from the fact that for the Mishnah no authorized vocalization exists that could be compared to the Tiberian vocalization of the Bible (which was accepted by all Jewish communities, except for the Samaritan, as the authoritative vocalization according to which the Bible should be read). Therefore, the reading of the Mishnah in most communities is based upon a text which does not possess vocalization signs and which, for many words, represents only their consonantal skeleton. The reader supplements those phonological entities that are not represented in the orthography according to the oral tradition of his community. In other words, in reading the Bible the reader gives each grapheme (that is, the letters and the vocalization signs) the phonetic value it has in the traditional pronunciation of the community. In reading the Mishnah, on the other hand, in addition to giving each grapheme the phonetic value it has in the traditional pronunciation of the community, the reader also supplements, according to the oral tradition of his community, those phonological entities which are not represented by the orthography. Since the oral traditions of the various communities differ from each other, it follows that the various reading traditions of the Mishnah disclose different forms of the same word. To illustrate the difference between the traditional pronunciation of the Bible and that of the Mishnah, the word qereaḥ (קרח) "bald" may be taken. This word appears both in the Bible and in the Mishnah. In the Bible this word is spelled קרח and vocalized qereaḥ, and the communities differ from each other in the phonetic values they give to the consonants and the vowels. Thus ק is pronounced as [g] or [q] by the Yemenite (the exact pronunciation depending on the district from which the individual reader comes; for further details, see below), as [q] by the Iraqi readers, and as [k] by the Ashkenazi and some of the Sephardi readers; the (the vowel sign ṣere) is pronounced as [e] by the Yemenite and Sephardi readers, and as [ey] or [ay] by the Ashkenazi; the ח is pronounced as [ḥ], an unvoiced pharyngeal fricative, by the Yemenite and some of the Sephardi readers, but as [x] by the Ashkenazi. The situation is utterly different in the reading of the Mishnah. Here the word is spelled קרח, as it is in the Bible, but no vocalization signs appear in the text which would make a certain form of the word binding for a specific community. Therefore, there are differences among the communities as to the very form of the word – the Iraqi reading it as [qareyaḥ], the Yemenite as [qereḥ], all other communities as qereaḥ (that is, a form identical to the biblical) – and not only in the phonetic values given to the consonants and the vowels.
1. First, a differentiation must be made between the Samaritan pronunciation and all other pronunciations. Due to its specific features, the Samaritan pronunciation occupies a unique position within the bulk of the traditional pronunciations, and is of particular importance. We shall mention a number of the features typical of the Samaritan pronunciation: (a) the total disappearance of ח and in many cases also of ע; (b) the realization of the letter ו as [b]; (c) the existence of four degrees of length in the realization of the vowels, two of which are phonemic; (d) the realization of historical [i:] as [ә] (for example, the word הַבַּדִּים – Ex. 25:13 – is pronounced as abbaddәm – Z. Ben-Ḥayyim, The Literary and Oral Tradition of Hebrew and Aramaic Amongst the Samaritans, vol. iii pt. 1, p. 40); (e) the distinction between the counterparts of the historical vowels ū, u and those of ō, o disappeared; in most cases, the realization of these vowels as either [o] or [u] depends on their position in the word. Thus Tiberian סוּס, "horse," is [sos] in the Samaritan pronunciation (Ben-Ḥayyim, ibid., p. 37), whereas Tiberian שְׁמוֹ, "his name" is [se:mu] in Samaritan (ibid.). These features, and a number of others, make the Samaritan pronunciation unintelligible to the members of all other Jewish communities. In this connection it should also be mentioned that the Samaritan reading tradition of the Pentateuch is not based on vocalized texts. Such texts have not been used by the Samaritan community in the teaching of the traditional pronunciation, and in the reading of the Pentateuch. The Samaritan reader supplements the Pentateuch with the missing phonological entities according to the oral tradition transmitted in the community. In this respect the Samaritan traditional pronunciation differs from the traditional pronunciation of the Bible extant in all other Jewish communities, which use the Tiberian vocalization for the reading of the Bible.
2. The traditional pronunciations of the communities except the Samaritan are to be classified into three major groups: the Yemenite, the Sephardi, and the Ashkenazi.
Geographically isolated for generations, the Yemenite community has preserved a traditional pronunciation possessing a number of peculiar features. Some of these features, it is true, resulted from the influence of the pronunciation of the Yemenite Arabic vernaculars on the pronunciation of Hebrew. This influence is disclosed, e.g., in the realizations of גּ (when having a dageš) by members of the community who originally came from various regions of Yemen as an affricate [gˇ], a velar [g], or a palatalized [g']. These realizations correspond to the realizations of the Arabic phoneme [g] in the respective regions of Yemen (for which see below). Other features of the Yemenite pronunciation – particularly as regards the vowels – reflect, however, the traditional pronunciation of the Jewish community of geonic Babylonia. The correspondence between the present-day Yemenite pronunciation and the Babylonian pronunciation may best be proven by two phonetic phenomena: (a) the identity of the realizations of pathaḥ and segol; (b) the realization of ḥolem as ṣere by members of the communities of southwestern Yemen and of Aden. Both these phenomena are attested by the Babylonian vocalization, that is, the vocalization which reflects the pronunciation of Hebrew in Jewish communities of geonic Babylonia. However, whereas the former phenomenon is a regular feature of the Babylonian system of vocalization, which has only one vowel as the counterpart of both Tiberian pathaḥ and segol, this is not the case with the latter. The Babylonian system has signs for both ḥolem and ṣere; but in certain Babylonian manuscripts the signs for these two vowels interchange freely, and this indicates that in the pronunciation of the vocalizers of these manuscripts the two vowels were identical. Evidence for the identity of the realizations of ḥolem and ṣere by the Jewish communities of some provinces of Babylonia in the first half of the tenth century c.e. is borne also by a literary source, mainly al-Qirqisānī's Kitāb al'anwār walmarāqib. It is, therefore, clear that the identity of the realizations of ḥolem and ṣere by members of some Yemenite communities reflects a feature of the pronunciation of Hebrew in geonic Babylonia, and the Yemenite community is the only community to have preserved the Babylonian pronunciation of Hebrew. In fact, Yemenite Jewry has been the recipient of the legacy of geonic Babylonia in other fields as well: for centuries the Yemenites have used the Babylonian vocalization, in periods when it has been completely unknown to other Jewish communities; the Yemenites have used the Babylonian recension of the Bible at least until the beginning of the 13th century c.e.; the Yemenite reading tradition of post-biblical Hebrew resembles in many of its morphophonemic and morphological features the Babylonian tradition of post-biblical Hebrew, which is reflected by manuscripts of the Mishnah and Midrashim possessing Babylonian vocalization. The Yemenites have preserved a stable tradition of the vocalization of Targum Onkelos and Targum Jonathan, a tradition which most probably received its final shape in geonic Babylonia.
The fact that Yemenite Jewry has been the recipient of the Babylonian traditions in a number of fields is to be seen in the light of the close relations that existed between the Jewish community of Yemen and that of Babylonia in the geonic period. Historical evidence for these relations is found in some Genizah documents.
The Yemenite tradition of biblical Hebrew is to be clearly differentiated from that of post-biblical Hebrew. In reading the Bible, the Yemenites use the Tiberian vocalization and masorah; however, they give the Tiberian vocalization signs the values they had in the traditional Babylonian pronunciation. Thus, no distinction is made in the Yemenite pronunciation between Tiberian pathaḥ and segol since in the Babylonian pronunciation of Hebrew these two vowels were identical. In other words, the Yemenite pronunciation of biblical Hebrew discloses phonological features of the Babylonian tradition, but the morphology of biblical Hebrew in this tradition is Tiberian. Hence it follows that in their pronunciation of biblical
Hebrew the Yemenites differ from other communities in aspects which are purely phonetic, but not in the morphology.
The situation is, however, different in the pronunciation of post-biblical Hebrew. The Yemenite reading of post-biblical Hebrew is not based upon vocalized texts, but upon an oral tradition of the vocalization. Therefore, the reading of these texts discloses in the morphology, and not only in the phonology, a number of specific features. Many of those morphological features are originally Babylonian. This is readily proved by comparing the morphology of post-biblical Hebrew as reflected by the Yemenite pronunciation with the morphology of Hebrew as represented by manuscripts of Mishnah and Midrashim possessing Babylonian vocalization. Thus, for example, both the Yemenite pronunciation and the manuscripts have הַם for "they" (versus הֵם in biblical Hebrew) and זוֹג for "pair" (versus זוּג in other pronunciations).
It should be noted that the Yemenite pronunciation is not homogeneous. Within what is usually called "Yemenite pronunciation of Hebrew" five major groups may be distinguished, each group representing a different geographical zone of Yemen. These groups are:
(a) central Yemen, around the capital Ṣanʿa;
(b) northern Yemen, the region of Ḥaydān ashshām – Ṣaʿda;
(c) southwestern Yemen, the region of Sharʿab;
(d) Eastern Yemen, consisting of the communities of Ḥabbān and Ḥādina;
(e) the city of Aden.
The differences existing among these groups in the consonantal aspects of the pronunciation of Hebrew mostly correspond to the differences which exist among the Arabic dialects of the respective geographical zones. This is not the case, however, as regards a number of variations in the pronunciation of the vowels, which do not reflect the influence of the Arabic dialects on the pronunciation of Hebrew. These variations probably disclose inner varieties of the Babylonian pronunciation, which, as seen above, is to be considered as the source of the Yemenite pronunciation of Hebrew. Of these features we shall mention the following: 1. the realization of ḥolem as ṣere by groups (c) and (e); 2. the realization of ḥireq as a central vowel in group (d); 3. the realization of qameṣ as a back low unrounded vowel [ ] in group (d).
This term denotes a rather large variety of pronunciations used by the Jewish communities of the Orient (except the Yemenite community, whose pronunciation as observed above, differs basically from that of the other communities of the Orient) and of North Africa, as well as by the Sephardi communities of Europe (such as the Dutch-Portuguese, the Sephardi communities of Greece, Bulgaria, Yugoslavia, and Italy). Among these pronunciations there are considerable differences, which are mostly due to the influence of the vernaculars of the respective communities on the pronunciation of Hebrew. Two features, however, are common to all the pronunciations known as "Sephardi" or Oriental (but non-Yemenite): lack of distinction between pathaḥ and qameṣ on the one hand (except a qameṣ is a closed unstressed syllable; see below), and ṣere and segol on the other. These two features are characteristic of certain manuscripts possessing 'Palestinian' vocalization. A certain variety of the 'Palestinian' pronunciation is, therefore, to be regarded as the source of the Sephardi pronunciation. By the term "Palestinian pronunciation" we denote the pronunciation reflected in manuscripts (mostly of piyyutim and of biblical texts) whose vocalization is called Palestinian, which was used by some communities of Palestine in a period approximately lasting from the sixth to the ninth century c.e. The Palestinian vocalization was apparently used in Palestine concomitantly with the Tiberian vocalization, but most probably not by the same communities. Each of these two vocalizations was based upon a different reading tradition, and in a general way it may be said that the Tiberian vocalization reflects a more classical and more pure reading tradition than the one reflected by the Palestinian. The latter vocalization is apparently based upon a more popular, or rather "vulgar," reading tradition. It may be added here that the Palestinian vocalization and pronunciation disclose a number of affinities with the vocalization and pronunciation of Samaritan Hebrew. The Palestinian pronunciation was adopted by many communities, far beyond the boundaries of Palestine, as the standard pronunciation of Hebrew, the pronunciation to be used in the liturgy and the teaching of the language. The Tiberian pronunciation, on the other hand, had, for some time at least, a classical standing. However, in spite of this classical standing of the Tiberian pronunciation, the Sephardi communities, some of which, particularly those of Spain and North Africa, may have used the Babylonian pronunciation during a certain period, finally adopted the Palestinian pronunciation. This may have resulted from the fact that the use of the Tiberian pronunciation was current only in a rather small group of Masoretes, whereas that of the Palestinian was far more common. The Palestinian pronunciation was probably first transplanted from Palestine to Italy, and later, when the influence of Italian Jewry on the Jewish communities of Spain became prominent, it was transplanted from Italy to Spain (this transplantation possibly took place in the time of Rabbi Moses ben Ḥanokh, in the second half of the tenth century c.e.). After the expulsion of the Jews from Spain, the Palestinian pronunciation, known by then as the Sephardi pronunciation, was transferred by the emigrants to the many communities in which they settled. These communities include the Dutch-Portuguese community of Amsterdam and the Sephardi communities in the Mediterranean countries (Italy, Yugoslavia, Bulgaria, Greece, Turkey, North Africa). The communities of the East (Syria, Palestine, Egypt, Persia) apparently used the Palestinian pronunciation for generations prior to the expulsion of the Jews from Spain.
The term "Sephardi" (or "Palestinian-Sephardi") pronunciation comprises in fact a number of pronunciations which differ from each other in a number of details, such as the pronunciation of the gutturals (for which see below). The features common to all varieties of pronunciations which are called "Sephardi" are, as stated above, lack of distinction between pathaḥ and qameṣ on the one hand, and between ṣere and segol on the other. The term "Sephardi" reading tradition when applied to the post-biblical literature (primarily the Mishnah) is a general term, covering a number of reading traditions, such as the Iraqi, the Aleppo, the Moroccan, etc. As yet the Sephardi reading traditions of the Mishnah have not been sufficiently studied, and an exhaustive description of their features is not possible. We shall therefore present here only a few phonological and morphological features of the "Sephardi" reading traditions of the Mishnah.
A number of Sephardi reading traditions of the Mishnah have a quiescent šewa in forms which in the reading of the Bible have a mobile šewa. This occurs when a medial šewa follows a qameṣ or a ḥolem, in forms like שׁוֹמְרִים ,שָׁמְרָה. In such forms all the "Sephardi" pronunciations have in the reading of the Bible a mobile šewa (with primary stress on the last syllable, and secondary stress on the syllable preceding the šewa), that is, šameʹra, šomeʹrim. In the reading of the Mishnah a number of "Sephardi" communities have in these forms a quiescent šewa (with primary stress on the syllable preceding the šewa), that is, ʹšamra, ʹšomrim. Another phonological (or, to be more precise, morphophonemic) feature in which the "Sephardi" reading traditions of the Mishnah differ from the "Sephardi" reading traditions of the Bible is disclosed by the distribution rules of the hard and soft realizations of those consonants of the בגדכפ״ת series for which the traditional pronunciations of the respective community have a double realization, hard and soft (for the realization of בגדכפ״ת consonants, see below). In the reading of the Bible the distribution rules of those /bgdkpt/ consonants for which the Sephardi communities have a double realization agree with the distribution rules of the Tiberian vocalization. This is not, however, the case in the reading of the Mishnah. Thus, for example, a number of Sephardi reading traditions of the Mishnah have a hard realization of a בגדכפ״ת consonant when this consonant follows an initial preposition whose vowel is a šewa. Such a realization – e.g., in the form bekerem, "in a vineyard" – stands in contradiction to that extant in the reading of the Bible, which is, in the form given here, beḵerem.
Morphologically, the Sephardi reading tradition of post-biblical Hebrew is not homogeneous, that is to say, there are differences in the forms of words in the various traditions, e.g., in the Iraqi tradition as compared with that of Aleppo, Morocco, and other traditions. There is lacking, for the time being, sufficient information as to the morphological structure of post-biblical Hebrew in the various Sephardi traditions; therefore, we shall mention here only two features which are shared by many of those traditions.
(1) The use of the pausal forms of the third person, fem. sing. and masc. plur., in the hof ʿal stem (e.g., huqama, "she was raised," huqamu, "they were raised") as the usual forms, that is, the forms which appear both in pausal and contextual positions.
(2) The appearance of -aḵ as the pronominal suffix for the second person masc. sing., as in e.g., kevodaḵ "your (masc. sing.) honor" (sometimes, in some traditions side by side with kevodeḵa).
The term "Ashkenazi" pronunciation is used to denote a variety of pronunciations used by the communities of Eastern and Central Europe and by immigrants from these communities who settled down in other parts of the world. We shall first survey briefly the salient features of the Ashkenazi pronunciation in the consonantal and vowel system. As regards the former system, the Ashkenazi pronunciation possesses two main features, which are shared by all its varieties, namely:
(a) the realization of ע as א that is, as "zero" (in some varieties of the Ashkenazi pronunciation [n] appears sporadically as the reflex of ע; this [n] may appear also as the reflex of historical א, e.g., kanšer, "when" (Heb. כַּאֲשֶׁר)).
(b) the realization of the soft ת as [s], e.g., 'bayis, "house."
The vowel system of the Ashkenazi pronunciation is far from being homogeneous. Thus the ḥolem is pronounced as [ey], and its realization is identical to that of the ṣere, in the northeastern variety of the Ashkenazi pronunciation (this variety is more commonly called the "Lithuanian"); as [oy] in the southeastern and central variety; as [aw] in many of the subtypes of the western varieties; as [ow] in those varieties of the Ashkenazi pronunciation used in the last generations in English-speaking countries. By and large the geographical distribution of the main varieties of the Ashkenazi pronunciation agrees with that of the dialects of Yiddish.
To quote another illustration of the heterogeneity extant in the Ashkenazi pronunciation as to the realizations of the vowels: the qameṣ is realized, in different varieties of the Ashkenazi pronunciation, as [o] or [u]. Of all the three major groups into which the traditional non-Samaritan pronunciations of Hebrew are divided, the Ashkenazi pronunciation is the only one to possess distinct phonetic realizations for all seven vowel graphemes of the Tiberian vocalization system: šureq-qibbuṣ (the qibbuṣ in the Tiberian vocalization is an allograph of the šureq and does not denote a different vowel phoneme), ḥolem, qameṣ, ḥireq, ṣere, segol, and pathaḥ. In the Yemenite pronunciation, which reflects the Babylonian, there is no distinction between pathaḥ and segol (see above); in the Sephardi, which continues the Palestinian, there is no distinction between qameṣ and pathaḥ on the one hand and between ṣere and segol on the other. In the course of time, however, some varieties of the Ashkenazi pronunciation developed a leveling of the realizations of two of the seven Tiberian vowels. In the northeastern ("Lithuanian") variety the realization of the ḥolem had been equaled with that of the ṣere, both becoming consequently [ey]; in most, if not all, subtypes of the southwestern and central varieties the realization of the šureq qibbuṣ had been equaled with that of the ḥireq, both becoming consequently [i] or rounded [i]. These developments within the Ashkenazi pronunciation resulted from parallel developments in the Yiddish dialects of the regions in which the above varieties of the Ashkenazi pronunciation were used. The fact that the Ashkenazi pronunciation had originally possessed distinct phonetic realizations for all seven vowels of the Tiberian vocalization system led some scholars to surmise that the Ashkenazi pronunciation constitutes, in its vowel system, a direct continuation of the Tiberian pronunciation. This opinion, however, cannot be accepted since there is evidence that until the 13th century c.e. the Sephardi pronunciation was used by the Ashkenazi communities. The above evidence is borne out by various sources (vocalized texts, transcriptions of Hebrew words in Latin character, notes in grammatical treatises), which show that in Ashkenazi communities until the 13th century c.e. the qameṣ was realized as pathaḥ and the ṣere as segol, and these two features are typical of the Sephardi pronunciation. The evidence is reinforced by certain Hebrew loanwords in Yiddish in which historical qameṣ is reflected by pathaḥ and historical ṣere by segol.
How can the fact be explained that until the 13th century c.e. there prevailed in the Ashkenazi communities (or, at least, in a number of them) the Sephardi pronunciation, making no distinction between qameṣ and pathaḥ on the one hand and between ṣere and segol on the other, and later these communities developed the Ashkenazi pronunciation, which differentiates between the above vowels? Max *Weinreich suggested that the pronunciation which is known as Ashkenazi was formed, in its main features, in Central Europe approximately in the 13th century; until that period the pronunciation used by the Ashkenazi communities was rather close to the Sephardi. The formation of the Ashkenazi pronunciation at that time in Europe resulted from the introduction of the Tiberian pronunciation (mainly as concerns the vowel system) into the Ashkenazi communities, this introduction being made by "Babylonian" scholars and teachers who immigrated to Central Europe from Babylonia (Iraq). By that time, and in fact for generations before, the original Babylonian pronunciation (for which see above) had been superseded in the Babylonian communities by the Tiberian. The Babylonian teachers and scholars were the carriers of the Tiberian pronunciation, and its transplantation into the Ashkenazi communities is due to their activity in these communities. The introduction of the Tiberian pronunciation into the Ashkenazi communities, which, according to Weinreich, played a major role in the formation of the Ashkenazi pronunciation, is a part of a more general process that took place at that time in Ashkenaz, and which Weinreich calls "The Babylonian Renaissance." Weinreich's theory is weak in that there is not sufficient historical evidence for the transplantation of the Tiberian pronunciation into the Ashkenazi communities through the medium of "Babylonian" scholars and teachers. The explanation offered by Yalon, with whom Weinreich agrees as to the very existence of a Sephardi pronunciation (or a pronunciation close to the Sephardi) in the Ashkenazi communities prior to the 13th century, is more plausible. Yalon's opinion is that the development of the distinction between qameṣ and pathaḥ, as well as between ṣere and segol – that is, of the main features in the vowel system in which the Ashkenazi pronunciation differs from the Sephardi – is due to the influence of the vowel system of Yiddish of that period (the Judeo-German dialects of the 13th century) on the pronunciation of Hebrew current in the Ashkenazi communities. In these dialects there was a sound shift a > o (cf. e.g., German das, Yiddish dos, "this"), and this sound shift brought over the realization of the qameṣ (which had been realized before as a pathaḥ, that is, as [a]) as [o] in the Ashkenazi pronunciation of Hebrew. The rise of the differentiation between ṣere and segol followed a similar course.
The classification of the pronunciations of Hebrew is presented in two charts, the first showing the historical aspects of the classification (See Table 1: Hebrew Pronunciations – Historical Classification), the second indicating the present-day ramifications (or, to be more precise, the ramifications that existed,
especially as to the varieties of the Ashkenazi pronunciation, until the extermination of the Jewish population of central Europe during World War ii). (See Table 2: Hebrew Pronunciations – Present-day Classification.)
Chart: Hebrew Pronunciations – Historical Classification shows that whereas the Babylonian pronunciation was continued by the Yemenite and the Palestinian by the Sephardi and (indirectly – see above) by the Ashkenazi, there is no direct continuation of the Tiberian pronunciation in any of the pronunciations that were adopted by the Jewish communities. The present-day classification of non-Samaritan Hebrew pronunciation is as follows:
The Yemenite ramifications divide into Central, Northern, Southwestern, Aden, and Eastern.
The Ashkenazi branches are Western and Eastern. The Eastern further subdivides into Northeastern, Southeastern, and Central.
Among the Sephardi pronunciations there are those of the European communities – Ladino-speaking communities, Italian-speaking communities, and Dutch-Portuguese pronunciation – and the pronunciations of the Asian (Yemenite excepted) and African communities – Arabic-speaking communities, Aramaic-speaking communities, Persian-speaking communities, and Georgian-speaking communities.
the realizations of the consonants and the vowels and the stress patterns in the various pronunciations
First, a word of comment on the term "realization." In describing the sounds extant in the traditional pronunciations of Hebrew we have to deal with a pronunciation of a literary language transmitted from generation to generation through a reading tradition. In each of these reading traditions specific phonetic values are given to the letters that represent consonants and vowels, as well as to the vocalization signs which represent vowels and some other phonological entities. These
phonetic values are called the realizations of the letters and the vocalization signs. (See Table 3: Realization of Graphemes.)
The realization of these letters as plosives [bgdkpt] in initial position and medially after a consonant and as fricatives [vgdkft] is attested in the Middle Ages in all Arabic-speaking communities, including Spain, and in France (I. Garbell, in: Bulletin de la Société de Linguistique de Paris, 50 (1954), 232). Of the pronunciations of Hebrew extant today, only the Yemenite pronunciation, and that of the Aramaic-speaking communities of Zakho (near Mosul, in Iraqi Kurdistan) differentiate the plosive series from the fricative series in the realizations of all the six letters. Other communities have fricative counterparts for the plosive realizations of some of the letters only, and the details will be given here. The realization of each of the בגדכפ״ת letters in the various communities will be presented according to the phonetic order of these letters (that is, in the order, כ ,ד ,ת ,ב ,פ, and ג).
For the letter פ, there exist in all communities, but for the Georgian-speaking community of the Caucasus, a plosive, [p], and a fricative, [f ], realization for the hard and the soft p respectively. The Georgian-speaking community has only the plosive realization [p] for both hard פּ and soft פ. Members of Arabic-speaking communities tend to replace [p] by [f ] as the realization of hard פּ, but all these communities have both [p] and [f ] as the realizations of hard and soft פ respectively.
The letter ב is constantly, whether hard or soft, realized as the plosive [b] by many Arabic-speaking communities (those of Iraq, Syria, Egypt, Libya, North Tunisia, and Algeria), as well as by the Dutch-Portuguese community of Amsterdam and by some Italian communities, such as the community of Leghorn. In the Yemenite community, the learned members distinguish between the plosive realization, [b], for hard בּ and the fricative realization, [v], for soft ב; other members of the community have [b], or both [b] and the bilabial fricative [β], as the realizations of the letter ב, hard or soft. Such a free variation between [b] and [β] is found also in other communities, namely those of eastern Kurdistan, the island of Djerba, southern Algeria and some of the provinces of Morocco. In the Persian-speaking communities, hard בּ is realized as [b], soft ב as either [β] or [v], and, in the vicinity of a back vowel, as the semivowel [w]. This realization of soft ב as [w] also occurs, when it comes at the end of a syllable, in some communities in the northwestern part of Italy. The Georgian-speaking community has only [b] as the realization of both soft and hard ב. All the communities not mentioned here have [b] and [v] as the realization of hard and soft ב, respectively.
The letter ת, when hard, is realized as voiceless (or fortis) dental (or alveolar) plosive by all communities but not a few exceptions: in some communities of Yemen, affricate [ts] occurs (mostly as a free variant of [t]). When it is soft, it is realized as [s] by the Ashkenazi communities, as interdental voiceless [ṯ] by the Yemenite community, the Arabic-speaking Iraqi community, and the Aramaic-speaking community of Zakho in northern Iraq; as [d] by the communities of Italy (but for the Sephardi communities of this country); and as voiced interdental or postdental [ḏ] by the Cochin community of India. All other communities have [t] as the realization of both hard and soft ת, but for some communities of Morocco which have the affricate [ts] for both hard and soft ת.
Only relatively few communities differentiate the realization of soft ד from that of hard ד. Soft ד is realized as an interdental [ḏ], whereas hard ד is [d], by the Yemenite community, by the Aramaic-speaking Zakho community of northern Iraq (where [z] appears as a free variant), and by a number of communities in the Balkan countries. The Arabic-speaking community of Iraq has [ḏ] as the realization of soft ד only in a few words: in the divine name [ʾaḏonay] and in the word [ʾeḥaḏ], "one," but only when occurring in the first verse of Qeriat Shema. In all other words [d] is the realization of both hard and soft ד in the Iraqi community.
כ is realized as a voiceless velar stop, [k], when hard, and as a voiceless velar fricative [ḵ], when soft, by all communities but for the Samaritan, who has [k] for both hard and soft כ.
ג, when soft, is realized as a voiced velar fricative, [ḡ], by most Arabic-speaking communities, as well as by some communities of the Balkans. The Dutch-Portuguese community has a voiceless velar fricative, [ḵ], as the realization of soft ג. All communities not mentioned here have only one realization, [g], for both hard and soft ג. This is also the realization of hard ג in the above communities who do maintain this differentiation, except for certain Yemenite communities. The communities of central Yemen have a voiced prepalatal affricate, [ğ], as the realization of hard ג; those of extreme eastern Yemen, and some of those of northern Yemen – a voiced prepalatal plosive, [g'].
The letter א is realized as a glottal stop, [ʾ], in most pronunciations. It should be noted, however, that in the Ashkenazi pronunciation, and occasionally, but much less frequently, also in some Sephardi (including the Italian) and Yemenite pronunciations, it may be realized as zero, that is, it is not represented by any sound. In some Ashkenazi pronunciations the contact between two vowels caused by the elision of א in medial intervocalic position gives birth to a glide. In the Georgian-speaking community as well as in some North African communities, and in the community of Cochin (India), [h] appears as free variant of א.
The letter ה is realized as a glottal fricative [voiced or voiceless] by most communities. In the Italian-speaking communities its realization is, however, zero, as that of א. In some Ashkenazi communities, as well as in some communities of the Balkan countries, and in a great number of communities located in the area stretching from Libya to southeast Morocco, it varies freely with the realizations of א, that is also with zero.
Final consonantal ה) ה with a mappiq) is realized as [h] in the Yemenite pronunciation and in those of some other Arabic-speaking communities. In the Ashkenazi communities and in most Sephardi communities of Europe it is generally realized as zero. In the Dutch-Portuguese community it is realized as [aha].
The realizations of the letters ח and ע as voiceless and voiced pharyngeal fricatives, [ḥ] and [ʿ], respectively, are found in all Arabic-speaking communities and in most Aramaic-speaking communities. Most of the Persian-speaking communities, however, have [h] as the realization of ח, and zero as that of ע .ח is realized as a voiceless velar fricative, [ḵ], by all European communities, both Ashkenazi and Sephardi, and by the Georgian-speaking community. ע is realized as [ʾ] or zero by all of the European communities, with the exception of those of Italy and the Dutch-Portuguese of Amsterdam. The latter communities have a voiced velar nasal, [ŋ] – the sound of ng in the English word "king" – as the realization of ע. The Georgian-speaking community has for this letter a voiceless glottalized uvular plosive, [qʾ], in initial and final position; intervocally it is realized as a voiced laryngealized uvular fricative.
ט is realized identically to its non-emphatic counterpart, ת in the pronunciations of the Ashkenazi, Italian, Dutch-Portuguese, and Sephardi communities of Europe, as well as in the pronunciation of the communities of Persia and eastern Kurdistan (in the pronunciation of the Kurdish communities, however, a historical ט is reflected in the quality of the phones of the word, which became emphatic). In the Arabic-speaking communities, the Aramaic-speaking community of Iraqi Kurdistan and in the Georgian-speaking community the realization of ט differs from that of ת: it is a velarized dental (or alveolar) voiceless (or fortis) plosive, [ṭ], in most Arabic-speaking communities and in the aforementioned Aramaic-speaking community; either [ṭ] or [ḍ], the voiced (or lenis) counterpart of voiceless (or fortis) [ṭ], in the Yemenite community, and [t'], a voiceless glottalized dental plosive in the Georgian-speaking community.
צ is realized as a voiceless dental affricate [ts] by the Ashkenazi, Italian, and Dutch-Portuguese communities; as [s], a velarized hissing sound, in the Arabic-speaking communities and the Aramaic-speaking community of Iraqi Kurdistan; as a non-emphatic [s] – its realization being identical to that of ס – in the Persian-speaking communities, in the Aramaic-speaking communities of eastern Kurdistan and Azarbaijan, in the Georgian-speaking community, in the community of Cochin (India), and in some communities of the Balkan and North African countries.
ק is realized identically to its non-emphatic counterpart, כּ, in the Ashkenazi, the Italian, the Dutch-Portuguese, and the other Sephardi communities of Europe, in some Algerian and east Moroccan communities, and a number of Persian-speaking communities. In the following communities ק is realized in a way different from כּ: in the communities of central, northern,
|Hebrew grapheme||Samaritan pronunciation||Yemenite pronunciation||Sephardi pronunciation||Ashkenazi pronunciation|
|Central||Northern||South-western||Aden||Eastern||European communities||Asian (Yemenite excepted) and African communities||North-eastern||South-eastern||Central|
|Ladino speaking||Italian speaking||Dutch Portuguese||Aramaic speaking||Arabic speaking||Persian speaking||Georgian speaking|
|א||, "zero,"||, "zero"||"zero"||"zero"||"zero"||, "zero"||, "zero"||, "zero"||, h||"zero"||"zero",(h)||"zero"|
|ע||, "zero,"||"zero,"||ᵑ||ᵑ||, , "zero,"||qʾ||"zero,"||"zero,"||"zero,"|
|ק||q,||g,ǥ||g,ǥ||q||q,||g,ǥ||k||k||k||q,||q||ǥ, , (k)||k'||k||k||k|
|qameṣ||a,å||å||å||å||å||ɒ||qameṣ (except when coming in closed unstressed syllable: "qameṣ qaṭan"), is identical to pathaḥ, both being realized as a; "qameṣ qaṭan" is realized as o.||o||u||u|
|pathaḥ||a,å||a,ä||a,ä||a,ä||a,ä||a,ä||Exception: some Persian speaking communities realize the qameṣ as å||a||a||a|
|ṣere||ɛ10||e||e||e||e||e||In all varieties ṣere and segol are realized alike, as e ˜ ε||ey||ay||ay|
|segol||ɛ10||realization is identical to that of pathaḥ||ɛ||ey,e||ey,e|
|mobile šewa||basic realization = ǎ; when preceding y = ǐ; when preceding ʾ,h, ḥ or ʿ = ultrashort vowel, identical in quality to the vowel of the following consonant||e ˜ ɛ||e ˜ ɛ||e ˜ ɛ11||e ˜ ɛ||e ˜ ɛ12||a,(⋀)||e||e,||e,||e,|
|ḥatef-segol||ǎ||ǎ||ǎ||ǎ||ǎ||e ˜ ɛ||e ˜ ɛ||e ˜ ɛ||e ˜ ɛ||e ˜ ɛ||e ˜ ɛ||e||ɛ||ey,e||ey,e|
Notes to Table 3.
1. Two or more phonetic signs are given as the realizations of any letter (or of a vowel sign) in a specific community, when the letter has two or more realizations, the relationship between them being that of "free" or "positional" allophones (that is, sounds whose articulation depends on their position in the word, the preceding and the following sounds, the structure of the syllable, etc.).
One of the phonetic signs that represent the realizations of a letter comes in brackets when it is less common than the other realizations, or when it is used only by a part of the community in question.
2. C stands for "consonant." CC = a geminated, double, consonant.
3. The sign ˘ above a letter representing a vowel, denotes that the vowel is ultrashort, e.g. [ă] = ultrashort [a].
4. A long vowel is denoted by a line above it, e.g. [ā] = long [a].
and eastern Yemen it is realized as a voiced velar or uvular plosive [g] or [ǥ]. In the communities of southwest Yemen and Aden it is realized as a voiceless uvular plosive, [q], which has, particularly in the community of Aden, a voiced uvular fricative, , as its variant in certain positions. In some Persian-speaking communities ק is realized as a voiced uvular stop, [g], which has as its variant a voiced uvular fricative ; other Persian-speaking communities realize ק as [k]. Many Arabic-speaking communities and those speaking Aramaic have a voiceless uvular plosive, [q], as the realization of ק. In the community of Aleppo, as well as in some communities of Algeria and Morocco, ק is realized as a glottal stop, [ʾ]; in these communities, however, this realization of [ʾ] exists alongside with [q]. In the communities of southern Algeria and Cochin, ק has a velar, or uvular, voiceless fricative, [k], as its realization (along with the realization [q]). In the Georgian-speaking community ק is realized as a velar-uvular affricate, [k].
No community maintains the distinction that existed in biblical Hebrew between the phonemes represented by the letters ס ,שׁ, and שׂ. In the Samaritan community the realization of שׂ is identical to that of שׁ; in all other communities the realization of שׂ is equal to that of ס. Most communities realize שׁ as an unvoiced hushing sound, [š], and ס(as well as שׂ) as an unvoiced hissing sound, [s]. Exceptions are the following:
(1) In the northeastern ("Lithuanian") Ashkenazi and some North African communities, no distinction was made between the realization of שׁ on the one hand and that of ס (and שׂ) on the other, the exact articulation of the sound representing both שׁ and ס (and שׂ) – whether [š] or a sound intermediary between [š] and [s] – varying locally. This pronunciation, which corresponds to dialectal features of northeastern Yiddish, tended to disappear after about 1930. A similar phenomenon occurs in some Moroccan communities in which both שׁ and ס (and שׂ) are realized as a sound intermediary between [š] and [s], or as [s].
(2) In some communities of Greece the realization of שׁ as [š] has a variant [s].
(3) In the communities of northwest Italy, and to a lesser extent in the communities of northeast Italy, שׁ is realized as [s].
ז is realized by most communities as a voiced hissing sound, [z]. In some Italian-speaking communities it is realized as a voiceless hissing sound, [s]; in other communities as an affricate, voiced, [dz], or voiceless, [ts]. In both categories of these communities, the exact realization of ז depends on the position of the letter in the word. In some communities of Morocco it is realized as a sound intermediary between [z] and [ ž].
The letter ר is realized as an apical flap or trill, [r], by most of the Arabic-speaking, Aramaic-speaking, and Persian-speaking communities as well as by the Sephardi communities of Europe. Most of the Ashkenazi communities, on the other hand, realize it as a voiced velar fricative [ḡ], or a velar frictionless continuant; some Ashkenazi communities, however, realize ר as an apical flap or trill, [r].
The letter ו is realized by most Arabic-speaking communities as the semivowel [w]; in some communities of Syria and Egypt, as well as in northwest Morocco, it is realized as a labiodental voiced fricative, [v]; in the communities of northeast Morocco the realization [w] has the variant [v]; in some communities of Algeria [w] is realized as a bilabial voiced stop, [b]. In the Aramaic-speaking communities ו is realized as [w]; this realization, however, has a bilabial voiced fricative, [β], as its variant. In the Persian-speaking communities the realization of ו is identical to that of soft ב: it is either a bilabial voiced fricative, [β], or a labiodental voiced fricative, [v]; in the environment of a back vowel it has as its variant the semivowel [w]. Some Italian-speaking communities realize as [v] when it comes in initial and medial position, but as [w] in final position.
All Ashkenazi communities have [v] as the realization of ו.
The letter י is realized by the great majority of the communities as the palatal semivowel [y]. Some communities have for this realization the variants [i] or [ʾi], the particular positions in which these variants occur differing for the various communities. In some communities in northeast Italy, י, when occurring at the beginning of a syllable, is realized as the voiced prepalatal affricate, [gˇ]. According to the medieval grammarians Profiat Duran and Abraham de Balmes, such was the realization of geminated י in Provence and Italy.
The dageš forte is realized by doubling the consonant in the Arabic-speaking communities and the Aramaic-speaking communities of eastern Kurdistan. However, a single consonant occurs in some of these communities as the realization of a letter which has a dageš forte, particularly in an unstressed syllable. Most of the Italian-speaking communities also have a doubling of the consonant as realization of the dageš forte; the communities of northeast and northwest Italy, however, tend to realize a letter with a dageš forte in a way identical to that of a letter not possessing a dageš forte.
All the Ashkenazi communities, as well as the Persian-speaking communities, disregard the dageš forte in their traditional pronunciations.
The basic features of the vowel systems of the three major groups of the traditional pronunciations, in their relation to the Tiberian system of vocalization, were described above. We shall now present the realizations of the vowels in the various pronunciations.
־ֻ ,וּ. The Ashkenazi communities of the regions in which southeastern and central Yiddish was spoken – that is the Ukraine, Poland, Western Hungary, Western Slovakia, etc. – realize the šureq and the qibbuṣ as [i], or, some of them, as [ü] (rounded [i]). A realization of the šureq and the qibbuṣ as [ü], or as a centralized variant of this vowel, [u̵], also exists in a number of other communities – in some communities of Yemen, in Shiraz (Persia), Azarbaijan, in western Kurdistan, in some communities of northwestern Morocco, in northwestern Italy – but in most of these communities the realizations in question, rounded [i] and [u̵], are in fact variants of [u]. Since qibbuṣ appears in the vocalization of the Bible mostly in closed unstressed syllable, some scholars stated that in these communities a realization of the qibbuṣ is found – rounded [i] or [u̵] – which differs from that of the šureq, which is [u]. But the situation is not so. Today a consistent and regular differentiation between the realization of the šureq and that of the qibbuṣ does not exist in any community.
For communities in which the difference between the realizations of the šureq (and the qibbuṣ) and that of the ḥolem was neutralized, see below.
־ֹ ,וֹ. In two groups of communities, which were geographically located quite apart, the realization of the ḥolem was identical with that of the ṣere: in the communities of the regions in which northeastern ("Lithuanian") Yiddish was spoken, and in the communities of southwest Yemen (as well as in the community of Aden): in the former group both the ḥolem and the ṣere were realized as [ey], in the latter as [e]. From a historical point of view, however, there is no relation between the realization of ḥolem as ṣere in these two groups of communities. In the "Lithuanian" communities this realization apparently resulted from an interference of Yiddish in the traditional pronunciation of Hebrew. In the pronunciation of the aforementioned Yemenite communities, the realization of ḥolem as ṣere constitutes a feature of the pronunciation that prevailed in some Babylonian communities of the geonic period (see above).
The communities of the regions in which central and southeastern Yiddish was spoken realized the ḥolem as [oy], some German-speaking communities as [au], English-speaking communities as [ou]. In the communities of central, northern, and eastern Yemen – that is, in all Yemenite communities but for these in which ḥolem is realized as ṣere (see above) – the ḥolem is realized as a lower-mid rounded central vowel, [ö] (quite similar to the realization of eu in French peur). A similar realization of the ḥolem is attested in the Aramaic-speaking communities of Persian *Azerbaijan.
־ַ ,־ָ. All the Sephardi communities of Europe, the Italian communities, the Dutch-Portuguese communities of Amsterdam, and all the Asian and African communities – but for the Yemenite and the Persian (to some extent – see below) – do not differentiate between the realization of qameṣ gadol (that is, a qameṣ not occurring in a closed unstressed syllable) and that of a pathaḥ: they are both realized as a low front (or, in some communities, low central) vowel, [a]. This is a feature typical of the Sephardi pronunciation. In the Aramaic-speaking communities of eastern Kurdistan and Persian Azerbaijan a historical qameṣ gadol (which is realized as [a]) is reflected by the emphaticization of the phones of the word in which it occurs. All the above-mentioned communities realize qameṣ qaṭan, i.e., qameṣ which comes in a closed unstressed syllable (and which historically reflects the phoneme [u]) as [o], that is as a ḥolem. It should be noted, however, that the communities in question make two exceptions to the realization as [o] of qameṣ which historically reflects the phoneme [u], namely:
(1) A qameṣ preceding a ḥatef-qameṣ, e.g., in נָצֳמִי, is consistently regarded by these communities as a qameṣ gadol, and is realized as [a]. This realization of the qameṣ originated in the fact that these communities regard the metheg that follows a qameṣ preceding a hatef-qameṣ as indicating that the *qameṣ is a qameṣ gadol (whereas in fact it indicates that the syllable is open and that it has secondary stress).
(2) In the word כָּל, but only in two instances in the Bible (Ps. 35:10; Prov. 19:7) the qameṣ is realized as [a], that is, this word is pronounced [kal]. The reason for this lies in the fact that in these two instances the word כָּל has an accent, and this has been taken by the communities whose pronunciation is Sephardi to indicate that the qameṣ is a qameṣ gadol, which is realized as a pathaḥ, namely [a] (see Z. Ben-Ḥayyim, Studies in the Traditions of the Hebrew Language (1954), 71–72).
The Ashkenazi, the Yemenite, and some Persian communities differentiate between the realizations of qameṣ (qameṣ gadol as well as qameṣ qatan: no distinction is made in these communities between these entities) and that of pathaḥ. In the Ashkenazi communities the qameṣ is realized as [o] or [u] (the latter realization prevailing in communities of the region in which central and southeastern Yiddish is spoken). Most of the Yemenite communities realize the qameṣ as a rounded lower-mid back vowel, [å]; the communities of Ḥabbān and Ḥāḍina in Ḥaḍramaut realize the qameṣ as an unrounded low back vowel, [ɒ]. Some Persian-speaking communities realize the qameṣ as a rounded lower-mid back vowel, [å].
־ֵ ,־ֶ. All the communities which follow the Sephardi pronunciation have one realization for both ṣere and segol (see above).This realization is a front higher-mid or lower-mid vowel, [e] or [ε]. In some North African communities no distinction is made between the realizations of ṣere (and segol) and that of ḥireq. This applies also, to some extent, to the Iraqi community. In the Yemenite pronunciation the realization of the segol is identical with that of the pathaḥ. The Ashkenazi pronunciations are divided into two major groups as to the realizations of ṣere and segol:
(1) The communities of the area of northeastern Yiddish (the "Lithuanian" communities), as well as some other Ashkenazi communities, realized the segol (in a stressed syllable) as a front unrounded higher-mid vowel, [e]; the ṣere is realized as the diphthong [ey].
(2) Many communities of the areas of central and southeastern Yiddish realize the segol in a stressed syllable as [ey] and the ṣere as [ay]. No community makes any distinction between the realization of a defective ṣere and that of full ("plene") ṣere, as well as between the realization of a defective segol and that of a full ("plene") segol.
־ִ. All communities realize the ḥireq as a high front vowel, [i], with some positional variants. In eastern Yemen the ḥireq is realized as a central vowel, [ә].
־ְ. In presenting the realizations of the šewa two points should be considered: the principles that guide the various communities in differentiating a mobile šewa from a quiescent šewa; the realizations of the mobile šewa in the various communities.
The Yemenite community and the communities that adhere to the Sephardi pronunciation, including the Italian community, regard the šewa sign as denoting mobile šewa when it belongs to one of the following categories:
(1) when it appears in the beginning of a word;
(2) when it follows another šewa;
(3) when it comes with a letter that has a dageš forte;
(4) when it comes with the second of two identical letters, as in the word הִנְנִי (to this rule, however, there are exceptions);
(5) when it follows, in medial position, qameṣ, ṣere, ḥolem, šureq, or ḥireq which do not come in a syllable that has a primary stress, but may have a secondary stress (the reading traditions of many communities regard the metheg as a sign denoting secondary stress), and after which a dageš forte does not come (this applies mostly to the ḥireq). It should be noted that there is no general consistency as to the realizations of the šewa of this category as a mobile šewa. In Hebrew grammar this mobile šewa is known as "a šewa following a 'long vowel'"; in the traditional pronunciation of these communities, however, there is no consistent differentiation between "long" and "short" vowels in accordance with the opposition between these two categories of vowels in Hebrew medieval grammatical theory (primarily, in the grammatical theory of the school of the Kimḥis). Therefore, for the actual pronunciation of the communities in question this kind of mobile šewa cannot be defined in terms of "a šewa following a 'long vowel.'"
In some Yemenite pronunciations the second of two final šewas is regarded as mobile.
In the reading of the post-biblical literature the communities who adhere to the Sephardi tradition disclose some deviations from the ways they follow in differentiating the mobile šewa from the quiescent šewa in the reading of the Bible. The most prominent among these deviations is the realization as a quiescent šewa of the šewa in forms of the qatěla, qatělu, qotělim, and qotělot patterns.
In the pronunciations of the Ashkenazi communities a šewa which historically should be regarded mobile is in fact quiescent in many cases. This is always the case with a šewa coming with a letter that has a dageš forte (see above, category (3)); these communities do not geminate the consonants) and with a šewa that follows a so-called "long vowel" (above, category (5)). As to an initial šewa (above, category (1)), its realization either as a vowel (that is, as a mobile šewa) or as zero (that is, as a quiescent šewa) depends on the phonological rules according to which initial clusters may or may not exist in the vernaculars of the various communities.
In the Yemenite community there exist three categories in the realizations of the mobile šewa:
(1) when it comes with a letter which is not followed by י or by a guttural (that is, a laryngeal – א ,ה, or a pharyngeal – ח ,ע), the šewa is realized like an ultrashort pathaḥ, namely, [ǎ];
(2) when it comes with a letter which is followed by י, the šewa is realized as an ultrashort ḥireq, namely [ī];
(3) when it comes with a letter which is followed by a guttural (but is not itself a guttural), the šewa is realized as an ultrashort vowel, whose quality is identical to that of the vowel of the following guttural. Thus, when the vowel of the following guttural is a qameṣ, the šewa is realized as an ultrashort qameṣ namely, [oˇ]; when it is a šureq, the šewa is realized as an ultrashort šureq, namely [uˇ]; when it is a ḥolem the šewa is realized as an ultrashort ḥolem, namely [ö].
A šewa followed by a gaʿya ("šewa-gaʿya") is realized by the Yemenite community as a short (not an ultrashort) vowel, its quality being determined by the nature of the following consonant as stated above, in categories (1), (2), (3). The realizations of the mobile šewa in the traditional pronunciation of the Yemenite community disclose complete agreement with the realizations of the mobile šewa in the pronunciation that prevailed in the Tiberian school. To these latter realizations a number of medieval grammatical treatises bear evidence. Very few communities except the Yemenite have preserved the above realizations that the mobile šewa had in the Tiberian school: they exist – but much less consistently than in the Yemenite community – in the Aramaic-speaking communities of western Kurdistan. Some Persian-speaking communities, as well as the Aramaic-speaking communities of Persian Azerbaijan (see above) have the realization [a], or [^], a lower-mid unrounded back bowel, for the mobile sewa. This realization is identical, in quality, to category (1) of the realizations of the sewa in the Yemenite community; but whereas in the Yemenite community this realization occurs when the šewa comes in certain positions only (see above), in the above-mentioned communities it is the regular realization of the mobile šewa. The Dutch-Portuguese community realizes the šewa -gaʿya as [a] and a šewa preceding י as [i], an ordinary mobile šewa as [e]. All the communities that follow the Sephardi tradition – except the Dutch-Portuguese community and the above-mentioned Aramaic- and Persian-speaking communities, which can be regarded as following the Sephardi tradition in as much as they realize the qameṣ as pathaḥ and the ṣere as segol – realize the šewa as a phone which is qualitatively identical to the realization of the ṣere and the segol. Quantitatively, the ṣere and the segol, when they come in stressed syllables, may be realized as long vowels, whereas the šewa, which does not come in a stressed syllable, is always realized as a short vowel.
In the Ashkenazi communities, the šewa, in those instances in which it is realized as a vowel (see above) is realized as [e] or [ә]. The Samaritan community has usually a vowel as the counterpart of a mobile šewa of the Tiberian vocalization.
The Yemenite community is the only one to have preserved the quantitative difference between the realization of a ḥatef and that of a vowel which is its counterpart, that is between ḥatef-pathaḥ and pathaḥ, between ḥatef-segol and segol, between ḥatef-qameṣ and qameṣ; the ḥatefs are realized as ultrashort vowels which are qualitatively identical with the vowels which are their counterparts – ḥatef-pathaḥ and ḥatef-segol as [̌a], ḥatef-qameṣ as [́å]. The reason for the identity of the realizations of the ḥatef-pathaḥ and the ḥatef segol is that in the Yemenite pronunciation the realizations of the pathaḥ and segol are identical.
In the Sephardi and Ashkenazi pronunciations the ḥatefs are realized as the vowels which are their counterparts: ḥatef pathaḥ is realized as [a], ḥatef-segol as [e] (or as a variant of this phone, in accordance with the variants of the realizations of the segol in the various communities), ḥatef-qameṣ as [o].
No community maintains, in the realizations of the vowels, the distinction between a "long" and a "short" vowel ("tenuʿah gedolah" and "tenuʿah qeṭtanah"), a distinction prevalent in later medieval grammatical theory. In most communities long realizations of the vowels occur in stressed syllables. The Yemenite community maintains the distinction between ultrashort and ordinary vowels, the mobile šewa and the ḥatefs being realized as ultrashort vowels.
The communities which follow the Sephardi tradition follow, in reading the Bible, the Tiberian rules of stress distribution, as regards both primary and secondary stress. In reading the post-biblical literature there are, in these communities, quite a few cases of deviations from these rules (detailed studies as to the stress distribution in the reading traditions of post-biblical literature in the Sephardi communities are, as yet, missing).
The Yemenite community generally maintains in reading the Bible the Tiberian rules of stress distribution in words which have disjunctive accents; words which have conjunctive accents, on the other hand, quite frequently have stress patterns differing from those of the Tiberian tradition. This latter phenomenon is manifest in the fact that words, which in the Tiberian tradition have an ultimate stress ("milleraʿ"), have in the Yemenite pronunciation the stress on the penultimate syllable ("milleʿl"), and occasionally on the antepenultimate, when they come with a disjunctive accent. In the reading of post-biblical literature the number of the words having a penultimate or an antepenultimate stress, and which according to the Tiberian rules of stress distribution should have an ultimate stress, is greater than in the reading of the Bible.
The Ashkenazi communities do not adhere to the Tiberian rules of stress distribution. Quite frequent is the occurrence of penultimate (or, in some communities, antepenultimate) stress where the Tiberian tradition has ultimate stress.
Samaritan Hebrew has, as a rule, penultimate primary stress (with concomitant secondary stress on the second syllable preceding the one which has the primary stress; secondary stress may fall on the syllable directly preceding the syllable which has the primary stress – this is the case when the former syllable has a long vowel). It may be, however, proven that the actual stress patterns of Samaritan Hebrew are rather late, and that the stress patterns that Samaritan Hebrew formerly possessed were identical with those of Tiberian Hebrew (Z. Ben-Ḥayyim, Sefer Ḥanokh Yalon (1963), 149–160).
The text of Genesis 47:28–31 is given here in the traditional pronunciations of several communities
1.´= primary stress; ͵ = secondary stress. Both signs appear before the stressed syllable. When a word has no stressed syllable, no space is left in the transcription between this word and either the preceding or the following word. Since in the Samaritan pronunciation, stress usually falls on the penultimate syllable, it is not marked in specimen text no. 1, unless it occurs in the last syllable.
2. A colon that follows a letter representing a vowel denotes that the vowel is ultralong, e.g. [ā:] = ultralong[a].
1. The Samaritan Community
(transcription kindly provided by Professor Z. Ben-Ḥayyim)
Gen. 47:28: wyī y˚ā:qob b˚ārәṣ miṣrәm š˚āba ʿ˚āš˚āra šēna wy˚āyyu y˚āmi y˚ā:qob šēni ʿayyo š˚āba šēnәm waŕbīm wm˚āt šēna.
29: wyiqrābu yābu y˚āmi yišr˚āʾәl almot wyiqra albēno alyūsәf wy˚āʾūmәr lū am nā m˚āṣātti ån bīnәk šim nā yēdåk tēt yirki waššītå nā immādi ēsәd w˚āmәt al nā tiqb˚ārinni b˚āmiṣrәm.
30. wš˚ākåbti am ˚ābūti wn˚āš˚ātt˚āni mimmiṣrәm wq˚ābårt˚āni afqēbirrātimma wy˚ā'ūmәr ˚ān˚āki ēšši k˚ādēb˚ārәk.
31. w˚āyʾūmәr iššāba li wyiššāba lū wyištabbi yišr˚āʾәl ʿal rēʾoš ammēṭå.
2. The Yemenite Community (Ṣanʿa)
28. wayh´ḥi ͵yaʿágöv bă´ʾāraṣ miṣ́rāyim šă´vaʿ ʿaśre šå´nå waýḥi yă´me ͵yaʿă´göv šă͵ne ḥaýyow ´šāvaʿ šå´nim wă´ʾarbå'im ʾum´ʾaṯ ͵šå´nå.
29. Wayyigră´vu yă´me yisrå´ʾel lå´mūṯ wayyig´rå liv´nö lĭyö´sef waýyömar ´lö ʾim´nå må´ṣāṯi ´ḥen beʿénāḵå ´simnå͵yåḏă´ḵå ´tāḥaṯ yăréḵi wå'å´sītå´immåḏíḥāsaḏ waʾă´maṯ ´ʾalnå ṭigbă´rēni bamiṣ́rāyim.
30: wašåḵav´ti ʿim ʾavö´ṯay ʾunså´ṯāni mimmiṣ́rāyim ʾugvaŕtāni big, vūrå´ṯåm waýyömar ,ʾånö´ḵíʾaʿăsa ḵiḏ´vå rāḵå.
31: waý yömar hiš,šåv˘´åʿå li way, yiššå´vaʿ ´lö wayyiš´taḥu yisrå´ʾēĺʿalröš ͵hammī´ṭṭā.
3. The Iraqi Community (Baghdad)
28: way´ḥi ´yāʿáqob be´ʾēreṣ mәṣ´rāyәm šәbaʿ ʿәs´re ša´na way´hi ye´me ͵yāʿa´qob še´ne ḥay´yaw ´sēbaʿ ša´nim we-ʾarbáʿim ʾum´ʾaṯ šā´nā.
29: wayyiqrébu yéme yisra´ʿēl lámūṯ wayyәq´rā lәb´no leyósef waýyomeŕlóʾimna ma´ṣāṯíḥen beʿénēḵa sim´na yadéḵa ´taḥaṯ yeréḵi weʿásiṯa ´ʿәmmadi ´ḥesed weʾe´mētʿaĺna tәgbérēni bemәśrāyәm.
30: we͵šāḵab´ti ʿimʾabóṯay wunsáṯāni mәmmәs´rāyәm wuqbaŕtani biq͵burátam wayyo'maŕʾanóḵi ʾeʿ´se ͵kәdbáreḵa.
31:waýyōmer hәš͵šābéʿálī way͵yiššábāʿ ´lō wayyiš taḥu yisrå´ʾēĺʿalrös ͵hammī´ṭṭā.
4. The Aleppo (Syria) Community
28: waýḥi1 ´yāʿáqob béʾēreṣ mәṣ´rāyim ši͵vaʿeśre šána waýhi ye͵me yaʿáqov šéne ḥáyav ´ševaʿ šánim weʾarbáʿim um´atšánā. note: Var.: vayhi [v] and [w] both appear in the pronunciation of the Aleppo community as the realizations of Heb. ו.
29: wayiqrә´vu ye͵me yisráʾēl lámut wayәq´rā lĭb´no leyósēf wáyomer ´lo ʾәm´na máṣati ´ḥen ͵beʿénēḵa ͵sim´na ͵yadéḵa ´taḥat yeréḵi veʿásita ʿәmmádi ́ḥesed ͵veʾémḍṯ ʾaĺna ṯәqbérḍni bemәṣ´rāyim.
30: we͵šaḵav´ti ʿәmʾavótay wunsátāni mәmәṣ̌´rāyim wuqbaŕtani bәq͵būrátam wayómar ͵ʾanóḵi ʾeʿ´se ͵kәdbárēḵa.
31: wáyomer hiš,šaveʿaĺli wayiš͵šávāʿlo wayiš´taḥu yis-ráʾel ʿaĺroš hammәṭ´ṭā.
5. The Community of Lithuania
28: vaýḵi ´yankev beyeḡets mits´ḡaim ´švaesrey ´šono váyi yemeýyankev šney ´ḵayov ´ševa ´šonim veaḡ´boim úmeas ´šono.
29: váyikḡvu yemey yiśḡoel lómus váyikḡo liv´ney léyeyey seyf váyeymaḡley im´no mótsosi ´ḵeyn beýnēḵo ´simno ´yodḵo ´taḵas yéḡeyḵi vey ósiso ímodi ´ḵesed ve´yemes ´alno tigbéḡeyni bemitsḡaim.
30: vešókavti imáveysay unsósani mimitśḡaim ukvaḡ´tani bikvúḡosom váyeymaḡ óneyḵi eése kidvóḡēḵo.
31: váyeymeḡ íšovoli vayísovaley vayiš´taḵu yiśḡoel aĺḡeyš ámito.
i. general (includes works dealing with the pronunciations of more than one community): J. Cantineanu, Essai d'une phonologie de l'hébreu biblique, in: bsl, 46 (1950), 82–133; I. Garbell, Quelques observations sur les phonèmes de l'hébreu biblique et traditionnel, ibid., 50 (1954), 231–43; idem, Mesorot ha-Mivta ha-Ivri shel Yehudei Asya ve-Afrika la-Ḥativoteihen, in: Fourth World Congress of Jewish Studies, Papers, 2 (1968), 453–4 (Eng. summ., 212); G. Garbini, Il Consonantismo dell' Ebraico attraverso il Tempo, in: Annali dell' Istituto Universitario Orientale di Napoli, 14 (1964), 165–90; S. Morag, Ha-Ivrit Shebbefi Yehudei Teiman (1963); idem, "Oral Tradition as a Source of Linguistic Information," in: Substance and Structure of Language, ed. by J. Puhvel (1969), 127–46; G.M. Schramm, The Graphemes of Tiberian Hebrew (1964); S. Morag, Review of Schramm, The Graphemes, in: ks, 42 (1967/68), 78–86; M. Ẓ. Segal, Yesodei ha-Fonetika ha-Ivrit (1918); M. Sister, Probleme der Aussprache des Hebraeischen (1937). ii. specific communities: (1). The Samaritan: Z. Ben-Ḥayyim, Ivrit va-Aramit Nosaḥ Shomron, 1–5 (1957–77); for pronunciation see particularly 3, pt. 1 (1961), 13–27; idem, "Some Problems of a Grammar of Samaritan Hebrew," in: Biblica, 52 (1971), 229–52; R. Macuch, Grammatik des samaritanischen Hebraeisch (1969). (2). The Yemenite: S. Morag, Haivrit (see i; Bibliography pp. 295–305); I. Garbell, Review of Morag, Ha-Ivrit, in: ks, 40 (1964/65), 323–30; E.Y. Kutscher, Yemenite Hebrew and Ancient Pronunciations, in: jss, 11 (1966), 217–25. (3). Sephardi, Italian, and Oriental Communities (except the Yemenite): I. Garbell, Mesorot ha-Mivta (see i); M.J. Premsela (Perat), in: Ha-Olam (Sept. 11, 1941), on the pronunciation of the Dutch-Portuguese community; A.S. Corré, The Anglo-Sephardic Pronunciation of Hebrew, in: jjs, 7 (1956), 85–90; E.S. Artom, Mivta ha-Ivrit ezel Yehudei Italia, in: Lešonenu, 15 (1946/47), 52–61; idem, La pronuncia dell'ebraico presso gli Ebrei della Tripolitania, in: Vessillo Israelitico, 70 (1922), 5; H. Zafrani, La lecture traditionelle de l'hébreu chez les Juifs arabophones de Tiznit (Maroc), in: glecs, 10 (1964), 29–31; I. Garbell, Mivta ha-Iẓẓurim ha-Ivriyyim befi Yehudei Iran, in: Lešonenu, 15 (1946/47), 62–74. (4). The Ashkenazi: M. Altbauer, "Meḥkar ha-Masoret ha-Ivrit haAshkenazit ve-Zikato la-Dialektologyah shel ha-Yidish," in: Fourth World Congress of Jewish Studies, Papers, 2 (1968), 455; D. Leibel, "On Ashkenazic Stress," in: The Field of Yiddish, ed. by U. Weinreich, 2 (The Hague, 1965), 63–72; M. Weinreich, "Prehistory and Early History of Yiddish," ibid., 1 (New York, 1954), 73–101; idem, "Reshit ha-Havara ha-Ashkenazit be-Zikatah li-Veayot Kerovot shel ha-Yidish ve-shel ha-Ivrit ha-Ashkenazit", in: Lešonenu, 27–28 (1966/67), 131–47, 230–51, 318–39. iii. historical problems (in addition to the bibliography given in i): I. Garbell, "The Pronunciation of Hebrew in Medieval Spain," in: Homenaje a Millás-Vallicrosa, 1 (1954), 647–96; Y.G. Gumperz, Mivta'ei Sefatenu (1952/53); S. Morag, "Sheva Kefulot bgd kprt," in: Sefer Tur-Sinai (1959/60), 207–42; H. Yalon, "Shevilei Mivta'im," in: Kuntresim le-Inyenei ha-Lashon ha-Ivrit, 1 (1937), 62–78; 2 (1938), 70–76; idem, "Hagiya Sefaradit be-Ẓarefat ha-Ẓefonit," in: Inyenei Lashon (1941/42), 16–31; "Al Hagiyat ha-Kameẓ ve-ha-Kameẓ he-Ḥatuf be-Ashkenaz," ibid., 31–36; idem, "Le-Toledot Hagiyat ha-Ivrit be-Ashkenaz," in: Inyenei Lashon (1942/43), 52–58. See also the series Eda ve-Lashon and Massorot.