The Pronunciation of Hebrew in the Western Sephardic Settlements (XVI-XX Centuries). First Part: Early Modern Venice and Ferrara (1)

In and around 1552, Isac Cavallero published in Venice three or more bilingual prayer-books (in Hebrew with Spanish translation). Simultaneously an anonymous scholar gave to light a translation of the seliot, penitential prayers, printed in Latin characters in the city of Lagoons. 1 Around the same time in Ferrara Yomtob Atias and Abraham Usque published a complete translation of the Bible, several liturgical works in vernacular, and two editions of a Spanish Psalter. 2 It is possible that other prayer-books (or reprints of the first ones) were published. If so they were lost without leaving any trace. The Spanish siddurim printed in both cities were clearly meant for synagogue use. Continuous annotations enabled the former Marranos, still unfamiliar with Hebrew, to follow step by step


§ 1. THE AIMS OF THE RESEARCH
In and around 1552, Isac Cavallero published in Venice three or more bi-lingual prayer-books (in Hebrew with Spanish translation). Simultaneously an anonymous scholar gave to light a translation of the seliot, penitential prayers, printed in Latin characters in the city of Lagoons. 1 Around the same time in Ferrara Yomtob Atias and Abraham Usque published a complete translation of the Bible, several liturgical works in vernacular, and two editions of a Spanish Psalter. 2 It is possible that other prayer-books (or reprints of the first ones) were published. If so they were lost without leaving any trace. The Spanish siddurim printed in both cities were clearly meant for synagogue use. Continuous annotations enabled the former Marranos, still unfamiliar with Hebrew, to follow step by step 1 See infra, § 2.1. 2 See infra, § 2.2. * aleoni@alice.it I have greatly benefited from the encouragement, guidance, precious and generous help of my teacher and friend, Prof. Iacob M. Hassán, ‫.ז״ל‬ This great scholar was the first to carry on a systematic study of the Sephardic idioms. It is thanks to his pioneristic stud-ies that the present work could be accomplished. a service conducted in that language. For this purpose, both Cavallero in Venice and Atias and Usque in Ferrara, transliterated some passages of the Hebrew ritual into Latin letters. By perusing these transcriptions, it was possible to study the features of Hebrew pronunciation in the XVI th century Sephardic settlements in Northern Italy.
A comparative study of the vulgarizations printed in Venice and in Ferrara led me to the conclusion that, not only the Spanish translations, but also the transliterations of Hebrew prayers were the original works of separate and independent authors who acted without knowledge of the work of the others. As a matter of fact, these authors pronounced Hebrew in an identical manner but sometimes represented the same sounds by different graphemes.
Additional research was carried out in order to ascertain when and where in Italy the pronunciation of the consonant ayin shifted towards the nasal sound represented by the graphemes ng, ngh, gn, n. It was possible to state that, in the Italian and Ashkenazi communities, this change took place during the course of one and a half centuries, roughly between the end of the XVI th and the beginning of the XVIII th century. The Sephardim adopted this pronunciation later than the Italiani. We shall expose the results of this particular investigation in the second part of this work. § 2. THE AVAILABLE SOURCES The present research is based on the analyses of: -Hebrew names, words and texts transliterated into Latin letters by Jewish scholars in Spanish and Portuguese vulgarizations of the Bible, siddurim, maazorim and haggadot.
-Spanish words and texts transcribed into Hebrew characters mainly by Jewish translators and compilers of siddurim and haggadot.
-Transcriptions of Jewish names and Hebrew words in deeds drawn up by the Chancellors of the Portuguese Nation of Ferrara, in contracts, letters, last wills and other documents written by Sephardic merchants in Ferrara, Venice and Ancona in the XVI th century.
-Jewish names and Hebrew words transcribed into Latin characters by Christian notaries and other public officers in several Italian cities as 91 well as similar transcriptions registered by Inquisition's notaries and of-ficers in Italy and in Portugal.
-Statements and annotations explicitly made by grammarians, schol-ars and rabbis.
-Transcriptions of Hebrew words in the XVII th century Registers of Deliberations ‫הסכמות(‬ ‫ספרי‬ Livros das hordems of Pisa, Leghorn, Reggio Emilia, Venice and of the Western Portuguese Nations).
In the following pages we provide a concise description of these sources. § 2. 1

. The Venetian Vulgarizations
Towards the middle of the XVI th century Venice was the capital of Jewish printing, 3 not only for the quantity of its production, but especial-ly for the hitherto unequalled typographical perfection reached by several printing houses. 4 3 I borrow this expression from C. Roth, EJ 13, cols. 1096-1114, s.v. «Printing, Hebrew.» Before dealing with the XVI th century vulgarizations of the Hebrew ritual, it is perhaps advisable to recall that the Spanish prayer-books were anteceded by several editions of Hebrew siddurim printed in Venice ac-cording to the Sephardic rite. In 1524, Cornelio Adelkind 5 edited the vol-ume ‫ספרד‬ ‫ותפילות‬ ‫תחנות‬ ‫תמונות,‬ (Images, Supplications and Prayers of Sepharad) 6 for Daniel Bomberg who had it printed yet again in 1544. A new edition of this siddur was prepared in 1546 by A. Benveniste for M. A. Giustiniani with the title ‫השנה‬ ‫מכל‬ ‫תפילות‬ ‫סדר‬ (Siddur for the whole Year). 7 Zuan de Gara 8 reprinted it in 1581 «with Bomberg types» 9 (just to quote a few of the many editions).
In or around 1552 Isac Cavallero published in Venice, at least, three bilingual prayer-books with a literal Spanish translation printed side by side with the original text in Hebrew. 10 Unfortunately for a long time these siddurim did not receive proper attention from scholars. Some au-thors found it easier to quote only the Spanish part of their titles with-out mentioning the Hebrew wording. Owing to these inaccuracies other scholars were induced to believe that the Venetian siddurim included only the Spanish translation (without the Hebrew text). The complete title is ‫ופיוט‬ ‫מזומן‬ ‫ולאכול‬ ‫בשלחן‬ ‫ערוכות‬ ‫ספרד‬ ‫תפילות‬ ‫תחנות‬ ‫תמונות‬ ‫דניאל‬ ‫בבית‬ ‫אדילקינד‬ ‫קורנילייו‬ ‫ידי‬ ‫על‬ ‫נדפס‬ ‫מסומן.‬ ‫ותמצא‬ ‫תחפש‬ ‫בסופו‬ ‫מפורד‬ ‫בתוכו‬ ‫ופזמון‬ ‫ר"פד‬ ‫בשנת‬ ‫יצ"ו‬ ‫.בומבירגי‬ 7 A copy of this siddur is preserved at the British Library (Sh.m. 1972.b.25). It was merely a reprint of the siddur published by Bomberg a few years before. From a legal point of view this was made possible by a parte issued by the Venetian Senate in 1517, according to which the validity of all the privileges and copy-rights was limited to a ten years period. See  The first of these books, the Orden de Oraciones, 11 was printed be-tween 1550 and 1552, a short time after the expulsion of the Marranos from Venice and its dominions. As a consequence of political circum-stances both the translator and the publisher of this book kept a low pro-file. Neither the name of Alvise Bragadini nor his printer's mark appear on the title page. 12 For some reasons the siddur was published without a prologuo (introduction), without a tabla (index), without a colophon and without any expression of thanks for the completion of the work. Even the date of the edition was withheld. No explanation was provided on the style of the translation that followed word for word the form and order of the Hebrew text 13 , even if the elegance of the modern Spanish language was lost in the process. 14 Other accidental circumstances withheld precious information from scholars. The Bodleian Library in Oxford owns a copy of the Ordenança de Oraciones published by Cavallero in 1552. This volume was however incomplete as the (first) title-page was lost. 15  tion, its date and the name of the printer remained unknown. It was the great merit of Cowley 16 to understand that this siddur was printed in or around 1550, and not towards the end of the XVI th century as formerly believed. 17 The incorrect dating of these prayer-books led many scholars to underestimate the importance of the Sephardic Settlement in Venice in the fifteen-fifties. 18 After the Second World War the Library of the Jewish Theological Seminary of New York acquired a complete copy of the Ordenança de Oraciones formerly belonging to E. N. Adler. 19 Margherita Morreale, who consulted it, described the style of the Spanish translation, but scarcely paid any attention to the Hebrew text. She even neglected the Hebrew title of the book and ignored the information therein supplied. 20 Her inac-curacies led some scholars to the wrong conclusion that Spanish was the main language used by former Marranos for their prayers in Venice and in Ferrara the XVI th century.
In 1552 Alvise Bragadini published also a Spanish vulgarization of the seliot (penitential prayers) without the Hebrew text. 21 The name of the print-er was provided only in «Raši» characters. 22 As the name of the translator does not appear in the title page, he thus remains unknown. By carrying out a comparative study of the vulgarizations of the bakašah 23 ‫ומטה‬ ‫מעלה‬ ‫ברואי‬ ‫כל‬ made by Isac Cavallero (in Venice) and by Yomtob Atias (in Ferrara), Salomon could state that these translations were produced independently by two separate authors who acted without knowing about each other's work. 24 By comparing the text of the Venetian Celihot 25 with the analogous Orden de Silhoth 26 published almost contemporaneously by Yomtob Atias in Ferrara, I came to the conclusion that these editions were also produced separately by different scholars.
In his bilingual prayer-books Cavallero gave a rigorous translation of the Hebrew text. In order to convey to the reader the exact meaning of every single Hebrew word he tried to maintain the structure of the sacred language even in the form of the verbs. This author arranged the Spanish translation opposite the Hebrew text in perfect symmetry, page by page, line by line, and word by word so that every Spanish word was facing its Hebrew counterpart. The reader could thus find an immediate explanation of every word of the original text. These siddurim provide a valuable in-strument for the study of the Hebrew language and the Sephardic Liturgy.
In Cavallero's Orden de oraciones the blessings for the reading of the Torah and the ‫הגומל‬ ‫ברכת‬ (to be said after having escaped a danger) are not translated into Spanish. Instead of the vulgarization, the Spanish side of the siddur bears a transliteration of the Hebrew text into Latin charac-ters. 27 Former Marranos, not familiar with their fathers' language, were thus enabled to take an active part in the service.
The siddurim printed by Cavallero  the worshipper throughout the book and explain which sections of the prayers are to be recited by the azzan and which and when responses are to be given by the congregation. These annotations are generally given in the Judeo-Spanish ver-nacular, both in Hebrew characters (on the Hebrew side of the text) and in Latin letters (alongside the Spanish vulgarization). 28 They pro-vide a precious instrument to appraise the phonetic value of Hebrew consonants in the XVI th century.
In the Orden de Oraciones, the annotations on the Hebrew side of the book are generally printed in small cursive «Raši» types. These texts are not immediately intelligible to readers unfamiliar with this particular writing. It is probably for this reason that some of the Hispanists who studied the style of these translations failed to note a fascinating feature: there are in these texts some characteristic ar-chaisms probably derived from pre-expulsion liturgical literature. In effect these annotations include Hebraisms and expressions typical of the Judeo-Spanish jargon that found their origin back in ancient Sephardic traditions.
An analogous phenomenon can be noted in the Ferrara vulgarizations, in the Venetian Haggadah con su Ladino 29  of salvation) by Ysac Abravanel. Samuel Sarfati was closely related to the Abravanels. 40 This link was strengthened when Yacob, the son of Don Samuel and Benvenida Abravanel, married Benvenida Sarfati, the daughter of Don Samuel Sarfati. 41 In 1552, shortly before handling his printing press over to Yomtob Atias, Samuel Sarfati produced a Hebrew maazor for the peniten-tial days. 42 In 1552 Atias published a Spanish vulgarization of the Jewish prayer-book for the whole year: the Libro de Oracyones de todo el año. 43 In the course of a few months this volume was fol-lowed by two other liturgical works in the vernacular: the Sedur de Oraciones de mes 44 (comprising the daily and saturday service and the prayers for Roš odeš, anuka, Purim and Fast days) and the Orden de Silhoth (penitential prayers). 45

99
Unlike the bilingual siddurim published in Venice by Isac Cavallero, the Spanish prayer-books printed in Ferrara by Yomtob Atias and Abraham Usque consisted only of the Spanish vulgarization with just a few occasional passages in Hebrew characters. For instance, in Atias' Sedur de Oraciones de mes, the Hebrew text of the blessings for the read-ing of the Torah is provided alongside the Spanish translation «en hebray-co y español». 49 Some words in Hebrew characters occur also in Usque's maazor 50 and in his Orden de mes. 51 Cecil Roth was perhaps the first modern scholar who paid extensive attention to the vulgarizations printed in Ferrara. 52 Convinced as he was that all Conversos of Jewish origin had remained faithful to their fathers' religion, Roth assumed that these translations were based on a (supposedly) widespread traditional vulgarization used in Spain by sev-eral generations of Marranos. 53 Following Roth's tracks, other scholars assumed that the translations of the Jewish ritual printed in Ferrara and in Venice were mainly meant for the benefit of those Marranos who did not know Hebrew and prayed in the privacy of their homes (rather than in synagogues).
According to some authors, Spanish had become a sort of sacred language that replaced Hebrew even in the religious ceremonies of the new Sephardic settlements. It was not so. It was the merit of Herman P. Salomon

100
In the Ferrara vulgarizations there are continuous references to the course of the synagogal service. In Yomtob Atias' Libro de Oracyones (Ferrara 1552) the title (or the first words) of many prayers are given in Hebrew, transcribed into Latin letters 55 and we can easily imagine that, in the synagogue of the Portuguese Nation of Ferrara, the azzan stressed his voice at the beginning of each section for the benefit of those former Marranos who still encountered some difficulties in following a service conducted in their fathers' language.
In his siddurim Yomtob Atias did not translate the initial passage of the šema but transcribed it into Latin letters, thus enabling all the worshippers to recite without difficulty the first words of this prayer in Hebrew. 56 He also transcribed the main congregational response to the Kaddiš. These short transliterations were inserted also into Atias' Sedur de mes and into his Orden de Silhoth. 57 Abraham Usque was well aware of this problem. In 1553, he trans-literated, in his maazor, the first lines of the Šema and several other Hebrew words in Latin letters. 58 Usque devoted a special chapter of this book to the repetition of the Amidot where he explained the somewhat complicated order of these prayers and provided a transliteration of some of the main congregational responses to the Kaddiš. 59 A few years later, in 1555, Abraham Usque inserted at the end of his Orden de oraciones de mes the blessings for the Talit, the Tephilin, the Torah reading and the Kaddiš fully transcribed into Latin letters. 60 It is from these and other similar transliterations that we can try to understand how the Sephardim in Venice and Ferrara read Hebrew at that time.
Beyond any doubt, the Biblia en Lengua Española 61 represents the most important achievement of the Sephardic press in Ferra- 55 fig. 9 and doc. 3 in Appendix. 60 See fig. 6 and doc. 4 in Appendix. 61 Biblia en lengua Española (for the complete title, see above, note 14). This pre-cious work is now available in the magnificent Edición facsimilar ed. by I. M. HASSÁN (Madrid 1992), published together with the volume Introducción a la Biblia. ra. 62 Several authors already described the merits of this precious vol-ume 63 and I shall not deal with them at this time.
No Hebrew passages are transliterated into Latin characters in the Biblia, nor there any liturgical reason for doing so. This volume is how-ever a helpful source of information on the Sephardic pronunciation of Hebrew as it bears scores of personal and geographical names transliter-ated into Latin letters.
Alongside XVII th century. 74 I took up its Portuguese title from a note written by the anonymous compiler. 75 The siddur contains the translation of the daily and Saturday service and the essential part of the service for Roš odeš, anuka, Purim, and the Holydays. The documents gathered so far offer a wealth of precious information on the legal status, communal organization, financial, commercial, industrial and cultural activities, as well as on the socio-economical structure of these communities.
Government decrees and judicial records provide names or lists of names of Spanish and Portuguese Jews compiled by chancellors and public officers who obviously did not know Hebrew and transcribed the sound of foreign words by ear. In Ancona the Levantine merchants used to draw up the terms of their commercial transactions in Hebrew and had them written by the chan-cellors of their community or by rabbis. The Portuguese merchants, still unfamiliar with Hebrew, made frequent recourse to city notaries. Both the Levantine and the Portuguese communities had their deliberations recorded by city notaries when they had to appoint delegates who would represent them in front of the Papal government in Rome. 81  The immense notarial archives of Venice have been only partially ex-plored, so far. However it was possible to gather enough evidence on a hitherto unknown Portuguese community established in the Ghetto Nuovo in the middle of the XVI th century. 82  provide scores of Jewish names and the transcription of several Hebrew words.
Other lists of names are found in the records of the Inquisition. A fundamental contribution was provided by the papers of the Venetian Santo Uffizio relating to Jews and Judaizers, integrally published by Ioly Zorattini. 83 This author published also the texts of several proceedings of the Portuguese Inquisition. 84 I have traced and studied the depositions delivered in Portugal by sev-eral people who had belonged to, or had visited Sephardic communities in Italy. In many cases these witnesses described the Jewish ceremonies and sometimes recited passages of prayers, using Hebrew terms and words that the notaries of the Inquisition transcribed on the spot, by ear, into Latin letters. Thirteen years later the same Duke awarded the Spanish and Portuguese Nation 87 authority to elect Massari and Deputati ‫וממנים(‬ ‫,פרנסים‬ guard-ians and appointed officers) with power to enact ordinances, to deal with any dispute among the members of their Nation, to give judgement with-out appeal, to impose and collect taxes, to fine and excommunicate the 83 P. C. IOLY ZORATTINI (ed.), Processi del S. Uffizio di Venezia contro Ebrei e Giudaizzanti (1548-1734), 14 vol. (Firenze 1980-1999). 84 87 This was the name of the Sephardic community of Ferrara in 1555. It was later called «The Portuguese Nation.» transgressors and to enforce punishments. However the jurisdictional au-tonomy of the Nation was not extended to criminal cases.
One of the first ordinances enacted by the Spanish and Portuguese community of Ferrara established the penalty of niddui (a form of ex-communication) for those who infringed the jurisdictional autonomy of the Nation by appealing to non-Jewish courts. A plenary assembly unani-mously agreed upon this provision 88 which became a binding Ascamah (agreement, approved deliberation), as the Sephardim used to call their ordinances. Unfortunately the first Register of the Deliberations of the Spanish and Portuguese Nation of Ferrara was lost (or however is not available to scholars). We know about this Ascamah from a Responsum by R. Yeiel Trabot. 89 The Chancellors and Rabbis of the Portuguese Nation of Ferrara acted as public notaries and drew up deeds and statements, mainly in Portuguese but also in Hebrew, at the request of members of their community. Copies of some of these documents are preserved in the city archives, attached to Christian notaries' deeds. These papers are important also from the linguis-tic point of view as they contain expressions lifted from Judeo-Spanish and many Jewish names and Hebrew words transcribed into Latin characters.
In the second half of the XVI th century, Portuguese former Marranos were alllowed to settle in Pisa and in Leghorn (in Tuscany), in Savigliano (in Piedmont), in Nice (in Savoy), and in several minor Italian cities where they established new Jewish centres. They obtained from the Dukes of Savoy and from the Grand-Dukes of Tuscany the identical privileges that Hercules II had granted to the Spanish and Portuguese Nation of Ferrara in 1555. 90 The new communities enjoyed wide administrative and jurisdictional autonomy. Their communal organization was governed by statutes and ordinances (Escamot) drafted by the Mahamad (governing board) and 88 JTSA Library, Ms democratically approved by the plenary assembly of the yeidim. 91 These regulations became a binding code for all the Sephardim and encompassed the whole range of Jewish life. The Escamot stated the principles accord-ing to which the «Men of the Nation» had to behave inside the commu-nity, in the privacy of their homes and in their relations with Gentiles. The Mahamad had full authority to impose taxes for the upkeep of the Nation and to inflict punishments to transgressors and rebels. 92 In Ferrara the Portuguese Jews were proud of their religious, adminis-trative and jurisdictional autonomy and regarded it as the most important of their privileges. The main, or at least the most conspicuous, duty of the Mahamad was to settle the disputes among the «Men of the Nation.» Three parnassim were always available (standing, ‫)עומדים‬ for this purpose at the Residence of the Nation, in the building of their synagogue. Thus, the Mahamad was metaphorically called «The Standing Residence». 93 The Spanish and Portuguese Nation of Reggio Emilia was perhaps the only Sephardic community in Italy who called its governing board Vaad ‫,ועד[‬ committee, council]. 94 The The tremendous importance of this documentary corpus cannot be overestimated. Strangely enough the scholars, who studied the minute books compiled in Italy, were mainly Hispanists interested in the pe-culiarities of the vernacular(s) spoken by the Portuguese Nations. 104 A comprehensive study on the rendering of Hebrew words and names in this literary production is still among the desiderata.
These registers were compiled partly by the chancellors and officers of the different Nations and partly by the Parnas Presidente in charge, 105 during his mandate. As these officers were not necessarily scholars or literates they generally had no academic interest in the rules of gram-mar. They drew up these minutes, and wrote the Hebrew terms therein occurring, transcribing the Hebrew (as well as the Portuguese) words by ear, according to their sound, as they heard it during the service in the synagogue or in current conversation, thus providing direct unamended evidence of the actual pronunciation.
Of great interest are also the Statutes of Charitable Confraternities ‫)חברות(‬ such as, for instance, the [ The importance of these instruments is obvious and does not need a special discussion. Of particular interest is a grammar book published in Spain in 1526 by the converso Alfonso de Zamora. This author claimed that he had access to the archives of the Jewish communities, seized at the time of the expulsion but still extant. 108 Thus he stated, in rather explicit terms, that his grammar reflected the linguistic traditions -and the pronunciation-of the Spanish Jewry. The scientific value of this work was stated by S. D. Luzzatto. 109 A grammar book compiled at the beginning of the XVII th century by the Catholic Humanist Blancuccio is particularly interesting as it proves that the /gn/ pronunciation of ayin was already taking place among Italian Jews. 110 The author was a trustworthy witness of this phenomenon. As we shall see, the perusal of grammatical works is es-sential in documenting the progressive shift towards a nasal-guttural pronunciation of the ayin by the Italian Jews between the XVII th and the XIX th Centuries.

§ 3. THE RENDERING OF HEBREW CONSONANTS AND VOWELS BY THE SEPHARDIC SCHOLARS IN THE XVI TH CENTURY
Of paramount interest is Yomtob Atias' translation of the book of Lamentations, 111 which is read during the Yom Kippur service. As it is known, this book consists of five poetic compositions: the first four are alphabetic acrostics where every verse (or cluster of verses) begins with letters of the Hebrew alphabet in progressive sequence, from alef to taw. The repeated succession of verses in alphabetical order was meant to stress the manifold sufferings of the Jewish people. 112 It was also a way of expressing the hope that the tribulations and misfortunes of the sons of Israel were completely fulfilled 113 and had now come to an end. Yomtob Atias was well aware of the importance of this message. 114 In order to convey it unaltered through the process of the vulgarization, he resorted to an expedient: he used the names of the Hebrew letters in Latin characters as headlines for each (cluster of) verse(s) in alphabetical order. Thus the names -and the pronunciation-of the Hebrew consonants are provided, one by one. 115 The same device was used in the Biblia Española, in the translation of Psalm 119, 116 where each stanza is preceded by the name of the initial letter of the Hebrew text transcribed into Latin characters. The names of several Hebrew consonants provided in the Biblia are slightly different from the ones given by Yomtob Atias in his Libro de Oracyones. This adds new evidence to the fact that the translation of the Biblia was undertaken by a team of scholars 117 rather than by a single unknown author. 118 111 ATIAS, Libro de Oracyones, ff. 209-217v. 112

111
In his Psalter (Ferrara 1553), Abraham Usque used the Hebrew con-sonants in alphabetical order as headlines for the stanzas of Psalm 119. However, for this purpose, he employed Hebrew fonts and did not tran-scribe the name of Hebrew letters into Latin characters.
In 1622 Abram Netto published an enlarged edition of Isac Cavallero's bilingual Orden de Oraciones. The new volume 119 presented a fascinat-ing novelty. Netto inserted at the end of the siddur an appendix with a list of the Hebrew letters and vowels, giving their names in Latin characters. Furthermore he provided a short Declaration delos Puntos (Explanation of vowel-signs) where he explained how every diacritic symbol 120 was to be pronounced. 121 Until now the Declaration delos Puntos was ascribed to Isac Cavallero 122 along with his Orden de oraciones. 123 However, I did not find any such explanation in the vulgarizations printed in Venice in the middle of the XVI th century. Thus, in my opinion, the attribution of this «work» to Isac Cavallero is completely groundless. Moreover, in the appendix of the prayerbook sponsored by Abraham Netto, the names and the sounds of several consonants are not consistent with the pronunciation rules (tac-itly) adopted by Cavallero seventy years before. The Declaration delos Puntos was therefore the work of an independent author: either the same Abram Netto or a scholar employed by him for this purpose. . 121 See fig. 15. 122 See, for instance, KAYSERLING, Biblioteca Española, pp. 37-38, 59-60. 123 As we have pointed out, the title of the Ordenança de Oraciones was not available to scholars until the middle of the XX th Century.
The names of the Hebrew consonants, given by several scholars in the XVI th century are shown in the Table to be published in the second part of this work. The changing names of some letters show the progressive variations occurred in their sound in the span of four centuries.
The pronunciation of each consonant is discussed in the following pages. Unless otherwise stated, these notes are related to the pronun-ciation of Hebrew by the Sephardim in Venice and in Ferrara in the XVI th century. In particular cases I paid attention to further develop-ments in the rendering of specific Hebrew consonants.
Additional research was carried out in order to ascertain how the Italian communities of different origins pronounced the consonant ayin between the XVI th and the XX th century. The results will be pre-sented in a following article to be published in a further issue of this journal. § 3.1. Alef In the sixteenth century the phonetic value of this letter was zero. In most cases it was left without transcription, especially when it occurs at the beginning of a word, e.g.  153 In the following centuries the almost general lack of distinction be-tween soft and hard bet remained a typical feature of the Sephardic pro-nunciation in Italy and in the Western Portuguese Nations, e.g., tebà, tebah ‫,תבה[‬ podium], 154 175 We also find the forms g ui , g ue and even gh i , gh e or gh a . The former graphemes were probably originated by their affinity and similarity to the Spanish forms q ue , q ui , e.  219 This form occurred also in the XVIII th century Register of Deliberations of the Portuguese Nation of Bordeaux where the et was transcribed also by <j>, e.g., ebra, 220 jebra or even jebera; 221 bedajaim ‫החיים[‬ ‫,בית‬ cemetery]. 222 These transcriptions were probably due to persons of Spanish mother tongue. § 3.9. et This consonant was uniformly pronounced /t/ identically with the taw. In the transcription of Spanish texts into Hebrew characters, the letter et was used to graphically represent the sound of /t/, e.

124
Abrahan, 265 Effrain, 266 Yon Tob, 267 Menahen, 268 Menachin, 269 Cain, Cayn and Hain ‫,]חיים[‬ 270 Yeoyachin, 271 but also queducin ‫,קדושים[‬ mar-riage celebration]. 272 Final mem was often merged with the preceeding vowel and ren-dered by a velar nasal sound, as it appears from such Lusitanian forms as Abrahão, 273 Abrahaõ. 274 The final diptych of this word was undoubt-ely pronounced with a marked nasal accent by the Portuguese Jews of Ferrara. § 3.14. Nun  281 In Ladino the alternation between /m/ and /n/ occurred also in other word positions as it is shown by such transcriptions as ‫אינפיסאר‬ [empe-zar, to start], 282 ‫איסקומדאראן‬ [escondaran, they will hide]. 283 In the transliteration of Spanish words, final nun was frequently left untranscribed, e.g., ‫ליאמרא‬ [llamaran, they will call], 284 ‫דירא‬ [diran, they will say]. 285 This happened occasionally in the transcription of Hebrew names, also when nun does not occur at the end of the word, e.g., Beyamin ‫,בנימן[‬ p.n.]. 286  In Italy, in the first half of the XX th century, ayin was realized as a voiced velar nasal phoneme commonly transcribed as gn. Its sound was similar to the Spanish ñ in the word España and to gn in the Italian word sognare or in French agneau. Typical was the rendering of Šema as Scemang or Scemagn. According to popular belief, this particular sound was introduced by the Iberian exiles that arrived in Italy after the expulsion from Spain (1492) and the general forced conversion in Portugal (1497). We could ascertain that it was not so. In the second part of this work we shall deal with the progressive shift in the pronunciation of ayin a process which took place in the course of a few centuries. § 3. 16

.2. The Pronunciation of Ayin in the Early Modern Sephardic Settlements of Venice and Ferrara
The perusal of a wide range of transliterations of Hebrew words and (mainly liturgical) texts led me to the conclusion that, in the XVI th century Sephardic communities of Venice and Ferrara, the phonetic value of ayin was zero or tending to zero.
In the vulgarizations of Hebrew prayers printed in Ferrara and in Venice this consonant was most frequently transcribed as <h>. For 287

128
The fact that in many cases the ayin was left without notation cannot be assumed as an absolute proof that its ancient sound was completely lost and abandoned by all Sephardic scholars in early modern Northern Italy. According to H. P. Salomon, already in the fifties of the XVI th century there were examples of ayin represented as <hg>. 312 I did find one only similar case in the works prublished in Ferrara. In his Libro de Oracyones, Yom Tob Atias gave in 1552 the name of this consonant as hgain. 313 This perhaps implies at least a reminiscence of an ancient pro-nunciation. However, rather inconsistently, this author did not use this grapheme in his siddurim and he normally transcribed the ayin as h or even ignored it. As far as I know the grapheme <hg> does not appear in other vulgarizations of the Sephardic liturgy printed in Italy during the XVI th century. § 3.17. Pe The pronunciation of this consonant calls for no particular attention.
Hard pe was regularly transcribed as p and soft pe as f or ph, e.g., sepher, 314 336 Italian notaries transcribed the sound of the qof as c or ch, for instance, chetubbà, 337 cheducin. 338 § 3.20

. Reš
This consonant does not require any particular attention. It was uni-formely rendered as r. According to the Converso Alfonso de Zamora, acquainted with Jewish traditions, Spanish Jews pronounced the reš in word initial position with a marked, doubled sound. 339 We are not in a position to confirm -or to exclude-that this char-acteristic was still present in the early modern Sephardic settlements of Venice and Ferrara. However in his Compendium Grammatices Linguae Hebreae (Amsterdam 1676) Spinoza stated that this consonant was «lene in medio, asperum in initio dictionis». 340 It is worth mentioning that when reš occurs immediately after a dor-so dental fricative consonant, such as šin, ade or zayin, a dental plo-sive is inserted in the transcription, between the two phonetic ele-ments, e.g. Isdrael, 341 Esdra, 342 Azdriel ‫,עזריאל[‬ p.n.], misdrah ‫,מזרח[‬ As we have seen, Abraham Usque transcribed both hard and soft taw indifferently as t or th. However, in his Orden de Oraciones de mes (1555), he provided a transliteration of the Kaddiš where the soft taw was sistematically represented as th. 368 It is difficult to understand why Usque suddenly adopted new rules, inconsistent with those followed in the pre-ceeding pages. Probably the distinction between hard and soft taw in this section of his book was due to the presence of an Aškenazi proof-reader in the Ferrara printing press. We also wonder whether Usque lifted this transliteration from another siddur, perhaps no longer extant and unavail-able to us.
In rare instances taw at the end of a word was left untrascribed. In his Orden de Oraciones de mes, Abraham Usque rendered ‫טּלית‬ as Tale. 369 The same transcription is found in the 1519 Venetian edition of this siddur 370 and in many later Amsterdam editions, as well as in the Orden de Oraciones published in Venice by Abraam Netto in 1622. 371 We find «Tale» even in an inquisitorial document written in Rome in 1586. 372 In 1622 Abraam Netto transcribed the soft taw in final position as <d>. In a table published at the end of his Orden de oraciones, 373 he gave the names of Hebrew letters as follows: Alef, bed,... daled,... ched, ted... He also declared the name of the vowel pata as padach. Netto was not consistent with his own explanation. In the last pages of 136 instance, in the 1630 Amsterdam edition of the Biblia we find an «Orden delas Aphtoras». 389 This form has survived in Bayonne where the discen-dants of the first Portuguese émigrés still pronounce this word as Aphtora. 390 § 3.23.2. The /e/ Sounding Vowels In the XVI th century Venetian and Ferrarese transcriptions of Hebrew texts no distinction was apparently made between «long» and «short» vowels. In any case, even if the authors did appreciate this difference, the typesetters were not equipped with fonts suitable to express the shades of phonetic values. As a matter of fact they had some difficulties even to cope with the accents of the Iberian languages.
No graphical distinction was made in the transcription of /e/ sounding vowels (segol, ere, šewa, segol atef) were all transcribed as <e> 391 and pronounced as in the English word set. Šewa was always pronounced when it occurs at the beginnig of a word, when it accompanies a letter with a dageš, as for instance,  395 Šewa was pronounced also in other instances which we do not list here. In fact it is almost impossible to ascertain whether the authors of the Venetian and Ferrarese vulgarizations followed precise principles in distinguishing between a mobile (pronounced) šewa from a quiescent one. In any case, on the basis of the available documen-tation, we cannot state whether these authors were consistent in following Yod occurring after vowels sounding /a/, /o/, /u/ was transcribed as i or y, e.g., Sephatay  We carried out an extensive graphematic analysis of the translitera-tions of Hebrew texts published by Sephardic Scholars in Venice and in Ferrara towards the middle of XVI th century, as well as of later documents. We can state that both the XVI th century Spanish translations of Jewish liturgy and the transliterations of Hebrew prayers were the original works of separate scholars who acted independently from each other. These au-thors pronounced Hebrew in the same way but, due to the lack of graphic norms, they represented the same sounds in different ways. The same was true also for the vernacular works of the Sephardi émigrés who carefully respected the sounds of the Iberic languages but often represented them by different graphemes.
It was possible to ascertain the features of the Sephardi pronunciation of Hebrew in early modern Northern Italy, as follows: 417  a) The almost general lack of distinction between hard and soft bet. In most cases the latter was pronounced as a plosive /b/ in the same way as hard bet.
b) During the XVI th century the phonetic value of the ayin was zero or close to zero. This consonant was transcribed as h or even ignored. c) There was almost no distinction between šin and śin. Both were transcribed as s or ç and pronounced /s/. d) Both soft and hard taw were pronounced as t. As we shall see better in the second part of this article, in the course of the XVII th century the pronunciation of soft taw started to shift towards /d/. This was due to the influence of Italian rabbis and scholars e) No distinction was made between kame and pata. Both vowels were read as an open a. This feature was also common to the Italiani Jews. * * * The Documentary Appendix and the Figures will be published in the next issue of Sefarad. [Continuará]